<%BANNER%>

Trajectories of Two Racine Birth Cohorts

University of Florida Institutional Repository
Permanent Link: http://ufdc.ufl.edu/UFE0021608/00001

Material Information

Title: Trajectories of Two Racine Birth Cohorts A Theoretically Integrated Model for Explaining Offending
Physical Description: 1 online resource (181 p.)
Language: english
Creator: Jennings, Wesley G
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2007

Subjects

Subjects / Keywords: multi, offending, trajectories
Criminology, Law and Society -- Dissertations, Academic -- UF
Genre: Criminology, Law, and Society thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Macro-level research evidencing the relationship between communities and crime dates back to the early ecological work of Park and Burgess and later further developed by Shaw and McKay into what is commonly referred to as social disorganization theory. This research has since shown that poverty, ethnic heterogeneity, and residential mobility lead to the inability for members of the community to develop mechanisms of informal social control, which in turn leads to higher rates of crime. In contrast, micro-level theories such as Sampson and Laub?s theory argues that individual pathways of offending can be altered by life transitions such as marriage and employment. Comparatively, Moffitt?s taxonomy asserts that there are two theoretically distinct groups of offenders, where one group displays an adolescent-limited type of offending and the other group are considered to be life-course-persistent or chronic offenders. Moffitt has also argued that the life-course-persistent type is predominantly a male phenomenon. This study will provide one of the first attempts to explore the possibility of developing a theoretically integrated model for explaining group offending using trajectory analysis and relying on data from two Racine Birth Cohorts. Results from this study relying on official data yielded varying levels of support for all of the theoretical frameworks mentioned above. Specifically, all of the trajectory models discovered groups consistent with Moffitt?s typology, as well as uncovering groups not necessarily consistent with her typology. Also consistent with Moffitt, sex was shown to be a significant covariate for predicting group assignment. Bivariate support was found for Sampson and Laub's age-graded theory of informal social control. Lastly, multivariate support was revealed for the impact of living in a socially disorganized area. Yet, once mlogit models were estimated incorporating the individual and neighborhood effects there was evidence of mediation. It appeared that individual and family variables at times completely or at least partially mediated the effect of social disorganization as a distinguishing feature of group membership. Implications for the current findings and directions for future research are also discussed.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Wesley G Jennings.
Thesis: Thesis (Ph.D.)--University of Florida, 2007.
Local: Adviser: Lanza-Kaduce, Lonn M.
Local: Co-adviser: Piquero, Alexis R.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2007
System ID: UFE0021608:00001

Permanent Link: http://ufdc.ufl.edu/UFE0021608/00001

Material Information

Title: Trajectories of Two Racine Birth Cohorts A Theoretically Integrated Model for Explaining Offending
Physical Description: 1 online resource (181 p.)
Language: english
Creator: Jennings, Wesley G
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2007

Subjects

Subjects / Keywords: multi, offending, trajectories
Criminology, Law and Society -- Dissertations, Academic -- UF
Genre: Criminology, Law, and Society thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Macro-level research evidencing the relationship between communities and crime dates back to the early ecological work of Park and Burgess and later further developed by Shaw and McKay into what is commonly referred to as social disorganization theory. This research has since shown that poverty, ethnic heterogeneity, and residential mobility lead to the inability for members of the community to develop mechanisms of informal social control, which in turn leads to higher rates of crime. In contrast, micro-level theories such as Sampson and Laub?s theory argues that individual pathways of offending can be altered by life transitions such as marriage and employment. Comparatively, Moffitt?s taxonomy asserts that there are two theoretically distinct groups of offenders, where one group displays an adolescent-limited type of offending and the other group are considered to be life-course-persistent or chronic offenders. Moffitt has also argued that the life-course-persistent type is predominantly a male phenomenon. This study will provide one of the first attempts to explore the possibility of developing a theoretically integrated model for explaining group offending using trajectory analysis and relying on data from two Racine Birth Cohorts. Results from this study relying on official data yielded varying levels of support for all of the theoretical frameworks mentioned above. Specifically, all of the trajectory models discovered groups consistent with Moffitt?s typology, as well as uncovering groups not necessarily consistent with her typology. Also consistent with Moffitt, sex was shown to be a significant covariate for predicting group assignment. Bivariate support was found for Sampson and Laub's age-graded theory of informal social control. Lastly, multivariate support was revealed for the impact of living in a socially disorganized area. Yet, once mlogit models were estimated incorporating the individual and neighborhood effects there was evidence of mediation. It appeared that individual and family variables at times completely or at least partially mediated the effect of social disorganization as a distinguishing feature of group membership. Implications for the current findings and directions for future research are also discussed.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Wesley G Jennings.
Thesis: Thesis (Ph.D.)--University of Florida, 2007.
Local: Adviser: Lanza-Kaduce, Lonn M.
Local: Co-adviser: Piquero, Alexis R.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2007
System ID: UFE0021608:00001


This item has the following downloads:


Full Text
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID E20101203_AAAAFE INGEST_TIME 2010-12-04T03:07:22Z PACKAGE UFE0021608_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES
FILE SIZE 7527 DFID F20101203_AACFLA ORIGIN DEPOSITOR PATH jennings_w_Page_164thm.jpg GLOBAL false PRESERVATION BIT MESSAGE_DIGEST ALGORITHM MD5
fc0f4f335f5fb4f7eeb9a804926d6caf
SHA-1
dc4258b53c2d331976e7d9173e8ccee7c0ddfd72
2067 F20101203_AACFKM jennings_w_Page_084.txt
cfb8cf0be8bbb80e20f5b4842881c9d3
896023836af77a58a54bb5ac2bfe3fdf6977b876
23913 F20101203_AACEHK jennings_w_Page_092.QC.jpg
229f92397130bbb8dc7ddb1a38f83f35
3488bbdfe22514409ea020e48ffde2a8dec99414
1053954 F20101203_AACEGV jennings_w_Page_026.tif
8bc3c55adf55e86dbaefef2229c5d07b
0ca94760269db1c6a7bad2559b13bbb07c2d468e
25714 F20101203_AACFJX jennings_w_Page_142.jpg
68684fa18b718caaaff9e7e071d2a60e
a66309db5d03855be6c3349eac390fad232a22bd
26487 F20101203_AACFLB jennings_w_Page_170.QC.jpg
046a60695c742a23373e0f27eaf273df
8a8f301debe64fd6c42277e542b3459146e6c757
6489 F20101203_AACFKN jennings_w_Page_079thm.jpg
e05e6adef4327e00f84f594f72f121b3
eb9acbe1924159aae880206f80ca92ec84345bef
16851 F20101203_AACEHL jennings_w_Page_005.QC.jpg
c5ac7bbf0089f26324e42c1d84f81558
5c973fe35846439a956549cfe7c45bd1c8888b4f
767 F20101203_AACEGW jennings_w_Page_158.txt
e2d552e84510fe1c9b9def795985bb10
a9bf9ec64cd00a5f90a995d7eac27a45096d7359
130566 F20101203_AACFJY jennings_w_Page_179.jp2
fc219184f54ee1c9eac4a37fd95713b9
2a2d239c675894d8f80a59c55dc50ec59352beec
56744 F20101203_AACFLC jennings_w_Page_047.pro
51a8541afe48882121c853240689a74b
55f79aa7d85856cdb5d007e086ab68bac1abb88f
2101 F20101203_AACFKO jennings_w_Page_093.txt
fa86efdb461efab74ac9b77e4a8b8341
558efe616996e1cf9ff4b22ff65fe1907b2f4f51
79504 F20101203_AACEIA jennings_w_Page_035.jpg
6907b140d2a6ff0802dcbd67aaabaff9
24768edc608e16577ed462102e9a0b8283cf6c71
2132 F20101203_AACEHM jennings_w_Page_108.txt
810828c41bbbc6b70428de80c50d8fd4
52b3d4a8f9434bd7e28e5778d6ac3c5450c482f2
2592 F20101203_AACEGX jennings_w_Page_168.txt
796bc9e7117035187c0f1c461156dd35
4e39101553d29848a6f74f11e9223178d5fe93d7
49295 F20101203_AACFJZ jennings_w_Page_077.pro
0c8ae3a75353d5a1ba6388bc88883a67
7cbd9da8f5ac9b003d2a27138b42a5d14638b51c
4545 F20101203_AACFLD jennings_w_Page_136thm.jpg
149643951471d09d9e8e37a70e921acf
26edd3ddc2be3edc4d932be924ff0ed839b4c24a
50937 F20101203_AACEIB jennings_w_Page_017.pro
525330fc1ba38f649dcbc54c8c1eaea0
cae34383c90c521a6948f8ce1552a41462589f12
1976 F20101203_AACEGY jennings_w_Page_155.txt
3e31bbd8aeafc5e4d45fafa97c307cc6
2597f824fb2fdcc5a547f9ce32b6096404768b10
52949 F20101203_AACFLE jennings_w_Page_062.pro
6b42dcf421ff2cccf64c6ca5a8c5d449
eba4a549ba10d7a9beb209832feb8f8b9c7795a6
F20101203_AACFKP jennings_w_Page_177.tif
734d8aff9d974c087a0e24f67702d0c8
f1f9e82b57c309c0d4028f6ce1ad88fcaaef0d2c
F20101203_AACEIC jennings_w_Page_178.tif
e98b1d8cb3788a47795212d8a2aa8c53
2b9d83feaed45881165aa4f3135e3a10a480c99b
85282 F20101203_AACEHN jennings_w_Page_167.jpg
df2ede603b8762954c6a575be63bd03e
8445cffee9a7d904e58051689b8a059d6a856655
2159 F20101203_AACEGZ jennings_w_Page_043.txt
b45b49759a260449d348e60afdb434a6
6fbdee2abe81799c4bf9368e634c9f2434a91b1e
6992 F20101203_AACFLF jennings_w_Page_145thm.jpg
3c2f1d268ca3cdabd53129df8ff95c6f
0c0df5b0317984ed632734e97399952330d61bbc
106037 F20101203_AACFKQ jennings_w_Page_015.jp2
897e752c533e4268e59cef95c81086ef
108aee3d3d475e2eab59bc02aa84055b6b422b73
2049 F20101203_AACEID jennings_w_Page_028.txt
79a6198740406e3058a11e6bbb6bf81e
ea2e41aa742e697756e62bc6c6b291830b24bb2f
23954 F20101203_AACEHO jennings_w_Page_019.QC.jpg
8fa340baba73709e08a9f61c5185adce
75d2d28ce534dcfbfe9b478c279136a3230dba85
9928 F20101203_AACFLG jennings_w_Page_002.jpg
6bcd653ece388ba78c04b9d2ad009f4b
bd0932bbf1bd67c2ea17e9bb712deccb734437a6
54118 F20101203_AACFKR jennings_w_Page_126.pro
effb1d4b5eb7f2067376934e84fcfc6b
05c4a755452ef6bbb8e745f30e8008f329e40b47
6665 F20101203_AACEIE jennings_w_Page_062thm.jpg
19bc7f926751a380707a55a2197d40fa
34ac0851448a8d1605dc25055b83531fcea3c42c
6343 F20101203_AACEHP jennings_w_Page_011thm.jpg
bb63aa9fb70bb2917deb5efec21e9e7a
b61f7de9890d4e860fb4f44128aabfee9373695d
54609 F20101203_AACFLH jennings_w_Page_076.pro
ba7c6954b44b620df6eeba800580d3d4
3a6d28a6784ba857ca08c1eaec04993f06d3aefb
F20101203_AACFKS jennings_w_Page_106.tif
2e8f9b0cfbc97986782528ee44d08552
2603274e8f5925b7bc791391ec5c5f2983ee3a02
F20101203_AACEIF jennings_w_Page_052.tif
e36bdb5f1ba49d8b2826c51469a46f13
f6f92f880a96f1eb6cf1d9c3f38ebcd8efe27df0
24311 F20101203_AACEHQ jennings_w_Page_089.QC.jpg
8c235ee790c9808cfd326f570f7996d9
86a58438359d159e3c45e345ba9b5429067f1224
2279 F20101203_AACFLI jennings_w_Page_050.txt
5cf8df87350431eb26a54402fbba7674
4d95fa9711089f573788fc1c165c507633dee580
75717 F20101203_AACFKT jennings_w_Page_059.jpg
105bc1362ec739db1bf11be7d84d6bf1
a2d453d84cdeed322ab708b70638f84e06a82331
81934 F20101203_AACEIG jennings_w_Page_050.jpg
2572a8963e03bd1cd3f63df9bef0f240
b481138eb4c8b7cff63447c03a0a2596b4499df8
7180 F20101203_AACEHR jennings_w_Page_119thm.jpg
220837995ab9dbcbb8f375e81a7b862a
91e6a4ece10ad20291bd96c9b8bc7d29c1a8c41a
26091 F20101203_AACFLJ jennings_w_Page_046.QC.jpg
1e478c7b3fae88591d9c39c74c40cb51
cb16f58c0734521e3af48d3a0afa3afc04e4018f
54419 F20101203_AACFKU jennings_w_Page_153.pro
486e155976beacb9b7aaf133642517e2
4dd061853376d324a3900792824c52859b0a5134
6968 F20101203_AACEIH jennings_w_Page_048thm.jpg
d5e8f4d2685db422dde9dbbb88f540ef
c10e5fb4f9d0733630785d6b8369c855c2f0e33d
25845 F20101203_AACEHS jennings_w_Page_150.QC.jpg
2f53fc4f9844649f1da4505fab39090e
3dc0be1b8f9395171161060ca3c956f11a7bc1fd
6374 F20101203_AACFLK jennings_w_Page_067thm.jpg
f4fc50157d74efbff87ecd380b404ead
7f3a511b9a2e298955d9ee1b60b3e722a147cb2c
111749 F20101203_AACFKV jennings_w_Page_032.jp2
06c6554fe25165a07af06a367298c829
3193d01503673ab9ce89ced7d7a7ddcc6ee5ad45
64008 F20101203_AACEII jennings_w_Page_162.pro
d90538ffa69753e339c99461cb4b5f5f
d5a040ca9e4599f1b5db0c9126b664dcecd6052a
117528 F20101203_AACEHT jennings_w_Page_064.jp2
2f6be6d41d22882bce8c71e1a78ff160
732ddeef954220139b62b01459ef510838ca1cca
F20101203_AACFLL jennings_w_Page_174.tif
8ed5f5ec4541d2971c89be156c35f118
df74d27af894d75590e304839f3e99f0c2025fcb
80687 F20101203_AACFKW jennings_w_Page_023.jpg
4c4fb462598c15a49b0354c6a515febf
d9bba8114da522dff443cdb82aca8ed2b6fc0cbc
F20101203_AACEIJ jennings_w_Page_060.tif
1510e3dabe472b97dd7e8f06b8c9bcb7
a96c352b3c1433186933ebe280a0ecf5ff9133e7
26516 F20101203_AACEHU jennings_w_Page_008.pro
28931ab5534557bcdc02bdb3a6c8f3ea
17c6fce40a5aba3c6ee5556f4b5d356f3a556a9b
52738 F20101203_AACFMA jennings_w_Page_120.pro
67d53f1c17051d598a7ccd88d88434ee
a156388f0b27c2b78fb9959251f576eb4a5febbd
F20101203_AACFLM jennings_w_Page_127.tif
541653fbac64df828f45a63813ae2118
8d30c9f776e658a2b745cf5fade643a4e05d5ef7
24996 F20101203_AACFKX jennings_w_Page_105.QC.jpg
25bd9861a9eb6a79b81b88ce6a05e2ab
437a11f7aa1d2e7ca489483bfd9ea798fb11cdec
15609 F20101203_AACEIK jennings_w_Page_102.QC.jpg
5645e705aabd522416330c19b6734040
b41cc6836ea5f47dad8b10c5ca3ba3c000f0bbdf
78575 F20101203_AACEHV jennings_w_Page_063.jpg
e9c567cb9a968deb8f2473f318108dc9
392c8c3bacd1c5cbb30886d028e965828d9c99a9
F20101203_AACFMB jennings_w_Page_038.tif
ef3ae180b749b1d7b634db4723303b4b
e125d04e19732a2f766d1b4273c84fd0606a8613
F20101203_AACFLN jennings_w_Page_116.tif
98c934749bb143b395c46d83ee2f4ad0
bcfdaae1169f8f562f9d2f49442e368eb7d36041
123665 F20101203_AACFKY jennings_w_Page_068.jp2
372f5c4618b762ebb1294febbc9f572b
7db24eedf28916a0dae3b2a6c9a91d2d7f7b77f9
59141 F20101203_AACEIL jennings_w_Page_044.jpg
ef385ab658750e38c2b98a0c3b4dab69
2d802b1836119aa31ff251e6e681ed0d4b76a607
2036 F20101203_AACEHW jennings_w_Page_032.txt
8bc30e4dba109d3c216837cb52277cb0
5284620521e9181f16bbb6da336e7878118a343e
73763 F20101203_AACFMC jennings_w_Page_028.jpg
58f5515eee6f40c9dfacf1994df84fda
a8999633011936f9c051b8692db53ebc423bf152
118675 F20101203_AACFLO jennings_w_Page_101.jp2
42409f1aa9d4a72a161d86fc8d298921
2e413f502d6272675683b16989326e2e3f212f8d
1985 F20101203_AACFKZ jennings_w_Page_009.txt
59405b94d979acf382fffd0ab7e242a4
fe64f21a9a50e1e45ca20bc376c0378426463297
F20101203_AACEIM jennings_w_Page_096.tif
4451193aebfef444bea0fb8a10ade13b
a817aaf36403564839096665f446f601cfc6df67
F20101203_AACEHX jennings_w_Page_114.tif
241c3d8846abe6b9a88c5fc728b004c5
3556c3911b3b8fc62650de3191ce1611dad3dc83
2349 F20101203_AACEJA jennings_w_Page_039.txt
1b3c1eebbbfc35081af937715f9544a4
90c070e87bd4fc2b9b7ff16e2015c4287a03783f
F20101203_AACFMD jennings_w_Page_051.tif
1bad277f90b087797f844400d071a698
ec699d39b828dc270d5a9590122b454d33998621
25768 F20101203_AACFLP jennings_w_Page_066.QC.jpg
2c69e8504c0d2d21f3ab05bc646e039b
e9a8d068955539176d3dab6e613a5846dc44fbfb
112433 F20101203_AACEIN jennings_w_Page_018.jp2
97d5ddc852e2999d0115c2d7496d7851
0c93b5e92e82e6ea896cdef9684a46e1f72abe61
11645 F20101203_AACEHY jennings_w_Page_008.QC.jpg
3c3c8e58e8adf161e0fa52c98ccae3fa
f142dee233666105c9aa27bc986142e90a93e8c8
2080 F20101203_AACEJB jennings_w_Page_029.txt
173d7474fb54edb026271498c39dcd6a
ee1c165f26e9b7585dabfc2822dfa3cd3e43a29c
44614 F20101203_AACFME jennings_w_Page_132.jp2
6e0f91c3811c9f17e6350ccdff0f76e7
0bdc8b7c0f03171e4c363875c46eccf5ea73ff3c
94488 F20101203_AACEHZ jennings_w_Page_006.jpg
813daf0055708b18e72cf71b46fdaf69
0feddb08be2c077e5866d06431d2b390312e0fba
1216 F20101203_AACEJC jennings_w_Page_180.txt
5b6496b255f306052e0280a2e8a6300e
c290ce3799467b72b29fac9e261a7521ac0a73e1
2155 F20101203_AACFMF jennings_w_Page_055.txt
f131a70febfbe5a3d0c6bc1c70b55276
5a5466db9e95a9353cffc5d8c8a41fd5a4df3089
2507 F20101203_AACFLQ jennings_w_Page_081.txt
2f2875a72379752e202c6e02ca6de8ee
f1a8583c68e9b7fa1e892972e0657c33be31f270
2111 F20101203_AACEIO jennings_w_Page_048.txt
15c1bca65afd60d92967c737c39d5db3
b61af5bcda97145a5c66b46e41f98e6c90ec01b1
F20101203_AACEJD jennings_w_Page_101.tif
0126474231fcf0b8766bde4128bb49e8
d008cb5b2573a63e1fbad56d2a63a896e0dc3faf
80635 F20101203_AACFMG jennings_w_Page_005.pro
2a8d11898f7056cfb5f4e7ebf241fd16
aa85cd867b10f996be778f2ee85116aa66c35f63
74893 F20101203_AACFLR jennings_w_Page_085.jpg
af68eefe202bc032c0850fa6554ef889
2ede2b9383dd7bf152c19a54b28cc70cd388bc26
F20101203_AACEIP jennings_w_Page_164.tif
609f3779ade224eebcaa28fa6d82b70c
540e48fe02c74306dba4a0294fdb2502273e59d8
55103 F20101203_AACEJE jennings_w_Page_014.pro
26ec8ec52b4e668b570b3de97f1d3438
f4c5708fa17671eb63946b0da3d8b094a2d44fef
51874 F20101203_AACFMH jennings_w_Page_045.pro
3059727b639db4f51141043a76a07e65
5292657b0150edfff2be80e0d23d11923cb63791
768 F20101203_AACFLS jennings_w_Page_157.txt
ee7219634d3fdb918d1491f0681cdfe6
fd53aa6f983014424af2e6dbd973b804eaccfbb7
135665 F20101203_AACEIQ jennings_w_Page_116.jp2
9e013cb7b83b17dee755b6d5e6eaa81c
4e52066e3d69364e5e33103e8b68c5d901425499
24984 F20101203_AACEJF jennings_w_Page_146.QC.jpg
676a03c921d1f6bc0c4fc39094876f3f
fafb032c645c846be30bdda7ec94aac49db0b683
28585 F20101203_AACFMI jennings_w_Page_133.pro
81fbc63410865efe6480d61750ee9572
22a24c8efc0da393b5dc838b4183bfa661a21fd7
86671 F20101203_AACFLT jennings_w_Page_117.jpg
d0baa31e76808a1c83de34e2906cda48
e78f183253fdb05dbe6e4edc4120e9d8fd002fe9
120496 F20101203_AACEIR jennings_w_Page_049.jp2
fe6a785ab3fd9fbbe3dc40271c423618
c01d3e2598e3ade29a20db5cfe8efba779a4a6c9
6957 F20101203_AACEJG jennings_w_Page_030thm.jpg
21f94b56a3e5b1c5a96234d5f4b5f628
fa0bd3063a6798c5998871316a39fb4b699d0c8a
120718 F20101203_AACFMJ jennings_w_Page_014.jp2
8ac539e393f8c9c12e0d4654d51af135
d8e57d2039dc00b6eae923bf1582344166c8a2a0
25097 F20101203_AACFLU jennings_w_Page_098.QC.jpg
6415c7a068bc18e667554f2d58bba8cc
2cdf6d51658459bcf90ff61ea9e11343ef9b4884
F20101203_AACEIS jennings_w_Page_104.tif
ca0f1e649665176fde9841c1d775fbec
854402814bd36fde11c2d3bd4c12d00f2d112924
6908 F20101203_AACEJH jennings_w_Page_179thm.jpg
34759c1c127addc0fe9619d2dd860572
f5b957221520829ee51eaf764111141e5b65bab7
27011 F20101203_AACFMK jennings_w_Page_007.QC.jpg
24df7b499356bc9610ef23dc7188815b
dbd83a0ef71ff67fc074ed6f53c3d2135092aabe
52501 F20101203_AACFLV jennings_w_Page_084.pro
8bd2123bd3a88a19949a1968e0fe3592
ec88feea99a70160bc3970202853f496cf7d8b10
7047 F20101203_AACEIT jennings_w_Page_153thm.jpg
02d3e6e893e9b574ecfd27d75d6eb356
724a9c68252578bb2510cec1eb02965e21df41da
27062 F20101203_AACEJI jennings_w_Page_116.QC.jpg
257c132e6534309b864913fca87af0fb
4d77bddb84324278a701ca858057cb503ec8b5a1
28093 F20101203_AACFML jennings_w_Page_174.QC.jpg
16fe6d0bbbf9886f3f6947104bc9bdcc
6dbbd0e8cb31fb1fb3919745accb5b2e146b2dee
57625 F20101203_AACFLW jennings_w_Page_122.pro
ebbb443031d0bfedf13dc606541ecae2
e752409b3d80710cc436456d666fad4e53b4ec11
68801 F20101203_AACEIU jennings_w_Page_180.jp2
6c992e429b46391a78b3e974a6365b4c
4a5801d16afd80782d118c106d080221c278a63d
486 F20101203_AACEJJ jennings_w_Page_001.txt
b55b41a8e467976ffee0346ce740a934
8687f093cf8f96d3a518ae0406ddb8d20ddd41c7
6814 F20101203_AACFNA jennings_w_Page_045thm.jpg
ef1caf3f6f63506ea77a21228cbd3585
e6fe1bb728d000539288ba94b881e5452f3fd05a
58623 F20101203_AACFMM jennings_w_Page_172.pro
287695e6d7c58bcac73f5edb9b36981c
586dca1890791f2cbd40b7acaea5488831f520a9
24490 F20101203_AACFLX jennings_w_Page_038.QC.jpg
af815232e436564562e79d01aa2d6417
3ab3ee4f05b5f5607370541a14723c1a55eebce6
2164 F20101203_AACEIV jennings_w_Page_018.txt
a19d515bf8fd22553d7f33e30b98ae27
36e2fc65e4d30577d389c7311ff4c56ffa8a7c69
F20101203_AACEJK jennings_w_Page_111.tif
8b8b2179ea0e9e5104fcbb072c537e7f
0b1aac9406232ba133b72bb07359b4b114a87aff
60669 F20101203_AACFNB jennings_w_Page_179.pro
2a2f9db1af7d82c8aada24ecea34c7d9
e9b4f18823b8fb856e7143f0ed65c442960904ae
F20101203_AACFMN jennings_w_Page_018.tif
75c993f24e24811dde31d9fa6ac20a4a
3487a3895da528434c2e8e2fb834788337c1a97d
52202 F20101203_AACFLY jennings_w_Page_019.pro
fcdb0242bd703662e2f43e0bbcd9775b
7b8e09e67395b9d3c527dae3271d8e804c789c4b
116577 F20101203_AACEIW jennings_w_Page_056.jp2
d81eff16ae7bdc7239d4a41154a05ef3
5867b503b908c8eb9afe601b33676f9ec6872e32
54888 F20101203_AACEJL jennings_w_Page_148.pro
c74766108038bdcd33535a83d229a2fc
db274e9ae7d3e65bf7b16eeabe667dd0995e1a62
50780 F20101203_AACFNC jennings_w_Page_092.pro
defad5372cdafec50a9d1d074f0d7d83
0b085e4e3f8643bf6b7775c905caa55e71fc5d75
2879 F20101203_AACFMO jennings_w_Page_007.txt
abdaec9ba940b2c67dbc4f4e6b84062c
72f86dcc0046c0b317a2336c2e6a95f300ddc31d
27966 F20101203_AACFLZ jennings_w_Page_164.QC.jpg
9ab2746645927da44a3514212c05ef9d
d108142f6fbd5562bc00591a9ed676ebf42e99c8
2233 F20101203_AACEIX jennings_w_Page_115.txt
0e66d00c8e37ff03ace9b8b0331113ec
beed7baa7cdd0b578da9e1af38232b688c0ddd18
7421 F20101203_AACEKA jennings_w_Page_080thm.jpg
d28a81739b13c87dc9afa8cf4041c341
2a2c50d746dc5cce7ab93644a3c5ba744428cbef
F20101203_AACEJM jennings_w_Page_072.tif
bcfc84a5c412eb2a14ae1664cfb5b7af
19266b7b6947799bd2d14b5383dd98a43bf0da7f
2239 F20101203_AACFND jennings_w_Page_147.txt
dbb8e1c21981cdd37116e94e8a19b668
ab2b2bda0272bd5490a8589aee537b9b93962351
76543 F20101203_AACFMP jennings_w_Page_104.jpg
e316bd285912431500ad47ff3c57a580
c89d1eaac436039a2256227b91e7f6a0ab95d3e6
54764 F20101203_AACEIY jennings_w_Page_049.pro
db35da6a6c55dbed55b2a2e3c2d9071d
26d9aebb73e1db55983c9d2de49ee5d23c92ae3f
7169 F20101203_AACEKB jennings_w_Page_116thm.jpg
1709456bc2552ee6b58e9f9348328674
02b5196e93d5481009e65a191b552674425e5cd8
23623 F20101203_AACEJN jennings_w_Page_095.QC.jpg
85efd099df6d82328507a7fc5312b259
155cdc6da4046ef64bdc4cbbd2250b556d560ef2
F20101203_AACFNE jennings_w_Page_076.tif
7537d1e117bb87d98b4d8ef9e7ba667f
d73200c184ee80781d88dfee5bed355fc8fed0f5
118869 F20101203_AACFMQ jennings_w_Page_150.jp2
2bc0dc6854a93731fd97c8f2fbe6bfc0
8edd92d1e187f5adb9f81d61fe9ae94bc07d35bb
72687 F20101203_AACEIZ jennings_w_Page_024.jpg
1d927a5f9f44528e6ef07ab9619f80ec
69924ccc6a38da4655cb3d001c879b3174f2d647
50641 F20101203_AACEKC jennings_w_Page_156.pro
8332155646dae435b4de3ae8441c20fb
c1e2ff88ddb63c8b044b9c936714801caa98f24a
1399 F20101203_AACEJO jennings_w_Page_002thm.jpg
2a10de28e209354061b0edb3264e7a37
fdb18160748be790939fe768a6914c73155e65c4
F20101203_AACFNF jennings_w_Page_102.tif
f748710bfd99d95724d48a38e3b3c351
c1c244e12bcb37c37efeda4400bd752fae799830
25334 F20101203_AACEKD jennings_w_Page_141.jpg
64e475e7d72cf29e3c529b94395711f2
1cbb3a6c6801519d8aeede45d1e367d6d6d4764e
50134 F20101203_AACFNG jennings_w_Page_155.pro
8df4c4719ca683ec2f8dea78aa7104e1
0dc6626c45f0a2052fd5e2ab6edb028b816ca848
62076 F20101203_AACFMR jennings_w_Page_169.pro
4715856e1607bf51aa280371fcba61da
f6f32bc47e9922db9c499434b51bff3f059fc30b
75560 F20101203_AACEKE jennings_w_Page_113.jpg
b4a474c052e59bf430f96413819a23ac
8b2bf1485f1bf59ca6547d076f232e5d7a5e55b7
F20101203_AACEJP jennings_w_Page_120.tif
3f30dc243f6607b8f13100b535cfba40
8e8fd6facac29d5e2f0d9ee1e4c3857f5fcc9389
24128 F20101203_AACFNH jennings_w_Page_079.QC.jpg
983f22b16c44ba67566ee377d638b21c
bc18552241ee272aea76f0d92486caeed388235c
75042 F20101203_AACFMS jennings_w_Page_018.jpg
97a7cb40f36269f7acf32a48e066cad5
16f664019a8839bf980f2f9ba6842f2ff671c1bf
72140 F20101203_AACEKF jennings_w_Page_165.jpg
33d0196a832af4bcb6770ece0c191b6e
0595a1261829c0a88ece844507e0879cfbf4c1b1
109513 F20101203_AACEJQ jennings_w_Page_099.jp2
a61fc589771af1ef0551df060fa1d6c1
ba81a8bc35fc005980716ee989f78f5749cf3b3c
27231 F20101203_AACFNI jennings_w_Page_163.QC.jpg
797007c6956c328b7e3c41c2eefa2b84
8cdf6bc807ba677774ef8ac38ac3cff788b878d3
6938 F20101203_AACFMT jennings_w_Page_007thm.jpg
8020aff18ea52c2bda68e5f314ef6105
6875614f34b807f0fc77e2a1f924686720a23542
126264 F20101203_AACEKG jennings_w_Page_081.jp2
e960441ca25a2fbf4123c032d1271b96
3743493c3c38da745f75d7ffbc48236b70334197
1584 F20101203_AACEJR jennings_w_Page_127.txt
39881b9b24f9cae23770f4ca4785d0bb
96059de7ec51964bc8dae218c9a3892ad639ea49
2207 F20101203_AACFNJ jennings_w_Page_150.txt
1448db212fe1a3ccfe9e9fe9652fd2a8
1d64600c96a243940b3aeed1db88856728a90557
2040 F20101203_AACFMU jennings_w_Page_031.txt
307e7f9c78d7c704afa3e327912c5656
dee56dbb73786465965481692657c126f4be2d2d
61679 F20101203_AACEKH jennings_w_Page_177.pro
42a42e808e7e31eabdff117e607b060b
c5322cbb92f17f52fedfac91f44888e8e4b525c7
6904 F20101203_AACEJS jennings_w_Page_074thm.jpg
caa9d0331151d1e4687362cf54edb287
3b10cf2ded7831c3f18569d0e647fe554ebc33ac
2404 F20101203_AACFNK jennings_w_Page_172.txt
981d9b2c204fa9465b33eb1ef780030b
91a62e40be977f0a19533438e5030c11433aaa78
F20101203_AACFMV jennings_w_Page_021.tif
f80bcddaba9d2580aabcf8b8723ff1e8
457626bfe5618062f7746fb58517e87595dc83cd
25796 F20101203_AACEKI jennings_w_Page_176.QC.jpg
4dd0c237576dff45fc889e9b3796ebb3
8cb4ed6579cc28bdb668cec9d316438d973f6e23
87252 F20101203_AACEJT jennings_w_Page_072.jpg
e61e4133c942f8f6ab59f12f624e5f3b
f165ddffb370894d45d9a807658689191484be94
6803 F20101203_AACFNL jennings_w_Page_149thm.jpg
e829de4cdacc9825ed18bf3055bfdc95
77f34e4d17fe8d846013fdea7f39985491e3a87c
1951 F20101203_AACFMW jennings_w_Page_077.txt
9aea27b8e0cba10190b63b98552eb370
0b2aab93d0414c166c039910acd17897d24e6842
53177 F20101203_AACEKJ jennings_w_Page_093.pro
b90b212af82b070f4703590b7f254a51
51e626e82611d266f754829a22fe80a5e48985ac
136117 F20101203_AACEJU jennings_w_Page_075.jp2
14ed8302357864c7e3473b1dff948f3e
1c2984f636b34c4978d443a13e13a9015ffaffd4
54626 F20101203_AACFOA jennings_w_Page_051.pro
9e45faed4648a5850752ac7ce297eb50
576c89ffe265efbbee96eebe504c3fb28a84021d
42210 F20101203_AACFNM jennings_w_Page_069.jp2
0074acfc44b28987152bf483fa9b14a9
d8e691190625d14f1fef9ed39fc0c462ff1b2101
6822 F20101203_AACFMX jennings_w_Page_101thm.jpg
71b71dbe7d53c53ed0543dfe66b705d7
c248450197192213efd6a48a1ee275e626d3f33c
75990 F20101203_AACEKK jennings_w_Page_110.jpg
854101700280c6686a202b0d56db8e6c
162a9a14b9a4617a1f7276f318d03e1c1b272efc
58584 F20101203_AACEJV jennings_w_Page_068.pro
e74a1b4b4fec4c78b093bbb8eb638277
bd672e482941cc1afd3cb90d8634eeae61b40c0c
2468 F20101203_AACFOB jennings_w_Page_179.txt
c15ff0d0db2439864deb93d8e9deec50
c02b688f89abef394c63bac0ecde926b71634b61
F20101203_AACFNN jennings_w_Page_048.tif
ebb3d5abba9398fbf106934c294f7000
7af542a1bf36e770382160063b65c43749ff55b3
122124 F20101203_AACFMY jennings_w_Page_057.jp2
cd3d7aa59efe84024a58971f17bbd5d0
3c663e479f55f0eed13639ab4278e6b81d1ed5e1
24621 F20101203_AACEKL jennings_w_Page_140.jpg
ed050e9ef68890583273b5177bd33741
b278ff1b3b4c61854650ad0a25d58b93dda4a3be
1772 F20101203_AACEJW jennings_w_Page_129.txt
cd13481ec4eca4603d702ba54ad0d4fd
004e915f65683f77557e6956d3df5479c55eb457
F20101203_AACFOC jennings_w_Page_055.tif
a38a15e6089587eed4727c37a2b29bba
30bfd9d5faf6e13e6297e50b4a13a619bf638540
F20101203_AACFNO jennings_w_Page_056.tif
49bb5e7665ca206053604823578cc4b4
6d915dadf2a3f682d0410e7af5f846de9f2d4a7a
4429 F20101203_AACFMZ jennings_w_Page_005thm.jpg
657103a9ec0cbc1ee76c5ec6bcefd9d5
a043f98927c06c00ffaf0466c0a97897a5919a96
53199 F20101203_AACELA jennings_w_Page_059.pro
d5d146429b4c572873ccf3b82d3ae1af
3b7622432c01abc9380d87a4a137e37cf030b261
116452 F20101203_AACEKM jennings_w_Page_055.jp2
df606180dbb643f31d0bd5dbc3071cd4
7b659c21c50db1609ca31080423748ea51c3667c
F20101203_AACEJX jennings_w_Page_124.tif
3890be24bec293228926514f7b2ff446
c311c4b9a5263d9c44c459cc9cd9505c8c51f2d5
2116 F20101203_AACFOD jennings_w_Page_146.txt
4b6dbebc08c33dfd2450676f63039e4c
495115a5ed6ffb18dca833448fdc341f50da99c7
120505 F20101203_AACFNP jennings_w_Page_154.jp2
769b3f43b88bfd73bf0c2a713ed4c7a2
40ee71a472ab175997e8b0bf67b0b4c764192122
74230 F20101203_AACELB jennings_w_Page_097.jpg
f960ba06374ed0b31f91622a36571b1f
8c3de6ee904f139233b3efe2e51590946db7dc25
80167 F20101203_AACEKN jennings_w_Page_016.jpg
47c39e5b704ae21df9a1175d5e9009a5
58deae8509d0796428cabac813c71c5d53fa7457
50235 F20101203_AACEJY jennings_w_Page_090.pro
0b3a0173268a75f0d9afebcd10a76c63
6fe9edf4a942bb1d7e94071c7fa2ad38288cf011
135130 F20101203_AACFOE jennings_w_Page_173.jp2
df64554b24ce5cf58b8c46714a06b323
4d1fd9f1820fe70274ab773e6e6f2bb13a463a3a
26263 F20101203_AACFNQ jennings_w_Page_178.QC.jpg
4ce9564406e9d87db227e23c98fd7960
0132a0a8a1f5bc188e7b20819cb507229b93155f
74310 F20101203_AACELC jennings_w_Page_144.jpg
fcac960bf4070ddaa2f6cd4e048983ed
7a3cb5f98a3de985cc01ec6c21173c6a83a21089
63539 F20101203_AACEKO jennings_w_Page_160.pro
8316251cbfc9d980351ab697b5285178
19254e1c80ae2f3f475ed5545303d8635a2b4049
F20101203_AACEJZ jennings_w_Page_153.tif
9f394e3f0f5b4ed80bf34bdcef5657c4
02223468c826d7abfea3dd83b0b1aab89f279781
137219 F20101203_AACFOF jennings_w_Page_175.jp2
511c7bc92f496710fc715c3d83cec904
e4dd7080c18ce2da8b875cf0b634ec38fe791af7
2616 F20101203_AACFNR jennings_w_Page_140thm.jpg
e0592e62c1d4e07a2791e96a47c36aa3
676630918477ec2c03ce82132915de046b93f752
2147 F20101203_AACELD jennings_w_Page_056.txt
f0e469398e0e996d35855dcfe43c08c2
d7daf3e72b275b6e3525b9f19dbdf86092727777
10622 F20101203_AACEKP jennings_w_Page_157.QC.jpg
4985596201b3130fff57693cb909a00d
f11269e86b68a2d3b33b2ed7a4883ab093dd9a63
2143 F20101203_AACFOG jennings_w_Page_064.txt
b9c77d66c82a4280f84c0a0c161a7720
1becf0072a2851fb3146704be80fc816db9c53d5
6970 F20101203_AACELE jennings_w_Page_152thm.jpg
d958d21809fc04db67639e16d972d1f3
8e1c5aa71a5d6b583e823b28d9a6f385100c9fa7
373 F20101203_AACFOH jennings_w_Page_142.txt
80b9cd231909fe9b1e4a4586007a7d9a
0a355cde1e7880d4d3a72e63592515ab105bda83
73551 F20101203_AACFNS jennings_w_Page_090.jpg
b6aad037108100ac2d027c2299af0ceb
24e4752b23d4c6d77b25f6485f89b17aa5e56895
2009 F20101203_AACELF jennings_w_Page_092.txt
dea67ca6180ff1a062004814f11468cb
8beb27ff3c21c5422c870c64b206c59fc7ffd82a
3229 F20101203_AACEKQ jennings_w_Page_157thm.jpg
78ddc2257d0d750e2fe3f1036279cded
f06e2c48b92e6fc1e97304e33958f58cf8024a4c
51726 F20101203_AACFOI jennings_w_Page_104.pro
fb4c46c2a6510b8321fe333156965e18
b19653e075eae707a4ae40076079ae2866f33eb4
113296 F20101203_AACFNT jennings_w_Page_028.jp2
6b7c23145ba231a43f7f299144e36000
6e9c603232a980312eae9a36105366566a64e347
2157 F20101203_AACELG jennings_w_Page_101.txt
51b31685a4c01950361209871dd0f76d
6e1e343c6c202fccf39e32b7018050eba1de9457
14760 F20101203_AACEKR jennings_w_Page_127.QC.jpg
f372c933ded0625781c469a81ca6b2e0
80a49dd8a193b5207dbe57c52e233eba90dbc9b8
6969 F20101203_AACFOJ jennings_w_Page_143.pro
c2433b815b386b3abb4258cc917e6807
fb450c3d31782d732069aa668f22ec35218959c5
24594 F20101203_AACFNU jennings_w_Page_144.QC.jpg
f095e1f619af94824dd1d8fc98a22556
a7f88370b76e086bcb35b77df63590e201650331
F20101203_AACELH jennings_w_Page_160.tif
7f60e73b484a9778d66a67af4848150b
86123ce90644d47173963539ac46167862ea683a
6782 F20101203_AACEKS jennings_w_Page_108thm.jpg
d7c8a489c625628b53ad042c7195d68a
13f2571d326c55093c50a0b441c25759faf706d1
F20101203_AACFOK jennings_w_Page_032.tif
3d83808696d7766c55adbbfca8ca13d1
d9635921799ca6bca4ed5fefbfa573fb8b13a70d
2103 F20101203_AACFNV jennings_w_Page_105.txt
13c1c7fc4499b0b11125b8f9494eae63
4d2dc96eac1d2a9bbd2eb8d628a6d1e21e3b10aa
73812 F20101203_AACELI jennings_w_Page_118.jpg
20b0e6e9561072b38bb965d61944e5ee
94ba9ab8abeb11517f7c6eeb1ffab06b08923710
3512 F20101203_AACEKT jennings_w_Page_135thm.jpg
22c4b2ed22c99a6bdb81994b8e271f2b
f9ac03db58631f9e5b954e73d0a2f4ae15ec16de
92284 F20101203_AACFOL jennings_w_Page_161.jpg
8b636ea96aa331e239cebcda814e179f
831020622bcaf77b33f89db8ed8c6602e2ecf2b9
85068 F20101203_AACFNW jennings_w_Page_173.jpg
b035e8cc8f04d45101b46ddb71c0ea7f
48235cc230df2dbc33d03396d37a241570f9f118
F20101203_AACELJ jennings_w_Page_098.tif
c11156ef678af7865b984c58e3a7c014
38ea47c363472e427b4bdadf1c06818a9a338100
21532 F20101203_AACEKU jennings_w_Page_009.QC.jpg
a8db61eb77478b067cec99814403d13d
8d18daab6b2814c6923fd94a51355f7ea4a2a32f
F20101203_AACFPA jennings_w_Page_103.tif
f044c5adf8d7d241242a38055a5685af
7951b687dc276f5e4bf97a7adbe50f01e275a659
115909 F20101203_AACFOM jennings_w_Page_153.jp2
a86126406aa1ac4372fce055275ffeb0
824f4ccc4880c09f1ee857b4bc085c958ea4045e
F20101203_AACFNX jennings_w_Page_042.tif
379d2db33028185d1d7a2d12c3667d6e
0fa1948e9ac0edd4ffea6d601b7da7f561fe57a8
55271 F20101203_AACEKV jennings_w_Page_109.pro
db1853286174e2e1145ee39c231a213f
0fade9b694f11a283366f7470c9a1237e06145e3
124473 F20101203_AACELK jennings_w_Page_115.jp2
16f7f49825d67bd58a5ce80a90776b6e
7776f3609bc2040cd551bd25e8223aa1e1672651
F20101203_AACFPB jennings_w_Page_023.tif
dc4f1e42579ba30181ee8374a58289fd
3d4779a4437c1b117c87daedf723db47202324fe
78079 F20101203_AACFON jennings_w_Page_066.jpg
56438ae06429ef1b2835c85221199894
8f670afacfd66c4389a16957a8e8d8c581128b6c
78794 F20101203_AACFNY jennings_w_Page_154.jpg
c060b7b0e1ba5d272cb7f62f6d9a091c
ab3971889a015decf50c7a5fa77e66945032aeb3
54075 F20101203_AACEKW jennings_w_Page_100.pro
5792d70ee24abf1ae0b08167d28ce33a
e66f9270fc2c839305e2a92a1ecc61025ce1be2f
F20101203_AACELL jennings_w_Page_044.tif
ae2ecc50691b6da7a7ea8fdb1a28aa42
f01a0fbe4b9373c0774fca90b833ae120acd7a3e
2064 F20101203_AACFPC jennings_w_Page_087.txt
99e570887cdaf3329c047054be500c3b
6d32ff5a8df1641949e7d26b3b0421a8ed59d0ac
26178 F20101203_AACFOO jennings_w_Page_125.jp2
2c3640c8bd5a462f89e3412f5e7ac145
ae482ef5e9fd3a5e596cd6290bc77ffb60a9913e
119360 F20101203_AACFNZ jennings_w_Page_165.jp2
135dc7b675698cb575651dbec43a4f6f
df74adf26585cb848ae5c529bc4133409b4ba587
2794 F20101203_AACEKX jennings_w_Page_164.txt
3abf5947114a9afff10167c31905b754
ae65e98fe8c518ce2fd961a0261a51db71e1c03c
26305 F20101203_AACEMA jennings_w_Page_124.QC.jpg
475ee98b15da87be92579d61c8309a48
05ecdfd73e9c920e930adf20b4d4cda35b47739d
2163 F20101203_AACELM jennings_w_Page_045.txt
29a669cc9863bea02b67f8777018acee
f81cad57af44bd4985971ed29bd5039bfa6c023c
8856 F20101203_AACFPD jennings_w_Page_138.pro
0dcdc933f614a767018aa620bdfe646c
a77b1cd94fb1153efebc664ad23ac6044cc93fd1
77806 F20101203_AACFOP jennings_w_Page_022.jpg
71402af535a42e372b67e75d647a0d64
c84e35ddf90500e1672a35e1ba91ced582632e7c
109364 F20101203_AACEKY jennings_w_Page_090.jp2
ae53ac636ad9a14131e0c18732d5cc9e
6b67af6e9dc699f94cd94212188f45ff95d55bec
F20101203_AACEMB jennings_w_Page_082.txt
18fb9ee6ddacc908f24afbfd2e70d57f
817a32946a6451e649298a1f97c75fe21005f89f
2739 F20101203_AACELN jennings_w_Page_161.txt
fc389efc3454e8fc4c93584fd406143b
1322846a459179a24c3182c9b46ed698ed4b744c
25893 F20101203_AACFPE jennings_w_Page_138.jpg
d83d4e0b98945dfb5c68851e4089d8ee
9461a3ef6ad3125aa5abcbfef16a95937427545e
46150 F20101203_AACFOQ jennings_w_Page_103.pro
605e68dd25c4a958f5ad61605da3b18b
442db16d051a2213aad3c1cac63921e0dc27830d
47974 F20101203_AACEKZ jennings_w_Page_015.pro
eef914f43102c43aaba940c62a97401f
5422e7820516de9d0fd7058cff7c7187ba9e5469
132628 F20101203_AACEMC jennings_w_Page_159.jp2
bd01fb00e2c17c7f0f31a0718d8f1caf
c0d95438d6ccb49dcbad89c4c64ec176a7ad7645
F20101203_AACELO jennings_w_Page_105.tif
994e6de39d618aeea2769861f038a7d7
e8a15a56df8fb98695196fe2ad6f187ff7d096a5
2152 F20101203_AACFPF jennings_w_Page_121.txt
9cdf68e794a8b396245bf723e34436a1
caa2dd067ac5e6511d2bab55ce2bfd32b4941882
75180 F20101203_AACFOR jennings_w_Page_091.jpg
096b2d56200fd290182da7043859d340
3ee68158ae5a2bd4fa0857d201c4f0ffd7174ddf
22675 F20101203_AACEMD jennings_w_Page_015.QC.jpg
d33fe64e9403877646bcc91a45ed53df
9915b3a0d92cbf0cbcc8c80950e9cd281c8560b9
53611 F20101203_AACELP jennings_w_Page_048.pro
1e8e1575c35d5ded722ce2e384e3db4e
42350dda0159e422067c954bc9bdb980d05b8353
F20101203_AACFPG jennings_w_Page_156.tif
f9189adc8a360c9f57def7827d22ec26
d563df87fefdbd855bf7afef80ef8f7d672e43ae
F20101203_AACFOS jennings_w_Page_040.tif
5042ac36f1d71212fa769fd952b5bb79
5c3867871165c90b936ad27ec1a10ab4e9d4c046
6976 F20101203_AACEME jennings_w_Page_150thm.jpg
ddd1e4bb1b467f419a0d766c1353e238
158bd59df32aa137ae1a7cdabbcd529ab362a727
81041 F20101203_AACELQ jennings_w_Page_151.jpg
c8b02b5e9dae5b601f58b9c6524ab909
50e15ee60bbcbbc7101c722a0d11089c84eab1de
2653 F20101203_AACFPH jennings_w_Page_170.txt
697aa103bc7263ec4d4837075afb2a34
1f5e151c11da94b9a6bc0c4560bcb45ccd0f09e2
15398 F20101203_AACEMF jennings_w_Page_129.QC.jpg
5ed954391ae8b7e8ed05b447438cb715
6ba84fafb3516e2dff578c89e7f86a11bff989bd
25997 F20101203_AACFPI jennings_w_Page_051.QC.jpg
8fffc37ef0d5ea1c1b3768139354b3f9
2a4e2807ce2a18f584517b7b652050f25d9cd152
54671 F20101203_AACFOT jennings_w_Page_055.pro
09b050bb7af85faab1e36e1bb819e338
bb4cb42f731b42b5953b5dfbe58f81182e3a757c
63189 F20101203_AACEMG jennings_w_Page_168.pro
6fefae82c1f73c363f935b1f4ead908b
8676a68cde61e332f611d8be26f0bfc63a1a0ddc
617849 F20101203_AACELR jennings_w_Page_141.jp2
0346f71f7c202b22ca47a50ad6448fcc
30b7764b615289c7160c829b31e74b801a47f3b1
25271604 F20101203_AACFPJ jennings_w_Page_007.tif
da85ff19baec3dec0a157febe23ffa0f
247fcbf2956e80f4d99deca7808fd7c53fb649ed
3927 F20101203_AACFOU jennings_w_Page_010thm.jpg
1268a5c9c97680936a3bb852b3214598
8958cad7e793b615c0cd38c7f0d7b5673c16595b
31504 F20101203_AACEMH jennings_w_Page_135.pro
a67158402d4a352e22b92848e4b73cd7
3455f929add94881d1c3598d1eec77dc435f40af
F20101203_AACELS jennings_w_Page_162.tif
3d25e5bc339a8762e685ad0f02373930
eaa4f7c29a6635ef96bd5e5e6c57e069f1517516
6722 F20101203_AACFPK jennings_w_Page_019thm.jpg
fed8d6258fa1394a83171e5fd8a33e0c
5a58ef2d01393051e6ec6ba5f3d3e4eb83ef9ae8
366 F20101203_AACFOV jennings_w_Page_143.txt
a1fda55478f62f0a026e6ec456136fa6
d1f91190a39a90f6db707ab282b8351cc5610795
78265 F20101203_AACEMI jennings_w_Page_027.jpg
f137d04bdd0fdcd56f777f7d6d6bb305
e629840a151cd16544abf1e252bd6c9153bac732
F20101203_AACELT jennings_w_Page_126.tif
7004bfac8eb0f47dfbe7f6419d43a72e
942cb6ce03e19a599fb53e1d5cdd0b953dfc2fe7
57108 F20101203_AACFPL jennings_w_Page_115.pro
e544b07d628033ad8f4098be0635cd6d
2a46785b2c5719242dc5e0989ea94ce04eb2387d
26587 F20101203_AACFOW jennings_w_Page_057.QC.jpg
4e09d15bc7c69a81839839f546ea5771
fb8575339b979b56ba09a733d99c55d3a7e0df0c
2238 F20101203_AACEMJ jennings_w_Page_023.txt
25ccf4fbd9adb323a2843d8e8c6df387
29fcf9e47d7726b3d665d8e46ff180b25ad7f46f
F20101203_AACELU jennings_w_Page_062.tif
261ce9df6fb893f05b4ff215a6fac1fa
de07b2add451f02d70b15354a2ace9e2832d08f1
120830 F20101203_AACFQA jennings_w_Page_030.jp2
99ca88c3d4a971dd893e8e0385d01949
fead336434048791d951d7d0d6006b49d00349c9
7118 F20101203_AACFPM jennings_w_Page_052thm.jpg
765968a3c5ae396d4dea3e32e5f8f586
f3c0daa6b1279773e49f5e69f437b87dfef2161e
76221 F20101203_AACFOX jennings_w_Page_062.jpg
c1e211bce2a864268fd7b41aea8015ea
4416d27f48b97045ddfef185703822e4b3b8836e
24367 F20101203_AACEMK jennings_w_Page_068.QC.jpg
0671ea032cbbc4b92a8c1be59e073009
542d68823a7edc18e6daaa145a189a3459a62d34
F20101203_AACELV jennings_w_Page_037.tif
ec3bd934b7f79704959d9e3022f00490
d8c2eae621dc641f266d6555f5f918074b3f5cb4
79430 F20101203_AACFQB jennings_w_Page_037.jpg
1c7506e2fa6398d0c5d36f5aa3ebbefa
a4d47b13f9363838b1b01c92ebaed330c73cde43
77529 F20101203_AACFPN jennings_w_Page_100.jpg
1af7cd68be7605107ce2821df057760e
1f1d602af65a65378b2eee39d882858df51906f5
F20101203_AACFOY jennings_w_Page_077.tif
d5f48024062fd27497142a54db89f28b
2309bd871ba69f6634c10bf1c9ec75a4987a0459
25073 F20101203_AACEML jennings_w_Page_172.QC.jpg
bd0d3a67215d0672faae77c3aa904026
b593073e7ba66e5187597070f4d679cf74c872db
4038 F20101203_AACELW jennings_w_Page_131thm.jpg
195f6280affd80bc4d6fb57bbf35121f
44297f49f403c946ebff0c37a7c5a36f0a2b663b
78341 F20101203_AACFQC jennings_w_Page_040.jpg
6eb5e1d0e3aac4549e27d17e49a0d54d
e37d971f2d6ccdc26c8db59c949abfda42e22a46
72456 F20101203_AACFPO jennings_w_Page_134.jp2
bb8af140f50a56e4031fa7681cf97cd7
563dddd92a79ba79af16a255b9b1f30ed0999614
5981 F20101203_AACFOZ jennings_w_Page_009thm.jpg
94f6d93cc0ced194f621a1256cc2a83e
e3390ab083e7df7b617e60dfd280cc38cd94492f
7225 F20101203_AACENA jennings_w_Page_039thm.jpg
4d24db4111d29b3612ea08f9a53ccab9
8dbcccfc2e0395397c31a07235163f79f28cb3f2
6996 F20101203_AACEMM jennings_w_Page_075thm.jpg
fa60017bebdba5a978acfd4fcd40b5c0
8f35073952d741c50d98939c50c2fb8895b424fb
68791 F20101203_AACELX jennings_w_Page_116.pro
0df272373b47cab05f5648e6195b6ef8
eda4e3e497caf48030721b15ddef8cede80e2a09
F20101203_AACFQD jennings_w_Page_020.tif
0b4d3dc6a798bd33b615476caab9f7d6
27997728f1e51dd250edee1a03ebb38496464af3
2098 F20101203_AACFPP jennings_w_Page_059.txt
508b3f5e14b0f3e49a406c0071f87d59
1e070d287327a31c4eec3c8110a6c0f4fd21d5c2
1620 F20101203_AACENB jennings_w_Page_137.txt
ee49661a36d1912ac5b036fdad6b883b
331fced7cf20688a4efa54a10d316a06a44553f6
2802 F20101203_AACEMN jennings_w_Page_126.txt
cd22433b18fe7df056869303bde9efd4
98e8b15d8368caf523017ede552895af8b99037e
26352 F20101203_AACELY jennings_w_Page_061.QC.jpg
27461e1c5cd7f898b7ab55c083e03b0a
134ecb9d1eb9ded0f745770dc4c65062d78ff225
1016 F20101203_AACFQE jennings_w_Page_010.txt
acc4885df73fb1ad9027a1adf99bf675
38a83898bcd44a61f77a9529e940b666772ef7eb
2137 F20101203_AACFPQ jennings_w_Page_149.txt
58d6957eb9e133cc779d15e1326ac78f
21fe3ea77da4fad341c0c372c363500c3703b541
242 F20101203_AACENC jennings_w_Page_069.txt
7021fd39162b5b098fa954f77ec3808c
c3ef789e10a07a1e4c225e06646319055e957488
6723 F20101203_AACEMO jennings_w_Page_078thm.jpg
ddf19eeaecb08bad3409c16cf93345f6
a7e5e55a395985d0f93c2f058784e236b5e967a3
122750 F20101203_AACELZ jennings_w_Page_151.jp2
033f0dcfb6680c07faad9f42ea14ef67
cad2e43b289f3f8a108fa784456078646f3a8970
51606 F20101203_AACFQF jennings_w_Page_031.pro
a4f406f281adae43d6f7c48e3bcf2a3a
77d5a09ad1f2981d6099a32551ac6edeb03e51a9
F20101203_AACFPR jennings_w_Page_125.tif
680039484a7af1c90db59b6b4d8407c3
d88106d547e7f6a6ac44b65fd69949b7116eec02
F20101203_AACEND jennings_w_Page_029.tif
0613d0ca1dea6349b3ab5356707304cc
b24a1767d5868c3fabf8734fa286002c64144d3f
72209 F20101203_AACEMP jennings_w_Page_077.jpg
7435e29a62a7b6a2af0820f6e294a460
dfa6816a9d86e74431f84289f6f48d6023c27d84
F20101203_AACFQG jennings_w_Page_027.tif
8278408a5cdaf72c0c0e9854365d56d3
611c44a796decf39dd3712fee6039f6636fbdd50
F20101203_AACFPS jennings_w_Page_033.tif
445905da4ef16777e2b96f695a58e957
e03aeebfebd5c7df971bbbed012496d82c496764
55604 F20101203_AACENE jennings_w_Page_123.pro
e9b37d8757e3fa3efd5324e8c60b2723
244c416c9fcd429de0ebab422544d78024b0f462
75156 F20101203_AACEMQ jennings_w_Page_074.jpg
c32b1c9c41cce9fb3cc79d6f88dc8dad
c75bccc46894d84816031c17c4729e669f135f2b
72210 F20101203_AACFQH jennings_w_Page_099.jpg
db11213e54e36ab838c85fe9df72d341
b3aefb9569e4e0a2f6f2d7091f232928766f79dc
111971 F20101203_AACFPT jennings_w_Page_031.jp2
6bc592b8fb3bda03df7866009f351834
abb367fc7a9004bcd52429b54720c44ae891b08f
129650 F20101203_AACENF jennings_w_Page_119.jp2
1afece5ade7b67723fc90a3a87c1677f
b929e2ea7b185228a5abf87ef9552f96fda1b2c3
2683 F20101203_AACEMR jennings_w_Page_181thm.jpg
aa0637ad13302a855b8042da16c53cbc
cf3d0c6a4f5c4b3f8fd576f71ccd528ad76071e2
6174 F20101203_AACFQI jennings_w_Page_103thm.jpg
f7c32c9312482aa526046f6003418913
bc9dafc3ea5fca021ca7289f48f6e9f6aff07e49
F20101203_AACENG jennings_w_Page_013.tif
46a0188c3ce8048b96704f47784d16e8
835f17da1f06c873ba78fd3f7a909a13a39d60cf
F20101203_AACFQJ jennings_w_Page_129.tif
82c4a1805bd0f74fee2dafd91af8ab8e
450657de28d30bdc64eb5fcda5c9ae9561d57ec9
75087 F20101203_AACFPU jennings_w_Page_033.jpg
4df433ed299849beb24469da37f2ed10
ad59d41b531e4fbc0c793b36336ecdb5b122fa7b
17420 F20101203_AACENH jennings_w_Page_130.QC.jpg
6fdbe1f8185e1864efb17400b53314d4
ec3b1804e7471f733ab631ddfdb557b0f41ec763
25026 F20101203_AACEMS jennings_w_Page_149.QC.jpg
1c48cda5588399cda575741d7e23edc8
cfad8908a6bfc7f294b3b098405b2641908a70da
15908 F20101203_AACFQK jennings_w_Page_126.QC.jpg
cd61294b93a94ff93d0a188d4f2fb1f4
7a62acc048879d7a277d6729823cc2819fc7dcf6
6346 F20101203_AACFPV jennings_w_Page_015thm.jpg
09cfc9d5c470a0801af81051b7a3ed99
8cd471cd4b35bc5ed549bc48246e743b47fbb1fd
15760 F20101203_AACENI jennings_w_Page_134.QC.jpg
f65add2970b3a2655277dd73835572f6
b06c0279ab7a9b9830a3ae166a8ff9e2da0e3ce0
31457 F20101203_AACEMT jennings_w_Page_131.pro
59df46075bf048af862f102a6e860ca3
ebc66a18f03bbbdc21cb3c2552985d7d18887c4e
7161 F20101203_AACFQL jennings_w_Page_072thm.jpg
a645d2289b4fee9a70f62edb6288cfa6
2556fd8910aa5753f18cdb7ced7cd0d8d1fcd973
75962 F20101203_AACFPW jennings_w_Page_146.jpg
1805c9efc197d362cc81acd4e66f537e
5c10de731273db0dd691740cb9504fb0d0bed6b1
77553 F20101203_AACENJ jennings_w_Page_148.jpg
24b457a6923071f82f842b114e2d5ac1
5a83ed1ecd92082f93bf07f81150136cb4511b47
53513 F20101203_AACEMU jennings_w_Page_065.pro
50892d7d9cccd346fceb13c8f00aae43
5c6067afc3c63fa67fd0b3a6d817c59d94e944e7
6332 F20101203_AACFRA jennings_w_Page_088thm.jpg
8c8bfdebc3e090e0f70f701b7c205329
53b6beef951938a2a633266829948480c5073441
76598 F20101203_AACFQM jennings_w_Page_149.jpg
ff8246d4e8d87a27db32b47dd8c58b89
a9b3689c6ac7b7c1cfd184c168105a44126354f5
24759 F20101203_AACFPX jennings_w_Page_085.QC.jpg
38dd52f58996d8de39fa32a3b6e78a14
8e8abb44fe1db14b1176fbab76649426017edd72
24627 F20101203_AACENK jennings_w_Page_013.QC.jpg
be63a316ac804add778765bb390a773e
32357a44058183d7f757e8520ac350770aaac2d7
F20101203_AACEMV jennings_w_Page_024.tif
166b0072ce4005c270d7a77e2608a6b8
74469b000bbb118057682e92cac8183508f8ed16
2010 F20101203_AACFRB jennings_w_Page_017.txt
0067b0e39485dd215e1f60e31a992b30
e81c6afcb35565dc5b38376cd4bf799c718a66d8
137336 F20101203_AACFQN jennings_w_Page_178.jp2
6284c92b6a0d8d6baaf080405f7597d8
b520f213015a489d05787aac8a52b16278d7b386
F20101203_AACFPY jennings_w_Page_171.tif
e36f8e63d76e944ce75f89022b0e4fc0
265e822e9cf32f62f90ea75101c4b5582856296d
6825 F20101203_AACENL jennings_w_Page_064thm.jpg
244c489bfc21195b60cd6b896f2c1c64
340532312c26e235e8e949cb3934534dfe66b35c
F20101203_AACEMW jennings_w_Page_090.tif
40822be10467d9da09badbda4aef21e4
bdca1c367b3dcb1fff14d662d50092362a4aad06
135392 F20101203_AACFRC jennings_w_Page_160.jp2
726bd2c10474a6254f9474a3f7bcf378
7f53961544312f04ad0b23c2e6fda5f32a0a4f67
54127 F20101203_AACFQO jennings_w_Page_126.jpg
f234a38dbc49570ccc5cd88547fb7abb
38d9c926b46fe55d7157770f3ad62c5c60198a53
14098 F20101203_AACFPZ jennings_w_Page_131.QC.jpg
9250ef447d22fff918a11bb4443d7faa
883a1ea09da7fdc5ac206a03b1ed3cab8fe42c2f
3257 F20101203_AACENM jennings_w_Page_005.txt
6fd363f983a627a5d7b5d3e775f8f0a0
fbe15eddd797245247856a6a38596b9df5d15a69
1637 F20101203_AACEMX jennings_w_Page_135.txt
1360b5da5f17023ed438f85ceacc7cdb
ed78985e9ac293c3bc9f3ba9af6ef4f73d085ec2
32927 F20101203_AACEOA jennings_w_Page_158.jp2
040e7fc683a572e88c073785b358a0fc
dd8de819e196f568dd0327cb1978bd875dbae653
53696 F20101203_AACFRD jennings_w_Page_071.pro
f2d5a437d2db06ea09f3c5740265820f
c0e0f9b02cc8ef27cc5354637194913f18d1e209
7316 F20101203_AACFQP jennings_w_Page_163thm.jpg
7479cfe2437ca1045a727dbfe3a717ac
cd9eb2297fe73697742ce4b98f968c0062540479
2214 F20101203_AACENN jennings_w_Page_134.txt
bca6caf96715b04b0eb75a73c062331b
7ad7fbef47d7d0a9ba6e538cfd433431a9dca788
116145 F20101203_AACEMY jennings_w_Page_110.jp2
a7852684c92ce5d281246adb02b820c8
c0af5ba5e0be3279e3a915eaffa703cc22aae477
54899 F20101203_AACEOB jennings_w_Page_101.pro
a13eb69c4c74311472da832e2d8f4e7d
fd6b2dc26b13da35165bc7f05f84422e5f7e731f
66156 F20101203_AACFRE jennings_w_Page_005.jpg
9862cb910f0d467de6014e15d45b811f
6d50b2d90b3749c802e44304ac912679474b1d59
24213 F20101203_AACFQQ jennings_w_Page_155.QC.jpg
9ab1f289ba2d441b6280c1cbabeb8275
32b14743ac685ada855f9ca5f69095dd385861a7
2077 F20101203_AACENO jennings_w_Page_118.txt
730fc0d9136cafdae9f2d42862208c14
3e213a1bf34d26f860b502b9fc706403f199df5d
2295 F20101203_AACEMZ jennings_w_Page_122.txt
a1f653b40deed4485f2491c52bf04ef1
7bd55b76136bbbaaf6166e86d6f3c44b7f28ed27
79681 F20101203_AACEOC jennings_w_Page_083.jpg
b555aa395f413e847f0ef0fa36a5f12c
84758e7c05a704922a5b1b4a4392d2c2cfaafc08
7063 F20101203_AACFRF jennings_w_Page_047thm.jpg
c87ef1fd6e0fad4b6ab8568d740bcf29
047045b30a87bd92248bbe7ebbc37ee547c09a02
6611 F20101203_AACFQR jennings_w_Page_082thm.jpg
5268d33616b7d165cba12c2add3983bd
b4d579d0859bb895cdec107066a626776027b602
F20101203_AACENP jennings_w_Page_085.tif
3b99fbeb073a169a7cc7d9a3ccbf923c
f1a7748860be667ad53bb83611b4a34af7988c1a
26244 F20101203_AACEOD jennings_w_Page_123.QC.jpg
53d9c50f53d75d80aa675e487a8a3ff5
7d48db5adb86eb7e5674a7de07ba8d1b7948814b
2522 F20101203_AACFRG jennings_w_Page_166.txt
aebeb1a7bd13e42f8f0afb2b458f23ab
8ef8999ad80105a30eea5ae7d51e6c044a530c68
27897 F20101203_AACFQS jennings_w_Page_133.jpg
f6f0a3d35412706ebe31cbe472290dce
8b6c23ea83dc4a2b47dd145c4833fc22160a032f
111638 F20101203_AACENQ jennings_w_Page_144.jp2
149f2a2375bc8ec55b439ec96fc93875
293ed41fe228d26b8cc8a93890f6f35c8ac162fe
114610 F20101203_AACEOE jennings_w_Page_094.jp2
be017783e35600cc105d3cf5cb95152b
d7b6eaced469069fbdc5b1360b778f790c414df2
1889 F20101203_AACFRH jennings_w_Page_088.txt
3e22fe3cbe44ef050451ff1a16c9c35c
788a2ac7be36c37bf10003656d140249c38bdcd8
2180 F20101203_AACFQT jennings_w_Page_041.txt
32e6ac673213336148f34b1b653ed37e
9cca58c0ef4b5b362e30cf6fe658f9f7d0060d0b
72421 F20101203_AACENR jennings_w_Page_136.jp2
7edf4345047c34ca26dfc95263fe1810
3516e5f65025ee4b54741f391855d603226a1dcf
100241 F20101203_AACEOF jennings_w_Page_103.jp2
7d7c4ca30a75372e3db98913217b6307
7810326e79fc1962d2df3d887353a4788438b3b8
27654 F20101203_AACFRI jennings_w_Page_181.jpg
c3edf1ab59b40da5f9609f4359ff607e
0d1829bb487071a5153c8385473712f7740c3587
85319 F20101203_AACFQU jennings_w_Page_073.jpg
446bceff1549c9839a9887017a5071ff
35a9871fdf54b521fde49a5c6bf571ad4ba42a26
4667 F20101203_AACENS jennings_w_Page_102thm.jpg
eda2fb61b8ef34b8856e2d73bceb9545
bc56b07b6a0bacaf5403433137cfa5ad66f2f95e
F20101203_AACEOG jennings_w_Page_172thm.jpg
df40f1008b684969cadd2fa5ab999a95
06dc463345431abc1c0e1e96170a046f8736ec08
2569 F20101203_AACFRJ jennings_w_Page_117.txt
8c9ad5c4d0948c6027bdf8506d2478a6
da7554b0e43d335282b518bb1cdcb47fdc655b77
116061 F20101203_AACEOH jennings_w_Page_146.jp2
794492c02665a6753b19110e6b1ed20b
49aefe9fe2ffd837cb7c108cdb46a4e8de9dde9e
F20101203_AACFRK jennings_w_Page_047.tif
c79f5bcb33a2310c9b24479235e4af31
7e139a126b50a7e0413311bee658c53e71a143b1
26636 F20101203_AACFQV jennings_w_Page_034.QC.jpg
9f81d092de68cea300afebf5d2dc601c
e03a3bc99fc2926a3d211826d720c289d7721ec6
F20101203_AACENT jennings_w_Page_011.tif
326e2f5fddccbac848f87b94f47c632a
ff1f0c94b331a31866b5eb5b0e6ffe789ae611f7
2114 F20101203_AACEOI jennings_w_Page_086.txt
bb2f0e5568867ab200143b3562571a5d
2ec894a8cef91b5e7f027ee91b23867fbb673956
6596 F20101203_AACFRL jennings_w_Page_144thm.jpg
e2559e53fe7d5b92d90eaa62bc5336a8
ad5539fde781d7a1acc9f1ff6de3a1a0270bc218
F20101203_AACFQW jennings_w_Page_096.txt
8ff55182c7f2cd38dec983ea64e57e9e
b5df3ec93d644b7978d850964dd274875ea1f5a5
29479 F20101203_AACENU jennings_w_Page_180.pro
d61bcd0ce1eee131b548f4c3eb8332f1
7c9a6c2fba6d323f3eda11712efbcc768d7acfb4
78638 F20101203_AACEOJ jennings_w_Page_020.jpg
5293242026f43bb9acb7207280237c7c
5b8a46f2cf08d68f2adc4315741d95c02b866027
87842 F20101203_AACFSA jennings_w_Page_159.jpg
fcc27366fec16a2424bc472cab8441b4
85c0cdd91607ecbc03d151996c52854d32f75c6d
573635 F20101203_AACFRM jennings_w_Page_138.jp2
0dafee704aa2e51bebd88a9515acaaf1
45b6a38d4b4a33341ea0293028b062f656da3789
44560 F20101203_AACFQX jennings_w_Page_136.pro
5a50767a5798e21aa5ea03bfc2624b9c
83e18da355bc309ecd7309204e3ee42fe22b583a
120651 F20101203_AACENV jennings_w_Page_114.jp2
77055084126d78d28509d24d7be56315
792705f258a141a5943ae1fdcd196b013fbeb432
F20101203_AACEOK jennings_w_Page_087.tif
7a03030eb13eb16f27e916a2b63d303c
e218fff49da9d8e21129192981418b2959a475a7
55266 F20101203_AACFSB jennings_w_Page_025.pro
03cf22f858ba2ba9a7afa11db86c38f3
42a47f5c5d02a3ee79d829971cba30a570bf4c40
2124 F20101203_AACFRN jennings_w_Page_021.txt
2fc33404609ef2859ff5b64c955c153e
40d0c3effb0565f87774c6351ddcd2246a316377
53647 F20101203_AACFQY jennings_w_Page_089.pro
b5db0e1dfab84a6616e700b7246cd8b6
b060f119f34eb06f1ab91cf3528d865d4d799e19
84956 F20101203_AACENW jennings_w_Page_178.jpg
570f92bae2e7a42706ad895f8f77823c
de602615d2f4fe526ff760e32f0fde0543fdad5b
122482 F20101203_AACEOL jennings_w_Page_016.jp2
bac6aef6d794621d441c0fefef14a08e
cc463657c1a59fc55559ed65b983c2daa7c17db6
210089 F20101203_AACFSC UFE0021608_00001.mets FULL
2880d1a4f6dd8e34061f114eb0d4a26e
2a12c35d1858c8a6af7c1d934e8d290a86080374
F20101203_AACFRO jennings_w_Page_146.tif
b8be77e8396e0ba5e799fdb59a026797
28b473c31ef6442628f2f88b925b47c547706f32
F20101203_AACFQZ jennings_w_Page_084.tif
dcf07146ea3b848e82b80c510d2b89e9
07598aa994a781f7e9231bde874e19baa049481c
69188 F20101203_AACENX jennings_w_Page_174.pro
bfae1b37d5782e8be9e6591f8f32e1cc
bd6e7201d18cdcca5acf3614224b6ed283c12477
2088 F20101203_AACEPA jennings_w_Page_058.txt
8383d756e737902e6ba3112725e7e3de
39056ae7e0a29ab80acef7a069cc2a1e9a5a5745
7765 F20101203_AACEOM jennings_w_Page_143.QC.jpg
d6aee0c55518a1909d4db54f0a011837
25e3827b24668a8aebbebf4c1e09fed29fed94d2
58152 F20101203_AACFRP jennings_w_Page_010.jp2
e9572f008fec779f48108187d95dde9b
50b2e7191ae7de0df3d0c291ae552eaf19b67d1c
2227 F20101203_AACENY jennings_w_Page_057.txt
82b9552bab1d7445554560082fdfad1a
febf3728f800bb69be1fe558616a7b108b221e25
F20101203_AACEPB jennings_w_Page_166thm.jpg
8180860e8c7504a0908f1373d495ad36
57cdde8ae584d97f02c63cd0ddd309b6cbf7272c
26010 F20101203_AACEON jennings_w_Page_030.QC.jpg
21c043dfd49625a97430b0b76a5edf86
cb99d77b030a0d76653325d1f7f69c7317f97cfa
114165 F20101203_AACFRQ jennings_w_Page_113.jp2
1170397dc04d021e7c4935eb14c2d65d
b312fe2f7601b16cc46202492abd31cfb63b883c
2282 F20101203_AACENZ jennings_w_Page_035.txt
1ea141c198a0cc6d32c2960d1de3ec83
7bf9ca0e3a92fb31075837adc17eff236e07c283
117911 F20101203_AACEPC jennings_w_Page_098.jp2
6f69d349b15c09dbe98ea63783388e35
3970a4d91667ec6dfe800d9239d3101ad3b5dfee
49492 F20101203_AACEOO jennings_w_Page_024.pro
204525e515dc4bcf2d7f832a10c15bbe
ef32baf74642649d57c955dca4dce46779eab76c
24148 F20101203_AACFSF jennings_w_Page_001.jpg
d3bf719727e4163ff639b4d61103a531
a7cfad623d3abb7d371caf3f5969a3ecdc0722d4
6569 F20101203_AACFRR jennings_w_Page_155thm.jpg
6fad3bd078a5064d21d048a4ea23b97a
e00ce27cc9e173a28c6bc35c75c896804a5d9ade
43287 F20101203_AACEPD jennings_w_Page_128.pro
87920b5d733faecdb598f5ad3222f5a0
9d69d32ad9a850c9a73ec0905203c481e40e9984
63450 F20101203_AACEOP jennings_w_Page_175.pro
e8fd701bc150bf0d37dbd6dc32c0b36e
075383c9d22f7118dee28f1d372be3b0f43f52d8
50210 F20101203_AACFSG jennings_w_Page_003.jpg
874c64277896da53ab22f598d704f1b7
3eed917bfa47fd065b37d168978c2d9e04aca510
54110 F20101203_AACFRS jennings_w_Page_165.pro
655574e8c55ef748158be1b433919e98
e6e10a5b82b7baef0b611341bc2291662bd05213
50074 F20101203_AACEPE jennings_w_Page_067.pro
23b637521cb9364ee935baba19a10fdf
296313e87ca32865440008448bab121166b31dae
7079 F20101203_AACEOQ jennings_w_Page_114thm.jpg
accb762869545231264f8ac2ce71579a
022ce48e50e92e60de2fb48d956a226b3908b720
81258 F20101203_AACFSH jennings_w_Page_004.jpg
0eb682c3bfedff2da0ff57c753ffda26
90433d1561332784307f8889849726f86d04a55b
7562 F20101203_AACFRT jennings_w_Page_141.pro
4f0ff538e06dba805f35c86f380063f7
4ac683f3da1647f1398a7fa6c3ff0be0886976f0
F20101203_AACEPF jennings_w_Page_100.tif
1fe394e27d8b787f6a215600b2ead6e4
f1de2fa5977e1123d75bf956e9e74af61485385d
73064 F20101203_AACEOR jennings_w_Page_032.jpg
c202eca71599ca9001a3f008efa9c2f8
794d987c86cdfcf24c699d37d91f97b43512c73f
40716 F20101203_AACFSI jennings_w_Page_008.jpg
c7d8bb9afd984b34da12382992dbf099
16c252eecea449ff180c95f0bf148dad2e8f8493
24248 F20101203_AACFRU jennings_w_Page_120.QC.jpg
f18714eda24d660177ea7e5a8301b0b3
696f9cb57736b96a3d6569812cb3daa0c199b7ba
56756 F20101203_AACEPG jennings_w_Page_037.pro
929d2da86140cf7efa6b6f7b108e4523
df4997b0ef3dcfe9ae584aea27be07f2e9092874
2339 F20101203_AACEOS jennings_w_Page_034.txt
4809995433bf8823e8a9096becb1bf43
de6270614612c04cbe3fb1835ba145617997d361
41054 F20101203_AACFSJ jennings_w_Page_010.jpg
12ccfe9fc5350a4ad715ec53b56a0c10
fd5247fcee93e4db76594043b6ae13b0efd92bd9
7046 F20101203_AACFRV jennings_w_Page_109thm.jpg
f98551327244bf7e78d82b6f7006ce97
7218243cdd1d5c826222db888c689a8cf2a64a52
14654 F20101203_AACEPH jennings_w_Page_180.QC.jpg
ebd10f4d1659fe80478864744c90f8b2
bc146ae2c0629bf6a83d83303759ff036d3b8982
109198 F20101203_AACEOT jennings_w_Page_087.jp2
f081ea6640da97bac06a2e43a44d4397
03a91207846779857efa22fb3e2c3da3d97c08c5
70036 F20101203_AACFSK jennings_w_Page_011.jpg
691d380ff29eff95345337af3938d30d
617c3c4cc571a43e68b7766354b09d2a17664932
26929 F20101203_AACEPI jennings_w_Page_122.QC.jpg
6a9d4f10f876b6bfcaf03e9217999b99
c936bb6f591379ea536caa895fd38492e3772e6a
78406 F20101203_AACFSL jennings_w_Page_014.jpg
757557ea6307826cc2abddd6efcbb25c
d8beae71a20a0266d7786d375287b823ca761803
6672 F20101203_AACFRW jennings_w_Page_146thm.jpg
68e96dfe7682858c07dd8b128773f2ff
301dc50782e415ad2590dff9a6eca5e46a396418
7222 F20101203_AACEPJ jennings_w_Page_016thm.jpg
2ac69feec416316170616c4412c2a70c
936a639206abfeb02dc6ad7b316379c7394eb20c
8910 F20101203_AACEOU jennings_w_Page_139.QC.jpg
95a46ef1541c6451a1661aaae0b6d8d6
c47b9577a6c27582b630106e397d3864b10ce969
75271 F20101203_AACFTA jennings_w_Page_089.jpg
88ba1fa84fdca0d9a9027bb923d65c50
2db6550324ac2a650902398a9f399ef305dd01d3
69474 F20101203_AACFSM jennings_w_Page_015.jpg
d13f53789c62c90e1c47916301c8c0b6
06e7b7890b82c4bb0b1ffc1790953dbb4717acfc
25369 F20101203_AACFRX jennings_w_Page_110.QC.jpg
b4ca4af95cee60a24a4b8bbf5c6ab9f5
2935a7d329bc214efa7a9c1a0dfbb2dee5a7ad4b
72591 F20101203_AACEPK jennings_w_Page_092.jpg
6f31948bed81dcebfb35847f8c89de7c
fa7ea0d095c6c32942c033c517e4f63a42097767
23986 F20101203_AACEOV jennings_w_Page_099.QC.jpg
dfb3f8eb5dd94e1bc2ea4fcff61a64d5
ee1ed95cbf04e01e7ff0712391d044273b9fdf87
73711 F20101203_AACFTB jennings_w_Page_095.jpg
7ba60653775aac0c01036c7271b1be57
2a49cf1d0981a9b1c85495bc0b616ae93bd8286a
73672 F20101203_AACFSN jennings_w_Page_017.jpg
8410d1023a33da9dce1723f626edb147
6574ac48c1a12418b68879fee211017148af1d2a
51190 F20101203_AACFRY jennings_w_Page_070.pro
6e4a7e0391e27eb6c168c27f8ae341c4
d243c35c9adce36156436fd858cb1db9ad7d116a
3006 F20101203_AACEPL jennings_w_Page_139thm.jpg
40585803cdc445cc75eab5be0c470f7d
69655bcae743a62d1f237c8b0f3e85a915885694
121648 F20101203_AACEOW jennings_w_Page_047.jp2
2c232b688376cba9a3a2c2bb8a1154de
1666777a7d6e72a5b6e433f553e52b361ce6c2f3
77955 F20101203_AACFTC jennings_w_Page_101.jpg
c56e4be6c9aa0b7205b63cad4ae05f8f
224ae08e8187e33a6160f56056d4f74c7558793d
74274 F20101203_AACFSO jennings_w_Page_019.jpg
4d32376d4df3f4560ba8c21ad91a8cb0
6e8d170c2cb67f574e4940c63b06f92462e12380
110535 F20101203_AACFRZ jennings_w_Page_070.jp2
9bb5894f96e20774a6e20e186c485192
c2ff4e369f17a7069939d23bc9beb0ebce376516
74468 F20101203_AACEQA jennings_w_Page_045.jpg
bad6a33d44f82f88ddd2e83ea21ed4ce
2b27d3b36edc2ba2d46e1faa19a5c453c7580d09
F20101203_AACEPM jennings_w_Page_069.tif
385162660703d5515ac166f823a68600
4a1d6882f102e742fc3abc2d206b748085c6af22
80649 F20101203_AACEOX jennings_w_Page_124.jpg
1ef4dd2695f63a834b6466789a181333
00332c896465d772e3bf59919ed6986e34acfbae
68016 F20101203_AACFTD jennings_w_Page_103.jpg
37b273758def65a6451f1c4283002b6d
f68d10b3d3a862c5ba83d7c12c3b515ec711e2d0
73150 F20101203_AACFSP jennings_w_Page_031.jpg
a288206381aaf9a6483b79d1d98f6bfa
4190a19a17ae3f5c12465083a68f8b07bae4a063
F20101203_AACEQB jennings_w_Page_063.tif
43d4e2849d05d6ef3b346aec344bbee0
bcca726a90ee68968d6f940df120dc3d5d0dba6c
90865 F20101203_AACEPN jennings_w_Page_175.jpg
8a7fe4da484165596a3775dd8ea39b4b
1a030d76f937350b50b6661472fb69525820b6bf
6975 F20101203_AACEOY jennings_w_Page_083thm.jpg
f305c7581f0a98ce6299255ad1acbddd
93464dd435be84cc6b1b7bd053d9774299fa2efd
75220 F20101203_AACFTE jennings_w_Page_120.jpg
9a6e2ffc0f03fb9a115bae96ed8379c5
266b496c165e6085b94240e17ce855a0beaf99e7
86436 F20101203_AACFSQ jennings_w_Page_042.jpg
92ab4524c1fb821d89d4a2d6fa83bbe3
efb8c267ec574a3afa509f437bf7e63361685e03
1298 F20101203_AACEQC jennings_w_Page_003.txt
7fa664937ef05409ac224d868a9d3ad2
82ba267cfd9bb9b09cdbb3778014c429bb956509
F20101203_AACEPO jennings_w_Page_161thm.jpg
fb86fdcfee7f3a398d34b89cf12582d9
58342d8bf2fbe344417df7e034227aef57c63e87
F20101203_AACEOZ jennings_w_Page_117.tif
1724cdb526c9c90388920e217ee175be
c156c4a57781f29a60f758c65a92fd84945d3683
21937 F20101203_AACFTF jennings_w_Page_125.jpg
0fdd5d0f83d17d72e07b10603336e94a
ba3a95aacb8074cf0a84beef6e3a6d770f8c9073
77242 F20101203_AACFSR jennings_w_Page_043.jpg
a6953fe3309036b191b1f2b7fcb5ffa8
7a331532964035b9de9405ae1c60bbb2654d12ae
76411 F20101203_AACEQD jennings_w_Page_056.jpg
c95fb3a584fee7e01858c122a99f4b36
95818800f5e6adac530aa4be59d1bc934087a9c0
50259 F20101203_AACEPP jennings_w_Page_099.pro
70d1bff70d6c708ec049295507e3eb54
6e45bc06fdfb6a02d5543a8937ad859a206769dc
52346 F20101203_AACFTG jennings_w_Page_129.jpg
05d55f0f22f0a00ceb2c25b2176fcb64
46e5272ef0c9beac0258cb919c04fb61a1fb6c57
78592 F20101203_AACFSS jennings_w_Page_046.jpg
5e2e27a779e62ac7b6e48a5be84bfcad
20312f91372471eeaa1340132f487fbc1e3f234c
6497 F20101203_AACEQE jennings_w_Page_077thm.jpg
744771f99d2869ff633e60f4e45aa0ea
4b31186b1b5721c0fe953ee6d2b14d96ad7fb7b0
80836 F20101203_AACEPQ jennings_w_Page_115.jpg
b65c9f9816ace1f21547ce9f72d320d8
a2e0b916c9f51a85c86d1a10de06c67502a5ba3e
58361 F20101203_AACFTH jennings_w_Page_130.jpg
a9a6a2251badb7472f847f0699e0dd07
74f1fbc607601749832eb3ae6a6c8899b8534e3b
76904 F20101203_AACFST jennings_w_Page_048.jpg
319684919c1e0d77698f8e5e7a86f9b6
c411102f9e3046cb438913bf5b3e2e40d9313e36
F20101203_AACEQF jennings_w_Page_168.tif
5bc998e14637081a9ebdae9e4ae50a94
837c01609e70c167a212b8b3b3c95bf1e9c83a95
26831 F20101203_AACEPR jennings_w_Page_039.QC.jpg
5b6008a67365074ef1c66bc54b35ccd4
92d25da9d2239bbed52fb96fcd6ef6a199649c6d
24483 F20101203_AACFTI jennings_w_Page_143.jpg
cdb0e175815e8a74f4b13ea0cd1bb21d
c8a12fc8c32a38ba58d9d9d3b8140c2b26422dd5
74016 F20101203_AACFSU jennings_w_Page_053.jpg
9aa7d5a67cc26130f55c998e0d0ae5cf
c5a15578a9223eb5779fe6119edde04568e0ca39
24131 F20101203_AACEQG jennings_w_Page_118.QC.jpg
631018a92c98d188becca3b0c8f94375
d721b5eadbee9e12682640878a642a8d85e5fe16
66359 F20101203_AACEPS jennings_w_Page_119.pro
2cc921f888e05ad6fc00244776b52bfa
5f1c0db45b036a2cbe2166ec29b335959b4852c2
78112 F20101203_AACFTJ jennings_w_Page_150.jpg
97a9bc119eff0606fb3e38cfa102984a
db50f82d0b09ea54d7cf73785b9db53f5e9abfa0
77995 F20101203_AACFSV jennings_w_Page_060.jpg
2cac6561d8df5fff6650c35ecb429f0c
ce484ab50f6c913b1246d9d3528de1371deba2d6
6931 F20101203_AACEPT jennings_w_Page_154thm.jpg
fef05af7e63706d64251395534698be5
73b0016f541c5cd4938b5e0370c3166f8c254fd2
6687 F20101203_AACEQH jennings_w_Page_053thm.jpg
824d7f083f6a183730a9fa30be9eb903
2e5dc6645aff9dae6d8358b48461b91eccc0c154
78752 F20101203_AACFTK jennings_w_Page_152.jpg
5d1422f6cad91e3123800b59842a6491
f81e962d090680ae72adfe3de9efe40479f6ecff
72025 F20101203_AACFSW jennings_w_Page_067.jpg
9a1c1c7891bc46dd564611e92619e5db
9fc7c04472e1fc732973e11052c1f45ccd61570c
23839 F20101203_AACEPU jennings_w_Page_082.QC.jpg
b470aecae531952442e9d3b246fa7949
a8680a00da8e61d28f7d6a0ad86b5f454bacf8cc
2146 F20101203_AACEQI jennings_w_Page_076.txt
75e61fd0dd86d7f86393cc78dea715bb
e42badb3d7088c761045b5a96ba16c9f91aa7edb
76704 F20101203_AACFTL jennings_w_Page_153.jpg
5a7af6b15e91466124f096ac82f54f30
6a3150cc6df3aac2cec9255a2c0a2d748556d47f
114026 F20101203_AACEQJ jennings_w_Page_071.jp2
d3b9ff7fe8e7df30bdae4a68cdafe450
914febcc2d6c7f0bd7a683b154374ef83c7ea28f
123598 F20101203_AACFUA jennings_w_Page_050.jp2
9f9e7fa693d45778a273d61fbc8ec482
7b5af85a4b5f77fcf6552066272ad6d91dd49261
73524 F20101203_AACFTM jennings_w_Page_155.jpg
77603f1defd8a181c8a47da6d5137715
fc9397ea33549205113893a682dc843c508bbda8
82292 F20101203_AACFSX jennings_w_Page_068.jpg
7d5deeb53d8574b8efa93711d794c844
0046a970cf9ea9aa656c11b65368f18b64b33a49
25265604 F20101203_AACEPV jennings_w_Page_142.tif
e00ff50a787774ff69d79ae930628b8e
89165dd7cae5a6a81e3734cc92030ab3e79e1c99
55947 F20101203_AACEQK jennings_w_Page_027.pro
263cb87c7f7aecefb709f54e923272f8
f458a402a65c73cd9961385fbd1c3eaa7d09405e
120159 F20101203_AACFUB jennings_w_Page_051.jp2
771e3c2699bf54ac3c0c03b2b63515a2
2eb6e60d577046ba7f0ed4385c47ed19f8b577f0
84554 F20101203_AACFTN jennings_w_Page_162.jpg
a95457fa28f8bccfffdd04753770a302
61e7e3526608510a7da2eb042a032aed24022ea4
104015 F20101203_AACFSY jennings_w_Page_080.jpg
7e8d0ecbe3222147bfc5971c44e264f6
40989ee7e4d8618b17cf156bcfe4bf02d03e9d9c
F20101203_AACEPW jennings_w_Page_078.tif
1449e881eebad6da413165f0ba373a07
0f176119fd2beff109adbf02e6788cbe8bc7c670
96727 F20101203_AACEQL jennings_w_Page_171.jpg
003199e5244d7467edbeebd85a6c31d2
0d090e2f2e2cc454d5995873d49a7207b14a920a
122181 F20101203_AACFUC jennings_w_Page_052.jp2
234e4871d7d33a5e7dd77a3a679b0e47
6ab66d0fcb6f9d311d3cb8e57d3b74c1f506a0a6
98204 F20101203_AACFTO jennings_w_Page_164.jpg
be072ebe2ea386839d3857865351caf1
ae0144343eeebf65280f4b912b8092b716ebbbdb
71708 F20101203_AACFSZ jennings_w_Page_087.jpg
623778e25545a100bca9f235c45de6de
a747569586fe7484361204cb1fa81c7adbe0f7d9
5156 F20101203_AACEPX jennings_w_Page_128thm.jpg
358c697bd17e275d555759f587d762aa
80438f72744ef44a6a46b5a231c07e8fb932c018
2154 F20101203_AACERA jennings_w_Page_060.txt
8533898338e7a8bf5b39e5b82eeb44ab
580a1382044aba5f6f28b1906a0c46251216501a
75206 F20101203_AACEQM jennings_w_Page_013.jpg
742d0511ce4d912ab7b26cf8c654296c
cd6bbd6d9f15a91bd11268f0c7b13941251eea68
115554 F20101203_AACFUD jennings_w_Page_062.jp2
c0186c411e2dfb9de3427aed14fdd0da
5d7d2beb867b5d739adee11c9773e019080add6b
101073 F20101203_AACFTP jennings_w_Page_174.jpg
3872acc19996a79ba1444adcc98eccb5
8dc2d9c635f63eb4f87514c266ab42afd59d516f
6688 F20101203_AACEPY jennings_w_Page_113thm.jpg
431359eee458c0267b949e552548f069
d9cec52d586499ce4e63c4b191b7932017afdd09
7157 F20101203_AACERB jennings_w_Page_034thm.jpg
7034d9362863b538d7a7ec2f06a75eee
0f75a6bd86c837132763a2508206af30905ec3c2
117754 F20101203_AACEQN jennings_w_Page_065.jp2
5682a6028178d9c82fb15f3d36eae930
f5d51ed161f3e09dd04f6bed6b1f2f4713077eb8
108266 F20101203_AACFUE jennings_w_Page_067.jp2
ca617e8461c0844623ca4d2be1f86dd9
d697814f83ace75de8fb443bc030cba948af6206
89479 F20101203_AACFTQ jennings_w_Page_177.jpg
02c7dd7118f6869d9e3b943b1cdce8dc
7a8023053addd91e955c1ed96c479a1372607786
2150 F20101203_AACEPZ jennings_w_Page_097.txt
a29032d9a1d6efe0fc9439b62b7c00cb
43fd5bbb99e6641db936c271c7a2e9cf423271bb
1051960 F20101203_AACERC jennings_w_Page_004.jp2
181e3f9fe46591c83a18b52bf0d5d5da
38d7df13dd28f21aae14d4e2c9ad45798a529032
44222 F20101203_AACEQO jennings_w_Page_130.pro
75a7c424ee14073db197dad1b10ec36b
874543fe092ae9f7260b470235c07075f63fe750
26144 F20101203_AACGAA jennings_w_Page_111.QC.jpg
1d84b9fbdbb723ed3823f99ef548dad7
fc298b0d5b8492831690a1d5c777a25ba2b57711
134082 F20101203_AACFUF jennings_w_Page_073.jp2
ada8dc558e2699ce5403f3c8a0bc1318
556dc0b4dffbff343bcff7cde89a7746525537a2
84836 F20101203_AACFTR jennings_w_Page_179.jpg
c390b0e334bf2c8b4d6a37ce3e37653e
55fea31ea630348cb8b47c05d3703dbb5a8dccaf
25562 F20101203_AACERD jennings_w_Page_179.QC.jpg
5342f50fdb6dcdbf61b7c0ffa064df86
7f684a0f9bf4d36631e7ee545204f7171414e8d2
26642 F20101203_AACEQP jennings_w_Page_073.QC.jpg
118c1e4d520ca310e1ad0fbc0ae3b9aa
fc95012999e9b7889e81520dcb3ee61bee1200b5
26392 F20101203_AACGAB jennings_w_Page_106.QC.jpg
71980682abc699468b7ed83727dda9c4
5604ee2fe746c3ad1b4b520a88fe1f0a2e8b5fef
113894 F20101203_AACFUG jennings_w_Page_074.jp2
f6a56e21415b2cc7ce21a626ca073509
33ea5cab07afe84a22b7eaaf31ce06e1840af66b
1046236 F20101203_AACFTS jennings_w_Page_008.jp2
619fc8d27714807646799173d823bae7
e241ad300336ef0a01c435f34f0c72ba1a176ee1
2604 F20101203_AACERE jennings_w_Page_162.txt
e0d0d831b6a3f63a55b6f5148def0da7
01feab9bb5f58042ea3dab3e90453f965e0dbcb7
2034 F20101203_AACEQQ jennings_w_Page_104.txt
26a27d91107c7dc99276949f2d3cc52d
132186d66ca7ec651becda55f10df24ef7420501
7002 F20101203_AACGAC jennings_w_Page_055thm.jpg
d602315bfe36b2171426f680e297bfb5
62e99e155760783ed0afabcb8ac4c1c1eadee446
108609 F20101203_AACFUH jennings_w_Page_077.jp2
428ad89fbbb973d4cf67242b541ed5b7
b1614d951daf88c50c568f67f09aca8c94f2781f
100254 F20101203_AACFTT jennings_w_Page_009.jp2
b8fe9ee7453e4556826ed738961673c7
955a3863218cccd672e421e23a48561931fd3f3f
56981 F20101203_AACERF jennings_w_Page_052.pro
b3dca872e8342e744d71fb124cfc34ef
5d8df28f0461ccb30cd243f09d96bbeebc5f56e6
F20101203_AACEQR jennings_w_Page_061.tif
e3225add3fb400475612806adde0b054
9ae7487b3f096550640cf66f63370af269138027
6827 F20101203_AACGAD jennings_w_Page_027thm.jpg
e2ed2d025eb01e0d778956761f499125
a463038e512e370d9aea7da2a8153f267a919c6c
117408 F20101203_AACFUI jennings_w_Page_078.jp2
27538a9b009d5f70fa80a00643687ec4
7c56dba0c1454f20f987dfde8095509180d92bd3
120394 F20101203_AACFTU jennings_w_Page_020.jp2
cd8c094ad2c856d50f3d8e8915a7880d
068e1d5a4f17c0eb6e25e976b03cea6b44aad911
114254 F20101203_AACERG jennings_w_Page_058.jp2
aca068aa9fe25267a73b2a6dbccda9b8
71a9b2a069646fa260fd40226a8d6245c03cb07c
25396 F20101203_AACEQS jennings_w_Page_022.QC.jpg
3ad2321a783bd8dcd744918cdd2e714e
e30d0a733f2d1ae972070737641f60ab691ada66
164223 F20101203_AACFUJ jennings_w_Page_080.jp2
9629f00b4b867349aec0791b23186203
aa565d17b0efbfae2612be61bf802a00915b2917
123254 F20101203_AACFTV jennings_w_Page_023.jp2
3fc6577969c184c793a9064313bc16d2
b9fa44441d914ee1cbb8389a9281f1f27092ae9a
80008 F20101203_AACERH jennings_w_Page_036.jpg
83f98845055d3c41a5e719961bf3d3e1
7c441f1ecefee850e2b44d232137d91ffc2f3f42
1051976 F20101203_AACEQT jennings_w_Page_005.jp2
b747ecc3dd8ba93cefc531988caa40f3
d07f868960d1cd78344a25bd5a172ecfe9cb70e9
13587 F20101203_AACGAE jennings_w_Page_010.QC.jpg
04a04f5055dde1a6653f06e17eab23e9
e7aa48e75d97ffa17335aaa5aa8e60b90a5614a4
121250 F20101203_AACFUK jennings_w_Page_083.jp2
977ad7d6dbbb4fe4d8bff4fe9782833c
62c3533f84b7ecc6dad4a55d757a102a989824ec
114401 F20101203_AACFTW jennings_w_Page_029.jp2
f486699e799a265d3114fac86e87deff
2bee71c3a52fe6d452578834ac04c96af477cd93
23659 F20101203_AACERI jennings_w_Page_077.QC.jpg
0a6526a05cf0373c57b276d1f58e5c39
a491bcad422b0a5644287aac41b4f4d54fc93bd4
7190 F20101203_AACEQU jennings_w_Page_175thm.jpg
d7f813bf435dcc848af35ada66ed5069
dc0866413470bb3252c4ac119a7570babcf5607a
6739 F20101203_AACGAF jennings_w_Page_029thm.jpg
80bf6c2ea28f224bf48646433465a8c9
9f22e38f264234c198a2b98e202db665a4512619
113078 F20101203_AACFUL jennings_w_Page_084.jp2
939f56efddb9c38a289c48b2eea640d7
0110e4723c8ecc357403b41f79d7e9ce69887654
123444 F20101203_AACFTX jennings_w_Page_034.jp2
9697472626eb7952685e198d3ac4a9dd
b22f5541f5ed581e4fc77326938b00e4b529b16b
47986 F20101203_AACERJ jennings_w_Page_180.jpg
736c1059a95b446ecb7e335ad0f6356c
3b406bf24c10f6e5fff5db962ec42d6a367c9a6a
54498 F20101203_AACEQV jennings_w_Page_056.pro
5b8e1ab23b4b90e5fccc5542dd4c1974
9a25b5fb4f6d9aeecc9f61c4ac57b9993d3c1e00
5030 F20101203_AACGAG jennings_w_Page_130thm.jpg
efa6368df09f1b32a695be46d278e45b
4fdbb505c5bb64d649c391abdf058a1a609c821f
552935 F20101203_AACFVA jennings_w_Page_140.jp2
bda3410833d4e873a94fb5ff3d290d91
3da87609afeb6bbe057480d097af8bb4701cf5c6
116084 F20101203_AACFUM jennings_w_Page_091.jp2
766cc73a0d1cd1fbb1b356145d7b77aa
de2507563aee2f7db0ccfe28699c4e95d71c47df
80583 F20101203_AACERK jennings_w_Page_169.jpg
f043eceee3359b77391175ed06dcd3bd
4601ad6374951de1b9e43bb026e98702b7174112
26198 F20101203_AACGAH jennings_w_Page_035.QC.jpg
f840e2556254395b6a5cd7d68814e8c4
0da9b417e3722184b17220ae9c60f3119ce5b45a
116310 F20101203_AACFVB jennings_w_Page_149.jp2
299c936d5514e5d2faf10020411d5983
88351179a5c1064b7fffd11958256784cc243644
111901 F20101203_AACFUN jennings_w_Page_092.jp2
deeb577bdb6b0a7c78b6a19e4ee139b2
812769334f8605f780f5eb1dcb94f754b210d545
121218 F20101203_AACFTY jennings_w_Page_036.jp2
ead05f0426f4f9dcb219b2f5465f2055
1cf94ef3ed4b6f7578436d354129da6a5fa68b5e
79497 F20101203_AACERL jennings_w_Page_030.jpg
1ce4f9982aa7c51b723d1ca5b91a1f3b
58b44ae0462f59bdb83644adde35d15c75488a14
119425 F20101203_AACEQW jennings_w_Page_043.jp2
dff7feb48737e55ff0ad03de0fd460b4
06700b1a7acfa92a55e5ce7d089a3dd8cd8b2212
272328 F20101203_AACGAI UFE0021608_00001.xml
51e572742a2abd83599248e9fd7300a7
d5c08f0cfa071fec69441aea6b411291b3e53a6f
136243 F20101203_AACFVC jennings_w_Page_167.jp2
660a7dd04d5484324e6186684761817e
17403f1c0678a52afe69051593790fd3b33ae2d0
111614 F20101203_AACFUO jennings_w_Page_095.jp2
3775b534c1aa9da08e3c9f82ff2158cd
edc803977fc9c24917986ad80029b2c17289f155
89369 F20101203_AACFTZ jennings_w_Page_044.jp2
47a3d1d296f4f9ce6697a55561ef4c2e
ec5dfb733a8bcbc89ac7413498bcdc8e68388800
73039 F20101203_AACESA jennings_w_Page_075.pro
78431e6f9c222ed3eab2db2bc9f8fae8
6b5a975665be2b552c7aea4e4199d376e3629044
26350 F20101203_AACERM jennings_w_Page_026.QC.jpg
f47baa21d9d21139265c0d70310419b2
06cbde71e254b350d2efc7917a25ac36ca67fec3
77290 F20101203_AACEQX jennings_w_Page_079.jpg
4bf80e848fcd4c58d0ea9adb85360aa0
03b2e0edd558717da5848a5f688be1a34c046208
3084 F20101203_AACGAJ jennings_w_Page_002.QC.jpg
9961c945052e17b1bc950fd68675ba04
6e07850be60cd73c7382eb37f861a0de42f9d437
137026 F20101203_AACFVD jennings_w_Page_168.jp2
b62192544bc58305da7bdc79b133b7ed
8e163a61719664ccc343bb2f666e0e4236d88f3c
114598 F20101203_AACFUP jennings_w_Page_105.jp2
ca5ebc580660a2790c8d11440f3b47ec
9c8647efc95cce0fec581de94071d955bf8a2824
2166 F20101203_AACESB jennings_w_Page_154.txt
616537fec5faa136f919b1d2f490e793
a8e983d7bda69df8569a1d2c61ba6d61ffabf314
25492 F20101203_AACERN jennings_w_Page_048.QC.jpg
62af29c6a910e37f1fa3f2d67a5f6b09
6ebe4f3a97cddbb1ce491b6d4374d38a259038a5
147925 F20101203_AACEQY jennings_w_Page_164.jp2
875e4a20918534cb15c22abe0d40dbce
400acebb9126a4ac1eeea4676f6389b86deaa114
4616 F20101203_AACGAK jennings_w_Page_003thm.jpg
dd8289a469ffbc06cf0fc6efb14cc15e
89fa2cf42b022ed9eac62774bfc75b59ac10d3a5
139787 F20101203_AACFVE jennings_w_Page_170.jp2
caf67a49a61e8d63f0866476bbd20d6e
c8e5265652316af37f7088df01e425dd04afcd66
116149 F20101203_AACFUQ jennings_w_Page_108.jp2
26724447c2a49aea7738ce66e9ff656d
19fa4b2b54b8c8a7c1fcd5628f26502f77799846
43749 F20101203_AACESC jennings_w_Page_157.jp2
d13fdbc35f045efac03f449f4a62c41d
c5747eac8b85ba6989e167731ee6755c98b73359
2384 F20101203_AACERO jennings_w_Page_016.txt
59975af64e52bbbddcd4f82209bd2196
8ef22deef831af07555a7e9c085bc7ef339067dc
F20101203_AACEQZ jennings_w_Page_122.tif
b0bde09cfd2e1aaea05b6236685f5a7a
52ccab4eb624273226c77ce5fa0a86f20d29c893
6995 F20101203_AACGBA jennings_w_Page_037thm.jpg
226cd8dcc1cc4cdcccb973202e5fafe8
dc1000fda9cf6b53b996acb63f20b420da90c042
5536 F20101203_AACGAL jennings_w_Page_004thm.jpg
5e20f6aaf89456cb425ef4348c6abe8c
b07b06e56c91ea25cf036a939366c4a870d67466
130074 F20101203_AACFVF jennings_w_Page_172.jp2
d9a070840497e164f88c185f044b336a
6f0879ac09737da7a9d2c78fa38a3b014b2afb1d
113480 F20101203_AACFUR jennings_w_Page_120.jp2
18f34903936adfd547a426676c7b7ffa
178ce3608273aab7ffcc22d127ca908e079c7e08
6913 F20101203_AACESD jennings_w_Page_111thm.jpg
c35378f428999923f7c954a7c58420af
d831deae9418a3d79d085849b0ee5c7d505f46fd
F20101203_AACERP jennings_w_Page_068.tif
c998ffb9f022e5c1b819d6f9840bb75d
52192ee2e34598a66206c9595c8ca179c83fe04b
6966 F20101203_AACGBB jennings_w_Page_040thm.jpg
5b79280f5ad794b3fa94e9b26b693856
58a8afa5677af678b42a45a6e62114e7e0729bd0
26115 F20101203_AACGAM jennings_w_Page_012.QC.jpg
6f348d75360c23353f257ba1408e44ed
82467062818465727adf6a0e5b694d814e8a2a1d
34454 F20101203_AACFVG jennings_w_Page_181.jp2
afaf41f99ff611ad81155a72e8640436
9b3150a562dc134d44616cea7bee99018a8a2643
118699 F20101203_AACFUS jennings_w_Page_121.jp2
54739203c98455b60d1df0c6dbdcdcfc
42485bae02df4caef13283df201677a6dc1a6996
6867 F20101203_AACESE jennings_w_Page_012thm.jpg
d4fa1e705d68268651bae6d041d191c3
9d29e4b0d5c4aed79712492dc26564edcfcaa866
26475 F20101203_AACERQ jennings_w_Page_036.QC.jpg
ffb1027d633817bb5f2caa28bb247dbf
dec1b9779edce8331ea05c58c3c04b9345a7a2d4
25532 F20101203_AACGBC jennings_w_Page_043.QC.jpg
7f4980da09c87500679ac6618c2a77e9
b7aac5e7d5547ea39b84adee985041896387ac54
25902 F20101203_AACGAN jennings_w_Page_016.QC.jpg
3c5c5a035abffcca2fbce360ff3a8692
fa47cb0ec346f130c8da2d1e951dedfd4dc161f0
F20101203_AACFVH jennings_w_Page_001.tif
748b06590887d00f4526ce38d19d0114
958374c6d6e134c25554d0d8faa84fa29340042a
123003 F20101203_AACFUT jennings_w_Page_124.jp2
5e09fe00dc9593cda21876c8bd619046
dfdd152603fa1004562484fd7996bdea08fcf49d
74662 F20101203_AACESF jennings_w_Page_058.jpg
ad3c31bb7eaaa43224c9188a3c0736d5
88c178c5801c62730ee474e93c9f879ee05280cf
24880 F20101203_AACERR jennings_w_Page_056.QC.jpg
bceb967273e52639eaacc163646fda02
60c5244e0c34ae09d3afe3d7b4ca32adc47624b4
26246 F20101203_AACGBD jennings_w_Page_049.QC.jpg
b10552c5b8e18d5625e8d237f48b770c
3c38cc7ee2d747c6a826328925c4fb82ea33a43a
24852 F20101203_AACGAO jennings_w_Page_018.QC.jpg
afc902e6281046b65116ec44be21880d
10b44dd7adae626e5866764b381969835708016f
F20101203_AACFVI jennings_w_Page_002.tif
6c96c7f6ee765cf1803e9dd176e0ee48
049539240d34a125d9f3be5e9e992cf5f0a8ed02
81902 F20101203_AACFUU jennings_w_Page_126.jp2
a88345f29d49834ab15df59e0fc522ad
7055ecf3b32af6635b8f21ea4aebbb4b05795cda
2104 F20101203_AACESG jennings_w_Page_078.txt
3ecaf088a57004299d9c211f392f0cbc
3feb99e24f671225587dd50e8e253196c330b35b
6910 F20101203_AACERS jennings_w_Page_043thm.jpg
36d30ce8e820555d0a75dd47e6883d56
997b0a11b0997aab94fc4451a9b3b8ce5c74cd83
6879 F20101203_AACGBE jennings_w_Page_054thm.jpg
4b73a0736f0c7d7b90bbc51f74a65682
edc4d7e995530a4fb8d16c3eeb66b39adfdc23af
6776 F20101203_AACGAP jennings_w_Page_018thm.jpg
3140340a0b09bf6bbbce578b9ec490bc
8b081d0f3befb605ab8b8fbb00f93a8f915994ed
F20101203_AACFVJ jennings_w_Page_003.tif
f8a831d7ea42f5dbb78017d8b0388948
e38c13c5b9c932c5abf395b4b0fc9daa9821bab2
71905 F20101203_AACFUV jennings_w_Page_129.jp2
91719403a135a053806fefa618e44ad0
1aa88ccad0a61aef31c6b4df615387ac0c41c74f
87269 F20101203_AACESH jennings_w_Page_116.jpg
3b778e70eeb55a3e7a9b947b96ef1531
5295b48839bfdc05ba05f7e749b6fb7de32ee94e
2595 F20101203_AACERT jennings_w_Page_178.txt
02dc7f9639a0371098a08e04255c9a0a
4b2e483e49133dcba1a12f4af4a03bee7342dd42
25257 F20101203_AACGAQ jennings_w_Page_021.QC.jpg
550cf09152f53df17409782b93afc4e0
0c118b4a01d720b772edbc7cf9e42aa646de6c95
F20101203_AACFVK jennings_w_Page_006.tif
720260a6aa54d6dd04aaa09b6afc8be6
662520983f66f71f4686e8eb13bf830927559260
81452 F20101203_AACFUW jennings_w_Page_130.jp2
72ade63d6fa4278e804e9b31118f8f5f
9d6413ac0505f6103adddda7254e5bc9992034ee
124404 F20101203_AACESI jennings_w_Page_122.jp2
e170b69e5d745bfe5504ec4644a2f642
68ede135b1c27363a092c92943cdd9ef9befada3
55105 F20101203_AACERU jennings_w_Page_114.pro
a7b2c37b7dffac29144bdff398b509c2
118e16d56929495392aeb23a36d134c2c00d7c18
25032 F20101203_AACGBF jennings_w_Page_055.QC.jpg
0f61dfe4d07855404286db13dcc4f587
a6af4baeb8a57b6a21d1b395d28737bb0c9ec83d
6988 F20101203_AACGAR jennings_w_Page_023thm.jpg
d50c344ac40b408df3674b04993648a6
9ad27bcf3d38c115b82e6f768edade050840235a
F20101203_AACFVL jennings_w_Page_009.tif
df643b2de7fc235e6509b2ad95e958ef
bbc2f650eff91d0fc150cc9b713b359977b4adc5
61055 F20101203_AACFUX jennings_w_Page_131.jp2
3e7b5db91e6aff6ad295e0a652ea3694
03bbf751f6e7bc1adb4483e42c2b2eaab0acefa5
2809 F20101203_AACESJ jennings_w_Page_075.txt
efa1fad29773ac93e619165a83de6461
610c17dbbbf759ef7a2f16a31fcdb0554190a4b7
7086 F20101203_AACERV jennings_w_Page_049thm.jpg
0e33f4ef9fd83786fdb6beb17af44334
c2ac927f37dde886a9fabbc30804831ef00093f3
F20101203_AACGBG jennings_w_Page_057thm.jpg
c0fe97cabfeac7aa81df664bc2f71783
3ee07fbfefecdfd6a4fe1f9e988482c62badf36d
25464 F20101203_AACGAS jennings_w_Page_025.QC.jpg
a54288ac7a105886a3eef6b608866371
78c49287c6b3b65454f8bbfa3459a8c6bf9f5c5d
F20101203_AACFWA jennings_w_Page_094.tif
9348f965eab8496b3ab4b570d6962d59
57bdf30f178a3bbbdf950cd2f11ab506eb0932d3
F20101203_AACFVM jennings_w_Page_010.tif
87be8ed40fbdde804143b48808947175
5daf1a86c9e75dc2d21193b80e1dfb201e24a60c
54480 F20101203_AACFUY jennings_w_Page_133.jp2
f3884289d89ee04eb184334bc6b255bb
44ff032ddaefae5a5a585bdeb989f92e50e395b6
52532 F20101203_AACESK jennings_w_Page_113.pro
ecd3c8104068ed0b979fcb772b41c0b5
bacbb70e300f3de94f614c5998241bced2653eb2
2195 F20101203_AACERW jennings_w_Page_027.txt
5a3047db4ec3da7938f62cc4b9b6c816
ab1c4410cf592a36c6ba879c0792162ba01ba8f4
25609 F20101203_AACGBH jennings_w_Page_060.QC.jpg
fc6a9439435139f0228263bacaaa6724
9cac04e99deb5bb59dc190e539fa00826329dc8a
6846 F20101203_AACGAT jennings_w_Page_025thm.jpg
3087900aa0ada4dd74c056f527c754e1
2402a9e206abbd84e2be019dc40370b2b4d1f762
F20101203_AACFWB jennings_w_Page_099.tif
bc11d0493b12e813d5743c4d713359cd
2d868829cec96d64e358598a5eb643ad5702cafc
F20101203_AACFVN jennings_w_Page_017.tif
167d6e0628b38b7e8f9f0aad3062b4d5
89c86ed4a2f5f57843a3700dd91ff094abf2035f
F20101203_AACESL jennings_w_Page_130.tif
c9f969ac85ce14fb6ec1d9697c0119e5
00ac3428a9c83f6523e447d49566b7675c243ff2
25995 F20101203_AACGBI jennings_w_Page_063.QC.jpg
bf267a6b63fd6d81399cbb99f26feb4e
e9638a66d7107e311c510eeb6011a54fd0db4b03
6716 F20101203_AACGAU jennings_w_Page_028thm.jpg
fb263c539fa42cb3d180449d97700b64
027ffbcede33be3ee93b5ed7306822030a9d27a4
F20101203_AACFWC jennings_w_Page_107.tif
81a843aa60a91898d244f19e592f84e9
ad42f5f2ef833c7f292e2a9d9f4c4a6f94bbf566
F20101203_AACFVO jennings_w_Page_019.tif
277486bb3b8aaa3527d62a65cb5ab21a
2174d8cd6ce348306da1e7af685072c780953329
49770 F20101203_AACFUZ jennings_w_Page_135.jp2
7e16686029b9721ef51b9aa00ce21d45
4f70ca7b4dad2008bb0e232e2df11659c0c8b151
25898 F20101203_AACESM jennings_w_Page_040.QC.jpg
f08b40f846aa06bb2959161da8decf4b
4f40bb663ad1a0c3a3a64138cbc21da2dbbe2590
F20101203_AACERX jennings_w_Page_179.tif
3be1e98671d0326d7b45550999e94df3
7f7c03cc88611ab3396734a504f0348a2f8881d5
F20101203_AACETA jennings_w_Page_033.txt
fb4c92ce0d45f4751bdeed32ef21b20d
ebc04cec5cffa2325cb872367731d11eda377899
7087 F20101203_AACGBJ jennings_w_Page_065thm.jpg
2094c30209f5ce10ba5b7d0405bf6b7e
5a789a1f5003ab681bebed813b7dec9151598136
24125 F20101203_AACGAV jennings_w_Page_029.QC.jpg
4124e1630ba1c99569976b067a54454b
47333738e23590ed084b285294ddfbf5fcfafd5a
F20101203_AACFWD jennings_w_Page_113.tif
a67db7b8fec49f1ed9f9c647035f8cab
d0fa1fecbeb0cccfbf96734a64743ca487d1597f
F20101203_AACFVP jennings_w_Page_022.tif
5b6106085371b08ca989fd2703d950b0
a6b89a9ee5dece1a169427278b5f680108c0ea30
F20101203_AACESN jennings_w_Page_036.tif
47637929c75c56659c9a1f30637b39fe
5df974397a24d94efe843424de24e9163f81238b
27185 F20101203_AACERY jennings_w_Page_117.QC.jpg
b40504dab057a4f09edd3b4270e33e0e
dc3bac773a32c2daa7c0a1f5c0282d38d27a256a
F20101203_AACETB jennings_w_Page_015.tif
7520026065cf75e1774c3e2f2b343a23
14e7225ec30217bb1f141a279d5b037fbc4ac58b
6728 F20101203_AACGBK jennings_w_Page_066thm.jpg
c40415ea230197204653547f7c2aba71
d907f974363108150bf9bf9b6f2b11c473516864
24208 F20101203_AACGAW jennings_w_Page_031.QC.jpg
d5796aa4d9fb96dde2d364e4ebb833b1
3a0b3f1909ddf210e2e84115cd8a000552fb07eb
F20101203_AACFWE jennings_w_Page_119.tif
7b0aead22e8ed2a492c42aca8be23fa4
d53e61030251d2ea740a9835e3e71ff923c5ee4a
F20101203_AACFVQ jennings_w_Page_030.tif
99bee15fbd6c811e9edb31a9d72cef0e
41916af237d9593133afdedd37d972e7805c5311
F20101203_AACESO jennings_w_Page_065.txt
c4a1f09ac55844665722afbe59f71188
cfbc9ab3b743d7371b4301ac6bd95f4611b48249
71764 F20101203_AACERZ jennings_w_Page_042.pro
3233afda70f712b8b57111cb8e86ebb3
8801fdca97acc5a316b0555fcdcd395f3e06866a
66590 F20101203_AACETC jennings_w_Page_117.pro
51b60655acf00724d9e2f9afceff53ef
061405f3ae4cf39b0e42b0ff14b8ee3b5ece90fd
25048 F20101203_AACGCA jennings_w_Page_113.QC.jpg
040320d050913edaee445a84d84aa37b
a5d5fcfc7fe44bba7f507548268d756d84b4c31a
6884 F20101203_AACGBL jennings_w_Page_071thm.jpg
5a1417b58372dbac5f4cd4b9dd158ce5
3e79712c4f2c7e2b33276f125e691f3c5d2d40c3
6395 F20101203_AACGAX jennings_w_Page_031thm.jpg
fd6a858056007e5050d8af1ebfc4de82
187afb5949352920666c202eb0ac370f4235ec11
F20101203_AACFWF jennings_w_Page_121.tif
a2fe4eebaff9d895b8e9ad9eb2789259
83118a6aa9fa498709f5350e67f44519e073d80c
F20101203_AACFVR jennings_w_Page_034.tif
7dc27845dc48d30867d30ff28020553a
12002bebd3abe9bd98ab20e4766efc2e6f44a5a5
6538 F20101203_AACESP jennings_w_Page_165thm.jpg
a78d00f323703d0666fae298cef6bc96
ddf6454a496984cc271c70c3262b82fe80d0380a
77069 F20101203_AACETD jennings_w_Page_054.jpg
dcea627cdb3300814814381823a3c3c9
6d3a4d100a52ee91a5a5ff8f529c11b69d070084
26187 F20101203_AACGCB jennings_w_Page_114.QC.jpg
578959b646201dd4d929dc2d56b1317b
1917dd1ffba06b00289efdd720138e8c56c91640
24002 F20101203_AACGBM jennings_w_Page_074.QC.jpg
3aa9a9fbdf86feb1835c713da230a6ad
0853325ba30c0b90e21630795574da373b79079b
24981 F20101203_AACGAY jennings_w_Page_033.QC.jpg
ccaf1b17447d1fc8bf1caff400d9b0d8
a8c32ec3ad055cda7e43afd2ad8179929350f433
F20101203_AACFWG jennings_w_Page_134.tif
0faa57570f95b6468b1ae1ccbbeb4954
8657e50e98df1d173c326e009203f1c52286c266
F20101203_AACFVS jennings_w_Page_039.tif
8208b8e3cdb62e358e3eff54f01e0c84
cdb9547971857753f1f7f85311a2306f4f3a9e43
79361 F20101203_AACESQ jennings_w_Page_114.jpg
8d116ea149d3d39618b47665a80e693f
7708605acb897839f1babacc22353ba8886d98be
F20101203_AACETE jennings_w_Page_067.tif
55d4331b4ed27875838f7d058f724fb9
f909736cf59119f6e02b8b0cc8706615fa7a4790
6766 F20101203_AACGCC jennings_w_Page_118thm.jpg
720b6b4897866f150eaa455e0f32aed0
9b86281328f07a5136149ca38e46321a2f923d49
F20101203_AACGBN jennings_w_Page_076thm.jpg
38b63fabc61b20a0c934db855b1358bc
003e58119435e3e2cd3a1ecb695424d04e4c3592
7081 F20101203_AACGAZ jennings_w_Page_035thm.jpg
1287bb4f88e7416e93962252f2d1731d
1241cda2db91ca5de440203283ba63ae4ee45436
F20101203_AACFWH jennings_w_Page_135.tif
9f3801d8acd0eac95622444582187a12
0c075ccbb9d9b866b5b940608e1f7698f2ed1d02
F20101203_AACFVT jennings_w_Page_041.tif
e1857f1ec750c7beebb24613cf7d236a
84161609632c4792e937234011cfd999c265acb2
99487 F20101203_AACESR jennings_w_Page_080.pro
9dced43b92a1004bff175360e4b0ca77
e94f2d03081a943b1433a9a59499b26e1659941e
52554 F20101203_AACETF jennings_w_Page_105.pro
c83941d01c6143fec8c68b7cbe29434a
13226cb42a7a6090c6e6665408f0c81c08e0d209
7024 F20101203_AACGCD jennings_w_Page_123thm.jpg
e1e25b20d59a104344f8b81382005cf7
3729333730fe1b58b8c378234cad4d26491b3afa
29579 F20101203_AACGBO jennings_w_Page_080.QC.jpg
5c9fe9c2130876629aa90cf1fbcafde8
7dab0cfb8da7a1efd3ac1dd41f33a809c59ffab7
F20101203_AACFWI jennings_w_Page_138.tif
fd7bc26e6029900ff92dea566ce5cd3f
a184bb8bd29010df85de9f77fcd9485d9a8d82dc
F20101203_AACFVU jennings_w_Page_071.tif
eb1e0cbd710f455638b0f04a3d53c0dc
fbac33ff19cba5cf183f62890be26ce45d8c5eb6
F20101203_AACESS jennings_w_Page_098.txt
c544ba0096f68f567b1cf5a26e0d7ed8
93aa3b4fc8670a5d32e3c0eaa2b077458337b6c3
2218 F20101203_AACETG jennings_w_Page_136.txt
1b2a5c056c61e55b35b10b67571b074b
806971d012678543f6bfc12e304b8a5a0207a279
7214 F20101203_AACGCE jennings_w_Page_125.QC.jpg
734644147e53791c670f7a001a04d696
25ec8c3c7a67a488649dc633373ac4db682dc015
7130 F20101203_AACGBP jennings_w_Page_081thm.jpg
acb4f846e9272947c1598b3f712174a8
59f96bd43083055fed6735bc477c61bf2899e611
F20101203_AACFWJ jennings_w_Page_140.tif
cb38170fdad27424ca2dbb44794a3ede
d58ddce8bbb21273c9d6194e70defffab36194b8
F20101203_AACFVV jennings_w_Page_081.tif
fcd958e208b51fad071e48ec3598e413
8ee3bd4565fc8e8f84b033d1a03a5f03edaedbbf
78878 F20101203_AACEST jennings_w_Page_145.jpg
727e091234778913d7ee9fcaaca3ce41
04f89c3d3dd5c5301e20a55279d8b82c1416c8eb
79570 F20101203_AACETH jennings_w_Page_026.jpg
f6234c782edb8974d43edc2bbba212d6
c77c0b1bc644fe5c6fbc141c4d955de58920dbf6
4434 F20101203_AACGCF jennings_w_Page_129thm.jpg
4174bab3d6f0c39cdf105f43a0761800
7bdf5a9ed03d0430d416cf4393824a84f9b523ec
25230 F20101203_AACGBQ jennings_w_Page_086.QC.jpg
503d67221ed51f45a3201d3c70d86c3f
668dd1765514db9559f873b1965d9ee1ca0b8170
F20101203_AACFWK jennings_w_Page_144.tif
aec0273e73c3baeefcbf324ecc8e43c6
4d731e6c15d3f89df23434ae31ca2bf6071f6360
F20101203_AACFVW jennings_w_Page_083.tif
3346f8b7eeee6ed896e38c1400a578fb
e7dd5e0cfee31285455e6a516ae9c91e6d5ed858
23599 F20101203_AACESU jennings_w_Page_053.QC.jpg
44c7d32b9cc3cefbb35a13916549e421
1551681ce5887d8859faef3d8c00bc1cc317a1e2
75513 F20101203_AACETI jennings_w_Page_071.jpg
d45d9c025179afe024b3df6b2e0f088d
bafd7c924a3ce222322110f60399c638b8f622dc
23049 F20101203_AACGBR jennings_w_Page_087.QC.jpg
49ad86fd7628641e6b5dc349353b0f10
cbaf2e65612df8c7ba97407eeff884ced8ec8e52
F20101203_AACFWL jennings_w_Page_145.tif
6e3cd4422632104e7d9a791bbd0ec982
4e0983e370b81dbfb75f21be54363c9e0aafdd4a
F20101203_AACFVX jennings_w_Page_086.tif
97ec59b86e3f4e2563aba5bbd2aa0cdd
81d09407f369c6fe6f84eafd90ae8d9a3b478fdc
F20101203_AACESV jennings_w_Page_112.tif
ac1b9dd153706156f418c94318cb5178
ea0230f33084c62ae6d089aa75d7275946e01070
385 F20101203_AACETJ jennings_w_Page_140.txt
70c1dd5e61e757b551066cf7dd6c078b
f31126bb06986b7b753c52df7035234de99b9261
6813 F20101203_AACGCG jennings_w_Page_132.QC.jpg
73124ade0aaf951e4a6bce1cad44a77a
c3af2320c70d43ea1644f8a81beea89ddbf5b25a
22550 F20101203_AACGBS jennings_w_Page_088.QC.jpg
da2e6fc8233be9e70401353c060ac525
c372a1c9b60e657d96c54852a93a27477b2af43f
54849 F20101203_AACFXA jennings_w_Page_012.pro
83365b05e52e13b2fc024241633f40e9
6bbab7bae9e470d5ec46e3798d7b648f9f1ccb4a
F20101203_AACFWM jennings_w_Page_152.tif
59188eb67a7d1522ba17c2c4e3da5753
6c144f50840cc497fd06bb1418b5275b37a6e8cf
F20101203_AACFVY jennings_w_Page_089.tif
bdac4d698638b2a7447ce2f3154bb4d4
db3dfce11d414c57cd841b0562fa6c2b8e7d8a7d
132135 F20101203_AACESW jennings_w_Page_117.jp2
afbd8457137d2dd3d3e78ecd55214693
c8e2b4cb3d90eeb027911a807204d1fb0bb7f43f
37125 F20101203_AACETK jennings_w_Page_135.jpg
2cd9f7126cdf488d9058c300c256d1a5
edf433b088427d9968cccfec4f38d5c7315dba26
15484 F20101203_AACGCH jennings_w_Page_136.QC.jpg
e9d74b34522503afbbacd21d6712c545
4048fcc60fa5c67c1e54c32dc14d9e645e4a1aa2
6520 F20101203_AACGBT jennings_w_Page_092thm.jpg
3c565921814184c8ff15c39c80bcff81
22d5f2cd53102afcd53b4758ff4d3bd06b68e405
52374 F20101203_AACFXB jennings_w_Page_018.pro
0757f50776f2dbb1ab7a06fb08655890
cffb220a0c304da94e577d21a123edb9a056704d
F20101203_AACFWN jennings_w_Page_154.tif
e4722ffe380eea2029311fb92dfcf861
d96ad6ea199da50f6c513a6e57463e4f7c6abf3c
F20101203_AACFVZ jennings_w_Page_091.tif
7b42d7379a2368ec6e85f6614ddb61af
d4a3e17f2da036df05484cef719f45d5a7cab176
21273 F20101203_AACESX jennings_w_Page_132.jpg
2c58290887bb4e72df1fa7044434d1cb
ad9a7dafbf5c8ae7ed949b65f41b65cd183543f5
133501 F20101203_AACETL jennings_w_Page_169.jp2
7aab79a3043fcce522d5327c95169336
f41b1f40e853bbc5b2d7a9e82763e1e3050e6606
11461 F20101203_AACGCI jennings_w_Page_137.QC.jpg
5b0aeced385415cd8eeb6bd8fe99f79b
7d1198312a808a09ff555a6d59bb08254cb14e00
24189 F20101203_AACGBU jennings_w_Page_097.QC.jpg
7c58942010254a54a39706daaf324cb4
c3ad8e30f2d883f9607b101472df30b1c889cc28
55608 F20101203_AACFXC jennings_w_Page_022.pro
1c7459ccbacb5056d211726d6d4534fe
04888b4bd17713ddd99f2deb5823e2e58d83018b
F20101203_AACFWO jennings_w_Page_157.tif
26e9e0fbc310b829a995fbf6a756e2fa
a66b4cf3e04aca8b943dd8b0ebd80fe1dba363b7
F20101203_AACEUA jennings_w_Page_099thm.jpg
5b9eecb0c8e01d3b94306a89b138014a
4601ded31692983cbce8d587e78e6b74fac4bb61
121070 F20101203_AACETM jennings_w_Page_106.jp2
6ef121ca925c76dfff7480a8beaeb991
d0120beba5bbfa0e9004fb127eddf05a55cde1c7
7951 F20101203_AACGCJ jennings_w_Page_141.QC.jpg
fcd5c55d58529b36b9869631816c1882
729ccb1371f36530a8d51d4ca696bda5fbe0558a
F20101203_AACGBV jennings_w_Page_100thm.jpg
c4ab15f38d9cc99258b95cda6260a3f7
178ca6f9e4511e184f4263a878dea686e3a9e8e1
56081 F20101203_AACFXD jennings_w_Page_026.pro
ec3403b1bb10cb61573d0d6a34b1a192
f287039a1e5ec624538d4a92e63807ddca086962
F20101203_AACFWP jennings_w_Page_158.tif
70152e5aecc89aa1a0cdbb64bb246b4d
d62d3f8a9529688dd30b9add33c7bcf2bd2b33a4
54308 F20101203_AACESY jennings_w_Page_149.pro
07d5be0e8a4b64e52337a17c85c43aca
b26f79c27736fdb488f809dc2eac71e510aed4f6
2226 F20101203_AACEUB jennings_w_Page_037.txt
8a5a2c9ad6379be7f95d5ce3a07c2202
19ed8635369c9e8ed935cc93b6320c4ed4034a65
57777 F20101203_AACETN jennings_w_Page_083.pro
f4044bfb08ca93482959bd7b3e46c01d
129080a0b8cf7d44a651cec6e9c3a73cdec299ee
8552 F20101203_AACGCK jennings_w_Page_142.QC.jpg
6de6a1ccdbf48912795e05a104482f81
214ae6b0d766ab801b03da0a43a10438478eb446
25245 F20101203_AACGBW jennings_w_Page_101.QC.jpg
f6efb653b4de8b07f99267814a79f2b9
38ea9232731b2f628905e75061b03bd657b5050b
52740 F20101203_AACFXE jennings_w_Page_029.pro
e96424e8478f39a85e4b753fac30860e
cebb34b23f154f4ce63831d54f613b47ff630901
F20101203_AACFWQ jennings_w_Page_161.tif
ed97b8804fe2762665821f2f36983c02
3fe9280826b24e14f984b93afc5998db0320c67b
7939 F20101203_AACESZ jennings_w_Page_140.QC.jpg
dca8f85649fc5f9c4a9b1adea7c370dc
79599871b68a1068450a3a0aeab95cf046fb74de
53601 F20101203_AACEUC jennings_w_Page_146.pro
1c0d4918453eb92702160eafeafc4c7d
5f0e7ad19d35d48364467333877861ac80c75ca2
116135 F20101203_AACETO jennings_w_Page_104.jp2
561fa0193e680e39425fbf06ff19600f
a07457d8dca6efc75272f2f72e3dac4caa95ffdc
25878 F20101203_AACGCL jennings_w_Page_145.QC.jpg
aa531579149f7049ad76cf2742b78619
35e2e1c98228b051289892bea539f93628bde313
F20101203_AACGBX jennings_w_Page_106thm.jpg
8156dd0c16c37f9c9d0f817b16536ecd
ee7a067c961d50191d18cd517a1f3af52fdd8e96
53332 F20101203_AACFXF jennings_w_Page_033.pro
86ac54228f9fe62c2a8598ba7f633bd5
fe9c52d218f03c91d88fda652b42415006423339
F20101203_AACFWR jennings_w_Page_163.tif
bccf02c6f6a22efc40c5a4bf55c84433
ab1c10c37be4c9fadb773fa34af9bcbc0f8c4837
4649 F20101203_AACEUD jennings_w_Page_126thm.jpg
4845a1d6991822edc1cb883a450d1d71
aa0d8fdc0507cb7ec9cf1cbee3dbdd5530c769f9
2079 F20101203_AACETP jennings_w_Page_120.txt
4ec076fc475eb35407cd2d31d2ff067d
9379672253f6e9ff60a60065c42e4e136dbcdf7d
26493 F20101203_AACGCM jennings_w_Page_147.QC.jpg
9ca798efee258f4f58cd385036ecc896
9a764a74c2da2962f4af7234a0025bf5883fe2a1
24701 F20101203_AACGBY jennings_w_Page_108.QC.jpg
ccb4a0b9c8339ec11d773091c0fb95f6
4a289d11757f0d0ed655610d6b194dbe85578701
59133 F20101203_AACFXG jennings_w_Page_034.pro
68781e027d58852591573d96b12f8dcd
ecffd18d315c2a18359f5b3213e1de77fd29140c
F20101203_AACFWS jennings_w_Page_167.tif
3ebf3a5862c785a076d819af9452dead
6a650f6007e398f20036a45cf75c274b823c0e2d
3699 F20101203_AACEUE jennings_w_Page_069thm.jpg
1f1ca442010a8cac0ce690aed5355abd
c46e474b85c39247b5ae1e8a7684c571540c921e
5636 F20101203_AACETQ jennings_w_Page_002.jp2
3f521a83994540c7f2fb354c4564bcf1
d88df4c5de3b93a9e2546f6978fd8d4564259f89
25312 F20101203_AACGCN jennings_w_Page_148.QC.jpg
cc2e59d12acf660a4535ca08ac279d49
ab74dd87c58d003e9c0bbb5e349e487445682f49
6832 F20101203_AACGBZ jennings_w_Page_110thm.jpg
49d6cb029ee154afdb1245b58ea5336c
51d41ff1f4d28b3a9a00d7b58b47c35afb2ad21c
55749 F20101203_AACFXH jennings_w_Page_036.pro
566c85744ba0b73d71789b0c5c39ada2
1d7a36d7f1b2c06665a3ff5b37fbf7bff56b5952
F20101203_AACFWT jennings_w_Page_170.tif
6bc3e4d7afcda0c02aaa7a648495fa7a
4c9e5f7f304c10ff6a4fd5c524bfbc31495784be
24859 F20101203_AACEUF jennings_w_Page_104.QC.jpg
e7637f8b0b96ad4918924b523986350c
9741ee287c9cadae3bec34e2ad339cb3d36c710b
49653 F20101203_AACETR jennings_w_Page_137.jp2
9c87db498b057ff17aada61b26ee81ac
1bfebdd2313b6cfe7a86b1d370d9280587bf7b6e
117474 F20101203_AACFAA jennings_w_Page_048.jp2
f3933917b0b45413154145a54e9ea086
1d346796b71489d0c5309bdaf14446710aad926e
26293 F20101203_AACGCO jennings_w_Page_160.QC.jpg
69d9860568a9c85fde8effe37134532e
d88cdaf412afd23fa4d1594b81cddb632cffa8d7
59460 F20101203_AACFXI jennings_w_Page_039.pro
6d0ead3e68e671cd5903d7b068f41c90
cc247c1b633c1958da8f561f87719ddde2ccca62
F20101203_AACFWU jennings_w_Page_172.tif
6c2db485a150e6ca2551733d379a43ef
a4f2a7e1024a22446f243ccaadcf2d4f38a8f7dd
26344 F20101203_AACEUG jennings_w_Page_047.QC.jpg
046bdcddab5b1dd664dba67b2f4ce74c
905b3a1856959117a4d59e9b1466039516704031
121942 F20101203_AACETS jennings_w_Page_123.jp2
69f36b71986b089234aa1e860137b47d
2c6614d23cc2cfb71af1d39cf413e053cfe4fef4
64853 F20101203_AACFAB jennings_w_Page_170.pro
603ec229af219609d7362984b3474527
b07fea203658e9ce5a8725fe547ca3f6fb2cb941
7093 F20101203_AACGCP jennings_w_Page_160thm.jpg
38e0a6993a30023b1e4b88517bd2eb3b
1004d737b81363c28ebd6ae5680a6f2ba8468408
55541 F20101203_AACFXJ jennings_w_Page_041.pro
bf46db7f9ed857229f6d5973caf7342b
18e16df8871739616fde02236b0413e4fe517817
F20101203_AACFWV jennings_w_Page_173.tif
6583d559481def7e0c7c348e7048543f
6f77c83bcac28e454a28c55b5e830647a2db7ff2
25163 F20101203_AACEUH jennings_w_Page_059.QC.jpg
3a0b76dbfedc80c2be7f84bad6bc99d8
4fee3f695ee8b0d2ee30849dd1dda9ceb9ac881a
6953 F20101203_AACETT jennings_w_Page_051thm.jpg
302b13ad3f9c3d3375abdbc067331e1a
87d5cae1ff75e072d1f6971f2e1c95b58fff5cd4
6798 F20101203_AACFAC jennings_w_Page_098thm.jpg
1216e32caac1a1fdb4c3113ba702101e
ef5a19999d1889385db10a275a02e6b7fd77c0e9
7123 F20101203_AACGCQ jennings_w_Page_162thm.jpg
65c0ad2cbc67c7af4cf76d69b40c8b83
6d33366872a61a9394ea91bb15bed3dbbaeb3e9e
40143 F20101203_AACFXK jennings_w_Page_044.pro
3058693b33c425586bc660af2b7a8ec4
78d29be7faae4aea9e15ca43152b71448d68fbe6
941 F20101203_AACFWW jennings_w_Page_002.pro
768decb982654a640d460b14f4d0b189
e01cc91d6465933fa2c764c382ea84bb110a9ac9
26332 F20101203_AACEUI jennings_w_Page_037.QC.jpg
73119525905ac96bf6419effa4b6226b
231b6447aba452bdac1053474f5617cf8222720f
66911 F20101203_AACETU jennings_w_Page_102.jp2
96502d86a2b07d30f1c4da25f604fe48
b223b0f1006997659d3b15bbb172f0b1f2454090
F20101203_AACFAD jennings_w_Page_181.tif
329e7d99d993003c59a1ffb1eb4ce5e4
92de9831f762ed57db161d66d85ed9806effa757
25949 F20101203_AACGCR jennings_w_Page_167.QC.jpg
05d2ddd3cc30477476f56d77d73b8e99
b2643abcc998d11ab747675bc064a3e24b0a19de
58088 F20101203_AACFXL jennings_w_Page_050.pro
55e368481b3fe0503384f21eea3e20ad
7a26292c8aba61203b5e409b9b3ad0b4c94de406
31789 F20101203_AACFWX jennings_w_Page_003.pro
69ddb41de5468ceaa9a12869bf5509b9
be606a33a42ca953c2f6f374d768513f710989a3
2539 F20101203_AACEUJ jennings_w_Page_176.txt
fb93ccd8de9fb58fa863a756731f8faa
163521611810bef09cb58f88fa2daca5cc9883f8
111769 F20101203_AACETV jennings_w_Page_038.jp2
847e11d3f0349993a2cde920aaf73689
f50ea58e3e933663f2603a544ee69a8f9d64e16b
F20101203_AACFAE jennings_w_Page_060thm.jpg
a74adce986b0724851a4e2b0c83d298d
01f5e0aa1926757806afba68d1963c2c2ae509fa
25784 F20101203_AACGCS jennings_w_Page_168.QC.jpg
4fdcfd82e3fbf801986ab58d3554c51f
5e06c1e0f2dc582ba374383625da8b6e91263249
55205 F20101203_AACFYA jennings_w_Page_145.pro
b9306b600acb2d719c6ccdb5f06028b6
f07b72c656fdee50ca54177c2985513abbfa2e35
51285 F20101203_AACFXM jennings_w_Page_053.pro
d0d67be2d641d1917ed2faff60be1a33
f09e4366dd7b14c29fc56c0536a995cfff2e3ae2
90675 F20101203_AACFWY jennings_w_Page_004.pro
3af31901f1453f4c5915369566dcae44
6fa7d9dec737f82258161da54cf38c8f3747a3eb
F20101203_AACEUK jennings_w_Page_106.txt
370607db715c725605b76fdc86e2f84f
b6ca3abcbc04bef0e29ef30a25a3ac2395addfc0
2601 F20101203_AACETW jennings_w_Page_160.txt
313b1bcba23b3c7f56f99df2a9e4573a
56bbb75a1b4352eb47a6479a13ef57e3d4c252c9
27730 F20101203_AACGCT jennings_w_Page_171.QC.jpg
5441bfe0975fcc41b797441f1c4d0e4e
debc28a38f2b68f70236727664d1ef0b921615c1
56151 F20101203_AACFYB jennings_w_Page_152.pro
27f46cd87b07be4a8f8daaf31c667688
c8bd98c012ff230c12f5a92fcbe0c05767e85a22
55434 F20101203_AACFXN jennings_w_Page_061.pro
35a6252863701ac0ce4db9f2c76a7f78
b0b34be432697b037534f7b3d052c4ad9df39e1a
71187 F20101203_AACFWZ jennings_w_Page_006.pro
e88b988d077f05f333510fd6d2c1dff8
2021578d4857de09911613ccf90db20977bb5fc3
F20101203_AACEUL jennings_w_Page_128.tif
246afdfc46dfd18df3da9a469925cec5
0fca726699a728013b5f19431f0b26f3b1ea895d
32049 F20101203_AACETX jennings_w_Page_157.jpg
f034e0b0dd8eb56d94ae5d46ba671e57
b5f6a90456a70e97d2370538674d59406f48315d
6733 F20101203_AACFAF jennings_w_Page_120thm.jpg
f34bbb7f40727145d39b6fc3674e8786
0b3e2fac7449cfa9cac90e4775967a5ccdf93fa9
27152 F20101203_AACGCU jennings_w_Page_175.QC.jpg
37e4d615aceb89b68368854504fe958b
9142f7db1915e4dafdf16cf99039e284ba4fff6d
15438 F20101203_AACFYC jennings_w_Page_158.pro
866188d465358f9b1dd2e0ff9175d579
8ad6cf8ecc73055c255c9d803eded03377d234cb
54722 F20101203_AACFXO jennings_w_Page_066.pro
3f8c586fd275bf0518481fbcc77ac5c9
f225a7c4cf1ebfd3f168cb2143e15e09cd9d849d
19317 F20101203_AACEVA jennings_w_Page_044.QC.jpg
27179af5e885ab18be2944ad54048619
49826f98212560ac8728e8fdd3751bf713427435
1153 F20101203_AACEUM jennings_w_Page_132.txt
f3d63c56a362d9dd65b779f912f77277
b6b89c4050c13da775f1ee545a9d60eb615d8632
73563 F20101203_AACETY jennings_w_Page_072.pro
dd28f8b8a3b9ee98a9533920feb75666
9af1b71914de7b7886ecc976245cb2929237ffd8
F20101203_AACFAG jennings_w_Page_115.tif
73011bb9819ac047724f94986a4bff34
bea4b123378a0b049409aadd86d62abddd07cb14
26313 F20101203_AACGCV jennings_w_Page_177.QC.jpg
795702869f89e2ba5bc49ce7f471c6bc
3aef080fc6b9c23c0ca5ebcbd6d65cbbc94ed3fb
65990 F20101203_AACFYD jennings_w_Page_163.pro
9c187df876a7507af1e45a5ed898af42
d7d920edc8f51af7124e08d1c6de4393ed2bccda
53561 F20101203_AACFXP jennings_w_Page_086.pro
5646dbd4073bbf32d703000b5ad13d71
27e628c72bfe2e1f38e30ee26d4f063d02a5310e
26819 F20101203_AACEVB jennings_w_Page_115.QC.jpg
e7e708d31e415d04c30ea9abbdd8f2c0
563eb0abdc3099f42afe7c9e6e33da7a3dd9b47a
2126 F20101203_AACEUN jennings_w_Page_074.txt
f99d314ead6630e792a575d446af31ad
8f09189063ac298500530538e51981e603cba68c
2129 F20101203_AACFAH jennings_w_Page_100.txt
f399c607157ded4f3b939f476e8ba66e
5210bedad384768f260f0d4dd14f5f52cbd98207
61729 F20101203_AACFYE jennings_w_Page_166.pro
19243b096d69f4aa690e35b1aaa59a36
cca7d14879f3e4a25c955e9ab62fe865aa259a1f
46087 F20101203_AACFXQ jennings_w_Page_088.pro
cd297fc53a5c14f9c2ae0dcb4bc646f9
6c4ed7e3d1e0faa10d3aa6c5a33a534aa1ffd8c7
114242 F20101203_AACEVC jennings_w_Page_013.jp2
f12868133843b5970b13e147acec4e7c
3281947bbd35b0d397857b11de1bca2f6d6c606a
25930 F20101203_AACEUO jennings_w_Page_014.QC.jpg
b1d3811ed64a9ec223e8087d59744ddd
a5b1e5ce5a78bc344bf345b8b97d1220ae66924b
26448 F20101203_AACETZ jennings_w_Page_001.jp2
545ff0d8172682e4b04c0240018d217e
22c6ec97738133caf4cd388c61b5b43941c7be6e
80844 F20101203_AACFAI jennings_w_Page_047.jpg
0f1b814a09e8e84d54fff9dda8d84a2d
91dff24abfad83f1769108bafbf0c90f41cf41db
3789 F20101203_AACFYF jennings_w_Page_004.txt
3316cbd67f0b5e3ddf307587c0aa49d9
6ebfbea98c96927c428719479c78ce68a85fb7cd
52670 F20101203_AACFXR jennings_w_Page_091.pro
a0204cd64b6f3b37e037fd76b1670caa
419ea2aa3ccd3c0d67d9b74f51ce6e1fd8bee2dd
7066 F20101203_AACEVD jennings_w_Page_124thm.jpg
9a629d81ff76e5e94312d099b6fd59c5
bd71c3c16c64cbddf29869d07828e2c707a384f4
54796 F20101203_AACEUP jennings_w_Page_121.pro
f4759f495b6bf75ae15af854b3992909
dee5c5b22dcff0807b3ef29d02c21c8ee23ae610
112607 F20101203_AACFAJ jennings_w_Page_045.jp2
3292d4f0dd6cfd0f6d5370c9bfa64092
af8e3c26b84f5bb72435bcd766362b6f5e869be5
1117 F20101203_AACFYG jennings_w_Page_008.txt
0edcb21358e8dc628747eff902f3e8ed
992de604a58f92e099c1c22fd3355aa39625bb21
54725 F20101203_AACFXS jennings_w_Page_097.pro
0507916c95097dd90d1ba701db02f3c6
6cd6aea05f572b8fb3ccdbae975c6d0f936f813e
6781 F20101203_AACEUQ jennings_w_Page_006thm.jpg
cb443298f4f072cf154000fd0d05bed0
1f45ca0a58935633b3b02511c44f93f62359107e
639434 F20101203_AACFAK jennings_w.pdf
102b5cc9589320e05e28ba0d48870d4c
d887b7f9e67987ddf2e235b0357aaf32e154b055
F20101203_AACEVE jennings_w_Page_177.txt
6e1dad91412766d2007522b6697ff7cb
2648dcc12e65d8af2c32a15fca259e2a8e7438c6
2169 F20101203_AACFYH jennings_w_Page_014.txt
6f0457f93e29f76535405ba45e247979
05477720d363ea89c061ea64b3bc61ff6444ecc5
31545 F20101203_AACFXT jennings_w_Page_102.pro
f2fa1e75796a242434f089b679c8ad19
c84ac165c9154bf2e397078629d6ff73881bcce8
50980 F20101203_AACEUR jennings_w_Page_095.pro
67aa64998a6a6b2a9f616b808a325117
559eff3fd0d83e4552521511d0d703eec93b4ead
F20101203_AACFBA jennings_w_Page_065.tif
4733a38e132f3f3c60bf6345398337ce
a3b5e6852a516a805869a980b712e37ff3e66292
25463 F20101203_AACFAL jennings_w_Page_100.QC.jpg
92dc8454913371428464b150157e36c9
67e261665cca7059e3ea2d40aa405540e55070db
104442 F20101203_AACEVF jennings_w_Page_011.jp2
983cd918bfafb6a7b8d5b9324ca9b5ec
f0920152bfd5dc2241b22777b843056a73b255e7
1931 F20101203_AACFYI jennings_w_Page_015.txt
cabb64910423c81177481b0b4a6a01d7
059253d8550cf23ef7028ed4150011331c2a6b6b
55694 F20101203_AACFXU jennings_w_Page_111.pro
2360da891565404dfe591f46b5f7bfee
7816e4c946581d2a12b095212f43b5f179fde182
F20101203_AACEUS jennings_w_Page_058.tif
d697c186e470dce6687f06bd0f4eee38
a080babf34b80f72f99d753ac4d874ed8a8400f1
F20101203_AACFBB jennings_w_Page_114.txt
c3c2a5ef1a3560c03e81f242e503c581
38cf2d9850a35d22bac873034bccfb42a8422b9f
46761 F20101203_AACFAM jennings_w_Page_127.jpg
871a078a0c532583fd6cf53271fb5ebe
6e0239465cd6686dd9900d0f878924f28213f95a
2015 F20101203_AACEVG jennings_w_Page_095.txt
0725b276d108d4541c38e248857cf233
0f669d4d2a26d2aee9e680044da453ac5c869fce
2055 F20101203_AACFYJ jennings_w_Page_019.txt
758523203d416f769479c27629081e9b
8e34e0f360013a4edf92cd37c670434f40abcbd4
56783 F20101203_AACFXV jennings_w_Page_112.pro
6ee8dc8905f07fe14ed14161807cd1f3
0f36be430e24190a2c421b7fff8673bc9e078998
79591 F20101203_AACEUT jennings_w_Page_012.jpg
773896f031fd753658c49233a98c7977
45f2cf8a7721d5a2d035cad80707ee4d50331a4c
24492 F20101203_AACFBC jennings_w_Page_058.QC.jpg
671cfc58ac61e057b7da2bd88b004581
49445501a11fd3ae8b55b204c010bfa3ed7849b3
25629 F20101203_AACFAN jennings_w_Page_121.QC.jpg
aa1b5ab11c20a5879f9309ddd4d5b52a
8a527086ff918657c9d9b988761a551f1f9367c1
6810 F20101203_AACEVH jennings_w_Page_139.pro
c3dbd794cc41ae8a68ef113e464e9528
964cd3fdf6a52c46bdacab40c34c929b32de0f09
F20101203_AACFYK jennings_w_Page_020.txt
deeb5b83ed5d699f6b61e108d2aad352
5300bbc837034335f4ba2b79addb8ed971b663dc
51599 F20101203_AACFXW jennings_w_Page_118.pro
ae0cc0cc38ba60d93d6776a9dadf1ff7
f9f0394206c68e6126a5dda0bf2c7ea2fbe6eaf4
54541 F20101203_AACEUU jennings_w_Page_064.pro
a6e638b6385c9143d3cc3956ae6905ff
7aebcd11e969d5e021d0eb5e59bd14b576393e49
7004 F20101203_AACFBD jennings_w_Page_148thm.jpg
7aefcfeebb27cdaa87c8c2206407f7fe
aa7ea3836efbe870d76022698ea5d07525cebd0c
79678 F20101203_AACFAO jennings_w_Page_111.jpg
e4a618bf6b27890ce3abda986126ac0a
6cde302ab788f50f93bf37c33bf65e7144334d17
F20101203_AACEVI jennings_w_Page_025.tif
e8e248eb4085e815703219c1261c5031
b0eaf753e6c568ecbd45f7ee9f13bfeb277b01f2
2197 F20101203_AACFYL jennings_w_Page_026.txt
6c226e8e8a135e197413997b7c5dfb51
c0ec106bb73db78751e964cb473270f3cff620ee
37738 F20101203_AACFXX jennings_w_Page_129.pro
72894b3fd4dc95e1f5335e9b481237fe
6503516f9efbf02fde8e2dba0b7040256f123f8a
63989 F20101203_AACEUV jennings_w_Page_127.jp2
85858d4e6d482d3a1ff22f85b499e55f
25e7a704dadad005822f3eae02deab40d79677e8
2294 F20101203_AACFBE jennings_w_Page_001thm.jpg
1a5ac58ef0ea6e5aa9703310c8e02d42
8e357024e6f79a86ddcd65a4e7cee957c1285651
55479 F20101203_AACFAP jennings_w_Page_150.pro
079cac7661c3d0d375c6cf4db598a774
0e5203a111dadb69c22ebecfbdaaf4676ad7e727
F20101203_AACEVJ jennings_w_Page_079.tif
e8cffe304e53d8be6b21b31a839df90c
1ccfbd01dfe48a3ecacfceef57573283c8f9177e
2179 F20101203_AACFZA jennings_w_Page_111.txt
332366e8e6e871a3414c3034f7c614fd
3a0e23f0e1bc0d447f5c285db321fae2e051c3cf
2219 F20101203_AACFYM jennings_w_Page_030.txt
a528eff1577da5a4d87778b302b85f72
352a42244c0be22494cc38c0dee26126c8d1a9e6
44511 F20101203_AACFXY jennings_w_Page_134.pro
7ac887756420fb18fe94f2471d2cb20d
cd49e40fbf72a1878a80adcba419afeb82b83ca3
F20101203_AACEUW jennings_w_Page_159.QC.jpg
d39c62ac15abf1c7e129967336f546b4
b23f39a55a8ae851c1834a42dc0c943de1be2741
25890 F20101203_AACFBF jennings_w_Page_173.QC.jpg
9b5f4ef772b6c244e5df9ca12d6d92ef
a1e6b0ff2be1463d8b75fab1cc98852c36dcf4b6
F20101203_AACFAQ jennings_w_Page_025.txt
16a85603020582ea308d7e5e519890db
42f1d64df3ac29a4e6cec98fefaf0c224819f694
6755 F20101203_AACEVK jennings_w_Page_093thm.jpg
0aba4af603a20d044f07a9b426be1fc0
92915c5f5cfe9d6847ea8ef7bfea55c72516fd4f
F20101203_AACFZB jennings_w_Page_112.txt
aebe2f5a65ac52fa7063ac40b3ce97b5
bad4ddd7a573fa31209f586c44ba62a80c718a13
2750 F20101203_AACFYN jennings_w_Page_042.txt
fbd3c353e04cb09739d44609857381be
6583bdc414bf748aecc0ca91e2c9ae9f0ebdcc8f
31283 F20101203_AACFXZ jennings_w_Page_137.pro
fab5585a97b221fd18c47d52f8172bd2
63987cec56f41f025e495a7f3f990f98003acb63
95013 F20101203_AACEUX jennings_w_Page_007.jpg
1b519c4f2c64e521faacbdad9eafea3c
a008ac79999b74850fd3644af76c86996a4caebc
77270 F20101203_AACFAR jennings_w_Page_064.jpg
fc200b603d2b463ab2a564c8186c47e9
f8e4545b3e56e72a08833b7a51b9ac71f9188099
25682 F20101203_AACEVL jennings_w_Page_041.QC.jpg
73c4f3ef553b7f8600290aa2c9538889
c1494013a4eb02cc69d1c75ae71a81b6ba0e3a7a
F20101203_AACFZC jennings_w_Page_113.txt
477b94206eab434ecb6d4879fce052e8
c160cd1e4dc70de1942389704337b187a0571eb6
F20101203_AACFYO jennings_w_Page_051.txt
49d353455690d73409e22c74b69c1c2b
bca57b05916f68a97b416b961db2476f99f05ba0
117833 F20101203_AACEUY jennings_w_Page_079.jp2
ffc7d4520524f95e97f2fad51fbe252e
824ffde7bab6e2469a8ebbadc1f791655de22c14
138461 F20101203_AACFBG jennings_w_Page_162.jp2
aae04078630c73d03867af6733828bf5
8f6c90ecea982e29851ed7c94c638b07cc7b9890
57715 F20101203_AACEWA jennings_w_Page_151.pro
3ffe4df8462d25fc2387076c8791babd
29af7d7f39f568736386cda07c7d604ab1e63626
603931 F20101203_AACFAS jennings_w_Page_142.jp2
664345a95846b7ff0ff62a0b3b2313ce
1a241e380de2dafec407faa5567b6e857790fe3f
6956 F20101203_AACEVM jennings_w_Page_046thm.jpg
152b4071c94773c5b9d5665394c21731
82c508a3c7104d288e51e71b210dceb3c50ba505
2681 F20101203_AACFZD jennings_w_Page_116.txt
ff57cf91c523feee562ec3b7707d97dd
143f405a9aae051fc301f9af88d6477e32a2c569
2174 F20101203_AACFYP jennings_w_Page_061.txt
f433f7e161dfefccf18bb6406d215d7f
0b7e05553cf61c352634ecc326e3650b78aac6a7
6609 F20101203_AACEUZ jennings_w_Page_032thm.jpg
9506eed1b99e7c0a8dc1e227d485522b
51da6c03197168ccf57dcde1e6d7186aee0a3334
7333 F20101203_AACFBH jennings_w_Page_170thm.jpg
975e80971e24028f5dbaedaacca61f91
93a0925e90ca03ae4ae1faf812ac850d9a7022ba
113155 F20101203_AACEWB jennings_w_Page_085.jp2
72ee03bc760a81c16142ea57838c1cc8
b88ef2cf1ecc2c6e9bfdcc13443157310806b326
84706 F20101203_AACFAT jennings_w_Page_119.jpg
e0e05c1a6cc8453b719c85f7384d3459
e935138577c61c314792d5fafaa39f0e93e20fff
2220 F20101203_AACEVN jennings_w_Page_047.txt
de00237e1fdbc9bbb13046cb350c4ce2
9e1dae1fe292f9075599495317d580bc66747fdf
2560 F20101203_AACFZE jennings_w_Page_119.txt
9408fecb4af41c792e69cf36ac546174
90cc87895700f9b81abf134a5ee2b6b897ea2428
F20101203_AACFYQ jennings_w_Page_066.txt
c81337ae2ec8bbbaa3316596d8ce6ec4
c5f3a3bbd17e291e3131a0b6fd8082aeb1bec7dc
1051970 F20101203_AACFBI jennings_w_Page_007.jp2
be80197df06f4baee5f74e8cd9000014
ca6f922864bb2663a39542178fc4457633aa991f
25634 F20101203_AACEWC jennings_w_Page_010.pro
b2b42e8130d8c197a089cf791be99240
149df64869ea1d79a88a0e484f46c5911e6bc8cd
27111 F20101203_AACFAU jennings_w_Page_139.jpg
30d7e7b327a7513d20254fcfbb77ebbd
f2a15534600ae76f71d2a8ac42fa780446101457
2186 F20101203_AACEVO jennings_w_Page_089.txt
fd8638b42d25c4eee6c707c033656bfa
144610422204e92b322feb2886523c2c75c1e9b4
432 F20101203_AACFZF jennings_w_Page_125.txt
e1114345de7a4e5b29c51957529ad5a5
6204e3f4bebef9f6c49b971646facbc989dbc094
2685 F20101203_AACFYR jennings_w_Page_073.txt
4a63052ed97a99647f07f900364b3711
34b674411cb00e042709a8cee8ddb84b673bdec8
13951 F20101203_AACFBJ jennings_w_Page_181.pro
6c4b8963d8e2833f6eda07b0cea44ee9
cf0228306f4c916c9f4d30812b4e25e073625a3e
77499 F20101203_AACEWD jennings_w_Page_076.jpg
b55d4584c70efb4add4091df34de019c
9f2df6170851054b3b95379f5f4e1b7cb68c7bbe
76647 F20101203_AACFAV jennings_w_Page_098.jpg
013d83f53a271e9c60c83f08042dc84e
96ec1de1cc720e991c03c23802696f196383ced4
2128 F20101203_AACEVP jennings_w_Page_070.txt
344f6eb8cfb4d013a50a9f606a6d30a3
cd6f198d0ca82c285e6dd71d824c40500bba4043
1490 F20101203_AACFZG jennings_w_Page_131.txt
733e20c6da346c007ef5f64407defdd7
bfbcf7086fb5d7d5c711d6b38b8d3c76536f6164
2296 F20101203_AACFYS jennings_w_Page_079.txt
0f71e4d5f86402bc1f498dce67b3bc3f
fde69f3c625bb17bfaef52c8ca5bdeb2c2ad467d
F20101203_AACFBK jennings_w_Page_136.tif
c32a2914416c920313a23d644fd35e5d
35fa7fc5ad5850336711bfe264532dae365d1fbd
F20101203_AACEWE jennings_w_Page_108.tif
8f82dfc008a9f870304df6b23d78afb0
e614eb8742c872ee9b1d71b2e542faf0ab684957
F20101203_AACFAW jennings_w_Page_028.tif
7eb897d944277fc76f8371a7a815e910
400f6770253d466e627a25b40b1d4f558b66da45
19218 F20101203_AACEVQ jennings_w_Page_157.pro
2293246903a19138988a0e392670de4e
2ae32aed832db44756be71f55a81aa27360320f1
367 F20101203_AACFZH jennings_w_Page_138.txt
3d8a9fd7f8b41b8a566961d57e61e6f7
3a3be45afe255021592261a696bcc82fff42ef1b
F20101203_AACFYT jennings_w_Page_085.txt
a6375de9a4f9fff50abe16064dce01aa
ae3306d3fe2027a6b7428f1745b377b2df1fc8e4
2149 F20101203_AACFCA jennings_w_Page_148.txt
9da45f397cf50404b66d103e7a35aa9d
5dc9d3702f48cb629fcc8f0cc8b961eeffbdbdad
F20101203_AACFBL jennings_w_Page_046.tif
5a3645552993547d2983008bc672027f
c55a275d7842790d4728b59ed2358b415d327c91
2097 F20101203_AACEWF jennings_w_Page_130.txt
7d179f95b6099df9f2e8a6612ec5a1ca
8844954aeabecb8b9c6e78476235b4352ae25802
4204 F20101203_AACFAX jennings_w_Page_069.pro
cd6471bb21a7c510dd806dee2d06551e
e0b96fe47595b30b79c8e44416d7a3af619c02ae
26153 F20101203_AACEVR jennings_w_Page_083.QC.jpg
dd3e5af37e2745e8fb08a3302f6b0ca5
50deecb8e44944bf36d686f15eb6ea65fc632ae7
288 F20101203_AACFZI jennings_w_Page_139.txt
ca2af5446b590452345e1dd5f4a42693
3bda1bf09bd0af14aedbb651081fa0220931a165
2037 F20101203_AACFYU jennings_w_Page_090.txt
6864fae1442a5dbf1524ee6c742578ce
7320b4ced71222319b96fef3341d13394a603160
36756 F20101203_AACFCB jennings_w_Page_137.jpg
56525c23b2f506d1bf7c3344310ddee7
10f237e74cacfd67187a2f46a2a5bf6efdedb6b3
6876 F20101203_AACFBM jennings_w_Page_020thm.jpg
0b9bc020087235087ca4269f64184715
01df216cacebb749f5fdab8e713b86e67cddd5ea
F20101203_AACEWG jennings_w_Page_150.tif
c09fedc70d29fdf1b42132305cef82a8
e71541ce077731943eff81bd59286328627d87a6
26982 F20101203_AACFAY jennings_w_Page_112.QC.jpg
6d92dbb7a67529a23cf56a6c37c6c3bc
572157fce075210d1c40acb17618914340160802
F20101203_AACEVS jennings_w_Page_054.tif
4a0e0cf249c365f3fd84f09e13cd0d89
0ba92532c465962d40db7b329bcbc4c43fd556b7
2122 F20101203_AACFZJ jennings_w_Page_144.txt
178d1f79012ac04dd3da9e369e027433
2bd9ae36e02bbda9687e7f81bc25bb7660357ccd
2123 F20101203_AACFYV jennings_w_Page_094.txt
3d63e4a577f78458eac8793189ac6c0e
eb971b92935cc41d0c3f872f0ff0260606da1b11
90227 F20101203_AACFBN jennings_w_Page_166.jpg
1415720627b2074958f7dc4c92358430
01963d483235ead136b1a84f32b32c3042352fef
2070 F20101203_AACEWH jennings_w_Page_038.txt
3bae45312b1131e0f2f0120e227a7485
0d5e5554654b1903bb508d7469f54e1805782e40
11625 F20101203_AACFAZ jennings_w_Page_135.QC.jpg
7e4bb3abd11132f76f2c46d0e2cf223c
4bc1c566f9b6796b701bba311c7b95e4e24159ae
7238 F20101203_AACEVT jennings_w_Page_173thm.jpg
5d54fc0dec50a46f42bd631d2a935a04
51403a294279180e570c37cdb7a3ad6b7d21908d
24673 F20101203_AACFCC jennings_w_Page_071.QC.jpg
b296864d49d9161a4f8883293cdedfe9
b8de5a53fa89e3a7606ca6043e1b6b7bc6bf0224
2160 F20101203_AACFZK jennings_w_Page_145.txt
c17c6c95776d30c97ff61858bf4d2997
6c7614b53b9b57a18e4127d225f08a0a2e1e6365
F20101203_AACFYW jennings_w_Page_099.txt
39a55986262ac2aa525f9c8ced33eea8
69e0b0ea83248a9084113185ce30de1fcf0eaa6a
56579 F20101203_AACFBO jennings_w_Page_030.pro
a289225c93735120d6ceb3ddbefbca68
aab7e4220dbbf91061e0948fa19933a0acc1bb6e
81107 F20101203_AACEWI jennings_w_Page_112.jpg
21afc0ad47499cdff2e63f360d421d30
424d03f61c2f5debed07c0fdc0104e274a1c10b6
2170 F20101203_AACEVU jennings_w_Page_109.txt
23318ef8a88c8f2472e3b818a8663d4a
bdc4c752a927d10e544abac3458ec12d3e87f92a
F20101203_AACFCD jennings_w_Page_004.tif
070396cbd17bc490d712341e588c8a8a
27a758a0faf29098a58846ca2ace8bd01968315f
2495 F20101203_AACFZL jennings_w_Page_159.txt
7a943f29f53d5df4ac57e1b3da7e15fa
1b457da3f84bc45de8b9a1af618decad3f8f85e4
1458 F20101203_AACFYX jennings_w_Page_102.txt
0ccd9fba8ac632fb9bbb2a96fa8014f9
d1dc3842a161a47d0e8925da2790cb70ed76dae5
2637 F20101203_AACFBP jennings_w_Page_141thm.jpg
22d57a632efd7b32dec3ab974bfcbc9d
5051606d4c5056cd82a3d0d898681b5b9ea6e1a6
57080 F20101203_AACEWJ jennings_w_Page_023.pro
91b9cc6fc8ab1774b1e56c0696014c3c
9ddfb2adee060dae372658f3264cb4dba3e9a147
6784 F20101203_AACEVV jennings_w_Page_094thm.jpg
74c580b9b47a610ff8a3adb3c2bb0ba6
6017a5594df9ae9d325b12ffdbe08bb4ef852e5e
F20101203_AACFCE jennings_w_Page_049.tif
92e0f98ba89fada545db0cdf2cb867c5
f6b8f5b0581e151321e9c7845925a6ca7012186c
2544 F20101203_AACFZM jennings_w_Page_169.txt
63bc79811adcb95ad35ac027cd974152
0cb8aff236abf0dc8cbf232c99f7466a6d575721
2327 F20101203_AACFYY jennings_w_Page_107.txt
5506c580209ca16910a86e7a820dd9ef
bf6707514ba6f0cf106085d8c499a740c62d34b7
113969 F20101203_AACFBQ jennings_w_Page_089.jp2
05c3df22932e385f0f54a08924cdb232
7262fca277f0689811aa51a0915d9297d39cf6b0
6905 F20101203_AACEWK jennings_w_Page_041thm.jpg
a6e897d009552a988c0638b47d0d7b3c
8472342b8dff26211cbae5e5a6adc9dcdb9f52c5
F20101203_AACEVW jennings_w_Page_166.tif
ee686ad469d1a282b6dabb70e58aa534
14da52b3e0eda7045177b7feb45b1b36d3999c0f
2406 F20101203_AACFCF jennings_w_Page_125thm.jpg
adc643262d7535f0489b6dc703849ac9
efeb9e54ffe7bd79b9b356e03e44db07986ed6aa
2759 F20101203_AACFZN jennings_w_Page_171.txt
9df4b67af117b193d4e8bcc9ef0b8cc1
d7c77e8a3ab1730cf8298db47a603af3fa6e1ae5
F20101203_AACFYZ jennings_w_Page_110.txt
cf6cb39a3298d111f860baea300b6584
992eb976155cc7f9abdb2e40e4e47638c9a10616
F20101203_AACFBR jennings_w_Page_011.txt
2c9c4c92b2af568d4c22b81b5087c74d
30b3fc2fa1511119de2041a4c86e3dcef7046aa3
24109 F20101203_AACEWL jennings_w_Page_090.QC.jpg
89ec52a76cecfdb962994e7e77770869
4800cefaccfc5f47152b5f5409d7325d94b64616
145861 F20101203_AACEVX jennings_w_Page_174.jp2
c6f9b8203669f0ef12e1bde76d350859
b2d4be9ec6641f924d8948f13c81f46e867abac2
79384 F20101203_AACFCG jennings_w_Page_147.jpg
c936cf082d18ca3698bb7aaa9efb370d
780ab0524a6e39b5fcfa3413ff416fe30ca2d199
595 F20101203_AACFZO jennings_w_Page_181.txt
6314c9a03a12ffb8e53bbedcfc0ff00e
b33d4e9e0ba80926f834b873a18803fb77bc37e8
6807 F20101203_AACEXA jennings_w_Page_056thm.jpg
22e5c5b27af524ee8089c8734248880d
eb13f80bdaa61c1fcfe132903dd34719938283f4
23702 F20101203_AACFBS jennings_w_Page_067.QC.jpg
4d7229f8e09525ce8f199646735dd79d
a61412eefdff34d7def94517992b66876ba26f72
56313 F20101203_AACEWM jennings_w_Page_147.pro
33472960487c006769941b02303e6b51
76bcf7967a022d41fd89e9521024ec64b6bd0413
51883 F20101203_AACEVY jennings_w_Page_102.jpg
ba6fdbec3a03ab7c1ba5b06de7b25f19
48b9bcc4dfcdaec561c33138f99fbd759b727492
21848 F20101203_AACFZP jennings_w_Page_004.QC.jpg
085466d2f7f0097e7093a6c01f3d281a
38807cb97dc77d8702e9b4ae645d43d89c992ece
121120 F20101203_AACEXB jennings_w_Page_147.jp2
31442ea92f8b4c24f741ee765623445a
fdaee9835ebb3608328f8b0eb8443973d5f08463
F20101203_AACFBT jennings_w_Page_075.tif
d01fbabe1643b4566d8abe9fb0249e42
e380b80259bd41ea0ee15e1971da955c743395ad
80214 F20101203_AACEWN jennings_w_Page_057.jpg
4138ae84f35670df0cd792aa3ed73ecd
4f3f890cf2c26a7f28841dda2a2756df8e4e7014
114985 F20101203_AACEVZ jennings_w_Page_033.jp2
d939901455adc67ecbd53f1914ae54e2
fdca9dd1d469ae05f353cb1f87ae4dcf928a556d
119176 F20101203_AACFCH jennings_w_Page_111.jp2
87b718a65fc77ec7ceb2f8b8e583b20f
1c801b571dd42e5a15f9c2b56db346358cb80333
25229 F20101203_AACFZQ jennings_w_Page_169.QC.jpg
c73fc957ed7b835f19cccb0f6c0e8258
3145e1afab45c8166bda052cad1471df3fd05d63
2882 F20101203_AACEXC jennings_w_Page_138thm.jpg
d49e92844c7a453e3e198dec3948358d
4509f3773d32880b01fb409e822b73ab02b53299
80180 F20101203_AACFBU jennings_w_Page_107.jpg
bcb8ef2b1eb3356f36e08b217ea6cae4
da4d0edebeefd4f9d79f4680aad3bd6dcf57260f
77853 F20101203_AACEWO jennings_w_Page_025.jpg
a72d52341f33979b69c98bdb979c0511
6925099a958f71244c30769135b27974bb0bfe2c
56947 F20101203_AACFCI jennings_w_Page_124.pro
becc44526ca07879783994ff630988ec
804292835218a2b12cd20e72de164a2a8cecf910
26223 F20101203_AACFZR jennings_w_Page_166.QC.jpg
c88fb453214c26c10508820bfdfa0f94
9f6f17b4293d817ad1eafe64b2b1429c47f8ef77
F20101203_AACEXD jennings_w_Page_104thm.jpg
700d6738cbbd74374788c972da39a489
4ee1569c02cd7d477a051601d62ab1677e055c09
51867 F20101203_AACFBV jennings_w_Page_028.pro
66a8d57c81386104935faf117dac0ea5
1d7440aae2db4f6f19b47c6d8efd3c279daeb3df
24557 F20101203_AACEWP jennings_w_Page_094.QC.jpg
087e455af5214f74b158234dc4d25958
0d1c97fac39516db699d61c1faee1ca68a6d05e9
7014 F20101203_AACFCJ jennings_w_Page_042thm.jpg
9ef18da3662035d136b837fdb68afc5a
20af3c8ef19f1577edd42d064a82cf67213243ea
4310 F20101203_AACFZS jennings_w_Page_180thm.jpg
7875e62a51cc45111de127f07a6af8ab
9fbd667197f9aff2d022c8ed16e600307b822573
8828 F20101203_AACEXE jennings_w_Page_133.QC.jpg
f1f63da96d04305c0f17f05d2c4f0b2d
3d8d1ed770eb9d1a079ee8d81c76148f8ac25451
25000 F20101203_AACFBW jennings_w_Page_093.QC.jpg
4b4c1ce276a8c96357e1d54efe439ea0
8834c4a48f1398eef7178b663f7137b8174b06eb
67652 F20101203_AACEWQ jennings_w_Page_171.pro
f48ca9db8bc392da05786114e5f6ab7f
c87048fdf5bd2c4e39c7cb9120bfc99f02f24510
6510 F20101203_AACFCK jennings_w_Page_024thm.jpg
8cc5fdf7eac9934bd96ab92515c6d8a9
473a1ea9b4ad1dbc5bc57985dfb4c88ac335e8de
F20101203_AACFZT jennings_w_Page_176thm.jpg
775f9d87f083b1dc15b98bd40294fd65
caed79a27881175a5da061d257af35f6e6b5454b
6760 F20101203_AACEXF jennings_w_Page_095thm.jpg
f370ab939c9be2dd2cb9139a67cc785f
819b55a5deefa0cdcc9d0603825ed94289b251b7
115583 F20101203_AACFBX jennings_w_Page_093.jp2
1c20f5c1f1fbfdf1cc5b2bccff71338c
ebfef1094a60f5fa6226d80ddcf6564edef6bb1c
F20101203_AACEWR jennings_w_Page_061thm.jpg
595174ea1b1b3d80a3bdce1cff8e3db0
bc11bdf0650c8c9a6f05d5b220ee7836e8589f54
7096 F20101203_AACFDA jennings_w_Page_073thm.jpg
347b899e6f37a748729c2c1ca9f66d20
c88fad194efec7adcc3fadd4152f62c5e228bd10
146648 F20101203_AACFCL jennings_w_Page_171.jp2
d4115b1460fa36cc4d444e7bd7f4cc08
c5b63dbeea1c1fddd9ee1708f5df580ded5ee998
5418 F20101203_AACFZU jennings_w_Page_044thm.jpg
18554f78c793fd26c8635bf6816d2bec
e3cc390c8fe83492fd21ea0080be4207db380a06
137718 F20101203_AACEXG jennings_w_Page_072.jp2
05f2be48ca0a20f6dba1655d159f3b9f
565fdcbccbcee65fdfc994bb58b313bcb1a118de
63581 F20101203_AACFBY jennings_w_Page_178.pro
37d1d589e82f0cd3dbffb2e854a7ef00
0c67ef2c1fdc1cc24498ffdc3d55801d00579e36
17283 F20101203_AACEWS jennings_w_Page_128.QC.jpg
da4c668f801d2917054dc9dc3ffab850
98f7559007db5baecb690bfed04df993e46357db
2029 F20101203_AACFDB jennings_w_Page_053.txt
30a69c2003943451103855685abc368f
311cd78ed86a8719b2baa6964a1fab4fe2209b6f
27584 F20101203_AACFCM jennings_w_Page_161.QC.jpg
f0f00619de61ca6653b877d298dfa389
740a9ef935fbe48e1c2f462370ed8ee9a8f4af1d
6578 F20101203_AACFZV jennings_w_Page_084thm.jpg
8bcb2908535efe2a9d8d77c9d97c379f
ad6f776babd231797673f45f8ddd905ee2e23e7b
57128 F20101203_AACEXH jennings_w_Page_128.jpg
64b80a4485ad90db913c4586c9db5954
b84d2bb9da1c862fc04ce7b63eb014315cb8cb2e
F20101203_AACFBZ jennings_w_Page_118.tif
a14d5bbf2b4a4bae01138463d517c051
6b4620c073105b84db95539a0b27d7e2eec7ba46
96 F20101203_AACEWT jennings_w_Page_002.txt
386b059d243d847d2a08e21d3de72cb1
f4c38c5f320f14e58b9305f80fda7c34e03b0e6f
76945 F20101203_AACFDC jennings_w_Page_055.jpg
520d500d83f856c796a474f2e6b32925
94856e5ef50b112068dac60a6545e83c3ea716f4
2753 F20101203_AACFCN jennings_w_Page_142thm.jpg
b774db0030d10becc690f2529f0c1f28
5bc9be75705cbd125b60eb798aa71966bf510078
6897 F20101203_AACFZW jennings_w_Page_022thm.jpg
eb9d07425ca51accb5296ae5af0de49e
4fe5aaa4833e4ea1d2d53a16bb3b0a73c0e7dc83
1603 F20101203_AACEXI jennings_w_Page_044.txt
68564eb395c2dc22f3587907c198a940
b080ed367810c2ff04eb86a8647a691be726adc7
6762 F20101203_AACEWU jennings_w_Page_085thm.jpg
c136c32d0fecc6936c85fb78127476e9
27cf71b14583c6c20842b56c22165c77a6bbacce
66373 F20101203_AACFDD jennings_w_Page_088.jpg
e5ddec706b54057cd7afd097a799a8d1
ce0da9707e97f9725872993b3955abe524cc207b
F20101203_AACFCO jennings_w_Page_147.tif
c4d1a1bc724ccbc8f11da8ef7e6e1190
9a0951ac47718c8f20629dd5854a8c1140f6efc5
24794 F20101203_AACFZX jennings_w_Page_045.QC.jpg
4b6da9308f371fb686859b1d7021a542
ab8b6a166870dda925927ea72097e067a1a0d10a
110397 F20101203_AACEXJ jennings_w_Page_156.jp2
d49a5d719de11716203ce279bd191bc3
799fa944c5a7f72322f745005f764e4a41cd4c24
76479 F20101203_AACEWV jennings_w_Page_065.jpg
b5a09ef28fd3d7853e8417822cf443c5
f5f4b90362733d3cf70d9b9059eef86dbc8d6133
69899 F20101203_AACFDE jennings_w_Page_073.pro
68ade4d616ca7b46e7ba3a16df5fb810
d00643277786f98bba0b45ade28246cceebd1459
F20101203_AACFCP jennings_w_Page_180.tif
35efb6b3d8c3a138644f32834353596d
e0cd6a75e642a656bd9e9eb7ffd1dc75a21544a6
25406 F20101203_AACFZY jennings_w_Page_064.QC.jpg
47b39d1118f1786c0de68da5159b879c
504d01f8b21f5a15093147389935cd97bfacc54e
25567 F20101203_AACEXK jennings_w_Page_078.QC.jpg
bcd53611197cd2ad2da4580627437146
bf193f2334ab0927771e4cc586d110ed6e53d3d3
F20101203_AACEWW jennings_w_Page_169.tif
935705e4bfaed8d0e2982cfdb6c6eb4a
1a7e99847242c59efe968b6c3966fa5b22d3882f
9026 F20101203_AACFDF jennings_w_Page_138.QC.jpg
4c0040a484a4b0985ed09ec618a2b7f7
580d73f7d3246ee3582104a4a93b982ac1107723
110323 F20101203_AACFCQ jennings_w_Page_017.jp2
92a12739bed2a85f8ca9c9d8ddfd688f
04c5bcc9cc1236582ba5e10ad6d478ff852be1e1
16375 F20101203_AACFZZ jennings_w_Page_003.QC.jpg
8fb2723a4fc8812320de18b55c1c70d3
c30eaac20651f8066d20dbc9964bf1d7460a9d1e
79112 F20101203_AACEXL jennings_w_Page_123.jpg
e0eda1f39f2bffbbcef6ceda92c40f77
3af0c7faf88ee7d5a0642b06cff02560a99282ef
50580 F20101203_AACEWX jennings_w_Page_087.pro
d98b5e242bc7435a02275751424e4b40
417873b791e978c8d48830266e295786f70321aa
4324 F20101203_AACFDG jennings_w_Page_127thm.jpg
f51c7d84f312589b8073fdbd1e6546be
0927a330ff6c9a353b219d9b821c417bf3f7651a
1971 F20101203_AACFCR jennings_w_Page_128.txt
e73c039f63ed90ecb93a4108931f48f0
425efed0279c3d9cc2ca705115d6d2092c7984f3
23496 F20101203_AACEXM jennings_w_Page_024.QC.jpg
d714a12ec3cf987224172c7586c1d192
e77bcf4a32e25248471aafb5fac7699d5af80db8
2509 F20101203_AACEWY jennings_w_Page_143thm.jpg
92a2d5887a00928aea44a4012a43c77f
8456116f925c40722410a6692bfe3653301c3270
122176 F20101203_AACFDH jennings_w_Page_061.jp2
3fe002ee5e7db9f7da22fd7e59e82593
b2c8d339cc9c2b3656061afbd3f0f50ded5a7a0c
55110 F20101203_AACEYA jennings_w_Page_154.pro
c16fa4039ff6cd010f1d5e46499467e4
4c988b3f40a7507c1578fa80a8cf9e6eb51ec5d8
119930 F20101203_AACFCS jennings_w_Page_037.jp2
ea8525554bb66103ac3817af848f0da7
cd5300949425c6235563710169df87f1b278a4b3
23748 F20101203_AACEXN jennings_w_Page_028.QC.jpg
6952c10cf3fe433bc5e9419cead26b11
bb33ce402afeaaf24b940c0f51e95816d1fef064
F20101203_AACEWZ jennings_w_Page_046.txt
b4405bc8c2dbd9818d055b510ee64fe5
8607a28c8c83467e2df45843e153aea535dee220
2234 F20101203_AACEYB jennings_w_Page_052.txt
2c26809f376f8de61d6c463c25f17f65
cfa0ed837f770bb1f8a6d3edb5193410453da0bb
2983 F20101203_AACFCT jennings_w_Page_158thm.jpg
e8b120d0240cccc5c703ade25f4a98cf
1c8110a7910a7ea2ee258e93a46596a3d69b43cc
25722 F20101203_AACEXO jennings_w_Page_109.QC.jpg
6921dad445b822a9d8b0ffbdba329286
29de260dac24971cace3864c425a0c77225aa443
2153 F20101203_AACFDI jennings_w_Page_012.txt
e41852e381f8001f634adfd5185bffee
7525098c55a652a9405ccd9c78b05f1dc8350f8a
F20101203_AACEYC jennings_w_Page_165.tif
1693705f2b8acf37c3a2b6b2ac5260c4
76df2e9e81027d6739ccefbcc2d0a062e2096f78
51461 F20101203_AACFCU jennings_w_Page_134.jpg
5d78a44a842a433dbb86ec3b34896c14
dc011338b4c88dcf9fb252fd5c7b7d5c2cc46170
7068 F20101203_AACEXP jennings_w_Page_014thm.jpg
0a104b2c3de023b7394bb25e6891c4d8
5679f60372f997e4320651d86f8015b99c99bcbf
120124 F20101203_AACFDJ jennings_w_Page_152.jp2
7cc46e95785be8c67210a2dca09df1ca
9cfa1bb9c3c5006d31be1392c8cfb1cf445dc07d
79939 F20101203_AACEYD jennings_w_Page_049.jpg
c4ed4974987f38e28893995715b6e68d
6f4ced4ffaee813e5b5ddb05a0524044869d2be5
33549 F20101203_AACFCV jennings_w_Page_069.jpg
f36fa7c13e187654e43363b5a72c1e49
83adcfc7481b6d8173c409a399e1158fc529478f
63051 F20101203_AACEXQ jennings_w_Page_173.pro
bc6066825257f276afbaca3f9c481bf5
48ae0b5515d703635bd538961f80a8534f4ec6ab
2690 F20101203_AACFDK jennings_w_Page_163.txt
ae6061a6fde08cd1c84cbb256ccc81b2
9b567ca6d564f811f7691902254a75d8803da9ba
51074 F20101203_AACEYE jennings_w_Page_144.pro
5dc616f937c6842eaf0186f44dcfecde
0b76c52b451cdd5749b956d021639e37ac4b0b2d
2583 F20101203_AACFCW jennings_w_Page_173.txt
c3ef8269791947639f3ce331311d2247
1587920f0d053d55fb0d43c331ba4b368e48dcea
120294 F20101203_AACEXR jennings_w_Page_022.jp2
20cf89767cf3bd8f7b9f74c02b669660
e6df3db626e110fd39e55c8f4ce813914a407d5f
121718 F20101203_AACFEA jennings_w_Page_026.jp2
5b4072350621883b638127467c010a0c
14225b6af34b846e594d8ca4c7c2d7e249139239
7034 F20101203_AACFDL jennings_w_Page_122thm.jpg
f3a9ed237bfd4758298858ce10a0d06e
d6e1b67df8b834763701c791b6f08503c759971d
79320 F20101203_AACEYF jennings_w_Page_061.jpg
c310f3452c8ebc497e79c1966bc2a1d9
6ab297871c367de02d056f31e47448551abb1305
74652 F20101203_AACFCX jennings_w_Page_029.jpg
67ed7d67ab3d3638cea1e4a3af710d01
0d1cfb8d90c991b8e3320ad6cc00c8cee4a7a255
56765 F20101203_AACEXS jennings_w_Page_057.pro
7dc9bdaeb693c3a68d981595969ae863
18dd395f453f51f74760db914a72c74710b7d78b
81973 F20101203_AACFEB jennings_w_Page_039.jpg
6e1f1418e1c0696c4e94b533ebe4d528
3d0b7f3fe89d3d2df5d28659ba4219c328c7c4d9
6940 F20101203_AACFDM jennings_w_Page_177thm.jpg
edb1ed309e59c97f1df4bbe4845129e5
44af3a0c816eaad7654b5ec6d9987235cacd8686
118603 F20101203_AACEYG jennings_w_Page_025.jp2
7a45d119b4fce1e1a81d3bd5149d87db
3e68da777b609c236c530dccebc8ee45b70b420b
45400 F20101203_AACFCY jennings_w_Page_009.pro
cfffbb3d8bd4d07b4479ff2b11bf4f34
01a90de80aa6e1d9a4540bff22afeb8b9b41f232
23251 F20101203_AACEXT jennings_w_Page_011.QC.jpg
ebd624193ff9160c037a5bdc567f788b
1166d6e336027bdf39c342c0b486c5738b689b7c
7191 F20101203_AACFEC jennings_w_Page_178thm.jpg
0140694e0f1c3303a50bb552d8f9d7a0
d9f2265640461954bb308a899932bc666750fc13
26136 F20101203_AACFDN jennings_w_Page_096.QC.jpg
44f23a151c2bd15954106817ca696115
22a36a30a51c91ef836b77603cc03b63b3ba2ac1
633662 F20101203_AACEYH jennings_w_Page_139.jp2
aba607005b9d69cc94108c39b90ff1de
1b0921e2fb8d2b62784ef1d43838e0390be5d918
26462 F20101203_AACFCZ jennings_w_Page_072.QC.jpg
80433c2bbc876d1d1e99967505ddb9b5
40e4576ca4993d96b14c96254fa57afa287bf950
119311 F20101203_AACEXU jennings_w_Page_096.jp2
f992b205a8337a7c19d058f002c31ca4
b62cd59bddd741955692ea90efc049eba7d1fc57
6629 F20101203_AACFED jennings_w_Page_038thm.jpg
e250dec1c3eb6f0d4e14995779ed27f3
a8444948347ec1847764c37d154b7d4cb3723112
6962 F20101203_AACFDO jennings_w_Page_096thm.jpg
9b8c9665ebd8e77432e712548c78663d
bffe7533d6558a691428161dd6bbba3db91b10d2
77755 F20101203_AACEYI jennings_w_Page_078.jpg
aee9b39ca54d2d99001556f8cbe57af3
8ece6a386154dd7bb63cf0e076057ddf90eff5d6
91848 F20101203_AACEXV jennings_w_Page_170.jpg
0c43adb45b3992c0f2574dd55aaf110c
371571d62e0b9b0a741e537485369704ca7040c4
10810 F20101203_AACFEE jennings_w_Page_069.QC.jpg
c0716d548e3f78c7dfcb9ab6bbc76184
10c213fa115537fa1dd124136ccae6de254b33ee
1982 F20101203_AACFDP jennings_w_Page_103.txt
d24899d898986ca8ecb2742a12a3c88a
6bd5c27b02c0f40ea4803df1a1f0eb0dbfc61e9d
2187 F20101203_AACEYJ jennings_w_Page_040.txt
d8b2c9a49ca968a8131285f72c184eb5
5d7289d278c9cbcb5e6565c9f6eb8e6a37f51f8b
F20101203_AACEXW jennings_w_Page_050.tif
8372428dd19c0e4320cc0b6f51102922
c2afc50f37cc25575178726e64537ec0e84d1060
26182 F20101203_AACFEF jennings_w_Page_152.QC.jpg
70172f438ac02b99d1932f3094022ee7
d02506b7adf7c0ace2be4c4419b2d91606dc8c37
120302 F20101203_AACFDQ jennings_w_Page_040.jp2
a097e1970a70e7813a8f32a8845fda66
7080ddea088093faaaf64cdcc4cfe89c43c4e848
75548 F20101203_AACEYK jennings_w_Page_021.jpg
5e0aa2ff27a2872a465afc577d4392d7
9ff1862e7041a457c4d076d6b84ea1a0990e8e44
1956 F20101203_AACEXX jennings_w_Page_024.txt
66be9f7930622c76d43d2c980ac59708
5597f36a9ac439c25f96269c2e08cc9313420e1c
54162 F20101203_AACFEG jennings_w_Page_108.pro
f4005b45b1cf416e699a3c1dd24fba3e
6b4892e6cdcea898bfc16dd1b637c9e9cde04763
2081 F20101203_AACFDR jennings_w_Page_091.txt
266c75c4bda99d80aded0658a7449850
c9d8d6753118411ae76076d3bd71211a211daf5a
115487 F20101203_AACEYL jennings_w_Page_059.jp2
0f7c98e2c91d2bdb99c1782ad9c2e23a
bc7576b5ab1977a06d75c9a86148b36534281103
26559 F20101203_AACEXY jennings_w_Page_081.QC.jpg
7cb51b5fa0893d4e27e13bf29bb55518
492efb55932f7e70d06b65e23a326b5ee88ac7af
7211 F20101203_AACFEH jennings_w_Page_171thm.jpg
8c642d6d6d4fe4a46c8865ccbd235c77
58a3381dac4e1ef5340f3826c0dbdbf52157f964
121611 F20101203_AACEZA jennings_w_Page_046.jp2
5850ec5f676b61638552b265411fb4d7
d89a581f88c3027a700a77d0e5469cde46ccb94b
F20101203_AACFDS jennings_w_Page_151.tif
29f5eab40bb157fad571d967aeb2eb68
89989c0c4f5d77eb9d4b2aaf9f32640646effe45
F20101203_AACEYM jennings_w_Page_159.tif
8ded59a8657827883303048e8787e15d
b590be98ef6b82459000ba709f74a02d7ed44f06
7563 F20101203_AACEXZ jennings_w_Page_174thm.jpg
b0fe65606aa9ff74fd29eddb70dd6882
3e29c266e333a45042732c53a8a62c9f6c798ee8
54021 F20101203_AACFEI jennings_w_Page_074.pro
75d0f75f4ef973dfd03756c9c44991e3
4ca7855087854e9ebb23413ccb388b8178191cdf
F20101203_AACEZB jennings_w_Page_175.tif
361fef31f2dd434ca68c6044fab8aa62
0835cc2ac0fe1d5b029f3c459d0981b1c3c5be93
115769 F20101203_AACFDT jennings_w_Page_100.jp2
5e011824b0e5004d5b0534b76db544bb
419c528b39932fd7d73b3c7afe1e394919b02163
2263 F20101203_AACEYN jennings_w_Page_151.txt
fceadaf90e7c62f5f35d56d2b8d93242
24a4ee70509b0a302186a6d077d84098cdf3aaab
26390 F20101203_AACEZC jennings_w_Page_023.QC.jpg
68e52069c63f36970b3b739602a24549
a7569177bea0ebbabb11679b806a2bb48d7a7966
6949 F20101203_AACFDU jennings_w_Page_167thm.jpg
8fab4eb9cbea569a248f717b23c8c35d
043e3af17a3d366e09e19c9b176841b23ed08a79
7020 F20101203_AACEYO jennings_w_Page_107thm.jpg
b852ce201f371a9b1d6756b1061ca42b
0bbb2b0251a134e72d33d254d45684759dabad48
6626 F20101203_AACFEJ jennings_w_Page_070thm.jpg
acfeeef88c7de141d06269fd604d01ea
4cfa2b376027dee29dd61056e677ceb805bfbc8b
6681 F20101203_AACEZD jennings_w_Page_086thm.jpg
75ad4399e9b807e1fe3dc32bf9f8a05f
fa369f5ad4b766bcb0515bd0c995910e3e792429
3517 F20101203_AACFDV jennings_w_Page_008thm.jpg
c27fe503adc77676b9af93577683a45e
4884ebab5f63bf2a0a54700066aeb4040107966e
6998 F20101203_AACEYP jennings_w_Page_026thm.jpg
8736c290b477a1abe89a21abad8a8e9b
fdbab18f7fd70e18cd090f4aa3babf8b794526ea
76055 F20101203_AACFEK jennings_w_Page_105.jpg
34ce35afd3b4b7a7c62eae0ad3ba0288
9b56a822946a7c00f924d53a06b003c32db19eee
6851 F20101203_AACEZE jennings_w_Page_033thm.jpg
9d9a70091c5352b871c5f74796dd86ba
5319ae76b30a47ce575dab44b09d7a4ca5f5f634
F20101203_AACFDW jennings_w_Page_155.tif
f68ad53d907a02735f28e01686514c5a
bb779280f15ee6c14190cf00cd4665b9f7d007ef
F20101203_AACEYQ jennings_w_Page_123.tif
9e1bdc79492af15760eea412c0feda5d
197f9acbdf52977fe728bfc24b3f0a968c17857b
3813 F20101203_AACFEL jennings_w_Page_080.txt
7ec483d5f157f2b791d264f7db81f3af
ab733543bc3a94bad93ad87cc3abc0716cc34b06
63153 F20101203_AACEZF jennings_w_Page_167.pro
b75b226c82663ebd5c47d28c7a4adff5
3ab16d5dca89ba91c9dcd8d6b03078961aa98f34
80177 F20101203_AACFDX jennings_w_Page_052.jpg
fe96b2c4e240e1d1c51307ffcd5642b7
c29a907f114b9f2928341dbb22bfab2ca126dbe0
74934 F20101203_AACEYR jennings_w_Page_084.jpg
4af6a82f69d7994841917e6267bc7ae4
f2740c4edd6fc28ddbcb9defb519df0c164c6603
122969 F20101203_AACFFA jennings_w_Page_107.jp2
ad5e8d0c3a6700736b23fa66ef51d2d9
710049d0dadc666eed360e38f8841f07347908f5
59225 F20101203_AACFEM jennings_w_Page_079.pro
f268653f5a65f0889159489a90038cf2
4586dc0f5ecd1339ad3a0fe1cd7991578762ed59
2816 F20101203_AACEZG jennings_w_Page_174.txt
f3d073cc12d9b7211b42b71d7b0d56ee
4f837562d4f31850505ae404b1c468eb91fdc004
F20101203_AACFDY jennings_w_Page_075.QC.jpg
7bde47ba81d84adb82e3366725233fdd
6a7849625b29081f218eb9dc3c1a65e0d5ebd0a6
7148 F20101203_AACEYS jennings_w_Page_050thm.jpg
ddfc145b37d3a9918cf0225ec974279d
eede7e66dafb69da62818f06eb01bf3d4e867d02
2817 F20101203_AACFFB jennings_w_Page_072.txt
f0d46d984407acec16bcd8e2be17136b
46fe58ff68d6e640a5a50a610dba95b36090fabb
F20101203_AACFEN jennings_w_Page_088.tif
e9ecc4371694d0049a2ab2a6d8b62e1c
9995e0f2023f95fe238c1de0aec68c922d99237e
53711 F20101203_AACEZH jennings_w_Page_085.pro
e41ba97db495418aad53903742fb385c
f4d013de823bc8d6acec33ba6db84e98fb934250
6926 F20101203_AACFDZ jennings_w_Page_021thm.jpg
a51a15bdedf6910d5869e9b8da1245b3
d880d2fb42bd22a021cca1508a9aaa34d58d4910
F20101203_AACEYT jennings_w_Page_059.tif
754a4a2c78ce9d83fc25ffcbc3184f45
a955ba2fdb5767aa9f42389a4f53b1d2e392399b
F20101203_AACFFC jennings_w_Page_143.tif
a73841e05101eb5293479104e1ad4111
7f22a9b793450c1524c7141fe7798c56aaa0eeec
116912 F20101203_AACFEO jennings_w_Page_021.jp2
bf1606a40e6efc5a10ba7b9215316659
ae31ef0f86b3753fa9d7790f941f110a15e3b475
62304 F20101203_AACEZI jennings_w_Page_176.pro
7b1a5da899a9495c92c7aa49724193e2
5d020f5ec5e1fa2fc676e4fad50398aa590fb318
25365 F20101203_AACEYU jennings_w_Page_065.QC.jpg
a3138f7a290ad5035bdcb961ce437da8
ad01efecc139ac0b5f76e9e7a38fba3f5515ef58
F20101203_AACFFD jennings_w_Page_148.tif
b2f5f2d3b378657b10f841c192bb82ce
d1a5ad49c5c52085462257e1cfbe6d5392d530c3
75750 F20101203_AACFEP jennings_w_Page_093.jpg
7c88b87a886118983046b1da6d170de4
6db931c21e723d9708c90cfd33f9ffa5e835efcb
2121 F20101203_AACEZJ jennings_w_Page_063.txt
448c545fd8fa3faa938a350b8e0b16c1
5ad61b46bdafdbd32afe5c36469925f0d52af4b2
7056 F20101203_AACEYV jennings_w_Page_159thm.jpg
f7ee105e1828a9dde2f5f3ba4545450f
24b68fff3b9e2fffcc916eb5da98a30f50aef520
55853 F20101203_AACFFE jennings_w_Page_046.pro
b2b83a57d28cef7d3f18470cfb80c890
608412a7bb9b5ef46fb203474ae05a0b092e2b03
2177 F20101203_AACFEQ jennings_w_Page_153.txt
470f610178ef42a74805233772956c37
28747ec3c1bea5536da71971495c42b3f3836b11
53109 F20101203_AACEZK jennings_w_Page_058.pro
8af36d43e6b4b5fc6a5020869c544032
e58ec1e4300385641e417a71ee260698fd605582
F20101203_AACEYW jennings_w_Page_156.txt
cf2c50569a482535f65b302c3ed4fe0b
0628f234c745a7b8dfb9bab5b8f212e7a030ee39
73323 F20101203_AACFFF jennings_w_Page_070.jpg
d83dcfb3dd9835ec50ffc235d58d618e
690aa5b791a79a89f207db68a39477abb80dbb38
F20101203_AACFER jennings_w_Page_106.pro
480c55168880fbf447a87fc3a5675f3f
a62360e1876342538ab9f8ad81b15487a9e7d22b
F20101203_AACEZL jennings_w_Page_109.tif
22a350f8b0ef6fd472be69a32e8bfb22
16e70b85471b2c9dc78b3602f9349d537e540ea5
23883 F20101203_AACEYX jennings_w_Page_017.QC.jpg
f28dbdbd3bb1463db1961f8786b60bc7
0c44f660b643ede8b48237f33126a6b0b71819a2
111942 F20101203_AACFFG jennings_w_Page_155.jp2
172135d8ac382b285d8b557bc50846c5
9527faf0db1af71a1dc82fcf4e3482904926ccdb
25900 F20101203_AACFES jennings_w_Page_107.QC.jpg
68b45ddd316f31bcb7a8b0b8b03903aa
0eac24bd3ea61cc188c3340a732a517df12f07ae
7252 F20101203_AACEZM jennings_w_Page_140.pro
f5ede7e8de444c6973bf894e233eccb2
d60d5f880daa15c7ab9adcaae7e6bb8133f91b3a
78902 F20101203_AACEYY jennings_w_Page_109.jpg
5689d654a50c3e1bb8c690dea95774e4
28b3e1dcae8bd49b126c4459a235bc5e0adac7ee
F20101203_AACFFH jennings_w_Page_008.tif
14e8b1bd5f656ede8959ca3313df4c36
e5aa29c724b2c9ec19a38e80a3f765d3a28af52a
71171 F20101203_AACFET jennings_w_Page_007.pro
cc30758e52d7efc8de77ac0542fcf40b
d8327ece27c56f618f26c52b3cb0949d19426951
6547 F20101203_AACEZN jennings_w_Page_068thm.jpg
c6791ac6e1d2ed17910ac32365d8c0b0
5d5044450089e8dab8795f30bd2f9247f1a0f0d3
120571 F20101203_AACEYZ jennings_w_Page_012.jp2
0a35eb5e12abf895599f657f1d821564
6ee2223be0df24732e2c720b5b33041c3ac44c9d
25800 F20101203_AACFFI jennings_w_Page_020.QC.jpg
ab449d05a50f347f0bc97f780d123f77
6380a8156d98e6af4f931275f4e2994c16bfc5e4
F20101203_AACFEU jennings_w_Page_062.QC.jpg
db23175e231c4d04fb50193a26cba75f
4a1d244ecf3ac246e55b8e076fde6282b624ee40
84548 F20101203_AACEZO jennings_w_Page_176.jpg
649ceb9fd966d870bf672289f122a884
9ad7db2efda3e52779491f62eb9e7ab772959f02
24885 F20101203_AACFFJ jennings_w_Page_153.QC.jpg
d47ce5ddb914c517405e31f58fac4902
b1fcfd342f9de6ea9d0798180707c2906c646a56
1054428 F20101203_AACFEV jennings_w_Page_132.tif
c59853baf7078393d9dc927dba92054b
135501d26686ef2f53ff6b23082423f5d77b42a3
F20101203_AACEZP jennings_w_Page_133.tif
63cc5de1d4f91c1107481dfd6e491ffe
cdf23aad4f2c6ac2cc593a1decafce8c00a48c99
6682 F20101203_AACFEW jennings_w_Page_087thm.jpg
ee70535a979de5aa0b5d5472c36fb878
b789cf4e6739771ecbe607b15d9d6b784a8bc904
6536 F20101203_AACEZQ jennings_w_Page_156thm.jpg
2fc4d0c4a64896b398b09dbe5c506d27
d30ab31a79dfd2a437ffbe5c654efa1a0ab62500
120821 F20101203_AACFFK jennings_w_Page_145.jp2
4b03fc5cee655bcb74ac3a7302de4594
b823a3eaa4737c8ca85d9d40d8a9a7c5ae8e3327
119240 F20101203_AACFEX jennings_w_Page_066.jp2
b48b1a93476b8a6699ed8e4df8595cc3
d4d53a83a11f4f53381ddb2466a173deeba897d6
F20101203_AACEZR jennings_w_Page_064.tif
304761c7a0f22ad65ea6b6f3d59cda80
a5de8be917bb682c6e0223f818789ffd99eca354
F20101203_AACFGA jennings_w_Page_043.tif
282ed82b4e0127aa8206733bac4ec575
99a7146799dbc2c355fa2edd43186f4835d7e58d
25551 F20101203_AACFFL jennings_w_Page_027.QC.jpg
a068b7de24f9229a9a30596ce7dd6f57
c7ea60c8c1ed2ce0b7c56d655b06e0a740339a80
53682 F20101203_AACFEY jennings_w_Page_054.pro
e49bac39c1305a087ff89e892a939789
07bc56276c1d465a5d020c322b5ce94761e4f0da
135289 F20101203_AACEZS jennings_w_Page_177.jp2
4207aa24e4a7eebc08cff7ee734331cb
3336441058ce63b112d95c9b69193d2ca3115421
2232 F20101203_AACFGB jennings_w_Page_165.txt
086b4da11db47eb83628a107d9c47620
82053eb7bf2015d176ddceb3fcff9873ed06a21d
53265 F20101203_AACFFM jennings_w_Page_110.pro
110bd96f93a26264147177a642832ce9
74d896d3df3812eae837767e1c97f8783f501b0a
67174 F20101203_AACFEZ jennings_w_Page_161.pro
01610871d6537452c2635e30196eccee
d61fb2036a49dc93cdd4da884f3cc0dd53d0762b
83146 F20101203_AACEZT jennings_w_Page_081.jpg
ceec0683d7999eb3b72b9167640d74e9
ce2e0d71dc0a142d09d023afcb638679c6619d39
8978 F20101203_AACFGC jennings_w_Page_181.QC.jpg
eab895d518f7339204bed24c21960363
330a9d546bec3acc75620e4de6b619a272deae28
53944 F20101203_AACFFN jennings_w_Page_094.pro
8daf6973401867456890d83de3d25609
53d2e01b44f5f84e942b26d26897ac8c7450f8a4
82613 F20101203_AACEZU jennings_w_Page_172.jpg
ddc5cb53789a70d8f244771b2c18d552
cd0f4962d008c931861d056ac70e978ff5c9bede
7083 F20101203_AACFGD jennings_w_Page_112thm.jpg
354c5e35ad6d1d752e24f85f75ddac77
521e345c50214f470605bed1af5d011ae80c6300
54834 F20101203_AACFFO jennings_w_Page_043.pro
ea27fc5ce6341180c8db59cd205bc5e0
e136198c79f3b6acb080f5d91df5d4dbb92d1c38
F20101203_AACEZV jennings_w_Page_016.tif
78ff0c0ae3541b6c01710376edad82ab
18a4e1c47405aba8f674bafe3b1b654f2bab81a6
6757 F20101203_AACFGE jennings_w_Page_121thm.jpg
54aabb1a33134a44ac131042e82bab60
c86638f3594df201a438ed554d46552e5e38073e
111126 F20101203_AACFFP jennings_w_Page_053.jp2
928c1baa1679c1d70809f466f8d04d15
86457d35ec8fa35cca9c908b214564ee9e4b6401
118201 F20101203_AACEZW jennings_w_Page_041.jp2
c6c9764e3faf037f7dca986e9a01a961
0fefc2ec5fdf6b01c98aeed4297f678f722a0861
7504 F20101203_AACFGF jennings_w_Page_001.QC.jpg
004f18f031032dad495e821909d110c4
0759964f3e7d02489bc8c44b8511aee3496a3cee
1417 F20101203_AACFFQ jennings_w_Page_133.txt
3fff678417be9825f6e728b2644f3890
40ebe2e03c4df12f8d1e083213ce1da311664c91
111922 F20101203_AACEZX jennings_w_Page_118.jp2
5769e5de8be491686fb5884ce747ec90
43b2e75b93cf5da844a715b68c646c937d1d56f4
118341 F20101203_AACFGG jennings_w_Page_027.jp2
38d4202a7ac3311a87a11c1601054621
19f74644c8b6be630331f2266ac46e1f5c6c1528
F20101203_AACFFR jennings_w_Page_031.tif
5fd3a0595c49e31261f560ace0dd50e3
a581d800315f141146fef5574af9fa52656d4bd9
33597 F20101203_AACEZY jennings_w_Page_127.pro
1893d6b959b6411a32e5237897b959a8
410f0d7c3b70a8a4c2854bbf0da16c1f0003afc3
108398 F20101203_AACFGH jennings_w_Page_024.jp2
8c1daf28d56618f0bcb72a457345132b
ac682d82c7f0baedd30211a72826b3cf9a3f760e
F20101203_AACFFS jennings_w_Page_080.tif
9aafc502f395f2999ccaf7a1c7b7907f
45e8d78b9b090f15eedc2a7ba91c970909126574
77882 F20101203_AACEZZ jennings_w_Page_041.jpg
538ac4676de8a682ac30f3407d0a13e5
1569a7a973d7d67263a41e3ab4e27f6787560f76
23994 F20101203_AACFGI jennings_w_Page_070.QC.jpg
a4f80c89ae08e4fa68f393f70b25e029
dfdbc5703d9fb841a2693c1ca95bb2763f4dde57
2902 F20101203_AACFFT jennings_w_Page_006.txt
fe7835fd58aab054a72253b6be57bd20
f6b6f8552f1131ae5e8b00c7dcb95d545cb9acd6
F20101203_AACFGJ jennings_w_Page_045.tif
d3b5d87065f85f2391398f576bb487a8
c5b2e23c342b6ccb3a4adb10e0ef7655b8d11e5b
23390 F20101203_AACFFU jennings_w_Page_032.QC.jpg
69dc5420713ccc34f50a90b882dffbfb
1a22dca6f6d560ddd61a10b5bc8144dfe6a72fe2
3404 F20101203_AACFGK jennings_w_Page_137thm.jpg
f52143db04d28f521393ca31b0a77307
0f88661591bc724575c9fae33e66276bcee41036
75081 F20101203_AACFFV jennings_w_Page_038.jpg
458cbb1df1bd1a5e325e2b5e4cb69a0f
fac6b1bda35fa8fcde4f483dcac51a41c9ef0f4e
7198 F20101203_AACFFW jennings_w_Page_117thm.jpg
6b0b1327b1a5964bb882b580c9aa8898
993bcec2d27b8eafd5ffa3e47861e6948fcfdbd6
122745 F20101203_AACFGL jennings_w_Page_035.jp2
115e4aaa5715ad3fcd96e54a96bcca76
1de3df88a05d0536ab6044c63ef6a04603d37501
26033 F20101203_AACFFX jennings_w_Page_162.QC.jpg
767c8cd0c365dcad11ae8f1f5b550ed0
0ca4efff5f9b472202b1d654303142724442dd1e
47978 F20101203_AACFHA jennings_w_Page_011.pro
e373efc00b63048ffc69ef009625bd94
b969f4165944ccbf9648b51ff9f7625b8b153320
F20101203_AACFGM jennings_w_Page_066.tif
9b96dac1911dd2dd26562dba15341b4c
add9b9e5bbf14e0374eb1a2df4ccaf600dc32989
76925 F20101203_AACFFY jennings_w_Page_096.jpg
5f58e6814ea6d479b09351571caa9c19
281e0928532e72a2047bf8072112284b48081b74
54026 F20101203_AACFHB jennings_w_Page_063.pro
331e4b5a8dca80bcdff73b5d8f791f42
90ee09e11aaecc3087ac433c8edb2487e7f9bb5e
134870 F20101203_AACFGN jennings_w_Page_176.jp2
86006ba53e8a8cf18daf996fd8b0077b
a2a9bf061566947fbdeb3b9d8e917fe4dc7c6a26
24075 F20101203_AACFFZ jennings_w_Page_084.QC.jpg
ca8a9f08e641e6e96024cd5cf963529e
d6433f214153846f2424af0e189125f622af8235
77800 F20101203_AACFHC jennings_w_Page_121.jpg
f6d2e9e9aea71e21485d61749ea425d0
45bdf05765b0cb5ab4822e05e8e12734c8fdf2e7
83371 F20101203_AACFGO jennings_w_Page_128.jp2
4ded51a6b1de7f395ea2d680766a31e2
20c33b7f00fb79563a454620e27fe825c8a3078c
119223 F20101203_AACFHD jennings_w_Page_063.jp2
4c1f41a651e401556719b1bc801043b0
a06f54f23d0170957ba5a55d421e18ca6cc6ec53
72345 F20101203_AACFGP jennings_w_Page_082.jpg
e48f224133b1ae14d564892203a743e1
2c7a6cf0eb24dd14d8ac26b6e37d3b43631829eb
59421 F20101203_AACFHE jennings_w_Page_107.pro
aea528f64f98b45304d5f938e6959081
f964d3501b2fbaa202b1ad4aaed38a2bd38e8bf2
119398 F20101203_AACFGQ jennings_w_Page_076.jp2
7007dfae095d0c9defbf911a158330d2
bc198cdb236c29e6e3f24f5eb02575806c78c798
26534 F20101203_AACFHF jennings_w_Page_052.QC.jpg
f9a9420796d1648613fd84c85a4f0b7d
0fbb64348a8482f6283b8824c9693db327a163a4
144622 F20101203_AACFGR jennings_w_Page_161.jp2
4058b635b063100cb9a0b0c3126ee3bc
dfa8082d4fe1b05a2ca709440db3f5b2689953e6
F20101203_AACFHG jennings_w_Page_082.tif
1883a6064ea36a2ddf9dbd6720b4f503
7378b5e91dfbb37cef3a5c3f485463cc04653dc0
6750 F20101203_AACFGS jennings_w_Page_089thm.jpg
5badc658038c3ac7c0ea0fd032f6b714
c607f1ca0d9a8bceb9992dddae1d18dfa10e1232
F20101203_AACFHH jennings_w_Page_057.tif
217041cd268933d2b132f08631c581d8
502dc575980d7d57c0a2486514d45a2dbaccbd64
6710 F20101203_AACFGT jennings_w_Page_091thm.jpg
c7ce5cedab0ae632a5f75beae61eb343
e2235cebed158d24fe53ee3ec8e3b3dc81969510
F20101203_AACFHI jennings_w_Page_095.tif
ae6bf29f42833d032ed5ee3a5b3d4c83
0a5fac8e6914ac9da595f271d405700530f82634
F20101203_AACFGU jennings_w_Page_005.tif
bc1e334f6a53d108dfb238acf7aee682
8223bcb50568c523563f23d4fae85858633d9268
6863 F20101203_AACFHJ jennings_w_Page_058thm.jpg
caa47961b014055c0d29a0270570bf8e
4a02676b107b17d58e4c28c576db26e1851c1400
61372 F20101203_AACFGV jennings_w_Page_159.pro
38fc4760edf9b2a9682a500e07a80621
c48a50f4ad01ed38d969566a4262bde6a83fd28e
54497 F20101203_AACFHK jennings_w_Page_020.pro
cab49256af8bf718dc86fd6c5e8ce141
eed5ded952940469bbe3d31ceb3e9e9b2932e9e2
F20101203_AACFGW jennings_w_Page_073.tif
9331a7aaa2325fada5f5868c88bfd4dc
0193b0b0e3c96f105ec516cc5052ee3c4b612428
2970 F20101203_AACFHL jennings_w_Page_133thm.jpg
f172e73b05681e0bccaf760a65cad692
099a9591a5b632fd8d12d84eec5778b7240b8f58
51001 F20101203_AACFGX jennings_w_Page_082.pro
eab23d48ffd4a970005040bee659aeed
830e704808aeb91b79e811a4fb142e5a51d81ed6
50802 F20101203_AACFIA jennings_w_Page_136.jpg
4dd74295f442aa9e320f1d4a9835746c
faafacc07a9da14b08e34837c65c90dfbf93175a
55171 F20101203_AACFGY jennings_w_Page_098.pro
332682abd04f8d735360a4faa9ca9fe4
af12ec7f3258187c216104e641f82a8a599ef80c
F20101203_AACFIB jennings_w_Page_131.tif
e289fcf0f86028855910db5eb7a39a5b
a35d878744d53b5995bdea8f09bed626d0081dcb
119655 F20101203_AACFHM jennings_w_Page_109.jp2
b19fd57bbeb8ed446f208490a235a800
fe7b06fc802989e9336c0253676fd10edc72c02c
F20101203_AACFGZ jennings_w_Page_053.tif
780f27ce98cb323921c7c829765daa5f
c595fbd39b50dcd80143bdb7372afd510d8a85ff
6914 F20101203_AACFIC jennings_w_Page_013thm.jpg
a50a74374fd81cf205eda614fb4de9a8
7fd065dd3e136931794e9be7584f9d2fe94fd079
76329 F20101203_AACFHN jennings_w_Page_108.jpg
7313e9cfcd3dd6c065b73a41e535e3f3
340b7293a1a10b11626b51d1fff269c255671e13
F20101203_AACFID jennings_w_Page_139.tif
77965ce4f32b02538a3b97031b481a65
549add1a47020c9490d8e0ed71602fcdb59663c8
F20101203_AACFHO jennings_w_Page_097.tif
baedb3573a6a1f3942af0b714a00a316
9b076b126f7996ff2bcd650c1e412a2dc9093a88
26941 F20101203_AACFIE jennings_w_Page_050.QC.jpg
5c569fe9a112abd25859e438b3d99efd
b2efe737f76177a9cb738e9bdc40a6b659eb85e0
528118 F20101203_AACFHP jennings_w_Page_143.jp2
da4d5ce9145aa0050ec5f2d667741c84
31507e541b944ddee65b69a3de85ae5d8aaf1917
F20101203_AACFIF jennings_w_Page_093.tif
82b782d0dfe6a7ae775cb3e915048e18
ea8a48fba93b4f9a5d448fbb25a42bf8ba883cd1
52377 F20101203_AACFHQ jennings_w_Page_013.pro
22aa05834a2eaf4d4c6980f17c266d4e
61d52b02edbaf0561bacc33b102b0963dda20f56
7044 F20101203_AACFIG jennings_w_Page_168thm.jpg
18e93b721aa536f15404713bd5625353
28e1968014165af22e437c558b3603e6095e7a41
F20101203_AACFHR jennings_w_Page_012.tif
b09a7eee2599c7d8c318e662d861d855
295bc331cc2310e0b030680aff7410f2a4d86940
10671 F20101203_AACFIH jennings_w_Page_125.pro
12b6ceab8ffaf5365d857205cb8bace5
d8c324417763a96a68ce51fbd1aaf059ec50351d
81988 F20101203_AACFHS jennings_w_Page_034.jpg
f8ffb9f66826b7cfe3b81ca9476ae21d
052fad16eaa40998f4a5d1ebedff44bb4f0984d0
6768 F20101203_AACFII jennings_w_Page_059thm.jpg
e4a86887625ab017bcf4438bca9ef940
739ae3872c3e041b67c8b89b1a04d725c9b47965
25493 F20101203_AACFHT jennings_w_Page_076.QC.jpg
622cf92fbcb7bf391c0e4cef1e0a67e6
91be64cdaf6f2abeeb6b94a2bb6964a7d618d79d
2191 F20101203_AACFIJ jennings_w_Page_036.txt
a96705858dc73daf15eba7235dffdcf6
5da54a380ce6fd1e09f6dff3d361edf1e1943532
119253 F20101203_AACFHU jennings_w_Page_060.jp2
87d5940a0cf534dc5d8b13708acbfa2d
b8ab8b46414ed49fcc3cc509dc8b268c22c451e9
7173 F20101203_AACFIK jennings_w_Page_147thm.jpg
e0f8d27b6d8b2dd5822e1bd0cc23f75a
6d77825ba7bfda75c2b1f05c5f53b5f51fcf0383
22232 F20101203_AACFHV jennings_w_Page_103.QC.jpg
0660996f43dbc86b65f2a8108353b89e
93f31506229aad79a7187f9e7aab8b0ade89033f
F20101203_AACFIL jennings_w_Page_110.tif
b71e35a07d71c3b51c488e7616f4ea98
ff33cf02d2f29e1b292604083aa6f11b4acbe1e1
54062 F20101203_AACFHW jennings_w_Page_021.pro
ffe86aa84568d3310e4c444cee89a292
3924460a5704903bb5bd2810570825b4ee0720d4
F20101203_AACFJA jennings_w_Page_006.QC.jpg
ab15423d04c1ff0edbe1e3010f3a6637
552bd9d1410fd5329d0d2fe19673a5811d441833
118837 F20101203_AACFIM jennings_w_Page_148.jp2
534686bf1cb8ddd34aaa2cc20ee4852e
26853544b634ac7401fdb3028033247e0a5f1b08
29854 F20101203_AACFHX jennings_w_Page_158.jpg
bf03f40c4acadd3dfd258189612f49ed
a931266e3a243d0c8e41c115a144083af8e13c31
9613 F20101203_AACFJB jennings_w_Page_158.QC.jpg
957ac915b5e60b6aabf8dc0665600d85
d26ca9265ea6c0fd2621b14bb0fd18515976e597
6849 F20101203_AACFHY jennings_w_Page_142.pro
59e6a189c65fbda48da8723fe12b2fa1
631f0d688eddff549f33b0380579a04115e9db95
51704 F20101203_AACEGA jennings_w_Page_032.pro
a0623c7f207313d7947e67dcfb5cc275
afedc3afeb45261e9d32425f65423846433114bc
6806 F20101203_AACFJC jennings_w_Page_063thm.jpg
94ff97192c108cff98cec0f31f8cfdfc
087615c80edd979d134cf0607085f4b3de894b8e
347 F20101203_AACFIN jennings_w_Page_141.txt
2f1880e47cf405a99b06a39fc3959c43
bd044477040889ea0e21d541594a5525ef7437ad
54556 F20101203_AACFHZ jennings_w_Page_096.pro
c897ada77d0b7e40b8592afb9a5c7665
1ca625e8c18bab71cfec9596a42336a08299cf0d
2086 F20101203_AACEGB jennings_w_Page_062.txt
3bd09c99da50467022c7b1a3ecb1aaed
4559b7280bfe52901139092ac40eef5973342163
81171 F20101203_AACFJD jennings_w_Page_122.jpg
6dc903f6b8d8db48f34cac4aba28b19b
74ed686b7f754bb41cd2131d21494ded86cbdf3f
F20101203_AACFIO jennings_w_Page_074.tif
dd5769f812b19d145768fbfed5cf3a1c
0c2798beac46796896f1cbd7ddeadb1d15a7ffc0
76648 F20101203_AACEGC jennings_w_Page_086.jpg
239966f4ae10762431b276d42bc4a2b2
c34c6b12c18204e2dbd13c228639a36b64dbdb29
60422 F20101203_AACFJE jennings_w_Page_016.pro
d54fa83370abbd4dd320ee5e63fe3012
39586f56131d33db7658a9e51abf23779c9f1157
2120 F20101203_AACFIP jennings_w_Page_054.txt
6039d583c1498de1a8baf55403ffbf04
14094bcaf8427bfe0957168fef5893e93ddde82d
8803 F20101203_AACEGD jennings_w_Page_001.pro
8db21954ca5fa56ed1ff07c7461fce0a
1351f429ae9e0745722385f370d04bd1eeb0d42c
F20101203_AACFJF jennings_w_Page_092.tif
fe961a86f36079bf971cef8da009eb51
ef022be9a4dc9d1d8138d041e6f45a752cf84d35
126548 F20101203_AACFIQ jennings_w_Page_039.jp2
d5d9dbf464b8c53a5964ac2cd455fc60
8d6606e7a96ae410ec124af6ba22d73c5c5b5f73
64892 F20101203_AACEGE jennings_w_Page_081.pro
4381f19a5d56bb9952c8a1158c77dcfa
428625c47ee1956d84ff01cbfd9d2aa2efdcd191
2203 F20101203_AACFJG jennings_w_Page_152.txt
f574dbef7ea7402f45458488487277d6
afd4f0ab8a8b501c4d6304a2c7f0943383bd329d
25236 F20101203_AACFIR jennings_w_Page_054.QC.jpg
f91f041264572598aaf16cca757ac3c6
1dc34ac4420c112a5c9c3ee46d1221a483369a49
23295 F20101203_AACEGF jennings_w_Page_156.QC.jpg
641c110873f16e501ae831c780c7a690
67a5e158428e7954fd0a5d8594e94fb5ebd3efb9
24943 F20101203_AACFJH jennings_w_Page_091.QC.jpg
6321468a211eec1b34885ec852e0e9f4
a552c36f33589ed185c4a5ee8833f14f1b3f461b
99371 F20101203_AACFIS jennings_w_Page_088.jp2
95f2f64dde04ffc33993f53a89a6ab09
204c269f5b7b80169d6637b525e327d73ecf7ef9
71163 F20101203_AACEGG jennings_w_Page_003.jp2
ea34619067f19ac19478fdc1aa3e7202
fcf4a123aa051ca90377333100e16d9a143227aa
140308 F20101203_AACFJI jennings_w_Page_163.jp2
5a471e2be82b4cc37927b926c7ff3b0e
69af9b431ffb0da6a20b261cb77fe39e69b6ee9c
F20101203_AACFIT jennings_w_Page_035.tif
954eb83982e55d0e1ab4cd7fa77a71bc
cf8d7cd3d3fc06cf60f5d4d0a914074009d84501
4501 F20101203_AACEGH jennings_w_Page_134thm.jpg
4f49b42c0b7995fa8903dbadbd22eb60
4572fc6d0262a0fa463141dd1dd7fea72c819fd3
112115 F20101203_AACFJJ jennings_w_Page_019.jp2
2cccfd6c9ee37f74bb7c616eaa2a2f1a
89afee51790702eef0c36efdac4aeba64018bb23
76092 F20101203_AACFIU jennings_w_Page_094.jpg
52b77819e8f4a5fcf48e8c868ee8e0b2
647a06c5a8160b587c7b091d73f30bc923cec051
2182 F20101203_AACEGI jennings_w_Page_123.txt
ac91133d896a00aaef11e60528e5b1d7
ca9ed95c930a8cf2b2b655e7541cfe5c725ae99d
68399 F20101203_AACFJK jennings_w_Page_164.pro
5eb5125f78c06849ffa9eeb68e372343
af6e387e1262e94c819b61d52f4fdac55b0d1f93
2341 F20101203_AACFIV jennings_w_Page_132thm.jpg
b6539c192e944d8e3ca23c03cfd440c7
7404ba52496baed8ceeaf08c1aa9b04d7318b36d
1974 F20101203_AACEGJ jennings_w_Page_067.txt
d00f154611bbcae55dfa9b764a43b970
a65d62b5503e9d359d17a4ad40e59dbc49ef155c
26511 F20101203_AACFJL jennings_w_Page_151.QC.jpg
cdf33171f616499314b5d1e500a35364
ea63634f750e87bda34689f0e00a17f84a8954e9
87856 F20101203_AACFIW jennings_w_Page_168.jpg
8b1bfad591d9df918f9f083663f25754
0ea611c1bc123531d897feabaddccd03d7773f19
133504 F20101203_AACEGK jennings_w_Page_042.jp2
b84b460af9ee4d9a1023209b651a9fc4
96919811377f9deb9a049584662a700c1a3bfc4b
96788 F20101203_AACFJM jennings_w_Page_163.jpg
c16a6d58dedebbb8d45895e1ba1c341f
a545fc3d3b47aac05e466a2a9bc37c3246788e79
133055 F20101203_AACFIX jennings_w_Page_166.jp2
6804cd3a75b1bff59320b985418796f7
b59a1263e2de7a8ba706f60959700bb383dc85d7
89924 F20101203_AACFKA jennings_w_Page_160.jpg
acdf0be12c27a3eb143e98bc0ac75575
34abb3c423ed2ac13de19412fa9d11e42ff18630
F20101203_AACEGL jennings_w_Page_014.tif
b34cafeb1028ffddd1eb20bd9e797cdb
d7e02981e39f3585040966d5a9f5dbaec3d51554
80035 F20101203_AACFJN jennings_w_Page_106.jpg
80913463080bf84d7b9be02a007fef44
b49a3e5a09a63960e49188950059e5685297af43
6650 F20101203_AACFIY jennings_w_Page_017thm.jpg
4512e8d299ac76a5b08310a02037e833
b7479e21a7e256589d490ba7bbe3714fb34b1715
F20101203_AACFKB jennings_w_Page_049.txt
bea34cbbad3aa16050a90b4df8d78000
fd8cd8bf059a7fc16f43c26333e926b955a9d97f
88271 F20101203_AACFIZ jennings_w_Page_075.jpg
0f4681651ba5345747ac2e64be5ce6c6
bbaaaeb6dce4bbc0a7a0f8c72c4043b573041ce2
F20101203_AACEHA jennings_w_Page_071.txt
f820810c13e660e6cfbfa7cebbd6c4d7
3db32f48908b744062161f1f301a2427c92dcc31
72303 F20101203_AACFKC jennings_w_Page_156.jpg
282af704c9749047a4853a7d92972e62
89b584750f9dcad51160ed2372c994b8129e8e32
58023 F20101203_AACFJO jennings_w_Page_035.pro
9acac117f86f4412f292b28eb2549a0b
e51c3c0e0b6eb774c829c76d9b867cde9a557b9c
F20101203_AACEFY jennings_w_Page_141.tif
5193c943f95e788390fdba8d3407a1c7
44c461ba0e591f755bfeee629351a9e51f62837d
113685 F20101203_AACEHB jennings_w_Page_097.jp2
147059e09d8cbab377d3d5ddac16f21e
1c0ab7bbc944e2502969ab7a2b8f3e30fbf03ffa
23557 F20101203_AACFKD jennings_w_Page_165.QC.jpg
e65036543d9cde1978f3e42135a04813
28c654f8bb9031ab566b28e6db4f0dd6b523bd65
45106 F20101203_AACEGM jennings_w_Page_131.jpg
ba022cbb08e65239d8ba01350c8b6d35
51245b077303932c8934924de2b6ff0974582bd7
F20101203_AACFKE jennings_w_Page_149.tif
18e0cad8493298e1649112621b76d528
071af956886ffa4183a63a4f4ca34fb2911a3800
2577 F20101203_AACFJP jennings_w_Page_167.txt
ccf2aa216db2c20bac0981d3cf2351a0
06b95ece7e52a0e6dd6a0e9624204c0b0b60f5e4
79358 F20101203_AACEFZ jennings_w_Page_051.jpg
93c7c5268ab4137d7e7264108e0ba202
4943c621bedba394423938acdb0944a0ed5c10dc
122918 F20101203_AACEHC jennings_w_Page_112.jp2
0331d1afafadf9d87e43fb36b7f9ee74
84942300e6c72078c1d81756d9b3f162c0e83a65
116479 F20101203_AACEGN jennings_w_Page_054.jp2
8efc77a025d31ef24453f67d6547c29e
6d2fb52ab64986d632fc64a97d5f8f9627f762e9
69426 F20101203_AACFKF jennings_w_Page_009.jpg
285d61682d0b0e416ab26f3e80b90c1a
85626f5b4b336ada514e931fa71d9241f5fba0c6
26241 F20101203_AACFJQ jennings_w_Page_042.QC.jpg
586894b947b712b143fa2193a457e94a
d5e0530a28806b0d6ee509859c1efc69ffd4bec5
7076 F20101203_AACEHD jennings_w_Page_151thm.jpg
23b2fc2ae79cf54bfa53a03cd1182bf1
32ceb9d9eb2d33d5ed98263e78a4c3fa06014a0a
F20101203_AACEGO jennings_w_Page_137.tif
69263d342f92e3769c53669f777ed020
9f4109fad6516f2e8de93e1750ee89b8116db81b
6689 F20101203_AACFKG jennings_w_Page_097thm.jpg
49b21e06264241fe2d9feb4c0869b76d
5038862ed38e746d5151df35e60d258ac82273d6
53340 F20101203_AACFJR jennings_w_Page_078.pro
91db93ee1d5032ecd058944b48370f8b
664110d8481e61aef0663f4092c0e2fb5c2ddb9b
6711 F20101203_AACEHE jennings_w_Page_090thm.jpg
840ca83bb7a3860b7c50fd006c2638dd
99e09464edea4581da0fe26ad8d8a6726c4d887a
F20101203_AACEGP jennings_w_Page_176.tif
095c996558a68aa002ac3a942238cbb6
285f2585fc7e99f99b808ba5d7c56069adcf05eb
6668 F20101203_AACFKH jennings_w_Page_105thm.jpg
ab31e87a2d0dad901c0fb25b13037536
8eafd6fff265eb1edd4f952637bc67a45671a6e4
2184 F20101203_AACFJS jennings_w_Page_022.txt
7fb733c9571508ec7e8313561c18f0f2
7419ad0cbb6e0b967bebf53b6895ff1ddfd9bf2f
7119 F20101203_AACEHF jennings_w_Page_115thm.jpg
0847bea7c82500a64055589c5c754232
e3d8448ff5f27c4f1b178c39d9acffde1442de8f
F20101203_AACEGQ jennings_w_Page_070.tif
2362b08424548b6b6bcdfa7cc30029e0
2b350caaa0656a313d62e74d0c1ef9dee88f50f9
26003 F20101203_AACFKI jennings_w_Page_154.QC.jpg
3ee445cd698433c2206e6c8888fcb9bb
14bd3df8eef91bd20d851288958b86c03f2296af
26829 F20101203_AACFJT jennings_w_Page_119.QC.jpg
affa0cac26c5357a014f634205e8e8fc
8c054ce8e2f4d55bbb1fbd2bbd3c71605ba60a43
1051983 F20101203_AACEHG jennings_w_Page_006.jp2
9e8ceda2567d18874c4db0bfe0335cca
1fefc6aea1df4b7792bf0347a3fc6ab0d86b240c
116944 F20101203_AACEGR jennings_w_Page_086.jp2
f3fd8b6176bbb0a24e922175d00f07f1
baec5487131795f92a7e3139c6617dcc9755d8f2
110560 F20101203_AACFKJ jennings_w_Page_082.jp2
adf32fbd7196d690d3c9b50a3b286299
fe62a76f9346884bea93d02c57accf640ea7dfa6
F20101203_AACFJU jennings_w_Page_124.txt
ab626a971d69bc7399768d83dc104f47
8113fd74fc8dc8344bfdbcb8356334093d7a856f
2593 F20101203_AACEHH jennings_w_Page_175.txt
e1104717990cb0a5e777bdffeae24bf6
6c1e70b6891964e6822a2f0269b43c09983e3a12
2292 F20101203_AACEGS jennings_w_Page_068.txt
a9876987d800df99cb88cc2a2a2dfcaf
ef2510b308e00dc3abce31cbb12d174f71a0384d
54632 F20101203_AACFKK jennings_w_Page_060.pro
686e584ff8568cee694e6fc16a8592c1
7045f3293b129ea98e786ed22cd6f9de2318d093
6991 F20101203_AACFJV jennings_w_Page_169thm.jpg
9ab8f3554821b2aa1aad26425a6f3ab0
07c02ea35653b771c8f3643fc9846f687e4fc110
2270 F20101203_AACEHI jennings_w_Page_083.txt
9f49162421865b225dc124610c1097a9
024dbd3856dbd3778a7b303c993c90f0ab2cb106
55860 F20101203_AACEGT jennings_w_Page_040.pro
d6b05c0fd39b942358a18aac97061e09
9526d89c30313050860ab4ee431962b79613a2be
22295 F20101203_AACFKL jennings_w_Page_132.pro
2d46617cc5ec9a519f2399bc22eb1d9a
324e607d9dac43b77b49c9f6d8efdf4fc6de5240
2061 F20101203_AACFJW jennings_w_Page_013.txt
735b411043fa5151edd5d445639f8364
2d9f638c1a930079e6f9bc92e2f81efc9c918eaa
52423 F20101203_AACEHJ jennings_w_Page_038.pro
a841d4dfbc5e46cc2f5823121db01cc3
e1ab18ed0cfa78ccfeb0c469abba3b26903b7ed7
7001 F20101203_AACEGU jennings_w_Page_036thm.jpg
ca4d058e94972cb2f83a261cdff7c681
b55849fefc30b5a88a871ca251104ecc66587e6a







TRAJECTORIES OF TWO RACINE BIRTH COHORTS: A THEORETICALLY
INTEGRATED MODEL FOR EXPLAINING OFFENDING




















By

WESLEY G. JENNINGS


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA

2007



































2007 Wesley G. Jennings









ACKNOWLEDGMENTS

I would like to provide a brief acknowledgement to all of those who have helped me along

the way to achieve this ultimate goal. First and foremost, I would like to thank my family for all

of their continued love and support throughout my many, many years of schooling. They have

been a source of strength and financial support. I would also like to sincerely thank my wife for

her willingness to understand how long this process may take and working hard to support us

financially throughout my graduate education. Lastly, I would like to offer an extreme amount

of gratitude to all of the faculty at the Department of Criminology, Law, and Society, and

specifically to my committee members who were steadfast in their support throughout my entire

graduate student career. Alex Piquero, Lonn Lanza-Kaduce, Brian Stults, Karen Parker, and

Chuck Peek played an extremely integral role in my success in completing this dissertation and

have provided me with excellent advice, criticisms, and thoughts that have certainly helped me

mature as an academic professional. For this I will always be grateful and I hope to continue

collaborating with all of these individuals for a long, long time.










TABLE OF CONTENTS

page

A CK N O W LED G M EN TS ................................................................. ........... ............. 3

L IS T O F T A B L E S ................................................................................. 6

LIST O F FIG U RE S ................................................................. 8

ABSTRAC T ...........................................................................................

CHAPTER

1 IN T R O D U C T IO N ............................................................... ...................................1 1

In tro d u c tio n .................................................................................................................... 1 1
B a c k g ro u n d ............................................................................................................. 1 1
P re sen t S tu dy ................................................................15
L ay ou t o f D issertatio n .................................................................. ..................................16

2 THEORETICAL PERSPECTIVES......... .............................................. ...... ............. 18

In tro d u c tio n ....................................................................................................................... 1 8
W hat is a C om m unity? ...........................................................................20
Social Disorganization....................................... 21
Age-Crime Relationship and Criminal Careers ........................................................26
Developmental Criminology/Life-Course Theory ...................................... ............. 33
Sampson and Laub's Age-Graded Theory of Informal Social Control ...........................36
M offitt's T axonom y ................. ..................................................... ................... ..... ..... 40

3 REVIEW OF THE LITERATURE .............................................................45

Introduction .. ....... ...... .................... .............................................. ...............45
Impact of Social Disorganization at the Individual Level ........ .................................. ....45
Prevalence and Incidence of Chronic Offending .................... .......................... 50
The Effect of Life Transitions on Offending .......................... .................... 54
The Saliency of Moffitt's Taxonomy across Sex ..............................................58
Prior Applications of Group-Based Models of Offending ...........................................61
Integrated Theoretical Model for Explaining Offending..............................................65
H ypotheses and E expectations ........6.. ........ .. .. .......... ...... ................ .............. 68

4 D A TA A N D M ETH O D S .................................................................................... 70

Introduction ................... .............................................70
D ata Sources ............................................. ..................71
Independent V ariables ..... ................................. .. ........................................... 73
M acro-level independent variables ........................................ ....... ............... 73



4









M icro-level independent variables .................................... ............. .................. 74
D ep en dent V ariab les ................. ............................................ ........ .......... .. .... 1
Study Lim stations ............................................. 82
M eth o d ology ................... ...................8...................5..........
G roup-B asked M modeling ..................................................... .................................. 85
A Seductive or Useful Statistical Method? ........................................ ............... 90
Analytic Procedure .................... .......................................96
T rajectory estim ation .......... .............................................................. .. .... .. .. ... 96
M ultinom ial logistic regression.......................................... ........... ............... 97
S tu d y S tre n g th s ........................................................................................................ 1 0 1

5 R E S U L T S .......................................................................................................1 0 3

In tro d u ctio n ................... ...................1.............................3
D descriptive Statistics ....................... ............................ .. .. ...... ............ 103
Independent Variables .................. ............................ ....... ................. 103
D dependent V variable ................................................................ .. ..... .. 105
Trajectory Analysis................ ......................... ............ 105
Analysis of Variance .................. ..................................................1..116
M ultivariate Analysis ................................................................. .... ........ 118
C onclu sions.......... .............................. ...............................................122

6 DISCU SSION AND CON CLU SION ...............................................................................144

D isc u ssio n ................... ......................................................................... 14 4
L im itatio n s ............... ... .. ................................. .................. ................ 14 7
D directions for Future Research .......................................................................... 150
Im p lic atio n s ............................................................................. 1 5 3
F in al T h ou ghts ...................................... ................................................... 156

APPENDIX: ZERO-ORDER CORRELATIONS BY COHORT..............................................158

R E F E R E N C E S ............................................................................... 159

B IO G R A PH IC A L SK E T C H .............................................................................. ... ............ 181









LIST OF TABLES


Table page

4-1 Resident Samples by Cohort, Size, Data Sources, and Offender Groups........................102

4-2 Observational Periods of Data Sources by Cohort ..................................... ..................102

4-3 Social Disorganization Index Factor Loadings by Cohort................................... 102

5-1 Descriptive Statistics for the Macro-level and Micro-level Measures by Cohort ..........126

5-2 Tabulated Bayesian Information Criteria (BIC) for 1942 Racine Continuous
R e sid en ts ...................................... .................................................... 12 6

5-3 Mean (Median) Posterior Probabilities for Group Assignments: 1942 Racine
C continuous R residents ........................................... ...... ...................... 127

5-4 Coefficient Estimates for Effect of Sex (being Male) on Trajectory Group
M membership: 1942 Racine Continuous Residents..........................................................127

5-5 Tabulated Bayesian Information Criteria (BIC) for 1942 Racine Continuous
Interview ed R residents .................. .............................. ............... ..... 127

5-6 Mean (Median) Posterior Probabilities for Group Assignments: 1942 Racine
Continuous Interviewed Residents ........................................................ ............... 128

5-7 Coefficient Estimates for Effect of Sex (being Male) on Trajectory Group
Membership: 1942 Racine Continuous Interviewed Residents ............ ... ................. 128

5-8 Tabulated Bayesian Information Criteria (BIC) for 1942 Racine Continuous
Interviewed Residents i/ll Block Data.................................. .... ................. 128

5-9 Mean (Median) Posterior Probabilities for Group Assignments: 1942 Racine
Continuous Interviewed Residents i/ll Block Data.................................................. 128

5-10 Coefficient Estimates for Effect of Sex (being Male) on Trajectory Group
Membership: 1942 Racine Continuous Interviewed Residents i, i/ll Block Data ............129

5-11 Tabulated Bayesian Information Criteria (BIC) for 1949 Racine Continuous
R e sid en ts ...................................... .................................................... 12 9

5-12 Mean (Median) Posterior Probabilities for Group Assignments: 1949 Racine
C continuous R residents ........................................... ...... ...................... 129

5-13 Coefficient Estimates for Effect of Sex (being Male) on Trajectory Group
Membership: 1949 Racine Continuous Residents........... .................. ......... ..... 130









5-14 Tabulated Bayesian Information Criteria (BIC) for 1949 Racine Continuous
Interview ed R residents .................. .............................. ...................... 130

5-15 Mean (Median) Posterior Probabilities for Group Assignments: 1949 Racine
Continuous Interviewed Residents ........................................................ ............... 130

5-16 Coefficient Estimates for Effect of Sex (being Male) on Trajectory Group
Membership: 1949 Racine Continuous Interviewed Residents ............ ... ........... 130

5-17 Tabulated Bayesian Information Criteria (BIC) for 1949 Racine Continuous
Interviewed Residents i /hil Block Data............................ ....... ................. 131

5-18 Mean (Median) Posterior Probabilities for Group Assignments: 1949 Racine
Continuous Interviewed Residents i i/il Block Data..................................................131

5-19 Coefficient Estimates for Effect of Sex (being Male) on Trajectory Group
Membership: 1949 Racine Continuous Interviewed Residents i, i/lh Block Data ............131

5-20 Trajectory Groups by Sample and Cohort.................................................................... 132

5-21 Mean Covariate Levels by Trajectory Group Membership and Cohort..........................133

5-22 1942 Racine Continuous Interviewed Residents ill/ Block Data: Odds Ratios and
Probabilities for Macro-Level Predictors of Offending Group Membership ..................134

5-23 1942 Racine Continuous Interviewed Residents ii il/ Block Data: Odds Ratios and
Probabilities for Micro-Level Predictors of Offending Group Membership ...................134

5-24 1942 Racine Continuous Interviewed Residents ii il/ Block Data: Odds Ratios and
Probabilities for Macro-level and Micro-level Predictors of Offending Group
M em b ership .............................................................................135

5-25 1949 Racine Continuous Interviewed Residents ii i/h Block Data: Odds Ratios and
Probabilities for Macro-Level Predictors of Offending Group Membership ..................136

5-26 1949 Racine Continuous Interviewed Residents ii il/ Block Data: Odds Ratios and
Probabilities for Micro-Level Predictors of Offending Group Membership ................... 136

5-27 1949 Racine Continuous Interviewed Residents ii i/h Block Data: Odds Ratios and
Probabilities for Macro-level and Micro-level Predictors of Offending Group
M em b ersh ip ...............................................................13 7

A -1 1942 C oh ort ...............................................................158

A -2 194 9 C oh ort ...............................................................158









LIST OF FIGURES


Figure page

3-1 Integrated Theoretical Model: Social Disorganization, An Age-Graded Theory of
Informal Social Control and Moffitt's Developmental Taxonomy .................................69

5-1 1942 Racine Continuous Residents Group-Based Trajectories.................................138

5-2 1942 Racine Continuous Interviewed Residents Group-Based Trajectories.................139

5-3 1942 Racine Continuous Interviewed Residents i, il/ Block Data Group-Based
T raje cto rie s ........................................................................... 14 0

5-4 1949 Racine Continuous Residents Group-Based Trajectories ................................141

5-5 1949 Racine Continuous Interviewed Residents Group-Based Trajectories.................142

5-6 1942 Racine Continuous Interviewed Residents i, il/ Block Data Group-Based
T raje cto rie s ........................................................................... 14 3









Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

TRAJECTORIES OF TWO RACINE BIRTH COHORTS: A THEORETICALLY
INTEGRATED MODEL FOR EXPLAINING OFFENDING

By

Wesley G. Jennings

December 2007

Chair: Lonn Lanza-Kaduce
Cochair: Alex Piquero
Major: Criminology, Law, and Society

Macro-level research evidencing the relationship between communities and crime dates

back to the early ecological work of Park and Burgess and later further developed by Shaw and

McKay into what is commonly referred to as social disorganization theory. This research has

since shown that poverty, ethnic heterogeneity, and residential mobility lead to the inability for

members of the community to develop mechanisms of informal social control, which in turn

leads to higher rates of crime.

In contrast, micro-level theories such as Sampson and Laub's theory argues that individual

pathways of offending can be altered by life transitions such as marriage and employment.

Comparatively, Moffitt's taxonomy asserts that there are two theoretically distinct groups of

offenders, where one group displays an adolescent-limited type of offending and the other group

are considered to be life-course-persistent or chronic offenders. Moffitt has also argued that the

life-course-persistent type is predominantly a male phenomenon.

This study will provide one of the first attempts to explore the possibility of developing a

theoretically integrated model for explaining group offending using trajectory analysis and

relying on data from two Racine Birth Cohorts. Results from this study relying on official data









yielded varying levels of support for all of the theoretical frameworks mentioned above.

Specifically, all of the trajectory models discovered groups consistent with Moffitt's typology, as

well as uncovering groups not necessarily consistent with her typology. Also consistent with

Moffitt, sex was shown to be a significant covariate for predicting group assignment. Bivariate

support was found for Sampson and Laub's age-graded theory of informal social control. Lastly,

multivariate support was revealed for the impact of living in a socially disorganized area. Yet,

once mlogit models were estimated incorporating the individual and neighborhood effects there

was evidence of mediation. It appeared that individual and family variables at times completely

or at least partially mediated the effect of social disorganization as a distinguishing feature of

group membership. Implications for the current findings and directions for future research are

also discussed.









CHAPTER 1
INTRODUCTION

Introduction

There exists a large body of research that has identified the association between the

variation in the rate of offenders and community structural dynamics, most notably in the social

disorganization literature (Bursik & Grasmick, 1993). Similarly, there has been a long line of

research that has linked individual-level factors to involvement in offending (Farrington, 1995).

Yet, few efforts combine these levels of analysis. On this score, Sampson (1997:32) provided an

accurate summary of the current state of research in integrated approaches to studying offending

in noting that "few studies have successfully demonstrated a unified approach to individual and

neighborhood-level dimensions of crime."

The primary focus of this dissertation is to present empirical data that will shed light on the

question of the extent to which macro-level, micro-level, or some combination of factors best

relates to offending over the life course. In order to answers these questions, this study relies on

extensive longitudinal data from two birth cohorts of Racine, Wisconsin residents (Shannon,

1988, 1991) and utilizes a novel methodology for identifying latent classes of offenders. After

classifying the Racine residents into offender groups, a series of multinomial regression models

will allow for an assessment of the relative strength macro-level and micro-level factors for

increasing and/or decreasing the likelihood of belonging to one trajectory group as opposed to

another.

Background

While a handful of studies have compared rates of offending with individual/family SES

and area (or school) SES and have found that the highest rates of offending generally were

observed in the lowest SES areas among the individuals with the lowest SES (Braithwaite, 1979;









Reiss & Rhodes, 1961; Wikstrom, 1991a), only a few studies have investigated the influence of

other individual and community covariates on offending. Lindstrom (1996) found that the level

of family interaction was significantly associated with juvenile offending and this relationship

was the most pronounced in more socially disadvantaged neighborhoods. Kupersmidt et al.

(1995) showed that juveniles living in single-parent families in low SES neighborhoods were

more likely to be aggressive than their counterparts in high SES neighborhoods. Relying on a

comparison of two different groups of Pittsburgh neighborhoods (i.e., underclass and non-

underclass), Peeples and Loeber (1994) indicated that the relationship between race and juvenile

offending was only apparent in the underclass neighborhoods.

Even less research has targeted the integration of individual and neighborhood effects on

offending under the auspices of the criminal career paradigm (Blumstein et al., 1986) and recent

developmental frameworks. Wikstrom (1991b) conducted one of the first studies to address this

deficiency in the existing research and found that individuals that resided in the most socially

disadvantaged areas had the highest levels of both participation and frequency of offending.

Loeber and Wikstrom (1993) also showed that not only did juveniles offend at higher rates (i.e.,

greater participation and frequency) in low SES neighborhoods when compared with those from

neighborhoods with higher SES a significantly larger proportion of the offenders in the low SES

neighborhoods became serious offenders.

In a more recent study, Wikstrom and Loeber (2000) examined whether the interaction of

risk and/or protective factors and community context had an influence on early and late-onset

offending among three samples of boys in 90 Pittsburgh neighborhoods. They found that

neighborhood socioeconomic context did not have a direct effect on the prevalence of offending

for boys who had a high number of risk factors, but for those who had balanced risk-protective









factor scores, neighborhood context mattered. In addition, they indicated that individual risk-

protective scores had a greater impact on the rate of early-onset offending (before age 13) than

did neighborhood context, but that a direct neighborhood effect was observed for late-onset

offending (after age 12).

Lynam et al. (2000) found that the effects of impulsivity on juvenile offending was

stronger in poor neighborhoods; however, living in a disadvantaged neighborhood did not

heighten nonimpulsive boys' risk for juvenile offending when compared with residing in more

advantaged neighborhoods. Lynam et al. (2000) also noted that the direct neighborhood effects

on individual juvenile delinquency were relatively weak, whereas the neighborhood effects were

stronger when using self-reported measures of the quality of the neighborhood obtained directly

from the inner-city youth.

In a more recent study, Piquero, Moffitt, and Lawton (2005) investigated the possible

interaction effects of individual, familial, and neighborhood-level factors on life-course-

persistent offending. Their data were derived from the National Collaborative Perinatal Project

(Baltimore site) and consisted of information from the original NCPP study that included early

childhood and census-tract measures along with additionally including follow-up interviews with

the children in the 1990s. Piquero et al. (2005) failed to find direct effects for either

neighborhood or individual risk for life-course-persistent offending; however, they did find that

individual, family, and neighborhood-level effects predicted life-course-persistent offending

patterns among African-Americans in disadvantaged neighborhoods. In addition, while Piquero

et al. (2005:230) concluded that their analysis demonstrated similarities across race in life-

course-persistent patterns of offending, there were marked differences in the neighborhood

contexts where the Whites and African-Americans lived.









With attention to the relative arbitrariness of how prior research has defined offender

groups, more recent research applying Nagin and Land's (1993) group-based modeling technique

has begun to identify latent classes of offenders that exhibit distinct pathways of offending (for

review see Piquero, 2007; see also Nagin, 2005). The consistent finding across these studies has

been the identification of three main types of offending groups: (1) non-offenders; (2)

adolescent-limited offenders; and (3) chronic or high-rate offenders. D'Unger, Land, McCall,

and Nagin (1988) were also able to discover similar offending types across a variety of

longitudinal data sources. It is important to emphasize here that the data relied on in this study is

official data not self-report; therefore it is likely that the non-offender group will be much larger

than it would be if the trajectories were estimated using self-report data instead.

Additionally, D'Unger, Land, and McCall (2002) found evidence of relatively comparable

latent offending classes using data from the Second Philadelphia Birth Cohort (Tracy, Wolfgang,

& Figlio, 1990); however, the adolescent-limited group they uncovered was characterized as

adolescent-peaked because of their elevated levels of offending in adolescence. This adolescent-

peaked grouped has also been noted by other researchers (for review, see Piquero, 2007). Yet,

other groups have also emerged such as low-level chronic or late-onsetters (see D'Unger et al.,

1998, 2002).

Prior applications of this methodology have also found similar offender groups in male and

female trajectories (Cote et al., 2001, 2002; D'Unger et al., 2002; Fergusson & Horwood, 2002;

Land, 2000; Piquero, Brame, & Moffitt, 2005) and other research has identified factors other

than sex such as low socioeconomic status (McDermott & Nagin, 1998) and peer and school

influences (Chung et al., 2001) as being strongly related to the likelihood of being classified as a

chronic offender. Land (2000) has also shown that sex was a significant predictor of chronic









offending patterns, although its effect was completely mediated by the inclusion of neighborhood

context.

Present Study

This current study contributes to the literature in following ways. First, by drawing from

the present advancements in the criminological literature regarding the application of trajectory

analysis for classifying offenders (Nagin & Land, 1993; Nagin, 2005), this study seeks to

provide another replication of the latent class groups that have been previously identified by

prior researchers (for review, see Piquero, 2007) such as the non-offenders, adolescent-

limited/peaked offenders and chronic offenders across two Racine Birth Cohorts. Second, this

study seeks to investigate whether sex is a distinguishing feature of group-based trajectories as

has been discussed in prior applications of this group-based modeling technique (see Cote et al.,

2001, 2002; D'Unger et al., 2002; Fergusson & Horwood, 2002; Land, 2000; Piquero et al.,

2005). Third, this research will include an exploration into the impact that social disorganization

has on individual offending patterns. Fourth, guided by two prominent developmental

theoretical frameworks, specifically Sampson and Laub's (1993) age-graded theory of informal

social control and Moffitt's (1993) taxonomy of adolescent-limited and life-course-persistent

offenders, this study seeks to explore the impact of key life transitions (marriage, employment)

on desistance from offending in adulthood and assess the viability of Moffitt's taxonomy.

Fourth, the saliency of Moffitt's (1994; see also Moffitt et al., 1994) claims that males are more

likely to be life-course-persistent offenders than females will also be explored. Finally, this

study seeks to address the rather limited amount of research integrating macro-level and micro-









level factors that may simultaneously affect an individuals' pathways of offending using data

from two of Shannon's (1988, 1991) Racine Birth Cohorts.1

Layout of Dissertation

This dissertation will consist of six chapters. Chapter 2 will begin with a discussion of the

various theoretical frameworks that provide the basis for the study. More specifically, it will

begin with a brief description of how scholars have defined a "community" over time followed

by an outline of the origins and theoretical tenets of social disorganization theory. Next will be

an explanation of the relationship between age and crime and how research on the age-crime

curve complements the emergence of the criminal career paradigm. Similarly, the narrative will

continue with a discussion of how the criminal career paradigm had a direct contribution to the

development of what has become known as developmental/life-course criminology. This chapter

will conclude with a presentation of Sampson and Laub's (1993) age-graded theory of informal

social control (see also Laub & Sampson, 1993) and Moffitt's (1993) taxonomy of offending, the

latter providing the theoretical rationale for hypothesizing why different classes of offenders may

exist and the former describing certain life transitions that may affect an individual's offending

trajectory.

Chapter 3 will continue by offering an in-depth exploration into the prior extant literature

that can be found examining each of the theoretical frameworks mentioned above. Most notably,

this chapter will illustrate the links between social disorganization and individual patterns of

offending, along with addressing the prior research on chronic offending, crime over the life-

course (i.e., Sampson and Laub's age-graded theory of informal social control), and offer a


1 This study relies on data from the 1942 and 1949 Racine Birth Cohorts, as opposed to all three of the cohorts
(1942, 1949, 1955), because there is not any self-reported interview data available for the 1955 cohort precluding its
use for investigating the particular individual-level factors addressed in this research. According to Shannon (1988,
1991), lack of continued federal funding prevented him from interviewing the most recent 1955 cohort.









description of studies that have supported Moffitt's (1993) proposed taxonomy and also

demonstrated evidence that life-course-persistent offending may be disproportionately found

among males. This chapter concludes with a presentation of the findings from several recent

applications of Nagin and Land's (1993) methodological technique for identifying group-based

offending trajectories.

Chapter 4 will provide a systematic explanation of the source of the data and their

limitations for the current research along with an in-depth description of the 1942 and 1949

Racine Birth Cohorts used in the analysis. The sample and data information will be followed by

a comprehensive identification of all of the measures incorporated in this study including the

macro-level and micro-level factors and the two dependent measures (number of police contacts

and offending group membership). The final section of this chapter will explain the stages of the

analysis as far as the ordering of the analytic procedures and the model specifications and will

conclude with a brief discussion of the strengths of the study.

Chapter 5 will present the findings from all of the statistical analyses including the results

from (1) the group-based trajectory estimations and model selection criteria; (2) a discussion of

the importance of sex in predicting group assignment; (3) description of factors that are

associated with group assignment; and (4) an examination of the impact of macro-level and

micro-level variables (both separately and simultaneously) for significantly distinguishing

between trajectory groups. The final chapter (Chapter 6) will begin with a further discussion on

the findings as well as limitations of this research. This section will be followed by an offering

of some suggestions and guidance for future research on mutli-level modeling and theoretical

integration and will conclude with a few final thoughts.









CHAPTER 2
THEORETICAL PERSPECTIVES

Introduction

The origins of macro-level research, specifically social disorganization, can be traced back

to the early work in the first half of the twentieth century by Park and Burgess (1925) and the

later expansions by Shaw and McKay (1942). These researchers and others that followed in their

footsteps have formed what is commonly referred to today as the Chicago School. These

researchers, most notably Shaw and McKay, were influenced by the deep-rooted sociological

tradition of studying the ecology of crime and focused on explaining why crime rates not

individual criminality was higher in certain areas of the community. This tradition of studying

and theorizing about crime in the aggregate persisted for some time until there became an

increased interest in studying individual-level factors that affect offending, which was strongly

influenced by the increased use of survey methodology in sociology and criminology. This early

individual-level research sparked some of the most influential criminological theories that exist

today including social learning theory (Burgess & Akers, 1966; Akers et al., 1979) and social

control/bonding theory (Hirschi, 1969).

By the late 1970s and early 1980s individual-level research had evolved in the sociological

and criminological literature and led to what is still considered one of the most robust

relationships in the social sciences, i.e., the age-crime relationship, or the notion that crime rises

to a peak in middle to late adolescence followed by a precipitous decline in early adulthood.

Hirschi and Gottfredson (1983) argued that there is no need for any theory to account for the

relationship between age and crime because it is invariant across culture, time, and space.

Although, there has been a relatively strong amount of debate over this early assertion (which

will be returned to later), this along with evidence from Wolfgang et al.'s (1972) research on









chronic offending fueled the development of the criminal career paradigm. This perspective

involves a series of dimensions that will be addressed below, but most importantly, this

framework highlighted the similarities and differences among offenders and their criminal

careers (see Blumstein et al., 1986).

Coinciding with this new research interest was the birth of developmental criminology or

life-course criminology. This theoretical perspective focuses on how involvement in crime

changes over the life course. It highlights factors such as life transitions or turning points that

can alter an individual's offending trajectory (Elder, 1985). In addition, Sampson and Laub

(1993) formulated an age-graded theory of informal social control that recognizes that different

types of individuals have different offending careers and that there are age-dependent or age-

graded processes that influence, shape, or alter one's pathway of offending. In comparison, one

of the most recognized developmental theories of crime (other than Sampson and Laub's age-

graded theory of informal social control) is that of Moffitt's (1993) taxonomy of adolescent-

limited and life-course-persistent offenders.

Moffitt argues that there are three theoretically and qualitatively distinct groups of

individuals including: (1) the non-offenders; (2) the adolescent-limited offenders; and (3) the

life-course-persistent offenders. Furthermore, she also argues that the patterns of offending for

the two latter groups (i.e., the groups who are actually involved in offending) are driven by

different causal processes, have different ages of onset, and are particularly distinct in their

frequency of offending and career lengths. All of the theoretical issues touched upon above will

be the focus for the remainder of this chapter, but it is first important to briefly visit the debate

among scholars (sociologists and criminologists alike) on what is a community or neighborhood.









This definitional disagreement over the "appropriate" unit of analysis has been at the heart of

macro-level research for some time.

What is a Community?

Community theory gained increased attention within the realms of sociology in response to

the ecological work of Park and his colleagues at the University of Chicago in the 1920s and

1930s. Their study of the dynamics of urban life in Chicago supported the idea that variables

such as population size, density, and heterogeneity interact to produce communities that can be

characterized as "urban." Following a brief decline in community-based sociological research in

the 1950s and 1960s, the 1970s and 1980s hosted a general resurgence in attention devoted to the

"community/neighborhood" debate. This movement was fueled by research that relied on using

local community areas (Heitgerd & Bursik, 1987; Bursik, 1984, 1986), census tracts or some

aggregation thereof (Schuerman & Kobrin, 1986; Shannon, 1988, 1991; Taylor & Covington,

1988), police districts (Shannon, 1988, 1991), and even electoral wards (Sampson & Groves,

1989) as the units of analysis. The consistency among the findings that were produced based on

these varying aggregations and definitions of community/neighborhood was that there existed

significant differences in local politics, economics, and lifestyles between communities and that

the world is no longer comprised of "small isolated groups of human beings scattered over a vast

territory" (Wirth, 1938:2).

An extension to the "community/neighborhood" definitional debate and its historical

origins was more recently provided in Bursik and Grasmick's (1993) classic work,

Neighborhoods and Crime: The Dimensions of Effective Community Control. The authors began

with the contention that although consensus had yet to be obtained regarding the definition of

neighborhoods, there were several characteristics that the majority of sociologists would indicate

that delineated a neighborhood including: (1) a small physical area within a larger area; (2) some









sort of common interest in the preservation of the inhabited area; and (3) a collective identity that

tends to be constant over time (Bursik & Grasmick, 1993:6).

Bursik and Grasmick (1993) further argued that social control may exist within

communities in different forms or what they identify as private, parochial, and/or public

networks. However, they did contend that all communities tend to operate with the same

underlying desire for informal and formal associations for eliciting social control in order to

minimize crime (bearing in mind that particular crimes are likely to be differentially viewed as

unconventional as well as differentially enforced across communities). Although, the

"community/neighborhood" debate may be unresolved, but there appears to be a movement

toward fleshing out the strengths of its definitional attributes and providing a more encompassing

characterization.

This concept of community and the process of identifying particular factors that are

associated or lead to crime were the basis for what has emerged as of one of the classic theories

of crime, or what is commonly known in the discipline as social disorganization theory. This

next section presents the historical roots of the theory and lays out the core assumptions behind

why crime should be higher in socially disorganized communities.

Social Disorganization

Early macro-level research investigating the importance of ecological characteristics for

explaining crime rates has its origins from the Chicago School and the work of Park and Burgess

(1925) and Shaw and McKay (1942). The work of the latter scholars focused on attempting to

offer a conceptual framework to explain the stability of aggregate crime/delinquency rates over

time in the city of Chicago despite the dynamic change in the racial and ethnic composition of its

neighborhoods (Shaw & McKay, 1942). Shaw and McKay argued that the majority of the crime

tended to be concentrated in pockets of the city that Park and Burgess (1925) characterized as the









zones of transition that surround the central business district. The negative inverse relationship

that Shaw and McKay observed between the neighborhoods' spatial distance from the central

business district and their crime rates provided the framework for their theory of social

disorganization. More specifically, their theory emphasizes the concentration of high levels of

population turnover, racial and/or ethnic heterogeneity, and poverty in the neighborhoods that

exhibit the highest rates of crime/delinquency.

At the fundamental level, social disorganization is assumed to be the causal factor between

neighborhood characteristics and offending patterns at the aggregate level because it emphasizes

the importance of the ability of a community's residents to realize shared goals and values and

regulate the conduct of its citizens (i.e., informal social control) (Berry & Kasarda, 1977;

Kornhauser, 1978; Sampson & Groves, 1989; Sampson, Raudenbush, & Earls, 1997). High

population turnover (i.e., residential instability) as identified by Shaw and McKay (1942), is seen

as a structural antecedent that may impede the ability of communities to organize effectively and

to self-regulate residents' behavior because of the constant out-migration that characterizes high-

crime communities. This high rate of residential mobility is believed to produce a social

atmosphere where many residents are "uninterested in communities they hope to leave at the first

opportunity" (Kornhauser, 1978:78). This situation leads to the potential failure of the

community's residents to forge interpersonal ties, or even weak ties which some scholars have

argued are the most important (see Bellair, 1997; Granovetter, 1973), because of the prevailing

opinion that either they or their neighbors will soon be leaving the area.

Racial and/or ethnic heterogeneity is the second structural characteristic recognized by

Shaw and McKay (1942) that may perhaps hinder the ability of communities to develop

mechanisms of informal social control. It is also causally linked to the high rates of population









turnover mentioned above. This connection is primarily due to the fact that in communities that

are characterized by rapid compositional change, it becomes difficult to promote a unified front

against the "invasion" of potentially threatening groups (Bursik & Grasmick, 1993:33).

Furthermore, heterogeneity can inhibit communication, especially in communities that are

plagued by language barriers. This will more than likely make it increasingly difficult for the

residents to solve or even express their common problems and goals (Kornhauser, 1978:75).

Shaw and McKay (1942) also indicated that high rates of poverty/low socioeconomic

status are characteristic of socially disorganized communities. According to Shaw and McKay,

this situation was attributable to the constant out-migration to the suburbs of residents that had

the ability to move, and the area that the former residents left behind still tended to be plagued by

poverty and increased crime/delinquency rates, regardless of which racial/ethnic group settled

into the area. These individuals who lack the opportunity and/or resources to "escape" this

socially disorganized area (i.e., the inner city) have been referred to by Wilson (1987) as "the

truly disadvantaged" because they have to live in areas of extreme concentrated poverty and with

a high density of subsidized housing. This density of subsidized and predominantly renter-

occupied housing has also been associated with crime rates (see Schuerman & Kobrin, 1986).

Schuerman and Kobrin (1986) have indicated that although little is known about how

neighborhoods evolve into high-crime areas, prior research has evidenced distinct stages of

neighborhood development that typify the reciprocal effects of crime and community decline.

"Emerging" is the first stage where neighborhood deterioration precedes the high-crime rates,

but once communities move through the transitional stage and into the last stage (i.e., enduring

phase), then the high-crime rates precede the further deterioration of the neighborhood. A host

of factors has been attributed to neighborhood deterioration including: movement from single to









multiple family dwellings, residential mobility, broken homes, rise in young adult population,

minority group relations, female participation in the work force, and even advanced education.

In addition, according to Taylor (1995), prior scholars have contended that community decline is

primarily the result of residential mobility or at least the residents' desire to move from their

current area. This resident turnover has been connected to many adverse neighborhood

outcomes including weaker attachments among residents, lower satisfaction with the

neighborhood, less local involvement, and decreased housing values.

Although not explicitly referenced by Shaw and McKay in the formulation of their

theory, the relative inability of researchers to perhaps appropriately interpret Shaw and McKay's

underlying assumptions and/or disentangle the reciprocal effects of social disorganization-

meaning that the factors identified above may both influence crime rates and foster social

disorganization-led to a shift in the discipline in favor of individual-level theories of crime

causation (Land, 2000). However, several researchers have made recent attempts to revitalize

the work of Shaw and McKay (1942) and better define the causal pathways that exist between

neighborhood characteristics and offending (see Bursik & Grasmick, 1993; Sampson & Groves,

1989).

Bursik and Grasmick (1993) put forth an attempt to define their systemic theory of

neighborhood organization for explaining crime. At the fundamental level, systemic theory is

predicated upon the ability of neighborhoods to control themselves and their surrounding

environment in order to minimize their likelihood of victimization and reduce crime. The

authors argue that formal and informal mechanisms of social control are equally important if the

neighborhood is to be successful in reducing crime.









The private, parochial, and public spheres are regarded as the three specific elements that

are important in eliciting the necessary controls on the behaviors of the neighborhood's

members. The private level of control is expressed through the strong intimate informal

networks that exist within the communities that are able to reduce crime by threatening or

withdrawing social support from those members who engage in inappropriate or non-normative

behavior. The parochial level of control is directly tied to the ability of social institutions such as

the school and church to influence behavior through the shared sentiments of the neighbors for

reducing victimization risk. The public level of control is perhaps the most difficult to obtain

and emphasizes the ability of the neighborhood to mobilize government resources at the local

level and acquiring crime controlling agents (i.e., police) as an external resource to reduce the

crime in their neighborhood. These three mechanisms are believed to manifest themselves

through neighborhood-level interactions over time; however, the greatest impediment for the

formation of the ties is residential instability, which presents a dilemma being that those

communities who need these three mechanisms the most will be the least capable of developing

them (Bursik & Grasmick, 1993).

Sampson and Groves (1989) also emphasized the importance of family disruption when

examining the effect of social disorganization on crime. They argued that marital and family

disruption (i.e., divorce and single-parent households) may decrease the ability for community

residents to develop informal social controls. This assumption is based on the belief that two-

parent households provide stability and increased supervision for their children. Two parents

also allow for more supervision over their personal property. Sampson (1987) has shown rather

large effects for family disruption on rates of juvenile crime. Although, Sampson and Groves

(1989) demonstrated that social disorganization (as measured by sparse friendship networks,









unsupervised teenage peer groups, and low organizational participation) mediated a rather

significant portion of the direct effect of family disruption on crime/delinquency rates as well as

mediating the effects of other structural characteristics such as classical social disorganization

measures (low SES, residential mobility, and ethnic heterogeneity).

In sum, the social disorganization framework provides a backdrop for explaining why

crime rates may be higher in communities that are characterized by high population turnover,

ethnic/racial heterogeneity, and residential instability, but just how social disorganization can

also be expected to have an affect on individual outcomes will become clearer in the next

Chapter (Ch. 3). The next section (along with the remaining sections in Chapter 2) illustrates a

complete shift in theoretical framework by not focusing on explanations of crime in the

aggregate but by presenting a series of criminological paradigms and micro-level developmental

theories that have been offered to explain individual variation in crime.

Age-Crime Relationship and Criminal Careers

Perhaps the most commonly accepted fact about crime is the association between age and

crime (Cohen & Land, 1987; Hirschi & Gottfredson, 1983; Hirschi & Gottfredson, 1985;

Farrington, 1986a; Greenberg, 1985; Obrien, Stockard, & Isaacson, 1999; Sampson & Laub,

1990, 1992; Shavit & Rattner, 1988; Steffensmeier, Allan, Harer, & Streifel, 1989;

Steffensmeier, Streifel, & Shihadeh, 1992). The progression of what has been referred to in the

literature as the "age-crime curve" is that the crime rate, as measured by arrest, typically

increases from the minimum age of criminal responsibility and peaks in the teenage years

followed by a decline in early adulthood (Farrington, 1986a). Although this general statement

accurately characterizes the age-crime curve, in order to fully understand the complexity the age-

crime relationship it is necessary to revisit Hirschi and Gottfredson's (1983) original article

discussing this phenomenon entitled Age and the Explanation of Crime.









Hirschi and Gottfredson (1983:54), in response to what they considered to be the

misapplication of the age-crime distribution by scholars in criticizing social theories of crime and

supporting longitudinal methods over cross-sectional designs, assert that the effect of age is

invariant across time, place, social and cultural conditions, and crime type. They indicated that

rates of delinquency have been comparable over the last 150 years and that the age distribution

of delinquency in England and Wales in the 1940s, Argentina in the 1960s, and the United States

in the 1980s are virtually indistinguishable. They further argued that although non-White

offenders may start their delinquency earlier and offend at a higher rate than Whites and that

while boys offend at higher rates than girls for particular offenses, the age-crime curve is still

invariant across race and sex (1983:556). In addition, they provided evidence against the

differences in the age-distributions of property and person offenses. They illustrated that while

official data may reveal different age-crime curves for person and property offenses, this

difference is not observed in self-report data; therefore, the effects of age are indeed invariant

across crime type.

Hirschi and Gottfredson's (1983) early claims about the invariance of the effects of age

across historical and cultural conditions and crime type were soon met by criticism. Farrington

(1986a) specified that the main limitation of interpreting the types of age-crime curves used by

Hirschi and Gottfredson (1983) is that there is probable ambiguity in differentiating between age

and period effects. Period effects are significant factors that are time-period specific, such as

eras of high unemployment, riots, and/or civil unrest, and these influences are only related to

changes in the crime rates of those cohorts that are aligned with that particular era in history. In

contrast, age effects are universal changes that occur with age (i.e., natural maturation) that lead

to physical deterioration for all cohorts during all periods of observation.









Farrington (1986a) further argued that there is also the issue of disentangling cohort effects

when interpreting the age-crime relationship. Cohort effects are influences that are due to an

offender's membership in a specific cohort, such as belonging to the baby-boomer generation.

Taken together, all of these matters present problems because an individual's actual age is

always measured in number of years and typically equals the current year (period) subtracted by

birth year (cohort) (Farrington, 1986a).

Furthermore, according to Farrington (1986a) the inherent characteristic of the age-crime

curve that receives the least amount of attention is that it reflects variations in the prevalence

(how many individuals are involved in crime) rather than the incidence (how many crimes are

committed) of offending (see also Blumstein et al., 1988). While age-crime curves for

individuals are typically unimodal (with only one peak in the teenage years), Farrington (1986a)

argues that these curves do not mirror the aggregate curve that was referred to by Hirschi and

Gottfredson (1983). Additionally, individual crime rates have been shown to be constant during

a criminal career, thus providing evidence that arrest rates do not always decrease with age for

all offenders (Blumstein & Cohen, 1979).

Farrington (1986a) also showed that aggregate age-crime curves can be comparable, but it

is their measures of distribution and central tendency that vary across time, place, sex, and

offense type. Nevertheless, despite the conflicting arguments discussed above particularly

between Hirschi and Gottfredson (1983) and Farrington (1986a), there is virtually no definitive

evidence refuting the strong relationship between age and crime. Given the robustness of this

relationship, recent attention has drawn researchers to focus on how criminal behavior changes

throughout the various stages of life (Sampson & Laub, 1990, 1992) and this approach has









become characterized as the criminal career paradigm (Blumstein, Cohen, Roth, & Visher,

1986).

Blumstein et al. (1986) pointed out that the bulk of prior research in criminology was

devoted to investigating crime using aggregate crime rates and/or crimes per capital in the general

population. This aggregate crime rate is the product of the participation rate of offending and the

frequency of offending among the general population involved in crime, which according to

Blumstein et al. makes studying individual criminality difficult. Hence, in an effort to better

enable researchers to study differential offending patterns, the criminal career paradigm is unique

because it divides the aggregate crime rate into two modules and has its own set of terminology

and dimensions.

The first module of the aggregate crime rate is participation in offending, or prevalence.

Criminal career research distinctly categorizes individuals into those who are criminally active

and those who are not (Blumstein & Cohen, 1987). Participation measurement is an expression

of both the offender's involvement in crime and the range of the criminal acts committed during

the observation period (i.e. serious or serious and non-serious offenses). Regardless of the range

of specified offenses or the particular observation period, active offenders always include new

offenders (i.e., first offense occurs during observational period) and offenders who began their

offending prior to the observation period but remain criminally active during the period of study

(Blumstein et al., 1986:18). Therefore, according to Piquero, Farrington, and Blumstein

(2003:246), "the longer the average duration of offending, the greater the contribution of

persisters to measured participation in successive observation periods."

According to Blumstein et al. (1986), national estimates of participation in offending can

range anywhere from a low of 25% to almost 45% depending on the operationalization of what









constitutes an offense. Visher and Roth (1986) conducted one of the most comprehensive

attempts to capture a "true" estimate of the participation rate of offending across the United

States and Great Britain. They found that across both countries, the average participation rate for

non-traffic related offenses was approximately 30%. However, the rate fluctuated based on

which type of criteria was employed, with the participation rate being the lowest when a

stringent official measure was used such as "conviction" (West & Farrington, 1973, 1977) and

the participation rate tended to be the highest when a more comprehensive official definition of

offending was utilized such as "police contact" (Shannon, 1988, 1991).

The second partition of the aggregate crime rate is frequency of offending. This refers to

the average annual crime rate contributed by a particular group of offenders. Participation

illustrates the pervasiveness of offending in the general population, whereas the frequency refers

to the intensity of the individual offender's criminal activity (Blumstein et al., 1988). The

frequency of offending is referred to as lambda (k), in order to discriminate it from the aggregate

rate used in the Uniform Crime Reports (UCR) (Blumstein & Cohen, 1987), and regarding

lambda, most studies have found relative stability over age, (Blumstein et al., 1986; Chaiken &

Chaiken, 1982), sex (Blumstein et al., 1986), and crime type (Peterson & Braiker, 1980).

The length of an individual's criminal career (duration) and the seriousness of offending

are also important concepts in the criminal career paradigm (Blumstein et al., 1986:13). The

duration of a criminal career indicates the total length of time of a criminal's offending

(Blumstein & Cohen, 1988). It represents the time that elapsed from the offender's initial

offense to the time of his/her last offense (Blumstein et al., 1986:13), or in other words the

interval between an initiation and a termination in offending (Piquero et al., 2003). Research has

shown that there is a significant amount of variation in criminal careers, where many offenders'









criminal careers are brief and end in their teenage years and others continue to commit crimes up

to and beyond middle age.

Piquero, Brame, and Lynam (2004) have argued that the relative inattention of research

regarding the duration of criminal careers is due to the difficulty in determining the true end of a

criminal career. This difficulty has been addressed by Blumstein et al. (1986:18) where they

stated the importance of distinguishing total career length from residual career length, wherein

the later refers to the estimated time remaining in an offender's criminal career given the specific

observational period. Attention to this element is what Piquero et al. (2003) referred to as

subsequently calculating estimates of career termination probabilities after an offender's arrest.

Numerous studies (for review, see Piquero et al., 2003) have attempted to measure offender

career lengths in actual number of years producing varying results ranging from a low of five

years (Blumstein, Cohen, & Hsieh, 1982) to a relative high of 17.27 years (Piquero, Brame, &

Lynam, 2004).

Most recently, Kazemian and Farrington (2006) found evidence of a general decline in

residual career length and offending frequency over time based on data from the Cambridge

Study in Delinquent Development. The decline in residual career length was observed in the

original South London male sample and among their sons, although offending frequency did not

decline as a result of conviction for the fathers. In addition, age of onset was associated with

both residual career length and offending frequency, but the relationship was noticeably weaker

among the fathers.

The seriousness dimension of criminal career research reflects both the number of offenses

committed, as well as the particular patterns of switching (i.e., versatility) among offenses

(Blumstein et al., 1986:13). There are numerous types of offenses in which criminals can engage









in across the course of their career, as is there are wide variation among offenders in their levels

of specialization and seriousness of offending. Offenders have been classified as either

"specialists" (concentrating in only one offense type or closely correlated offenses) or

"generalists" (active in numerous types of related and non-related offenses) and the types of

offenses can also fluctuate, either escalating or de-escalating in seriousness (Blumstein et al.,

1986:18).

In a summary provided by Piquero et al. (2003:250) the authors note that most research has

provided support for the generality of offending throughout the duration of a criminal career and

that specialization might occur more frequently among adult offenders versus juvenile offenders.

In addition, there is some evidence in favor of offense versatility, although this switching tends

to occur across clusters of logically related crime types (i.e., violence, property, and other).

Also, minimal research has addressed the issue of co-offending patterns and found that juveniles

tend to commit crimes in groups, especially regarding burglary and robbery, and males tended to

co-offend with other males.

Overall, criminal career research, comprised of all its relative dimensions, attempts to

address questions of if how, and why different factors influence the beginning of criminal

activity, the continued participation and prevalence of offending, and factors related to desistance

or termination of criminal activity. This criminal career paradigm also directly coincided with

the movement toward developmental and life-course perspectives for explaining crime; however,

the life-course approach differs from the traditional approach to criminal career research because

it recognizes that there may be different types of offenders that may have different types of

careers (Farrington, 1999). The next section is devoted toward describing life-course theory and









presents both sides of the debate regarding the persistent-heterogeneity versus state-dependent

explanations of criminal behavior.

Developmental Criminology/Life-Course Theory

The life-course perspective can be defined as a series of pathways throughout the

developmental process where there are varying age-dependent expectations and available options

in an individual's decision-making process. These decisions along with the natural course of life

events are seen to shape the trajectory of an individual's life and can be influenced by transitions

and turning points (Elder, 1985). Trajectories are viewed as pathways of development over the

life span such as work life, marriage, schooling, or involvement in crime. They are seen as long-

term behavioral patterns that are marked by a sequence of life events and transitions. These

transitions are then described as distinct life events that occur in shorter intervals that may alter

an individual's behavioral trajectory (such as first job, first marriage, first child, etc.) (Elder,

1985:31-32).

Sampson and Laub (1990) extended this framework by focusing on the strong association

between childhood events and adulthood experiences, and the idea that transitions can redirect an

individual's life-course behavioral pathway. They indicated that life-course analysis focus on

the length, timing, and the order of significant life events (including crime) and their effect on

future social development (Sampson & Laub, 1992). Attention to the impact of life events is

critical because prior research has shown that committing a crime has a considerable behavioral

influence on the likelihood of an individual committing crime in the future (Nagin & Paternoster,

1991).

Loeber (1982) indicated that children who initially exhibited high rates of antisocial

behavior were more likely to continue this behavior than those who initially showed slower rates

of antisocial behavior. Relying on data from 411 South London males, Farrington (1995:956),









concluded that "offending may increase or decrease over time, but the worst offenders at one age

still tend to be the worst at another age." This statement is essentially the basis of Gottfredson

and Hirschi's persistent-heterogeneity interpretation (or static approach) of the relationship

between past and future offending where they state that:

The best predictor of crime is prior criminal behavior... [and] research shows that
differences between people in the likelihood that they will commit criminal acts persist
over time... and these differences appear early and remain stable over much of the life
course (1990:107-108).

Gottfredson and Hirschi (1990) expand upon this viewpoint in their discussions regarding

their theory of low self-control. According to Gottfredson and Hirschi (1990: 97), the main

underlying cause of low self-control is ineffective child-rearing. They argue that it is the primary

caregiver's (i.e., the parents) role to teach the child self-control through monitoring their

behavior, recognizing deviant behavior when it occurs, and effectively punishing the observed

deviant behavior. These practices are believed to foster the child's development of self-control

and an ability to delay gratification. The formulation of this core trait will also better enable the

child to become more sensitive to the interests and desires of others, become more independent

and willing to restrain their behaviors, and have a lesser propensity for using force or violence as

a mechanism for resolving disputes or achieving their ends.

Gottfredson and Hirschi (1990:90-91) further argue that the six primary elements that

characterize individuals who exhibit low self-control are impulsivity, insensitivity, physically as

opposed to mentally oriented, short-sightedness, risk-takers, and displaying a low tolerance for

frustration. Although these six components are known to be distinct personality and behavioral

traits, these various dimensions are all believed to be established early on in childhood and

coalesce within individuals to manifest a single construct that remains persistent throughout the

life course. The fact that these elements fuse together to generate a unitary construct, what









Gottfredson and Hirschi refer to as criminal propensity, does not necessarily mean a crime will

occur. The theorists indicate that this propensity will only cause an individual to engage in a

crime or an analogous behavior once it converges with an opportunity to commit a criminal act.

However, this temporal convergence of propensity and opportunity is not meant to argue the fact

that it occurs without the individual's knowledge of some social, legal, or natural sanctions

associated with their behavior. It is meant to indicate that individual's with low self-control will

tend to behave criminally given the opportunity because they typically perceive fewer negative

consequences and have less to lose when compared with individuals with high self-control

(Gottfredson & Hirschi, 1990).

In contrast to Gottfredson and Hirschi's (1990) viewpoint, a state-dependence

interpretation (or dynamic approach) has also been presented to account for the positive

association between past and future criminal behavior. This perspective asserts that prior

participation in criminal activity reduces internal inhibitions and external constraints to future

criminal activity and intensifies the prior offender's motivation to commit future crime. Or in

other words, the prior commission of a crime causes a subsequent increase in the probability of

an individual engaging in a future crime (Nagin & Paternoster, 1991; Nagin & Farrington, 1992).

The state-dependence hypothesis also recognizes the impact of life events on criminal behavior

(Piquero, Brame, Mazerolle, & Haapanen, 2002). In support of this interpretation, Sampson and

Laub contend that:

Lives are often unpredictable and dynamic; exogenous changes are ever present. Many
changes in life result from chance or random events in individual lives...while other
changes in life directions stem from macro-level 'exogenous shocks' (1995:50).

Despite the existing points of contention between the two rival explanations, the results

from Piquero et al. (2002) based on data from 524 males released from California Youth

Authority (CYA) institutions indicated that criminal offending appears to involve a combination









of the persistent-heterogeneity and state-dependence explanations. This finding led Piquero et al.

to conclude that theoretical models must account for the relative contribution of both

explanations in attempting to describe the overall progression of criminal behavior. Other

research has also argued in favor of taking the "theoretical middle ground" when studying crime

(Paternoster & Brame, 1997; Paternoster et al., 1997).

Additionally, previous research has found evidence for the relationship between the

stability of offending and antisocial behavior, but indicate that the trajectory of offending will

differ across ages and vary over the life course (Nagin, Farrington, & Moffitt, 1995). Taking this

into account, Sampson and Laub (1993) proposed a theoretical model that addresses the impact

that life events and transitions can have on an individual's behavioral pathway and the following

section provides a larger discussion of their theory as contained within a developmental life-

course perspective.

Sampson and Laub's Age-Graded Theory of Informal Social Control

Sampson and Laub's (1993) age-graded theory informal social control can be viewed as a

merging of the ideas of state dependence (Nagin & Paternoster, 1991) and cumulative continuity,

wherein cumulative continuity assumes that adverse life circumstances accumulate and exert a

more significant effect as their frequency increases over time (Caspi & Moffitt, 1993; Moffitt,

1993) in order to create a model for explaining crime over the life course. They argue that social

capital (investment or stakes in conformity) and turning points (employment, marriage) are

important factors that may challenge traditional deterministic theories of crime that assume that

early childhood experiences lay the groundwork for later development (Bandura, 1982).

According to Sampson and Laub (1993:301), two books "rocked" the criminological field

and renewed researchers' interest in childhood: A General Theory of Crime and Crime and

Human Nature. In Gottfredson and Hirschi's (1990) work, A General Theory of Crime, they









argue that ineffective child rearing (including the parents' failure to effectively monitor,

recognize and punish their child's unwanted behaviors) in the early years of a child's life (prior

to age 8) fosters the development of low self-control. This latent trait is then viewed as a

relatively stable phenomenon that increases the likelihood of involvement in criminal and

analogous behavior throughout an individual's life. Comparatively, in Wilson and Hernstein's

(1985) book, Crime and Human Nature, these scholars state that crime could be best explained

by individual differences in impulsivity and temperament, along with an interaction of these

qualities with familial factors (see also Grasmick et al., 1993; Nagin & Paternoster, 1991).

Fueled by this renewed interest in childhood and drawing from social control theory

(Hirschi, 1969) and Elder's life-course approach (1975, 1985), Sampson and Laub (1993:303)

began to focus on age-graded life transitions and argue that institutions of formal and informal

social control matter and vary over the life course. They contend that age-graded informal social

control is essential for promoting interpersonal bonds that link individuals to the larger social

institutions in which they live (i.e., work, family, school). They also acknowledge that both

continuity and within-individual changes do occur over time and that specific life transitions and

other factors related to development may intervene in one's pathway of offending (i.e., also

referred to as the stepping-stone approach by Farrington, 1986b).

Sampson and Laub (1993) further developed their theoretical framework with attention to

what they characterize as a series of building blocks. The first building block focuses on the

intervening role of informal family and school social bonds. Their second building block

focuses on the continuity of antisocial behavior that begins early on in a child's development and

extends throughout adulthood. They also recognize that particular life events and social ties in

adulthood can alter an individual's behavioral trajectory and this viewpoint is central to their









third building block, i.e., the importance of adult social bonds. They propose that social bonds

developed in adulthood (employment, marriage) can explain involvement/lack of involvement in

crime regardless of any underlying individual criminal propensity. In addition, they emphasize

that it is not only the mere exposure of individuals to these bonds, the quality and intensity of

these bonds is critical. Drawing on the work of Elder (1975, 1985), Sampson and Laub also

hypothesize that different adaptations to key life events can affect an individual's trajectory, and

crucial turning points such as divorce or death of a love one can redirect life trajectories.

Two additional important concepts in Sampson and Laub's age-graded theory of informal

social control include the processes of self-selection and cumulative continuity. Self-selection is

the classic "birds of a feather flock together" argument, and is based on the belief that

individuals with low self-control resulting from ineffective parenting systematically sort

themselves into criminogenic environments. For example, delinquent youth will tend to

gravitate toward delinquent peers in adolescence, and similarly, will find delinquent spouses in

adulthood and continue their offending behavior (Gottfredson & Hirschi, 1990).

In comparison, cumulative continuity (Moffitt, 1993) integrated with the concept of state-

dependence (Nagin & Paternoster, 1991) refers to the idea that involvement in delinquency has

an incremental effect on the interpersonal social bonds formed in adulthood (i.e., labor force

attachment, marital cohesion). For instance, an arrest and subsequent incarceration in

adolescence may lead the youth toward dropping out of school. This practice would logically

affect their future job prospects and thus result in the failure of the individual to develop strong

adulthood bonds to the labor force, which inevitably increasing their likelihood of involvement

in crime (Tittle, 1988).









Furthermore, Sampson and Laub (1993) view change as occurring along a continuum

where change can be shaped by a shift in societal roles and environments and this individual or

environmental change can lead to the development of social capital (Coleman, 1988, 1990;

Nagin & Paternoster, 1992). Sampson and Laub (1993) further argue that the lack of social

capital is one of the central antecedents of weak social bonds. Similarly, individuals who lack

social capital are expected to be closed off from the informal networks among individuals in the

social system, which also has a negative influence on their future life prospects (Coleman, 1990).

However, Sampson and Laub (1993) also caution that just getting married or employed is

insufficient for manufacturing social capital. These social ties are only seen as important in

cases where they impose obligations and restraints on individuals that lead them to refrain from

criminal involvement. According to Sampson and Laub (1993), this process offers an

explanation for not only career desistance but late-onset criminality because regardless of an

individual's early developmental history, if they are able to form social capital or investment in

societal institutions, then they will always be less delinquent/criminal. Additionally, they argue

that social capital can be reciprocal in nature and that there are racial, ethnic, and structural

variations in social capital (see Anderson, 1990; Short, 1990). In short, according to Sampson

and Laub, their life-course theory seeks to:

Return development to where it probably should have been all along, conceived as the
constant interaction between individuals and their environment, coupled with purposeful
human agency and random developmental noise (2005:40).

Sampson and Laub's (1993) age-graded theory of informal social control soon sparked

researchers' interests; however, in the same year developmental psychologist Terrie Moffitt

(1993) unveiled a theoretical typology that claims that there are different types of offending

trajectories that exist in the social world and her taxonomy classifies offenders as either

adolescent-limited or life-course-persistent. The next section provides a discussion of Moffitt's









proposed taxonomy of offenders and highlights the features that distinguish an adolescent-

limited offender from a life-course-persistent offender. Also discussed is Moffitt's (1994; see

also Moffitt et al., 1994) more recent claims as to why life-course-persistent offending may be

limited almost exclusively to males.

Moffitt's Taxonomy

Moffitt (1993) identifies the adolescent-limited offenders as normative and believes they

include individuals who are generally well socialized and adjusted, do not begin their offending

early in life, and are simply acting out a passing phase in the developmental process of

adolescence. These individuals only offend because of the process that she identifies as "social

mimicry," whereby the youth achieve status and power in their adolescent social world by

mirroring the antisocial behaviors of others (most notably the behavior of their life-course-

persistent counterparts). Therefore, through direct observation and reinforcement, the

adolescent-limited offenders begin to engage in similar behaviors in order to assert their

individuality and emerging "adulthood" status. However, although their rates of offending may

converge with that of their life-course-persistent counterparts only briefly in mid-adolescence

(ages 14 to 17), the adolescent-limited offenders (by definition) desist from offending soon after

this developmental time period. These adolescent-limited offenders are also typified by

inconsistent, erratic, and brief offending careers. Moffitt (1993) further argues that they may

shoplift and smoke some pot, but at the same time follow school rules.

In contrast, Moffitt (1993) identifies the life-course-persistent group as offenders who

begin their criminality/behavior problems early in childhood and exhibit an ongoing pattern of

behavioral problems throughout their life. These offenders begin biting and hitting at age 4,

shoplifting and truancy at age 10, selling drugs and stealing cars at age 16, robbery and rape at

age 22, and fraud and child abuse at age 30 (Moffitt, 1993:679). Moffitt argues that there is no









reason to believe that these offenders ever extraordinarily abandon their antisocial tendencies and

assume a prosocial identity later in life. Although prior research has shown that while offending

may decline in middle to late adulthood among criminal psychopaths, their antisocial personality

traits (Cleckley, 1976) have been shown to persist until age 69 (Harpur & Hare, 1994).

Farrington et al. (1990) have also indicated that early-onset offending is strongly correlated with

the seriousness of offending and the length of an individual's offending career.

Furthermore, compared with their adolescent-limited counterparts, life-course-persistent

offenders' problematic behavior is not the result of social mimicry but it is believed to result

from neurodevelopmental impairments that lead to a lack of self-control. This inability to form

self-control (see also Gottfredson & Hirschi, 1990) affects their socialization and development

and thus makes these individuals vulnerable to committing crime. Moffitt (1993) has also

attributed this group's persistent offending to an interaction of living in a poor social

environment and neurological problems during early fetal development that inhibit

temperamental, cognitive, and behavioral development. A series of studies have shown that one

of the most robust predictors of antisocial behavior is neurological deficiencies (see Moffitt,

1990; Moffitt & Henry, 1991; Hirschi & Hindelang, 1977).

Moffitt (1993) continues to argue that a substantial portion of the antisocial behavior

exhibited by these life-course-persistent offenders can be attributed to heritability (or

intergenerational transmission of antisocial behavior) (Heusmann et al., 1984). Prior research

has shown that there is a strong consistency between parents and their children in their

temperament and personality traits (Plomin, Chipuer, & Loehlin, 1990) and in their cognitive

abilities and intelligence scores (Plomin, 1990; Loehlin, 1989). Loeber et al. (2000) indicated

that at-risk children are often exposed to adverse home environments and neighborhood









conditions because of their parent's vulnerability to a host of problems including their inability to

deal with difficult children because they lack sufficient coping resources (see also Synder &

Patterson, 1987). Furthermore, Moffitt (1997) has found that these life-course-persistent

offenders were always more likely to be involved in crime regardless of the status of the

neighborhood in which they resided.

Moffitt continues by asserting that age will exert a direct influence on the strength of the

association between an individual's underlying criminal propensity and their offending patterns.

According to Moffitt, it is expected that criminal propensity should be strongly associated with

antisocial behavior in early childhood, yet the strength of the association should be greatly

reduced in adolescence because the adolescent-limited offenders exhibit the highest prevalence

rates during this time period and these offenders lack a criminal propensity. The strong

correlation between criminal propensity and crime should again resurface in adulthood because

the life-course-persistent offenders become the predominant group once again in this time period

(see also Piquero & Brezina, 2001).1

Guided by prior research, Moffitt and colleagues have offered some theoretical refinements

to the two offender typologies to explain differential rates of offending across sex. The

adolescent-limited type exemplifies female offending patterns (Moffitt, 1994) and life-course-

persistent offending pattern is almost exclusively a male phenomenon because (1)



1 It is important to note here that there is a strong degree of overlap between Moffitt's life-course-persistent offender
and Gottfredson and Hirschi's (1990) characterization of an offender with low self-control. The main similarity is
found in their definitions of propensity, which is believed to be influenced by impulsivity, hyperactivity, and low
verbal ability (see also Bartusch et al., 1997). Similarly, in line with Moffitt et al.'s (1994) sex expectations for life-
course-persistent offending, Gottfredson and Hirschi (1990:148-153) argue that male and female differences in
opportunity or supervision can explain the disproportionate rates of offending across sex. Blackwell and Piquero
(2005) found support for the relationship between low self-control and offending among males and females, but
indicated that there were sex differences in their responses toward their parents' control. Therefore, it is likely that
Gottfredson and Hirschi would argue that sex differences in self-control can explain the concentration of life-course-
persistent offending patterns among males.









neuropsychological disorders are more prevalent among males (Rutter, 1983; Singer et al.,

1968); (2) childhood-onset conduct problems are very rare among girls (Werry, 1986); and (3)

female delinquent behavior tends to be more erratic than male delinquent behavior (Moffitt et al.,

1994:283).

While large scale self-report studies versus those based on official data have shown that

virtually every adolescent commits some illegal behavior (see Elliot et al., 1983), according to

Moffitt (1993) there is yet another group of individuals who are characterized as abstainers or

non-offenders. Similarly, Moffitt (1993) provides some theoretical speculations as to why these

individuals fail to become involved in crime and delinquency. Moffitt argues that perhaps these

youth experience late-onset puberty and therefore do not sense the maturity gap or perceive the

social assets associated with crime (1993:689). Prior research has indicated that females that

began menstruating after the age of 15 did not show any involvement in delinquency.

The abstainers may also lack the opportunities to engage in delinquency. Adolescent

delinquency rates have been shown to be higher in urban versus rural areas (Skogan, 1990), and

Sampson and Groves (1989:789) have also commented that one of the strongest correlates of

violence is the "presence of unsupervised groups of teenagers." In another study, Farrington and

West's (1990) research with their London male cohort showed that adolescents who resided in

criminogenic environments but did not engage in delinquency could be characterized as nervous

and withdrawn and also tended to entirely lack friends or only have a few friends.

Most recently, Piquero, Turner, and Brezina (2005) provided one of the first empirical tests

investigating Moffitt's (1993) abstainer hypothesis and found that there did exist a relatively

small group of adolescents (a slightly greater proportion of females) that could be classified as

abstainers. Piquero et al. (2005) indicated that teacher attachment was inversely related to









association with delinquent peers for the females and that less association with delinquent peers

was related to being an abstainer among the males. However, Piquero et al. also noted evidence

that was contrary to Moffitt's expectations. More specifically, they did not find that that the

abstainers were more likely to be sad and depressed or completely excluded from all peer

networks. Piquero et al. (2005) showed that the abstainers were still relatively active in the

dating scene and were not typically loners, but argued that they were just adolescents who were

involved in a more prosocial network of peers.

Overall, Moffitt's taxonomy provides a solid theoretical framework to explain not only

differential offending patterns of individuals, but also offers speculations as to why some youth

never become involved in delinquency. Furthermore, her offender typology can also account for

sex differences in offending.

This concludes the chapter on theoretical frameworks. It attended to the assumptions

underlying social disorganization theory, the age-crime curve and the criminal career paradigm,

the recent theoretical advancements of an age-graded theory of informal social control (Sampson

& Laub, 1993), and the existence of adolescent-limited and life-course-persistent offenders

(Moffitt, 1993) including the recent theoretical extensions to account for sex differences in

offending. The next chapter contributes a thorough discussion of what the prior literature has

discovered regarding these perspectives.









CHAPTER 3
REVIEW OF THE LITERATURE

Introduction

The theoretical perspectives presented above have been the focus of a considerable amount

of empirical testing. While, social disorganization theory has received a substantial amount of

attention for explaining crime rates at the macro-level (for review, see Bursik & Grasmick, 1993;

Sampson & Groves, 1989). The theory (particularly surrounding neighborhood SES) has also

been applied to accounting for individual variation in offending. This line of research will be

discussed in more depth below. It will be followed by an examination of the studies (most

notably that of Wolfgang et al., 1972) that have identified the existence of a relatively small

proportion of chronic high-rate offenders (between 2% and 14% depending on which study is

being referred to) who are responsible for at least half of the overall crime and even a larger

percentage of the violent offenses. The next sections will present the research evidence on the

impact that life transitions such as marriage and employment can have on an individual's

pathway of offending and the literature that has provided support for Moffitt's (1993) taxonomy

as an explanation of sex differences in offending. Finally, this chapter will conclude with an

illustration of the prior applications of Nagin and Land's (1993) methodological technique for

identifying groups of individuals with similar offending trajectories as well as providing several

hypotheses that will be tested in this current study.

Impact of Social Disorganization at the Individual Level

In general, community-level scholars (e.g. Bursik & Grasmick, 1993) have associated the

geographic concentration of socioeconomic disadvantage to the social disorganization of the

neighborhoods, and these disorganized neighborhoods have been argued to have adverse effects

on the individuals and families that reside within those areas. In contrast, other scholars contend









that the concentrated effects of poverty cannot be fully explained without including control

measures for both familial and individual-level characteristics (see Gephart, 1997:14). Although,

while social disorganization theory (as originally formulated) sought to explain crime at the

aggregate level, there are empirically-based reasons to expect that social disorganization may

also have an effect on individual behavior. Simcha-Fagan and Schwartz took a step toward this

possibility:

An aggregate-level theory cannot be expected to provide a comprehensive explanation of
individual delinquency or criminality, it should clarify those contextual effects which,
independently of or in interaction with the adolescent's predisposition, affect individual
behavior (1986:688).

Jencks and Mayer (1990) provided a review of the current extant knowledge on the effects

that social compositional measures (e.g., poverty) have on a series of developmental outcomes

including: educational attainment, cognitive skills, criminal activity, and economic success.

They indicated that the research is generally mixed at best regarding the influence of average

SES (in the aggregate) on the likelihood of an individual's chances for planning to go to college,

attending college, and/or graduating from college, especially for Whites. Research has shown

that youth from high SES neighborhoods are more likely to have higher educational attainment

compared with their counterparts who come from lower SES neighborhoods. In addition, Black

men's chances of competing in the job market and finding well-paid jobs were greatly reduced

when they resided in a heavily concentrated Black or welfare dependent neighborhood (see also

Aber et al., 1997).

Klebanov et al. (1997) provided an extension of the research on the influences of

neighborhood effects on crime by specifically testing whether or not these effects are mediated

and/or moderated (or interact with) by family processes such as maternal warmth, cognitive

stimulation in the home, maternal coping style, availability of social support, and the physical









environment of the home. They found that the positive effects of living in a high SES

neighborhood on academic achievement, higher verbal ability scores, and less behavior problems

and the negative effects of living in heterogenous neighborhoods on verbal ability scores were

mediated by the quality of the home learning environment. The effect of home environment was

found across all age groups, measures, and resource-rich and resource-poor neighborhoods.

They also found support for the interaction of the home environment and family characteristics

in each neighborhood type in ways related to criminal outcomes. Similarly, additional research

by Brooks-Gunn et al. (1993) showed that neighborhoods had an effect on childhood IQ and the

likelihood of becoming a teenage parent and/or dropping out of school. Thus, it appears that

neighborhood level effects may be both direct and indirect, but perhaps a substantial proportion

of their effects are indirect in that they operate through individual and family factors.

Simcha-Fagan and Schwartz (1986:671) argue that structural characteristics have an

impact on the organizational networks among members of the community, and these networks

can subsequently affect an individual's ability to form social bonds to the community.

Individuals who rent property as opposed to owning their own property are less likely to have a

stake in the community in which they live, and thus may be less inclined to participate in

neighborhood organizations that provide a mechanism of informal social control. Similarly,

individuals who rent their property may be less likely to care about the maintenance or up-keep

of their dwelling because they do not intend to be there long. This lack of investment could lead

to neighborhoods incivilities (Wilson & Kelling, 1982), which have been linked a breakdown in

informal social control and crime (Sampson, 1995). Dwelling units that are over-crowded such

as public housing units are also likely to lead to a sense of "anonymity" among the residents.

The failure of individuals to get out of their units and develop informal social ties with their









neighbors may be due to their exposure to violence and fear of crime (Roncek, 1981; Sampson,

1995). Sampson and Groves (1989) have also shown that high rates of residential instability

were associated with juvenile delinquency and crime.

Gephart (1997) reviewed the developmental impact of community effects on adverse

individual outcomes. In early childhood, children who reside in areas of extreme disadvantage

face a slightly higher risk for having a low birth weight and suffering from child abuse.

Comparatively, additional analyses have shown that the presence of affluent neighbors was

associated with higher IQs and the presence of low-income neighbors was related to a greater

likelihood of behavioral problems in early childhood.

Furthermore, research has also found that the influence of a neighborhood's socioeconomic

status on individual outcomes in early childhood and middle childhood were different across

races. At age 3, the benefit of affluent neighbors (versus middle SES neighbors) on children's

IQ scores was more salient among White children than African-American children (Brooks-

Gunn et al., 1993). Similar findings were also found among children at ages 5 and 6 for IQ

scores, verbal ability, and reading achievement scores (Chase-Lansdale et al., 1997; Duncan et

al., 1994). In addition, all of these studies revealed that the presence of low SES neighbors was

positively associated with early childhood behavior problems.

Gephart (1997) also indicated that research investigating neighborhood effects during late

childhood and early adolescence found that community impoverishment, the burden of child-

care, and residential instability were significant predictors of various forms of delinquency such

as violent crime, drug trafficking, juvenile delinquency, and teen pregnancy (see Gephart, 1997).

Two additional studies have also shown that neighborhood SES had a stronger effect for male

adolescents than female adolescents (Entwisle et al., 1994; Halpem-Felsher et al., 1997).









Research has also shown that living in a high SES neighborhood increased older

adolescents' chances of graduating high school and attending college (Brooks-Gunn et al., 1993;

Duncan, 1994; Ensminger et al., 1996; Halpern-Felsher et al., 1997). All of these studies

suggested that the protective effect of living in a higher SES neighborhood was more salient for

males. Crane (1991) has also provided evidence that males had the worst outcomes when living

among low SES neighbors. In addition, a series of studies reviewed earlier have all consistently

shown that living in a low SES neighborhood was positively associated with higher rates of

criminal and delinquent behavior (Braithwaite, 1979; Lindstrom, 1996; Loeber & Wikstrom,

1993; Lynam et al., 2000; Peeples & Loeber, 1994; Piquero et al., 2005; Reiss & Rhodes, 1961;

Sampson & Groves, 1989; Wikstrom, 1991a, 1991b; Wikstrom & Loeber, 2000).

The U.S. Department of Housing and Urban Development (HUD) initiated what is known

as the Moving to Opportunity for Fair Housing Demonstration (or MTO) in 1994 motivated by

the success of the Gautreaux Program that has been in operation in Chicago for almost three

decades. Rosenbaum (1995; see also Rosenbaum & Popkin, 1991) presented quasi-experimental

evidence that Gautreaux youth who moved to the suburbs greatly increased their educational

attainment and job opportunities when compared with those that remained in the poor urban

areas. Intrigued by these findings and the idea that residential mobility may benefit individuals

from disadvantaged communities, the MTO program became operational in five selected sites

(Baltimore, Boston, Chicago, Los Angeles, and New York City).

The MTO program involved randomly providing families vouchers allowing some families

with children in high poverty areas to move to less poor areas with the hope that this relocation

would increase their educational and employment opportunities. Eligibility for the program was

determined by the families that lived in Section 8 public housing units with at least a 40%









poverty rate according to 1990 census tract data. Prior MTO evaluations have indicated that

residential relocation significantly contributed to a reduction in violent arrests for males and

better overall health among the children that moved from the public housing units (Katz et al.,

2001; Ludwig et al., 2001). Katz et al. (2001) also noted significant reductions in criminal

victimization among individuals that moved out of the Section 8 housing.

Leventhal and Brooks-Gunn (2003) noted that the MTO program had profound effects on

mental health outcomes for the individuals and families that relocated. They reported that the

parents who moved reported less distress than did those that remained and the boys that moved

(especially those between ages 8 and 13) reported significantly less depression, anxiety, and/or

dependency problems when compared with their counterparts who were still in public housing.

Most recently, Kling et al. (2007) noted the substantial mental health and education benefits for

the adults and females, but commented on the adverse effects associated with the males who

received vouchers (for an extended review of MTO impact evaluations, see also Orr et al., 2003).

Overall, there is reason to believe that structural factors including those highlighted in

social disorganization theory can have an effect on individuals' outcomes and perhaps influence

their pathway of offending. It is likely that living in areas characterized by social disorganization

will increase an individual's likelihood of being a chronic offender. This concept of chronic

offending is the topic of the discussion presented in the following section.

Prevalence and Incidence of Chronic Offending

Relying on data from the well-known Philadelphia Birth Cohort Study, Wolfgang, Figlio,

and Sellin (1972) provided some of the earliest evidence of the existence of chronic offenders.

Wolfgang et al. found that there were a group of delinquents that represented only 6% of the

cohort and 18% of the cohort's delinquents but were responsible for committing roughly half of

all the offenses and about two-thirds of all of the violent offenses. In a ten-year follow-up study









of 10% of the original Philadelphia birth cohort, Wolfgang, Thornberry, and Figlio (1987) found

that these chronic, originally defined as having had five or more police contacts as juveniles,

had increased the seriousness of their offending into adulthood. Tracy, Wolfgang, and Figlio

(1990) found even higher rates of chronic offending among members of the Second Philadelphia

Birth Cohort Study. They found that the chronic represented 7% of the cohort and 23% of the

offenders, yet were responsible for committing 61% of all the offenses including 60% of

homicides, 75% of forcible rapes, 73% of robberies, and 65% of the aggravated assaults.

In a similar study, Snyder (1988) examined a cohort of Phoenix and Utah youths and found

that the chronic, defined as having had five or more referrals to juvenile court, represented only

5% of the cohort. However, like the original Philadelphia Birth Cohort, they were responsible

for 51% of all the offenses committed including 61% of murders, 64% of rapes, 67% of

robberies, 61% of aggravated assaults, and 66% of burglaries. Shannon (1988) studied three

youth cohorts in Racine, Wisconsin and found a slightly higher percentage of chronic (8% to

14%) than the previously mentioned studies, but still found that they were involved in an

overwhelming majority (75%) of all of the felony arrests.

Hamparian, Schuster, Davis, and White (1985) in a follow-up study of the original

Columbus Cohort (Hamparian, Schuster, Dinitz, & Conrad, 1978) found that 60% of the

previously identified chronic (representing only 2% of the cohort) had been arrested at least

once by their early-twenties for an adult felony offense. These offenders were also more likely

to have had an earlier onset of delinquency, more frequent and serious offending, and displayed

continuity in offending from their youth through their mid-twenties.

The high frequency offending of chronic that has been illustrated when examining official

arrest data is also evident in self-report surveys. Results from the National Youth Survey (NYS)









have shown that about 5% of juveniles at each age level from twelve to seventeen years old can

be classified as serious, chronic offenders. These offenders are characterized as having engaged

in three or more aggravated assaults, sexual assaults, gang fights, and/or strong arm robberies

within the previous year. In addition, on average these chronic committed 132 delinquent

offenses annually and 8% of these offenses are considered serious and violent. Perhaps more

alarming is the fact that 84% of the offenders identified as chronic in the survey had no official

record of delinquency and this disparity between official measures and self-reported measures of

offending cannot be underestimated (Elliot, 1994; Elliot, Huizinga, & Morse, 1986).

Further support for the existence of chronic offenders was also provided in a series of

RAND Inmate studies conducted during the late 1970s and early 1980s. Petersilia, Greenwood,

and Lavin (1977) conducted the first of these studies and focused on the criminal careers of 49

inmates in a California medium-security prison. All of these offenders were currently serving

time for armed robbery and were characterized as repeat offenders because they all had served at

least one prior prison term. Petersilia et al. (1977) gathered information on the frequency of

offending from structured interviews with the offenders and their official criminal records.

Overall, the 49 inmates self-reported committing 10,500 serious crimes even though the majority

of the offenders had spent the better part of their criminal careers behind bars.

Peterson and Braiker (1980) carried out the second Rand Inmate study where they

administered self-report questionnaires to 624 incarcerated male felons in five of California's

state prisons asking the inmates to report what types of offenses they committed during the three

years prior to their current imprisonment. Peterson and Braiker found that the majority of the

inmates did not specialize in offending and that most of the felons who committed one crime

type at a high frequency also reported significant criminal activity in other types of crime. On









average, these chronic offenders represented about 8% of the inmates but self-reported

committing over 60 crimes per year.

Chaiken and Chaiken (1982) conducted the third Rand Inmate study investigating high-rate

offending, wherein they analyzed the criminal histories and survey results from 2,190 prison and

jail inmates in California, Texas, and Michigan. Chaiken and Chaiken identified what they

referred to as a small group of "violent predators" or "omni felons" and these individuals

reported committing robberies, assaults, and drug deals at disproportionately higher rates than

the other inmates. In addition, they tended to commit burglaries, thefts, and other property

crimes at higher rates than even the offenders who self-reported to specialize in those specific

crime types.

Overall, all of the studies described above have shown strong empirical support for

"chronic offenders" and these offenders represent only 2% to 14% of the offending population

(depending on which study is being referred to), yet they are responsible for at least half of all

the offenses committed (Farrington, 1999). The fact that a sizeable fraction of the crime problem

can be attributed to a relatively small number of identifiable chronic offenders provides support

for preventative measures and early identification (see Piquero, Farrington, & Blumstein, 2003).

However, in contrast to the robust finding of chronic offending among various samples and

time periods and the evidence that these individuals tend to be criminally active for long periods

of time, Sampson and Laub (1993) argue that key life transitions such as marriage and/or

employment can have an impact on an individual's pathway of offending, even if they are/were

chronic offenders. The following section addresses the empirical evidence in support of the

effect that these various life events can have on behavioral outcomes, most notably desistance

from crime.









The Effect of Life Transitions on Offending

Sampson and Laub's (1993) age-graded theory of informal social control takes the position

that life key life transitions (including delinquency) can affect adult outcomes. Similarly, Caspi

and Moffitt (1993) have referred to delinquency (particularly serious) as a "knifing off" of future

opportunities, or in other words delinquency in adolescence is expected to impede the formation

of conventional social bonds that will continue to have an affect in adulthood

Relying on the Gluecks' data (1950, 1968), Sampson and Laub (1993) presented

preliminary support for their theory, whereby individuals who exhibited the strongest ties to

work and family had the lowest rates of delinquency. In contrast, incarceration in the adolescent

years was inversely related to job stability and positively associated with continued criminal

involvement. Nagin and Waldfogel (1998) have also noted the adverse effects that an adolescent

conviction has on future job opportunities among West and Farrington's (1977) London male

cohort. Additional research has shown that involvement in delinquency results in the worst adult

outcomes for lower class boys, particularly among lower class boys with a high frequency of

police contacts (Hagan, 1991). Jessor et al. (1991) have also indicated that delinquency is not a

major impediment for the formation of prosocial adult bonds for middle class adolescents.

Based on the results of Sampson and Laub's (1993) original test of their age-graded theory

of informal social control, support was found that delinquency was associated with crime in

adulthood as well as being significantly related to a series of adverse life events including being

AWOL from the military, economic dependence, and having had martial problems. Contrary to

what Gottfredson and Hirschi (1990) would predict, they also found significant independent

effects for marital attachment and job stability on involvement in adult crime net of the effects of

criminal propensity.









Sampson and Laub (1993:317-320) concluded by offering six key issues to guide future

research testing their theory. First, there is some element of luck, randomness, or chance in an

individual's life course (also referred to by Rutter, 1989 as adventitiouss happenings"). Second,

there is individual variation in the effects that life transitions have on behavior. Third, macro-

level opportunity structures for marriage and labor market activity are important (i.e.,

segmentation of the labor market, ethnic enclaves). Fourth, criminal propensity cannot

completely explain adult crime because life transitions can also affect opportunities and

motivations to commit crime (see also Cohen & Machalek, 1988). Fifth, the effect of cumulative

continuity is strongest among those in disadvantaged racial and economic conditions. Sixth,

historical context matters.

With attention to some of these suggestions, Homey et al. (1995) found clear evidence that

meaningful change in life circumstances was associated with desistance from crime based on

their interviews with 658 offenders in a Nebraska Correctional Facility. They also estimated

hierarchical linear models that allowed them to control for criminal propensity and still found

that the offenders' reported reduced involvement in criminal activity when they were living with

their wife. Overall, their results suggested that life transitions can alter behavioral patterns of

offending at particular periods of time, especially in the short-term. Laub et al. (1998) estimated

trajectories with the Glueck's data and found that the timing and quality of marriages were what

was important and that the inhibiting effects of marriage on crime were gradual and grew in

intensity over time. Similar to Homey et al. (1995), these findings were still observed while

controlling for criminal propensity and a host of other factors.

More recently, Piquero et al. (2002) developed a stakes in conformity index where they

combined the life circumstances of employment and marriage to gauge the possible cumulative









effects of local life circumstances on pathways of offending among serious offenders from the

California Youth Authority. Based on their results, Piquero et al. concluded that it is likely the

case the these mechanisms of informal social control (or life transitions) can have a stronger

effect on instrumental as opposed to interpersonal crimes because commission of the former

tends to involve a relative degree of planning.

Laub and Sampson (2003) estimated models that took into account within-individual

change and found that when controlling for age, individuals who were currently married had

lower criminal propensity scores when compared with their scores when they were not married.

They also found these differences in propensity existed in times where the individuals were

employed or in active military service compared with periods when they were not.

In another study testing Sampson and Laub's theory, Warr (1998) provided an analysis of

two waves of longitudinal data from the National Youth Survey of 1,725 individuals. Counter to

Sampson and Laub's (1993) position, Warr hypothesized that the link between life transitions

and crime might be better explained by the effect that these transitions have on altering the

relationships with delinquent peers (Akers et al., 1979, 1985; Sutherland, 1947). Warr's results

did reveal that marriage was strongly associated with desistance from crime, but he indicated that

for many of these individuals, getting married was a key life transition that signaled a severing of

heavy involvement with their delinquent peers and instead re-focused their attention to their

spouse and the desire to have children.

These findings led Warr to conclude that marriage not only reduces individuals'

motivation for continuing criminal behavior by re-directing their goals, it also greatly minimizes

their opportunities for offending by limiting the amount of time that they have to spend with their

former delinquent peers. In addition, Warr cautioned that there still existed the possibility of a









selection effect in that individuals who do in fact get married are perhaps qualitatively different

from those who do not get married, and that individuals who are not involved with delinquent

peers and/or engaged in crime may appear to be more suitable marriage prospects.

Maume, Ousey, and Beaver (2005) have since reexamined Warr's (1998) hypotheses and

conclusions using data from the National Youth Survey, but improve upon Warr's earlier study

by specifically investigating the effects of marital attachment on desistance from marijuana use

and control for the possible selectivity bias in the relationship between marital attachment,

delinquent peer associations, and desistance. In line with Sampson and Laub's theory, Maume et

al. found that marriage was important to the desistance, particularly for those marriages that were

characterized by high levels of attachment. Contrary to Warr's (1998) findings, marriage still

maintained significant directs effects for affecting desistance net of the effect of delinquent peer

association (measured as the change in the number of delinquent associates).

Uggen (2000) investigated the impact that employment (i.e., a turning point) has on an

offender's recidivism relying on data from over 3,000 individuals from nine U.S. cities, which

were originally part of the 1977 National Supported Work Demonstration Project that targeted

the underclass and the ghetto poor. The individuals who participated in this project were

randomly assigned to either of two treatment groups (construction work or service jobs) or to a

control group where no job placement was given. The work and arrests histories were collected

for all of these individuals every nine months for three years. The results from the event history

analysis yielded a variety of noteworthy findings. First, although the program failed to reduce

crime overall across the entire sample, its impact was definitely age-graded. Job treatment

significantly reduced recidivism among the older participants (particularly aged 27 or older)

whereas the job treatment had little to no effect on crime among the young adults (late









teens/early twenties). Therefore, it appeared that this life transition (employment) affected

desistance from crime only for a certain age group of individuals (older adults) providing support

for Sampson and Laub's (1993) theory.

In a recent review of the state of their theory and in response to some noted criticisms (see

Moffitt, 2005), Sampson and Laub (2005c: 18) refined their views toward the influence of life

transitions on adult offending by arguing that involvement in institutions such as marriage and

work provide both short-term and long-benefits. In the short-term, these life transitions can alter

an individual's motivation for crime by re-focusing their energy and attention toward

conventional outlets (spending time and caring for their spouse/family or their job) and can

produce the long-term benefit of fostering their development of prosocial adult bonds or stakes in

conformity.

Sampson and Laub (2005c:36) continued by arguing that marriage (or employment) should

not be viewed as a one particular point in time, but viewed as a dynamic and cumulative causal

factor that affects desistance from crime over time (see also Laub et al., 1998). Given the above

presented evidence in support for Sampson and Laub's age-graded theory of informal social

control, it is important also to visit the research providing support for Moffitt's taxonomy of

offending for explaining sex variations in offending and this becomes the focus of the following

section.

The Saliency of Moffitt's Taxonomy across Sex

According to Moffitt et al. (2001), females typically have fewer risk factors than males;

therefore, they are less likely to be early-onset and/or life-course-persistent offenders. In

contrast, the sex similarities in adolescent-limited offending are due to the process of social

mimicry where there is an equal-opportunity for offending among males and females in the

adolescent years (Moffitt et al., 2001). Moffitt et al. (2001) have also suggested that while









neurodevelopmental disorders affect the life-course pathways of offending similarly for males

and females, males experience early cognitive deficits far more frequently than females. Their

results showed that significant sex differences existed in almost all cognitive areas including

hyperactivity, impulsivity, and temperament.

Tibbetts and Piquero (1999) conducted one of the earliest tests of whether Moffitt's

taxonomy applies to males and females and reported that they failed to find enough females who

exhibited the life-course-persistent pattern of offending in order to provide sufficient statistical

power in a Philadelphia cohort. Subsequently, Fergusson et al. (2000) were able to show that a

single model could explain antisocial behavior across sex, but the ratio of involvement in early-

onset offending and late-onset offending was 4:1 and 2:1 for males and females respectively. In

contrast, Aguilar et al. (2000) found that the prevalence of early-onset offending was similar

across sex.

In another study, Moffitt and Caspi (2001) found that the sex ratio for life-course-

persistent offending was 10:1, whereas the sex ratio was virtually negligible for adolescent-

limited offending (1.5:1). Similarly, Kratzer and Hodgins (1999) found a rather large sex ratio

for life-course-persistent offending (15:1) and adolescent-limited offending (4:1) among a

Swedish cohort of 13,000 individuals. Mazerolle et al. (2000) reported that early-onset

offending was associated with persistent offending and a lack of specialization among members

of a Philadelphia cohort (n= 3,655). In addition, in a longitudinal study of 820 girls, Cote et al.

(2001) did observe an antisocial group characterized by persistent and frequent offending,

although this group only made up 1.4% of the entire sample of females.

Comparatively, Silverthom and Frick (1999) have argued antisocial girls tend to share

many of the correlates of childhood-onset offending and adulthood adjustment problems that the









childhood-onset boys express. The difference is that antisocial girls typically demonstrate a

"delayed-onset" pathway (do not begin their offending until adolescence). They propose that

this third developmental pathway is unique to females and while the cognitive and neurological

deficits and the dysfunctional family environment is present in these girls' childhood, they do not

produce overt antisocial behavior until adolescence.

Silverthom and Frick (2001) provided a preliminary test of their "delayed-onset" offender

typology among 72 incarcerated youth. They found that there were similar percentages of early

and late-onset offenders among the males, but that the females were more likely to show the

delayed-onset (late-onset) patterns of offending and these girls displayed similar risk factors as

the childhood-onset boys. Subsequently, two additional tests found that early-onset and delayed-

onset offending were similar among boys and girls and that the early-onset girls shared the same

risk factors as the early-onset boys (see Fergusson & Horwood, 2001; Moffitt & Caspi, 2001).

More recently, White and Piquero (2004) provided another test of sex

differences/similarities in early versus delayed-onset offending, except using a large sample of

African-American participants from the Philadelphia portion of the National Collaborative

Perinatal Project. White and Piquero (2004) discovered that females were just as likely as males

to display early-onset patterns of offending, but comparatively, the late-onset females shared

many risk factors with the early-onset males. Noting all of the research presented above,

according to recent review, Piquero and Moffitt (2005) argued that females do show less

involvement in criminal activities at both levels (adolescent-limited offenders and life-course-

persistent offenders) when compared with males.

Taken together, all of the previously reviewed research has indicated that (1) macro-level

factors can affect individual behavior, (2) there exists a small group of chronic offenders who









commit disproportionate amounts of crime; (3) life transitions can alter individual offending

pathways; and (4) Moffitt's taxonomy can be used to explain sex differences in offending,

particularly the findings that males are more likely to be found among the life-course-persistent

offender group than the females. Noting this evidence, the following section directs attention

toward examining the stability of offender groups across recent prior applications of Nagin and

Land's group-based modeling technique, along with a discussion of the factors that have been

found to be associated with offending group membership.

Prior Applications of Group-Based Models of Offending

Similar to the theoretical developments in criminal career and developmental research,

modern innovations in methodological approaches to studying group-based modeling,

specifically the technique developed by Nagin and Land (1993) (i.e., what has been referred to

by many names including trajectory analysis, finite mixture models, or semiparametric mixed

Poisson-based regression models). The technique permits disaggregating aggregate offending

patterns into group-based offending trajectories of individuals that exhibit distinct pathways over

time (Nagin & Land, 1993; Land, McCall, & Nagin, 1996; Land & Nagin, 1996; Nagin, 2005).

This method for modeling multiple pathways of offending provides a mechanism to identify and

distinguish between certain types of behavioral trajectories. It also enables the use of regression

models such as multinomial regression in order to examine which covariates are best able to

predict which offending group that an individual is assigned.

Nagin and Land (1993) provided the first application of this methodology based on data

from the well known South London cohort of 411 males (West & Farrington, 1973, 1977).

Relying on convictions as their measure of offending, they found evidence of four groups of

offenders including a non-offender group, an adolescent-limited group, and two groups of

chronic offenders with different degrees of high-rate offending.









McDermott and Nagin (1998) found support for the existence of three offender groups

based an analysis of male respondents from the National Youth Survey, which is a well-known

self-report survey. The first group was classified as the non-offenders. The second group was

referred to as adolescent-peaked. They tended to resemble the third or chronic offending group

in adolescence but then began to desist at the age of 18 until there was no observed offending at

age 24.

Additionally, trajectory analyses based on two distinctly different types of offenders

(Gluecks' Boston delinquents vs. California Youth Authority parolees) from different time

periods and contexts found that there seemed to be 4 to 6 trajectory groups that all began to

decline by adulthood (see Laub et al., 1998; Piquero et al., 2002). Laub et al. (1998) and Piquero

et al. (2002) also noted that marriage or stakes-in-conformity (that included marriage) was

related to lower trajectories.

A more comprehensive attempt to identify the consistency of latent offending groups was

performed by D'Unger et al. (1998) using data from the London cohort (West & Farrington,

1973, 1977), the Second Philadelphia Birth Cohort (Wolfgang, Figlio, & Sellin, 1972; Tracy et

al., 1990) and the three Racine Birth Cohorts (Shannon, 1988, 1991). They found support for the

considerable similarity of offending groups across the data sources, with the constant finding of

non-offending, adolescent-limited or peaked offending, and chronic offending groups.

Specifically concerning the Racine Birth Cohorts, D'Unger et al.'s results revealed the similar

four classes of offenders noted above for the 1942, 1949 and 1955 cohorts, but discovered an

additional fifth group for the 1942 and 1955 cohorts with a late-onset pattern of offending.

D'Unger et al. (1998) illustrated that this group showed relatively stable rates of offending

throughout their twenties. Broidy et al. (2003) also found three to four trajectory groups when









examining childhood physical aggression and later involvement in delinquency/crime using data

with various demographic groups (males/females) and six different study sites represented

including: Child Development Project; Christchurch Health and Development Study, Dunedin

Multidisciplinary Health and Human Development Study, Montreal Longitudinal Study,

Pittsburgh Youth Study and the Quebec Provincial Study.

More recent research has also observed these similar offending groups disaggregated by

sex (Cote et al., 2001, 2002; D'Unger et al., 2002; Fergusson & Horwood, 2002; Land, 2000;

Piquero et al., 2005), although some studies were unable to identify a chronic offending group

among females (D'Unger et al., 2002; Piquero et al., 2005). D'Unger et al.'s (2002) analysis

provided five offender groups for the males and only three for the females. Furthermore, they

indicated that while the age patterns of offending tended to be similar across sex, the frequency

of offending was higher among the males.

In contrast, Cote et al. (2001) discovered a life-course-persistent group of offenders among

the females, but noted that this group only comprised 1.4% of the sample of nearly 900 females.

In a more recent study, Cote et al. (2002) found that the broad patterns of development (as far as

groups) were similar across sex, but males were more likely to exhibit higher impulsivity

trajectories when compared with the females. Relying on data from 896 members of the New

Zealand Dunedin Birth Cohort, Fergusson and Horwood (2002) indicated that while identical

offending trajectories were found for the males and females, the females were more likely to be

of low risk and found predominantly among the adolescent-limited group, whereas the males

were more likely to be in the chronic offending groups. Most recently, Piquero et al. (2005)

discovered three classes of offending (low, medium, and high-rate) among males and only two

for the female models (low and medium-rate) using data from the Dunedin birth cohort. In









summary, Hipwell et al. (2002) concluded that the main difference between boys and girls

behavioral patterns of offending is that only a very small proportion of girls follow the life-

course-persistent trajectory.

Chung et al. (2001) also discovered another trajectory group that they termed desisters that

illustrated an atypical trajectory in that they were characterized by early-onset offending (before

age 13) but instead of showing the steady rise in mid-adolescence into young adulthood, the

group's trajectory declined and appeared to desist by age 21. According to Chung et al. (2001),

further analyses suggested that it was not individual and familial factors that could be attributed

to their desistance but rather the influence of less exposure to peer, neighborhood, and school

factors typical associated with offending. They also indicated that aggressiveness, antisocial

peers, and drug availability were consistent predictors of initial levels of offending measured at

age 13 and that individual factors including aggression, depression, and anxiety significantly

distinguished the late-onset offenders from the non-offending group.

Piquero (2007) provided a comprehensive review on the studies to date applying Nagin's

and Land's (1993) group-based modeling technique and concluded that there existed relative

consistency in the number of trajectory groups and the shape of the curves, but there was

variability across the studies depending on the length of follow-up and the ages that were

represented in the analysis. Additionally, in line with Moffitt these studies have shown evidence

of adolescent-limited and life-course-persistent trajectories across diverse data sets, cohorts,

countries and using various measures for offending including self-reports, teacher ratings, parent

ratings, and police records based on either contacts, arrests, or convictions. However, these

applications have also uncovered other trajectories not consistent with Moffitt's (1993)

predictions such as the low-level chronic and late-onset type.









The next section provides an elaboration of the theoretically integrated model using macro-

level and micro-level factors for explaining group-based offending trajectories. Specifically, the

discussion focuses on how the social disorganization framework can be incorporated with

Samspon and Laub's age-graded theory of informal social control and Moffitt's developmental

taxonomy.

Integrated Theoretical Model for Explaining Offending

Typically, criminological and sociological research focuses on either examining

community level or aggregate rates of delinquency/crime or investigating the differences

between delinquents/criminals and nondelinquents/criminals. Although, in recent years scholars

have begun to recognize the importance of multi-level research and integrated theoretical models

(see Reiss, 1986; Moffitt, 1997a; Sampson, 1997; Farrington, Sampson, & Lauritsen, 1994). In

an attempt to answer these calls for research in this area, the proposed model in this study

emphasizes the importance of three particular theoretical perspectives (social disorganization,

age-graded theory of informal social control, and Moffitt's taxonomy) for explaining group-

based offending trajectories. The proposed model is further elaborated on below and the

conceptual figure (along with additional variables included in the analyses) is displayed in Figure

3-1.

In the 1920s Shaw and McKay reported that the areas that were characterized by the

highest rates of crime and delinquency also tended to display comparatively high rates of infant

mortality, low birth weight, child abuse, along with other adverse factors for early childhood

development. According to Sampson (1997:43) the consistent finding of moderate-strong

continuity of early onset offending behavior and later criminal involvement and the long-

standing covariation of delinquency rates and social problems clearly emphasizes that "there is a

connection between healthful development of young children and community structure." The









quality of health care services has also been shown to vary across communities with the most

socially disorganized areas having the poorest services (Lash, Sigal, & Dudzinski, 1980).

Similarly, Moffitt (1997a) and others (see Bronfenbrenner, Moen, & Garbarino, 1984)

have argued nueropsychologically vulnerable children may be disproportionately found in poor

areas given the association between brain deficiencies and poverty. Moffitt (1997a) also

highlights the concentration of low birth weight, poor infant nutrition, and the incidence of

exposure to toxic agents as being greater in poor neighborhoods. Furthermore, children with

early developmental/neurological deficits are not typically born into protective and supportive

environments with in-tact families, but rather into families with parents who socialize their

children in a criminogenic environment where both parent and child are vulnerable to the adverse

neighborhood conditions (Moffitt, 1997a). These at-risk or "difficult" children are then likely to

receive negative reactions from their parents in efforts to control their problem behavior in early

development and this negative reaction is likely to be continued throughout adolescence among

peer groups at school and individuals in the community (see Goldsmith, Bradshaw, and Reiser-

Danner, 1986) whereby further exacerbating the development/continuity of antisocial behavior.

Caspi et al. (1987:308) summarizes this interaction as one in which "the child acts; the

environment reacts; and the child reacts back in mutually interlocking evocative interaction."

Thus, in the proposed theoretical model, living in a socially disorganized area should impact

group-based offending trajectories by increasing the likelihood of being assigned to a chronic

offender group.

Moffitt (1994) has also suggested that chronic offending can perhaps be regarded as an

exclusively male phenomenon because of the disproportionate concentration of risk factors such

as neurodevelopmental disorders and early behavior problems among males. Additionally,









Moffitt and others (see Moffitt et al., 1994) have argued that female delinquency tends to be

more erratic when compared with male delinquency. More recently, Piquero and Moffitt (2005)

have contended that males do tend to show more involvement in crime/delinquency regardless of

offender group, i.e., adolescent-limited or life-course-persistent. Consequently, it is expected

that a greater proportion of males will exhibit both types of offense trajectories and more

specifically, being male will be a significantly distinguishing feature of group-based patterns of

offending.

In contrast, Sampson and Laub's age-graded theory of informal social control argues that

social bonds or stakes in conformity in adulthood affect desistance from crime. It is expected

that individuals who exhibit a virtual non-offending trajectory will also display the greatest

amount of stakes in conformity since they are likely to be already bonded as youth and

adolescence since they show no involvement in delinquency. Similarly, this assumed early

attachment to institutions such as family and school is likely to continue into adulthood in that

they will also be more likely to be married and have full-time jobs which will further restrict

their involvement in crime. Second, while social disorganization is likely to affect group

assignment to an offender group, it is believed that the experience of life transitions such as

marriage and employment will increase the likelihood that individuals will express an

adolescent-limited/peaked offense trajectory compared with a chronic offending pattern since

these social bonds are theorized to have an effect on criminal career desistance. Or in other

words, while these groups may have relatively similar patterns of offending during adolescence,

these life transitions will be related to desistance among the adolescent-limited/peaked offender

group.









In short, recent theoretical and methodological advancements in criminological research

permits me to examine the possible generalizability of formerly discussed offender groups (non-

offenders, adolescent-limited/peaked offenders, and chronic offenders) and exploring a potential

integration of macro-level and micro-level theories for explaining offending. These issues along

with a description of the methodology and analytic procedures that will be utilized in this current

study will be the focal point of the following chapter (Chapter 4) and the results chapter (Chapter

5) will involve a systematic exploration into the several key hypotheses mentioned below.

Hypotheses and Expectations


Hypothesis #1: According to Moffitt's (1993) taxonomy and prior applications of trajectory
analysis (for review, see Piquero, 2007), it is expected that the semiparametric Poisson-based
regression analyses will produce three distinct classes of offenders including: 1) non-offenders;
2) adolescent-limited; and 3) chronic offenders.

Hypothesis #2: In accordance with the theoretical extensions provided by Moffitt (1994) and
Moffitt et al. (1994), it is expected that sex with be a significant covariate for distinguishing
group-based trajectories or in other words, males will be more likely to be in offending groups
than females.

Hypothesis #3: In line with Sampson and Laub's age-graded theory of informal social control, it
is hypothesized that individuals who self-report having experienced these key life events (or
stakes in conformity) will be less likely to be chronic offenders.

Hypothesis #4: It is expected that individuals whose neighborhood of origin is characterized by
poverty/low neighborhood socioeconomic status and residential instability (i.e., social
disorganization) will be more likely to be chronic offenders than adolescent-limited offenders or
non-offenders.

Hypothesis #5: Macro-level and micro-level effects on the participation and frequency of
offending will still be observed net of the effects of each other-or in other words, both macro-
level and micro-level effects will be significantly related to latent class membership. However,
the possibility of mediation will also be explored.









































Figure 3-1. Integrated Theoretical Model: Social Disorganization, An Age-Graded Theory of
Informal Social Control and Moffitt's Developmental Taxonomy









CHAPTER 4
DATA AND METHODS

Introduction

The former two chapters (Chapters 2 and 3) provided the theoretical framework for

explaining individual variations in offending and the latter chapter expressed how macro-level

and micro-level factors could affect an individual's pathway of offending and why males can be

expected to be disproportionately found in chronic offending groups. This information along

with the closing section of Chapter 3 that discussed prior applications of Nagin and Land's

(1993) group-based methodology all bear equal importance in framing the outline of this current

chapter.

This chapter begins with a description of the data and sample for which all the analyses of

this current study are based. This is followed by a section highlighting the macro-level and

micro-level independent variables incorporated in this study to empirically test Sampson and

Laub's (1993) age-graded theory of informal social control and Moffitt's (1993) taxonomy of

adolescent-limited and life-course-persistent offending. The macro-level measures used in this

study were developed with attention to those factors emphasized as having relatively strong

effect sizes according to a recent meta analysis by Pratt and Cullen (2005) and the micro-level

measures were theoretically-derived and based on prior empirical research. After the section

describing the measures, a discussion of the specific group-based modeling technique (Nagin &

Land, 1993) used in this current study to identify latent classes of offenders is presented.

Next, there is also a discussion devoted toward fleshing out the recent debate on the

usefulness of the technique that has occurred between the originators of the methodological tool

and some of its most vocal critics. This chapter will conclude with a description of the software

and statistical procedures that will be used in order to estimate the group trajectories and then to









assess the relative strength of macro-level and micro-levels variables for predicting the

likelihood of group membership.

Data Sources

The data used for these analyses were drawn from the first two birth cohorts collected by

Shannon (1988, 1991) in Racine, Wisconsin. Racine was a Midwestern industrial city with an

estimated population of 89,000 residents according to the 1960 U.S. Census. Shannon initially

gathered detailed information on three births cohorts, the first being 1942, followed by a second

cohort in 1949, and the third from the year of 1955.

Overall, official police contact data was collected on all of the individuals from each of the

three cohorts from "birth" up until the end of the follow-up period (i.e., 1974) which yielded

sample sizes of 1,352 for the 1942 cohort, 2,099 for the 1949 cohort and 2,676 for the 1955

cohort. In order to create a uniform baseline age of offending across cohorts, this current study

only included information on police contacts after the age of six. More specifically, Shannon

was particularly interested in adolescents that were defined as "continuous residents" of Racine

through their eighteenth birthdays (regardless of their mobility within the city's limits).

This continuous-residence data (which provides the data for the focus of this study)

generated sub-samples of 633 for the 1942 cohort, 1,297 for the 1949 cohort, and 2,159 from the

1955 cohort. Of these sub-samples, Shannon (1988, 1991) selected random samples for the

purposes of collecting in-person interview data for the first two cohorts that began on June 1,

1976 (when the 1942 cohort was age 34 and the 1949 cohort was 27 years of age).1 Group-based

offending trajectories will be estimated separately for the 1942 and 1949 continuous-resident

subsamples presented below. A later section in this chapter entitled "Trajectory Estimation"


1 According to Shannon (1988, 1991), the lack of continued federal funding restricted him from collecting interview
data for the 1955 cohort.










provides a more thorough discussion of the various models and a tabular display can also be

found below in Table 4-1. In addition, Table 4-2 illustrates the different years of observation

that can be attributed to each of the data sources along with the corresponding ages of the Racine

continuous-residents by cohort.

Thus, the final continuous-residence based samples previously alluded to involves taking

the police contact and interview-level data separately for the 1942 and 1949 residents and

merging this individual-level data with the limited available neighborhood-level data based on a

unique Census block identifier for each individual. Once the individual and neighborhood-level

data were merged and sample frequencies were conducted, it was observed that a number of

individuals were missing information on all of the neighborhood-level variables of interest.

Therefore, in order to create an accurate baseline and minimize missing data, we chose only to

include the cases that had existing 1950 Census block information.2 This selection criterion left

a total sample of 235 and 323 individuals for the 1942 and 1949 continuous-residents samples

respectively. These samples were predominantly White residents (92% and 85%), followed by

Black (6% and 11%), and the approximately remaining 3% of each sample were Chicanos.

There was an almost even split as far as the sex composition between the males (47% and 53%)

and the females (53% and 48%).



2 In order to incorporate change into the models (i.e., such as computing change scores for the Census measures by
using 1960 data as well) would have resulted in the loss of more than 60 additional individuals which would have
affected the stability of the trajectory models (see Nagin & Tremblay, 2005; Sampson & Laub, 2004). Preliminary
comparisons indicated that the 1950 and 1960 block data were relatively similar; therefore, the 1950 Census
measures were used with the intention to preserve any additional sample loss and to represent macro-level-level
characteristics that existed closer to the early developmental years of the Racine residents that Moffitt (1993) has
indicated as a crucial time period (see also Gottfredson & Hirschi, 1990). Prior research has shown the adverse
effect of living in a low SES neighborhood on factors associated with life-course-persistent patterns of offending,
especially in early childhood (see Brooks-Gunn et al., 1993; Chase-Lansdale et al., 1997; Duncan et al., 1994). The
original principal investigator Shannon (1994:7) also indicated that if a researcher was interested in the ecological
characteristics of which the cohort members spent their early childhood/juvenile years, then the 1950 Census data
should be used.









Independent Variables

Macro-level independent variables

Social Disorganization. Based on the factors highlighted by Shaw and McKay (1942) that

are believed to contribute to social disorganization and associated with crime (as defined as the

breakdown of the formal and informal social networks for regulating neighborhood resident

behavior), measures of neighborhood socioeconomic status (SES) and residential mobility were

included.3 Neighborhood SES was measured directly by the average housing values of the

residential properties on each block. Prior research has shown the positive effects of living in a

high SES neighborhood has on academic achievement, verbal ability scores, and behavioral

problems (see Klebanov et al., 1997). Residential mobility was indirectly measured by the

percent of housing units on a block that were renter-occupied compared with those that were

owner-occupied.4 Simcha-Fagan and Schwartz (1986:671) have argued that structural

characteristics can affect an individual's ability to form social bonds to the community.

Individuals who rent as opposed to owning their property are less likely to develop the necessary

ties (see Bellair, 1997; Granovetter, 1973) that are important for providing a mechanism of

informal social control. Finally, percent overcrowding was Census-based measure indicating

the percent of rooms within housing units on each block with more than one person per room.

Wilson (1987) has also attributed housing density to concentrated poverty, which has been


3 Because of the sparse amount of U.S. Census data available in 1950 and relative homogeneity of Racine, this study
was unable to include a measure for ethnic heterogeneity in order to provide a complete measure of social
disorganization incorporating all of the processes that lead to crime.
4 The original measure of "social disorganization" used by Land (2000) was a scale created by Shannon (1988,
1991, 1994) where a rank of 1-26 was assigned to an individual to represent their "neighborhood SES" based on the
average Census block characteristics of the area in which they lived. This measure was actually referred to by
Shannon as a "natural area code" with no reference to social disorganization and given that the measures used
(average dollar value of rent, average dollar value of owned property, percent lacking plumbing, percent
overcrowding, and percent renter-occupied) did not load on one construct it is likely that Shannon's (1988, 1991,
1994) "index" and Land's (2000) later use of this index is not an entirely accurate proxy measure of social
disorganization.









posited as a precursor of social disorganization. Dwelling units that are over-crowded also may

lead to sense of "anonymity" between the residents because of their failure to interact with their

neighbors (see Roncek, 1981; Sampson, 1995).

The results from principal components analysis confirmed that these three variables were

measuring a similar construct (i.e., social disorganization) across the two cohorts as presented

below in Table 4-3; therefore, the corresponding component loadings were utilized in order to

construct the appropriate indexes.5 This procedure allows for preserving the integrity of the data

by removing the potential of a partialing fallacy (Gordon, 1968) and reducing the potential for

multicollinearity which is common in macro-level data (Land, McCall, & Cohen, 1990). Thus,

positive z-scores were indicative of more socially disorganized areas.

Micro-level independent variables

Stakes in Conformity. Sampson and Laub (2005c) argue that involvement in institutions

such as marriage and work provide both short-term and long-benefits for individuals. These life

transitions can change an individual's motivation for crime by re-focusing their attention toward

spending time with their spouse and/or investing time in their job. These events are believed to

foster an individual's development of prosocial adult bonds. Prior research has shown support

for the impact that these transitions can have on desistance from crime (Sampson & Laub, 1993;

Warr, 1998) and other research has shown that the effect of these life transitions are age-graded

(see Uggen, 2000).

Therefore, in accordance with Sampson and Laub's (1993) theory and prior research (see

Piquero et al., 2002) a stakes in conformity index was created for the two local life circumstances



5 This procedure involves multiplying the raw variable scores by the standardized weights obtained as a result of the
principal components analyses and then summing these values together to produce a single measure (see Kim &
Mueller, 1985).









of marriage and employment. Each of these transitions are dummy-coded with 1 indicating if the

individual self-reported having experienced the particular life transition by age 25 and 0

indicating if the individual failed to report having experienced the specific life transition by age

25.6 Therefore, the index ranged from 0 to 2, with 2 indicating that the individual reported

having experienced both of these life circumstances.7

Sex. Sex has long been recognized as one of the strongest predictors of involvement in

delinquency and violent crime where crime rates have consistently shown higher prevalence

rates of offending for males when compared with females (Rowe et al., 1995). Daly (1998) has

also noted that self-report studies have repeatedly indicated that males report a higher prevalence

(involvement) and incidence (number of acts) of offending. Wilson and Hernstein (1985) have

commented on the strength of the relationship between sex and crime over time and cultures, and

according to Gottfredson and Hirschi (1990), anywhere and everywhere males tend to commit

more crime than females (with the exception, check fraud, shoplifting, and prostitution).

Additionally, Moffitt and some of her colleagues have suggested that life-course-persistent

offending may be observed almost exclusively among males because neuropsychological

disorders and childhood-onset conduct problems are more prevalent among males (Rutter, 1983;

Singer et al., 1968; Werry, 1986). Female delinquency also tends to be more erratic when


6 There were self-reported measures that existed in the data for yes/no responses to these life events; however, in
order to ensure that the actual life transition occurred within the observation period of the data available for both
cohorts, the experience of these life transitions after age 25 for the 1942 cohort were not included.
7 It is important to note up front that this measure is limited in that Sampson and Laub have argued that mere
participation in life transitions may not be as important as the quality of the attachment to the various social
institutions (i.e., marriage, work, etc.). Laub et al. (1998) also indicate that the inhibiting effects of life transitions
are cumulative over time and it takes a while for individuals' to express the necessary level of cohesion or
attachment to affect their behavioral pathway of offending, or affect their desistance from crime. In the end,
according to Piquero et al. (2002) it is likely that just being married or having a job is some kind of indication that
individuals' have developed some affective ties to their spouses and/orjobs or else they would not be likely to be
involved in the life circumstance to begin with.









compared with male delinquency (Moffitt et al., 1994). Prior research has shown that females

typically show less involvement in crime in both offender groups (adolescent-limited offenders

and life-course-persistent offenders) when compared with males. Recent applications of Nagin

and Land's (1993) trajectory analysis have also shown distinct offender groups among females

(see Cote et al., 2001; D'Unger et al., 2002; Fergusson & Horwood, 2002; Land, 2000; Piquero

et al, 2005). Provided with this theoretical framework and building on the prior research

mentioned above, a dichotomous variable representing the sex of the 1942 and 1949 cohort

members was included with males coded as 1 and females coded as 0 as a covariate in the

trajectory models as well as in the multinomial logit models presented in Chapter 5.

Black. Similar to sex, race has also been a long-standing correlate of crime (see

Hawkins, 1995) and Moffitt (1994) has argued that the higher prevalence of both adolescent-

limited and life-course-persistent offenders among Blacks may account for the racial variation in

crime rates because the theoretical origins of the latter pattern of offending disproportionately

affect Blacks because of the strong association between race and poverty. Prior research has

shown a higher prevalence of life-course-persistent offenders among Blacks (Piquero & Buka,

2002), and other researchers have argued that the risk factors associated with life-course-

persistent offending typically tend to be disproportionately concentrated among Blacks (see

Piquero & Moffitt, 2004; Piquero et al., 2005).

Furthermore, macro-level research has indicated that Blacks were more likely than Whites

and Hispanics to live in poor neighborhoods and be involved in more serious forms of

criminality. Prior research has also shown that Blacks who reside in these areas are

disproportionately affected by the burden of child-care and residential instability and these

factors were strong predictors of delinquency and violence (for review, see Gephart, 1997).









Additional research utilizing trajectory analysis has also noted the increased likelihood of Blacks

in chronic offending groups (see Land, 2000); therefore, it was necessary to include race. The

original nominal level variable indicating the race of the 1942 and 1949 cohort members was

recorded into a dummy variable with Blacks being coded as 1.

Low SES Household. This prestige-based measure will be included in order to provide an

individual-level SES complement to the macro-level measures of neighborhood SES. Prior

research has shown that individuals who are from low SES households in low SES

neighborhoods tend to have the highest rates of offending and comparatively those from the high

SES households in high SES communities have the lowest offending rates (Braithwaite, 1979;

Reiss & Rhodes, 1961; Wikstrom, 1991a). Using trajectory analysis, McDermott and Nagin

(1998; see also Land, 2000) have also shown that low SES is an important factor for predicting

group membership, more specifically the likelihood of being classified as a chronic offender.

Additionally, Brooks-Gunn and Leventhal (2000) have noted the importance of including this

family-level measure when estimating neighborhood effects.

This variable was measured as a scale indicating the prestige of the head of the

household's occupation in the family in which the individual grew up. It ranged from a low of

one to a high of nine with the following occupational types represented: (1) professional,

technical, managerial; (2) clerical and sales; (3) craftsman, foreman; (4) operative; (5)

maintenance and service; (6) private household labor; (7) industrial labor; (8) farm, agricultural

labor; and (9) unemployed. Consistent with the direction of the macro-level social

disorganization index, higher values indicated worse situations (i.e., growing up in lower SES

households).









Family Stability. This measure was used to represent the stability of the family dynamic

in which the cohort member was socialized. Most research has shown an established

relationship between family stability and delinquency/crime (Glueck & Glueck, 1950; McCord

& McCord, 1959; Goetting, 1994) and crime rates (Sampson, 1987). Studies have found that

adolescents living with both biological parents have lower rates of offending when compared

with those who reside in broken homes or reconstituted families. According to Gottfredson and

Hirschi (1990:104), single-parent-headed households (usually by a woman) may be unable to

provide the same amount of supervision because of the time constraints imposed on them and

thus may be more likely to have negative responses to their children's unwanted behaviors.

Furthermore, this variable along with the following measure listed below (family structure) have

been identified as key family-level variables when estimating neighborhood effects (see Brooks-

Gunn & Leventhal, 2000)

Originally, this variable was based on a scale created by Shannon (1988, 1991) which

combined responses to several items regarding family composition ranging from 1 to 20, where

according to Shannon (1988, 1991), 1 represented the most stable family dynamic (two parents

present during the entire childhood/juvenile period) and 20 the least stable family dynamic (lived

only with grandparents, never with parents). For the purposes of this study and due to the little

variability among the categories, the scale was collapsed into a dichotomous variable with 1

indicating the presence of two parents during the entire childhood/juvenile period and 0

indicating any other type of less stable family dynamic.

Family Structure. In order to account for the possible differential opportunities of

parental supervision affecting individuals' offending pathways (Gottfredson & Hirschi, 1990)

along with having a complementary individual measure of household density to the macro-level









household density measure (percent overcrowding), a second measure of family dynamics was

included to represent family structure. This was a measure of the number of children within each

of the cohort member's family (including the cohort member).8

Differential Association. Differential association theory (Sutherland, 1939, 1947) has

long stressed criminal and deviant behavior is learned through exposure to and acceptance of

definitions favorable to crime. Burgess and Akers (1966), and in later reformulations by Akers

(1977), expanded on this idea by specifying the process in which learning occurs. Favorable

definitions of crime and deviance are learned in a social setting where the imitation of criminal

and deviant behavior also occurs. Delinquency, like any behavior, generates either beneficial or

adverse consequences positive or negative reinforcement. If delinquent or criminal behavior is

reinforced either socially or non-socially, the behavior is likely repeated in the presence of cues

to which it has been paired. Or in other words, Akers (1998:50) argues that the probability that

individuals will engage in deviant behavior is increased when they differentially associate with

others who commit the behavior and express favorable definitions toward it. Thus, based on

research that has produced strong evidence for what is now termed the perceived "delinquent

peer effect" (Felson & Haynie, 2002; Matsueda & Anderson, 1998; Schreck, Fisher, & Miller,

2003; Warr, 1993; Warr & Stafford, 1991), a measure was included with 1 indicating that the







8 Overall, there was very little missing data at the individual-level (less than 1% for any of the variables used in the
analyses). The only two variables that had missing data were family stability and family structure. In order to
correct for the missing values, the method of group specific means were computed. In order to do this procedure the
sample was sorted into five categories: 1) White males; 2) White females; 3) Black males; 4) Black females; and 5)
Chicano males (there were not any Chicano females in the sample). For the few cases (n=4) where a missing value
was recorded, the sex/race specific group means were imputed.










individual self-reported "peer influence" as the strongest contributing factor toward their

delinquent/criminal behavior in junior high and high school.9

Formal Social Control. This measure was a scale of the individual's perception of the

level of police patrol in their neighborhood during their junior high and high school years.10

This indicator of institutionalized formal social control ranged from 0 representing "not at all" or

"no police patrol," 1 for "lightly patrolled," 2 for "moderately patrolled," to a high of 3 which

indicated that the individual perceived their neighborhood as being "heavily patrolled" during

their early to late adolescent years.1

Moved. In order to provide a contextual measure of residential mobility to complement

the compositional measure included in the social disorganization index, this variable was coded

as 1 if the individual had moved within Racine during their juvenile years. It is likely that

moving, particularly during the adolescent years, is likely to have an adverse effect on an


9 It is worth mentioning here that this is certainly a less than ideal measure of differential association and more direct
measure would have been preferable; however, due to the limitations of data this measure was included as a control
variable.

10 It is important to note that this measure is a weak measure of formal social control and that direct measures such
as prior conviction or incarceration are optimal; however, there was only two items that allowed the individuals to
report having received jail time as a result from their police contact and these questions were only in reference to
their first and second police contact. Because only 1 to 4 individuals across the cohorts actually reported this as a
response I made the determination to use the "police patrol" measure because there was a sufficient amount of
variation in responses to this latter survey item. Additionally, this measure is also limited due to the fact that it was
asking individuals to retrospectively answer this question based on their perceptions of how they felt as a middle and
high-school juvenile (which is at least 10 years prior and for the 1949 cohort and 20 years prior to the date of the
interviews in 1974 for the 1942 cohort) and it is also likely that those individuals who actually came into contact
with the police may have had a bias toward reporting higher levels of neighborhood patrol. Future studies would
benefit from incorporating more direct measures of formal social control such as those discussed by Sampson and
Laub (1997) and De Li (1999), which relied on convictions and on court appearances and convictions, respectively.

1 Shannon (1988, 1991, 1994) originally developed a scale to assess formal social control, but this measure had a
series of flaws, with the most serious being that the scale was constructed by summing self-reported responses of the
cohort member's perceptions of the (1) level of police patrol when they were in junior high and high school
(l=heavily, 2=moderately, 3=lightly; 4=not at all); (2) their and a couple of their friends' views toward the police in
junior and senior high, positivei, 2=negative, 3=indifferent); and (3) to report whether or not they got into trouble
with the police with a couple of their friends (yes/no). Even more oddly, this scale was constructed by coding an
individual's response as "1" if they had either negative or indifferent views toward the police, adding "2 points" if
the individual reported that they and their friends got in trouble with the police, and adding "4 points" if the
individual also responded that they perceived police patrol was either heavy or moderate during their juvenile years.









individual's pattern of offending by disrupting the ability to form/maintain prosocial bonds with

peers and their schools, which are important factors for inhibiting delinquency (see Hirschi,

1969; Akers, 1985).12

Dependent Variables

Delinquent/Criminal Involvement. The variable used to measure offending was the

number of official police contacts for each individual from age 6 to age 25.13 The police contact

measure was limited to those instances in which an individual came into contact with a Racine

police officer specifically for a crime that he/she was suspected to have committed or did in fact

commit. In addition, the official police contacts were also restricted to contacts that had true

crime codes with a felony/misdemeanor distinction as well as contacts for status offenses to

measure offending (i.e., excluding traffic offenses, and city ordinance violations) because (1)

they are a more accurate official measure of "true" offending when compared with arrests and

convictions; (2) they are closer in time to the actual criminal act than self-reported offending (see

Hindelang et al., 1981; Thornberry & Krohn, 2003); and (3) the inclusion of status offenses

increases the potential for identifying adolescent-limited offenders (see Moffitt, 1993). 14

Offending Group Membership. The results of the semiparametric mixed Poisson-based

regression models will guide the identification of the number of offending groups that exist

within these data and the shapes of their offending trajectories. The identified latent classes of

offenders will operate as the dependent variable for the estimation of the multinomial

12 It is important to note that this measure is limited in that is was impossible to discern from the available data if the
individual moved from a social disorganized area to a more or less disorganized area; therefore, it is important to
interpret the findings with caution and keep in mind that the expectation is that moving is seen as having an adverse
effect on offending patterns within this study.

13 This police contact data refers to the years that are consistently available for the 1942 and 1949 continuous-
residents or up to and including age 25.
14 Prior trajectory analyses with Shannon's (1988, 1991) Racine cohorts relied solely on felonies and misdemeanors
(D'Unger et al., 1998) or everything including traffic offenses and "suspicion" contacts (Land, 2000).









(polytomous) regression models (Models 1 and 2), with the non-offender group omitted in order

to serve as the reference category. The following sections describe the group-based modeling

technique designed by Nagin and Land (1993) that will be utilized in this study for identifying

the offender groups along with a discussion of the existing debate that has been ongoing related

to the usefulness of the methodology.

Study Limitations

There are a series of limitations of the data source itself and the measures used in this study

that are worth noting. First, the timeliness of the data is an issue since the cohorts themselves

were born in the 1940s, socialized as children and/or adolescents in the 1950s and 1960s.

Therefore, it is likely that these samples may not be generalizable to other cohorts from more

recent time periods because they were born during WWII (e.g., 1942 cohort) or are from the

baby-boomer generation (e.g., 1949 cohort) who grew up in the post-WWII era. Second, Racine,

WI was a Midwestern industrial city with an estimated population of 89,000 residents, and it is

not likely that the results from this study would apply to the larger metropolitan urban cities of

today (i.e., cities with populations greater than 100,000).

Third, particularly during the time of data collection, Racine was a relatively

homogenous city with limited numbers of minorities, and according to Shannon (1988), although

the overall population did rise to 95,000 persons by 1970, this was followed by a decline to

86,000 persons by 1980. He attributed this population decline to a growing number of young

people in need or work and increased female participation in the work force, but this was met

with a decreased number of individuals leaving for retirement. Taken together, the results

discovered here may not necessarily apply to areas with larger minority representation or

dissimilar population shifts.









Noting the limitations of the cohort members and the location, there is also a series of

limitations with the measures used in this analysis. First and foremost, the data available in the

1950 Census at the block level is very limited or in other words, there was not the typical "long

list" of measures accessible to macro-level researchers today for this research, and there also was

a noticeable amount of missing 1950 census data for the 1942 and 1949 continuous-residents.

With this is mind, the social disorganization measure used here is relatively weak in the

sense that it is only measuring two (neighborhood SES and residential mobility) of the three

elements that Shaw and McKay (1942) highlighted as important factors associated with the

breakdown in informal social control and crime rates. Once again, Racine was a relatively

homogeneous area, so it is likely that racial dissimilarity would be difficult to directly observe at

the block level. Additionally, Sampson et al. (1997) have argued that what matters more than

social disorganization is collective efficacy or what happens at the individual level as far as

developing community ties and informal social control. There were not an available individual-

level data that would allow for an investigation of this hypothesis, so caution should be noted

that these individual-level processes are not accounted for in this study.

In addition, as footnoted earlier, the stakes in conformity index used to assess Sampson

and Laub's (1993) age-graded informal social control theory is limited in that it is a dichotomous

index with a short range (0 to 2). More recently, Laub et al. (1998) have argued that mere

participation in local life circumstances may not be as important as the quality of the attachment

and that the inhibiting effects of life transitions on desistance are cumulative and gradual over

time. Therefore, while this measure does not capture the level of attachment or cohesiveness of

the relationship, it is likely that just being married or having a full-time job is indicates that some

degree of attachment has already been formed or else the life transition would probably not have









occurred in the first place (see Piquero et al., 2002). The differential association measure is also

noticeably weak.

Regarding the dependent variable (police contacts), there are also well known strengths

and weaknesses from using either official data (police contact, arrest, conviction) or self-report

data to measure criminal behavior. Under-reporting is one of the principal limitations of official

statistics because the majority of police activity is a direct result of the willingness of citizens to

report crimes (see Hindelang et al., 1981; Thornberry & Krohn, 2003). However, there are also

extensive disparities in victim reporting based on various factors including: requirements for

insurance companies, confidence or lack of confidence in police departments, etc. Furthermore,

there are many offenders who will never come to the attention of law enforcement, such as

shoplifters, burglars, or white-collar criminals (Junger-Tas & Marshall, 1999). This result might

be due to the fact that more skillful and shrewd offenders are less likely of being arrested and/or

may reflect the differences in police discretion and priorities across regions/agencies (Blumstein

& Cohen, 1987).

One key advantage of self-report measures is that they often offer more comprehensive

information than official data. They provide background and sociodemographic data on

individual offenders, as well as personal characteristics. Research has indicated that self-reports

are effective in answering questions of causality, motivation, change, and stability in criminal

behavior (Junger-Tas & Marshall, 1999:296). Self-reports also tend to include both a broader

range of deviant behavior and less serious types of crimes than official police records indicate

(Blumstein, Cohen, Roth & Visher, 1986:32). Some examples of the types of crimes that are

likely to be found in self-reports but not in official data are prostitution, drug abuse, public order

crimes, and drunk driving.









Notwithstanding the mentioned benefits, there are several limitations to the self-report

method. Under-reporting and poor recall of criminal activity are disadvantages because self-

reports tend to be retrospective in nature (as was the case with the interviews in these data). In

addition, some respondents may even inflate their actual level of criminal activity. Cross-study

comparisons of self-reports may also be hindered due to differences in the wording of the

questions, differences in the administration of the survey, or differences in time periods

(Blumstein et al., 1986:33). Taking into account the noted limitations of police records and self-

reported measures of offending, this study chose to measure police contacts because they are a

(1) more accurate official measure of "true" offending than arrests and convictions, and (2)

typically closer in time to the actual criminal act than self-reported offending which sometimes

asks individuals to recall their past behavior in their lifetime, last twelve months, last six months,

etc. (see Hindelang et al., 1981; Thornberry & Krohn, 2003).

Within the context of all of the above mentioned limitations surrounding the data, time

period, samples, and the location along with the measurement shortcomings, the following

section highlights the group-based methodology that will be used to estimate the group

trajectories.

Methodology

Group-Based Modeling

Longitudinal research in its most basic form (as applied to the study of crime), involves the

identification of a cohort at birth and following that group of individuals throughout their life, or

at least long enough to capture the termination of the majority of criminal careers (Blumstein et

al., 1986:104). The longitudinal method allows for a specified cohort's level of involvement in

crime to be traced as the cohort, itself, matures (Greenberg, 1985). This knowledge is crucial for

detecting causal factors or sequences that associate individual characteristics or certain events to









consequential offending (Blumstein et al., 1988). Longitudinal data also permits the use of

complex behavioral models because they allow one to examine the effects of missing or

unobserved variables that may be correlated with crime (i.e., also known as the unobserved

heterogeneity problem). It is often the case that even among people who have similar

characteristics (race, age, sex) there is a substantial amount of heterogeneity. Thus it is

important to utilize longitudinal models or run the risk of using misspecified models and

producing inconsistent parameter estimates (see Nagin, 2005).

According to Nagin (2005:1-2), longitudinal data also provide researchers with the basis to

study developmental trajectories. Most of the common methods for studying developmental

trajectories, however, are those that are calculated to account for individual variability, which

masks the possibility of observing meaningful subgroups within certain samples and populations.

Furthermore, this over-reliance on traditional longitudinal approaches has led to much ad-hoc

theorizing and classifying individuals on the basis of what could be random variation. Hence

important developmentally distinct groups could be missed or misclassified.

The particular method used in this study to model the group-based offending trajectories

was originally developed by Nagin and Land (1993) and is known by several different names

including: group-based modeling, finite mixture models; trajectory analysis, and semiparametric

Poisson-based regression analysis. It is important to note that this technique shares some

similarities with the traditional group-based modeling approaches such as hierarchical linear

modeling (Byrk & Raudenbush, 1987, 1992; Goldstein, 1995) and latent curve analysis

(McArdle & Epstein, 1987; Meredith & Tisak, 1990).

Still, there are distinct differences. For instance, both hierarchical modeling and latent

curve analysis assume continuous distribution functions. Trajectory analysis does not assume









this continuous distribution, but rather assumes that there exist distinct clusters of group-based

trajectories that exhibit different etiological processes, and the methodology approximates the

continuous distribution by identifying points of support (i.e., distinct groups or trajectories)

(Nagin, 2005). These trajectory models are also based on a maximum likelihood (ML) function,

which of course permits them to display the qualities of ML parameter estimates-they are

consistent and asymptotically normally distributed (Cramer, 1946; Greene, 1990). Given the

assumptions and statistical properties, this method assumes that individual differences in

behavioral trajectories can be described as polynomial functions of age or time. The following

formula presents the probability for belonging to a particular group:

J
P (Y,) = Z P(Y),


where P(Y,) is the unconditional probability of observing individual i's longitudinal sequence of

police contacts, Y, over time. It is equivalent to the sum across J groups of the probability of Y,

given i's membership in group weighted by the probability of membership in group (see

Nagin, 2005:25).

There are various forms and derivatives of the trajectory model, but the two most common

parametric forms that are likely to apply to the data used for this study are the censored normal

model (CNORM) and the zero-inflated Poisson model (ZIP). The CNORM model is appropriate

for data where the outcome variable of interest is normally distributed but there exists clustering

of the data at either the minimum or maximum (which typically occurs with offending data).

This model can also be applied to data without censoring by assigning the scale maximum and

minimum values to values that are respectively smaller and larger than any value observed in the

data. The equations listed below correspond to the statistical derivations of the cases when (1)









the data are censored at the scale minimum; (2) normally distributed between the scale minimum

and maximum; and (3) the data are censored at the scale maximum:



p' (yt = Smin) = D (Smin f x,) / o,

(yt) = 1 / C (y,t- f x,t/ o) for Smin < yt< Smax, and

p' (yt = Smax) = 1 P (Smax -ff xt) / c

where it is assumed thatyt is normally distributed with a mean of p3 Xt and conditional on age (t)

with a standard deviation of c and P represents the cumulative distribution of a normal random

variable. In addition, Smin represents the scale maximum and Smax represents the scale

minimum (Nagin, 2005:30).

In contrast, the parametric form of the trajectory model that is needed for modeling count-

based data such as the number of police contacts is the ZIP version, which was first presented in

Nagin and Land's original work and is based on the generalization of what Lambert (1992)

called the zero-inflated Poisson (ZIP) distribution. Count-based models have been more

frequently utilized in the social sciences in an effort to correct for the inefficient and biased

estimates that are produced when linear regression is employed for modeling data where the

dependent variable is measured as a count or frequency of an event (for review, see Long, 1997).

Although the Poisson regression model is the most basic form of the count-based models, it is

often not the most suitable model due to its primary assumption of mean variance equivalence, or

equidispersion. It is more likely the case that, in practice and especially in criminological

research, the distribution of the outcome variable is overdispersed, (i.e., the variance exceeds the

mean); therefore, a more appropriate modeling procedure is recommended (Long, 1997:218).









The ZIP model allows for the assumption of a dual-state-system. A dual-state-system

signifies two situations: one in a zero-contact state where the assumption is that a group of

individuals exists that can be regarded as police-contact free, and the second state or non-zero

contact state where the expected incidences of police contacts are assumed to follow some

known distribution, such as the Poisson (Long, 1997; Lambert, 1992). Several other studies have

also noted the practicality of the ZIP model when using count data with excess zeros (King,

1989; Land, McCall, & Nagin, 1996; Zorn, 1998).

The ZIP model takes the following form where the first term is the probability of no police

contacts and the second term is the probability of the combined event's being active and the

probability of zero police contacts given Xt > 0 (Nagin, 2005:35).

j y't
S(y,)= (1- a, ) (t e t /y) (t= 1, 2,....)

Once the parametric form of the trajectory model is established it is necessary to determine

the functional form of the groups' trajectories and there are three possibilities including a linear,

quadratic, and cubic form (or age, age-squared, and age-cubed). Depending on the ages and the

length of the observation period, it is unlikely that individual offending will have a linear form.

Similarly, although the quadratic model is more often used in research (see Bushway et al.,

2003:140), it only permits one major directional change in the trajectory, and it is likely that

cohort members will and have undergone periods of intermittency in their involvement with the

police. Intermittency periods refer to those points in time (i.e., years) where no official police

contacts were recorded for any particular cohort members. Therefore, it is likely that the cubic

model will demonstrate the best functional form of the trajectories. The CNORM and the ZIP

equations that allow for all three of these possible functional forms are presented below:

yt* = po' + P0Age,1t + P02Aget,2 + P3Aget3 + t, (CNORM),










In (X/) = 3P + P1Age,t + Age t+ P33Age, + ,,t (ZIP),

where Agel,1, Age2,t, and Age3,t are the cohort member's (i's) age, age-squared, and age-

cubed respectively, t represents time, and P,t is a disturbance parameter that is assumed to be

normally distributed and the 3s are the parameters that determine the functional form of the

trajectory (Nagin, 2005:28-33).

Finally, the exact number of groups is chosen based on an examination of the Bayesian

Information Criteria (BIC) in order to maximize model fit. This process is necessary because

conventional likelihood ratio tests cannot be used to assess whether the addition of another group

adds any explanatory power to the model (see D'Unger et al., 1998). In the end, the exact

determination of the number of groups is based on the ability of the model to find a relative

convergence using all of the available information and by the utility of the groups for indicating

distinct and homogenous trajectories, while also having a sufficient number of persons within

each group to allow for further statistical analyses (Bushway et al., 2003:140; see also Brame et

al., 2001). The BIC values can be relied on to determine the number of groups in both nested

and unnested models (see D'Unger et al., 1998; Nagin, 1999), and the BIC values are estimated

based on the following equation, where L is the maximum likelihood, n is the sample size, and k

is the number of parameters (Nagin, 2005:64):

BIC = -2 log (L) + log (n) k.

A Seductive or Useful Statistical Method?

More than thirty years has past since Hirschi (1969:53) initially provided the word of

caution to criminologists when using offender typologies by asserting that "it begs the question

of causal homogeneity by focusing exclusively on the question of behavioral homogeneity."

Criminologists have come a long way since Hirschi's earlier remarks regarding offender









typologies; however, this does not mean that Nagin's and Land's latest methodological technique

has been insulated from theoretical scrutiny. The group-based modeling technique has been the

source of recent criticisms and counterpoints between its originators and two prominent life-

course theorists, Sampson and Laub.

This debate began with the findings from an earlier study by Sampson and Laub (2003)

questioning the stability of offending patterns over the life course and more specifically with an

article published in the Journal of Quantitative Criminology, entitled "Methodological

sensitivities to latent class analysis of long-term criminal trajectories." The authors (Eggleston,

Laub, & Sampson, 2004) argued that the "enthusiasm" for the use of Nagin and Land's (1993)

semiparametric group-based models has not been equally supplemented with robustness and

sensitivity analyses that are often inherent in offending data.

They proceed by summarizing the obvious advantages of these models (see also Nagin,

2005) that include (1) the fact that they are designed for longitudinal data; (2) they lead to the

ability to distinguish groups of offenders that are homogenous within their trajectory but distinct

from other trajectories; (3) they provide an objective systematic criterion for establishing

offending groups by using the Bayesian Information Criterion (BIC) rather than presupposing

subjective sub-groups of offenders; (4) posterior probabilities are used to assess the precision of

group assignment; and (5) this method accommodates missing data, whereby offenders that have

unavailable information for several time periods are not excluded from the analysis (Eggleston et

al., 2004:2).

Relying on 500 juvenile delinquents' data from the archives of the classic Gluecks'

Unraveling Juvenile Delinquency research (1950, 1968), Eggleston et al. (2004:23) concluded

that the length of follow-up can affect the trajectory patterns observed and argue that data that









focuses on say age 6 to 15 versus 10 to 32 may produce trajectories with different curves (i.e.,

shapes), peak ages of offending, and possibly membership to different offending groups.

Additionally, they argue that research using trajectory analysis needs to attend to both

incarceration and mortality information because these factors can also affect offending

trajectories (see Piquero et al., 2001 for a similar discussion on the importance of accounting for

incarceration data).

This article discussed above prompted an initial response by Nagin (2004) and a

subsequent response by the Sampson, Laub, and Eggleston (2004) that were both printed in the

same issue of the Journal of Quantitative Criminology. Nagin's (2004) initial response indicated

that he had no criticisms regarding the analysis conducted by Eggleston et al. (2004) and

typically agreed with their findings; however, he did note some counterpoints to a few of their

conclusions.

Specifically, Nagin (2004:28) argued that the criticisms offered by Eggleston et al. aimed

at the trajectory model are limitations that plague all longitudinal data (i.e., projecting beyond the

range of the data and missing data resulting from incarceration and/or death). Other researchers

have also commented on the idea of career termination or false desistance resulting from death

(see Reiss, 1989) or based on data that does not track the individuals throughout their entire life

course (see Bushway et al., 2001; Piquero et al., 2003). Nagin concluded by reminding

researchers that the groups that result based on the trajectory models must be understood as an

approximation and this warning was already previously presented in Nagin and Land's (1993)

original work.

Sampson et al. (2004) reanalyzed the Gluecks' data in response to Nagin's comments and

subsequently provided several more suggestions for future researchers relying on the trajectory









method. They also felt that the majority of their criticisms should also be heeded by any future

criminological research incorporating longitudinal data. First, using random samples of varying

sizes, they found that the number of groups only began to stabilize with samples of over 200.

Other research has indicated that the number of identified groups is partly a function of sample

size (see Nagin & Tremblay, 2001) and that the number of groups tends to be consistent with

samples above this range (see D'Unger et al., 1998).

In their final remarks, Sampson et al. (2004:41) indicate that the larger problem with the

use of trajectory analysis and inferences made from longitudinal data with short follow-up

periods concerns the causal theory for using groups. They state that future research should resort

to grounding all methods in strong theory and be wary of models that provide evidence of an

existence of groups just because they are statistically discovered or what is referred to as the

reification of groups (see also Laub & Sampson, 1993; Sampson & Laub, 1993).

As if this above exchange was not enough to settle the debate and prompt future

longitudinal researchers to be mindful to theory and consider the estimation of groups in

trajectory analysis as approximate, this dialogue was continued most recently in a 2005 issue in

the journal of Criminology. Nagin and Tremblay (2005a:873) set forth with an attempt to

debunk three commonly held misconceptions regarding trajectory analysis that originate from

researchers' failure to understand that the method results in an approximation of groups

including: (1) the individuals actually belong to a trajectory group; (2) the number of trajectory

groups is immutable; and (3) the trajectories of group members follow the group-level trajectory

in lock step.

Regarding the first misconception, Nagin and Tremblay (2005a:882) state that trajectory

groups are referred to as is they are real/true entities only for the purpose of "literary









convenience" in that it facilitates the ability to clearly discuss groupings of individuals that

display similar offending patterns. Since the discovering of trajectory groups is a statistical

approximation, they are not meant to be interpreted as factual representations of reality (a

common limitation of all statistical models). Similarly, this false understanding of the fact that

the method results in a statistical approximation of groups also has fostered the development of

the second misconception, i.e., that the number of groups in the sample is immutable.

According to Nagin and Tremblay (2005a: 888), the number of groups and the shapes of

their offending curves are not a fixed reality either. Instead, the number of groups and their

respective trajectories are influenced by the quality of the longitudinal data that are relied on for

the analysis. In this sense, the quality of the data refers to the sample size and the number of

observation periods, and obviously larger samples and longer follow-up periods are always the

most optimal. The research presented above also showed that the number of groups and the

shapes of their trajectories were relatively stable in samples that were larger than 200 cases

(D'Unger et al., 1998; Sampson et al., 2004).

Nagin and Tremblay (2005a:892) concluded by offering a couple of comments concerning

the final misconception (that individuals will follow their offending group's trajectory in lock

step). First, the shape of the group's trajectory within which the individual is assigned is

intended to reflect a behavioral process over time versus possible individual variation across this

identified pattern of offending. There will always be individual variation, just as there is

variation in offending if a researcher relies on other units of analysis (households, cities,

counties, etc). It is highly unlikely that any one particular member of any these aggregations will

exactly display the average offending rates of the group. In addition, there are also statistical









reasons why individuals within a group trajectory will deviate from the process captured over the

entire period of observation, most notably involving sampling variation.

Notwithstanding this further attempt by Nagin and Tremblay to defend the methodology

and to caution its users about the limitations of trajectory analysis, this article was inevitably met

with a rejoinder by Sampson and Laub (2005a). In a departure from the previous dialogue of the

debate as depicted in the Journal of Quantitative Criminology, this article had a much more

forthright tone. Sampson and Laub (2005a:905) basically concluded by suggesting that

trajectory analysis is just another "sexy" statistical tool much like path and factor analysis long

ago, followed by event history analysis in the 1980s, and most recently preceded by LISREL and

HLM in the 1990s.

Although, they did argue that the methodology has some uses, they believe that the larger

problem is criminology's obsession with statistical tools without attention to theoretical

underlying processes that influence behavior. In a last attempt by Nagin and Tremblay to

provide a response to Sampson and Laub's critiques, Nagin and Tremblay (2005b) themselves

were surprised by the tone of the latest rejoinder, and equally cautioned criminologists to be

wary of "seductive" theories as well as "seductive" methodologies (especially when the critics

are promoting a life-course theory that they themselves developed).

In another work, Sampson and Laub (2005b:41) did concede that they were not arguing

that grouping techniques such as Nagin's did not have any place in criminology, they were just

trying to warn criminologists about the mistake of the reification of groups as a result of the

application of the method, which Nagin (2005) himself also mentioned as a possible outcome.

They argued that this mistake is comparable to researchers making errors in interpreting









regression results in causal versus associational terms (for further discussions also see Sampson

& Laub, 2005b, 2005c; Maughan, 2005; Nagin & Tremblay, 2005c, 2005d; Raudenbush, 2005).

With these criticisms and limitations in mind, this chapter concludes with identifying the

software that will be used to estimate the group-based offending trajectories. The final

discussion is on the use of multinomial regression for assessing the relative strength of macro-

level versus micro-level predictors for predicting latent class membership.

Analytic Procedure

Trajectory estimation

PROC TRAJ is the name of the computer software that will be relied on in order to

estimate the offending trajectories for the 1942 and 1949 Racine continuous-resident samples by

cohort and data source including sex as covariate. This is a special procedure that is obtainable

as a macro-level function in SAS (Jones et al., 2001) and is made available for free use by the

National Consortium on Violence Research. Although a relatively new procedure, PROC TRAJ

has become a widely used method for investigating trajectories of development (Piquero, 2007)

and has been applied to a variety of different research questions including criminal career

desistance (Bushway et al., 2003), offense trajectories (Chung et al., 2002; Fergusson et al.,

2000; Piquero et al., 2003), sex differences in mental health (Barrett & White, 2002), age-crime

trajectories (McDermott & Nagin, 2002), and aggression toward mothers (Pagani et al., 2004).

While there have been two previous studies that have used PROC TRAJ with macro-level

data, they relied solely on aggregate crime rates in Chicago census tracts (Griffiths & Chavez,

2004) or neighborhood street segments in Seattle (Weisburd et al., 2004). Therefore, this is

possibly the first attempt to date (see Land, 2000 for a possible exception) that has utilized this

method to estimate offending trajectories and then estimate models incorporating macro-level

and micro-level characteristics for predicting group membership.









The initial trajectory models will be estimated separately for the 1942 and 1949 Racine

continuous-residents including sex as a covariate in the trajectory estimation (Models 1 and 2). 15

This will be followed by two models estimated similarly by cohort with sex as a covariate but

only using the random subsamples of interviewed Racine continuous-residents from the 1942

and 1949 cohorts (Models 3 and 4). These models will allow a preliminary examination of the

stability of offender groups across the cohorts and by data source. In other words, they offer a

type of sensitivity analysis for the stability of group identification that Nagin and Tremblay

(2005a) have argued has rarely been done in the trajectory literature (for exceptions see D'Unger

et al., 1998; Sampson et al., 2004).

Finally, because the purpose of this study is to examine the importance of macro-level

and micro-level factors in predicting longitudinal trajectories of offending, the 1942 and 1949

subsamples will be merged respectively with the sparse 1950 Census block data and an

additional model will be estimated for each cohort with sex as a covariate (Models 5 and 6).

These final trajectory models with the 1942 and the 1949 continuous-resident subsamples

(merged with the interview and block data) are the basis for assessing the possible independent

effects of macro-level and micro-level measures for predicting latent class membership (see also

Table 4-1 presented above that graphically displays the trajectory analyses that will be

estimated).

Multinomial logistic regression

After identifying the latent offender classes and establishing their respective participation

and frequency rates of offending, multinomial logistic regression models will be estimated in



15 Ideally, race-specific models would also be estimated also; however, due to the homogeneity of Racine in the
1940s and the small numbers of non-Whites in the continuous-resident samples, race will remain a covariate in the
multinomial regression models estimated in the following chapter.









order to evaluate the relationships among the macro-level and micro-level measures for

predicting offender group membership within the context of a multivariate model with attention

to the hypotheses presented at the end of chapter 3. This multivariate procedure will be used

because the multinomial logit model is able to deal with dependent variables in which there is no

hierarchical order to the responses and can handle both continuous and discrete independent

variables, similar to ordinary regression models (Agresti, 1990). While dependent variables that

are interval scales are completely appropriate for the common multiple regression technique

known as ordinary least squares (OLS) and although this same technique is often used in cases

where the dependent variable has ordinal outcomes that range from one to four (i.e., strongly

disagree, disagree, agree, and strongly agree), this latter application is typically not correct.

There are a variety of assumptions associated with ordinary least squares regression

including the assumptions of linearity, mean independence, homoscedasticity, uncorrelated

disturbance (error) terms, and the assumption of normality (see Allison, 1999); however, the

most important assumption that relates to the inappropriate decision to use OLS with categorical

dependent variables is the assumption of linearity. If OLS is applied to an ordinal outcome

variable, then there is the expectation that an increase or decrease of one unit on the scale is

interpreted as having the same influence no matter where you begin on the scale. Similarly,

there are cases when the dependent variable outcomes have no order at all in that they are

nominal scales (Allison, 1999).

Although it is theoretically possible to assume some inherent degree of order to outcomes

such as being a non-offender, adolescent-limited offender, and a chronic offender (i.e., 1, 2, 3)

these offending groups are based on nominal typologies that are theoretically and

methodologically driven. There is also the potential outcome of where more than these three









groups can be discovered and there is no logical way to order these groups, for instance early-

onset adolescent-limited offenders, low-low chronic offenders, late-onset offenders, etc.

Therefore, in this case the most appropriate modeling technique for the form of the dependent

variable (i.e., group membership) in this study is what is referred to as the multinomial logit

model.

Fundamentally, the multinomial logit model (MNLM) can be referred to as an extension of

the typical binary logit model (0, 1) where binary logits are simultaneously estimated for all of

the potential comparisons across the available categorical outcomes (offender groups in this

case) (Long, 1997:149). For example, its results are similar to estimating binary logit models

hypothetically comparing group 1 (non-offenders) against group 2 (adolescent-limited

offenders), group 3 (chronic offenders) against group 1 (non-offenders), and then finally group 2

(adolescent-limited offenders) against group 3 (chronic offenders). This obviously can become

much more work and present difficulty for explaining the results across these models as the

number of groups increase. The multinomial logit model can be expressed in terms of the

probability of observing outcome m given x (assuming y is the dependent variable with J

nominal outcomes, i.e., adolescent-limited offender, chronic offender, etc.). The MNLM as a

probability model takes the following form (Long, 1997:152):

Pr (y, = m x,) = exp (x,pj) / J/1 exp (x,p,).

Similarly, the MNLM can also be presented in the form of an odds model, where the odds

of outcome m versus outcome n given x is illustrated as follows (Long, 1997:154):


Qm I n (x,) = exp (x, 3m n),

and, after several derivations, the odds of an outcome m (being an adolescent-limited offender)

versus outcome n (being a non-offender) are expected to change by a factor of exp (Pk, m n









multiplied by 6), controlling for the effects of the other variables in the model (Long, 1997:169).

This odds model allows for a cleaner interpretation. For example, given this formula we could

now hypothetically say that the odds of being an adolescent-limited offender relative to being a

non-offender are 2.9 times greater for males, holding all other variables constant.

For all of the following multinomial analyses, latent class membership will be treated as

the outcome variable with dichotomous variables representing each offender group and the non-

offender group (group 1) will be left out as the reference group. The results of the multinomial

logit models that provide the best model fit for these data will be assessed using the pseudo-R2

statistic which is bounded between the values of zero and one, with values closer to one

indicating better model fit (Long, 1997). The multinomial analyses will be estimated separately

by cohort for the 1942 and 1949 Racine continuous-residents (total=558 individuals) who had all

of the available information (interview and block data) necessary to assess a theoretically

integrated model for explaining group-based offending patterns.

The progression of the models will be as follows: Model 1 will be a baseline macro-level

model where the effects of the social disorganization index (neighborhood SES and residential

mobility) for predicting group membership will be determined. Model 2 will be estimated

similarly to Model 1, although this model will only include the individual-level independent

variables (life transitions, race, and low SES household, and sex). Finally, Model 3 will present

the full model where both the macro-level and micro-level measures with be estimated

simultaneously, which would allow for an exploration into the possible mediation effects of

individual and family processes on the relationship between living in a social disorganized area

and group membership. The results of all of these analyses will be discussed and elaborated on

in Chapter 5.









Study Strengths

Notwithstanding the limitations noted previously in this chapter, there are several

important strengths of this study that are worth mentioning. First, this study answers the recent

calls among criminologists regarding moving toward theoretical integration of macro-level and

micro-level levels of analysis (Sampson et al., 1997; Messner et al., 1989), specifically focusing

on social disorganization theory (Shaw & McKay, 1942), Sampson and Laub's age-graded

theory of informal social control (1993), and Moffitt's (1993) offender taxonomy. Furthermore,

it approaches this process by employing a relatively new methodological tool that has gained a

substantial recognition in the field as an appropriate modeling technique for identifying distinct

behavioral patterns (for review, see Piquero, 2007).

Second, as an addition to the trajectory literature, this study provides an exploration into

the importance of sex for distinguishing group-based trajectories. Third, prior studies that have

estimated trajectories with the Shannon's (1988, 1991) Racine cohorts have relied on only

felonies and misdemeanors as the measure of offending (see D'Unger et al., 1998); therefore,

this research also includes status offenses in the hopes to better capture adolescent-limited

offending patterns. This study also improves by Land's (2000) study by estimating a series of

trajectory models as a "sensitivity" analysis to discover if similar offending trajectories are

produced by cohort and data source. This is especially of concern once the Racine continuous-

resident subsamples are merged with the interview level data, and subsequently, once the even

fewer of the individuals are merged with the limited amount of available census data

characterizing the community in which they were socialized. Finally, this study is perhaps the

first one (see Land, 2000 for a possible exception) to use trajectory analysis along with multi-

level regression models to predict the likelihood of exhibiting a distinct pattern of offending over

the life-course.









Table 4-1. Resident Samples by Cohort, Size, Data Sources, and Offender Groups.
Continuous-Resident Samples Sample Size (n)


1942 COHORT
Continuous-Residents
Continuous-Residents (w/ Interview Data)
Continuous-Residents (w/ Individual & Block Data)

1949 COHORT
Continuous-Residents
Continuous-Residents (w/ Interview Data)
Continuous-Residents (w/ Individual & Block Data)


633
333
235


1297
556
323


Table 4-2. Observational Periods of Data Sources by Cohort
Data Sources Years 1942 Cohort 1949 Cohort
Police Contact Data
1942 Cohort 1948-1967 6 to 25 years of age 6 to 25 years of age
1949 Cohort 1955-1974 6 to 25 years of age 6 to 25 years of age
Interview Data 1976 34 years old 27 years old
Census Data 1950 8 years old 1 year old


Table 4-3. Social Disorganization Index Factor Loadings by Cohort
1950 Census Measures 1942 1949
Owner-Property Value -.774 -.818
% Renter-Occupied Units .794 .794
% Overcrowded .612 .665

Eigenvalue 1.61 1.75
% Variance Explained 53.47 58.04
N 235 323









CHAPTER 5
RESULTS

Introduction

This chapter begins with a description of the summary statistics and mean differences

across the cohorts for all of the independent and dependent measures. Following this discussion

is a presentation of the group-based trajectories for all of the samples of the 1942 and 1949

Racine continuous residents (full sample, sample with interview data, and sample with interview

and block data). Included in the explanation of the group-based trajectory results is also a

discussion of the effect of sex (being male) as a covariate for increasing the likelihood of being

in an offender group as opposed to a non-offender group. This chapter concludes with an

ANOVA presenting the mean covariate levels across groups for all variables used in the

multivariate analysis by cohort. This is followed by an illustration of the multivariate results and

a description of the impact of macro-level and micro-level factors for distinguishing between

group-based trajectories. The multivariate analysis concludes with a presentation of the

theoretically integrated model for predicting group membership.

Descriptive Statistics

Table 5-1 below presents the descriptive statistics that correspond to all of the macro-level

and micro-level variables by cohort described previously that will be used in the mlogit analysis

that follows and the zero-order correlations among the continuous measures can be found in

Appendix A.

Independent Variables

As can be seen in Table 5-1 there was a variety of measures that significantly differed

between the 1942 and 1949 Racine continuous residents that precluded the possibility of

combining the cohorts in further analyses. Regarding the measures that were combined to create









the social disorganization index, there was evidence that while both cohorts were characterized

by comparable measures of mean owner-property value (1942=$9,374; 1949=$9,196) and

roughly equivalent percentages of renter-occupied units (1942=34.87%; 1949=34.99%), there

was an indication that the 1949 cohort (M=1.89; SD=3.83) had a significantly greater percent of

overcrowded units on average when compared with the 1942 cohort (M=1.48; SD=4.40).

Turning to the micro-level measures, there were fairly comparable sex percentages

between the cohorts with 47% of the 1942 sample and 52% of the 1949 sample being male.

While equivalent percentages of 1942 and 1949 Racine continuous residents reported having

been employed prior to age 25 (94%), there was a significant mean difference in reporting

whether or not they had also been married prior to age 25 with more of the 1942 residents

(1942=77%; 1949=66%) reporting having experienced marriage. Once these measures were

combined into the additive stakes in conformity index (range 0-2 life transitions), the 1942

cohort reported having significantly higher stakes in conformity (M=1.71; SD=0.45) when

compared with the 1949 cohort (M=1.60; SD=0.63). Comparatively, the 1949 cohort was

comprised of significantly more Blacks (11% versus 6%) and were more likely to have reported

having moved at some point in their juvenile years (61% versus 49%). In addition, the 1942

cohort reported having been socialized in significantly lower SES households (1942=3.61;

1949=3.25) and were less likely to have grown up in a natural two-parent household (84% versus

89%) when compared with the 1949 cohort. Although, the 1949 cohort reported slightly greater

numbers of children in their families, perceived more succeptible to differential association, and

had higher perceptions of formal social control in their neighborhoods during their juvenile

years, these differences were not significantly different.









Dependent Variable

The prevalence and incidence of police contacts for the 1942 and 1949 Racine continuous

residents can also be found in Table 5-1. While the cohorts had relatively comparable

prevalence estimates of having come into contact with the police with 37.9% of the 1942

residents and 41.2% of the 1949 residents having an officially recorded police contact, the 1949

cohort had significantly greater mean numbers of police contacts than did the 1942 cohort. On

average, the 1949 residents had 2.33 police contacts (SD=6.71) whereas the 1942 residents had a

1.29 average number of police contacts (SD=2.63). Additionally, the 1949 cohort had

accumulated a significantly greater total number of police contacts by age 25 (754 police

contacts) when compared with the 302 accumulated police contacts for the 1942 cohort.

Trajectory Analysis

In order to aid in the substantive interpretation of the nature of the offending trajectories

identified and estimated for the 1942 and 1949 Racine continuous residents, the predicted and

average curves for each of the trajectories implied by the model are plotted in their

corresponding figures presented below. The trajectory results are described in order of

progression and by cohort, thus the first model is that which was produced using the full sample

of 1942 continuous residents, the second model displays the results from a model relying on data

for the 1942 continuous interviewed residents, and the third model illustrates the trajectories of

the 1942 continuous interviewed residents in ith block data. This same process is then repeated

for the 1949 Racine continuous residents (total=six models; 1942=3 models; 1949=3 models).

Each model is discussed in terms of the various model selection criteria (Nagin, 2005), or more

specifically: (1) the increase in the BIC value (as previously described in Chapter 4); (2) the

posterior probabilities for group assignment; and (3) the model's ability to demonstrate a









theoretically distinct and substantively meaningful number of groups. Finally, the models are

described with attention to examining the effect of sex (being male) on group membership.

The model estimations began by fitting the 1942 Racine continuous residents' police

contact data to a two group model and then comparing the BIC values from a three group model

to the two group model. The three-group model proved a better fit and the analysis continued by

adding groups and observing the improvement in the BIC values. The trajectory procedure

ultimately arrived at a four-group solution because despite several attempts the model was

unable to converge with a five-group model (see Table 5-2).

The average and the predicted curves are displayed below in Figure 5-1 and these plots

illustrate that the predicted (smoothed) curves closely followed the "true" mean offending

curves. The first group (G1) represents the overwhelming majority of the sample or 73.1% of all

of the 1942 Racine continuous residents. Individuals in this trajectory group exhibit virtually no

offending (as measured by official data) across the entire twenty years of observation (i.e., ages

6-25). The second group includes 8.8% of the sample and individuals in this group appear to

demonstrate the type of offending that would characterize adolescent-limited offenders as

described by Moffitt (1993) with their offending beginning in adolescence, reaching a peak by

age 16 before declining and desisting prior to young adulthood. No individual in this group

accumulated more than 10 police contacts by age 25 and the average number of police contacts

for this group was 3.91 (SD=1.98). The third group, representing 13% of the sample, exhibit a

low-level of chronic offending beginning in early adolescence (age 12) and continue to illustrate

a non-zero rate of offending into adulthood and are still offending at age 25; however, no

individual in this group accumulated more than seven police contacts during the twenty year

period. Group 4 (5.1%) presents a sharp contrast to the other groups in that individuals in this









group begin offending relatively early in life (some even as early as age 7) and they maintain a

rather high amount of police contacts throughout the remainder of the observation period but

does decline in young adulthood. Although, it does appear that these individuals peak in late

adolescence they still exhibit high levels of offending in early adulthood. No individual in this

group accumulated less than 5 police contacts and a handful of offenders in this group had even

amassed more than 20 police contacts before reaching age 26. Overall, this small group of

offenders was responsible for 352 total police contacts between the ages of 6 and 25.

When turning to an examination of the mean posterior probabilities for group assignment

(see Table 5-3), the mean assignment probabilities for each group are noticeably high suggesting

that there was little ambiguity in the model when determining group assignments. For example,

for G1, the mean posterior probability was .91, and the next highest was G3 with a mean

probability of .07 of being assigned to G1; for group 2, the mean posterior probability was .84

(median=.87); for group 3 it was .85 (median=.93); and for group 4 it was .91 (median=.99).1

The results of this trajectory analysis provided evidence that sex was a highly significant

covariate for determining group assignment (see Table 5-4). With G1 serving as the reference

group (i.e., the non-offender group), being male significantly increased an individual's likelihood

for being assigned to any of the three offender groups (i.e., adolescent-limited/peaked, low-level

chronic, and high-level chronic).

Table 5-5 illustrates the BIC criteria for assisting in model selection for the trajectory

analysis using the reduced random sample of interviewed Racine continuous residents and the

average and predicted group-based offending trajectories are displayed in Figure 5-2. These



1 All average posterior probabilities are above the standard cut-off of .70 (Nagin, 2005).While there is always the
possibility of mis-classification of individuals to particular trajectories none of the posterior probabilities in these
data were below .80 thus indicating a high degree of discrimination between groups.









trajectories were similar to those produced in the previous analysis, although only three distinct

groups were identified in this particular analysis.2

The first group represented the overwhelming majority of the Racine interviewed residents

with 86.2% of the sample being classified as non-offenders with virtually no offending during

the observation period. Comparatively, G2 represented 7.8% of the sample and individuals in

this group appeared to display an offending pattern consistent with Moffitt's (1993) adolescent-

limited offender type wherein these individuals began offending in adolescence before peaking

in mid-adolescence and showed evidence of desistance near age 19 with no individual in this

group having any officially recorded police contacts after age 20. Individuals in this group had

an average of 4.50 (SD=2.49) police contacts and no individual in this group accumulated more

than 12 police contacts prior to reaching age 26. The third group, representing 6% of the sample,

is recognizable by their early onset of offending (prior to age 10) and relatively high levels of

offending throughout the rest of the observation period. Although there is evidence that this

group's rate of offending peaked in late adolescence/early adulthood they still maintained

relatively high levels of offending through age 25. On average, individuals in this group had

7.85 (SD=3.42) police contacts and a handful of these offenders accumulated more than 10

police contacts between the ages of 6 and 25. Furthermore, this small group of chronic offenders

was responsible for a total of 157 police contacts.

Similar to the results of the previous trajectory analysis, the posterior probabilities for

group assignment were remarkably high providing evidence that the individual trajectories in

these data could be assigned to a group-based trajectory with a relatively high degree of



2 A four group model was not able to fit these data, thus the low-level chronic offending group was not identified in
this stage of the analysis.









probability (see Table 5-6). Only in G2 did the mean posterior probabilities fall slightly below

.90 but still remained well above the .70 cut-off (Nagin, 2005). Consistent with the previous

model, the results in Table 5-7 indicated that being male significantly increased an individual's

likelihood of being assigned to either offender group (adolescent-limited/peaked or chronic)

when compared with the non-offender group.

Table 5-8 presents the BIC model selection criteria for the trajectory analysis using the

data from the 1942 Racine continuous interviewed residents n i/h block data that provides the

trajectories and groups that will be of interest in the 1942 ANOVA and mlogit analysis that

follows. A three-group model once again provided the best fit for these data and the average and

predicted offending curves are plotted below in Figure 5-3.

Gl, representing 70.2% of the sample, are virtually a non-offending group with little to no

police contacts during the ages of 6 to 25. The second group, representing 20% of the sample is

similar to the adolescent-limited offender type proposed by Moffitt (1993) with the individuals in

this group beginning to offend in adolescence before peaking around the age of 16 before

declining and showing evidence of having desisted by 21. No individual in this group had

accumulated more than 5 total police contacts during the twenty years of follow-up. Also

consistent with the earlier analyses, the third group exhibited an early onset of offending (prior to

age 10), relatively high levels of offending throughout adolescence before reaching a peak at age

16; however, these individuals continued to maintain comparatively high rates of offending into

adulthood and through the age of 25, although the trajectory was declining. No individual in this

group had accumulated less than 4 police contacts and a handful of these individuals had as

many as more than ten police contacts prior to reaching age 26 (M=8.09; SD=3.30). This small

number of Racine residents amassed 186 police contacts between the ages of 6 and 25.









Table 5-9 presents the posterior probabilities for group assignment and consistent with the

two prior trajectory analyses using residents from the 1942 cohort, there was relatively little

ambiguity when the model was determining group assignment with mean posterior probabilities

as high as .91 (median=.99) for G3 and still comparatively high for G1 (M=.89; median=.97) and

G2 (M=.89; median=.97). Additionally, the results indicated that sex (being male) significantly

increased an individual's likelihood of belonging to either an adolescent-limited/peaked or a

chronic offending group when compared with membership in a non-offending group (see Table

5-10).

The next set of trajectory analyses involves data from the 1949 Racine continuous

residents following the same progression of models that were estimated for the 1942 cohort: (1)

continuous residents; (2) continuous interviewed residents; and (3) continuous interviewed

residents in ith block data. The results of the BIC model selection criteria for the 1949

continuous resident sample is presented below in Table 5-11. This was the only sample in which

a five-group model provided the best fit for these data, but this is not surprising given that prior

researchers have shown evidence that the number of groups typically increase with sample size

(D'Unger et al., 1998; Nagin & Tremblay, 2001; Sampson et al., 2004) and this was in fact the

largest sample in this study (N=1,297).

The average and predicted curves for the 1949 continuous resident trajectory analysis can

be found in Figure 5-4 and consistent with the analyses produced using data from the 1942

continuous residents, the predicted offending curves tended to closely follow the "true"

offending curves. The first group, representing 57.1% of the sample, is purely a non-offender

group as none of these individuals had any recorded police contacts throughout the entire twenty

years of observation. The second group, representing 32.7% of the sample, exhibited very low









rates of offending over the entire observation period. There was a tendency for these individuals

to peak at age 16 and then decline, but they still showed non-zero rates of offending through age

25. On average, individuals in this group had 1.98 (SD=1.24) police contacts and no individual

in this group had more than 7 police contacts. G3, representing 2.6% of the sample, is an

atypical group that is not identified in any of the other analyses presented in this study although

this offender group has been the point of discussion in prior research (see D'Unger et al., 1998;

Sampson and Laub, 1993). Individuals in this group are unique in that they had virtually no

officially recorded police contacts until they reached adulthood (age 18) and their rates of

offending increased rather noticeably in young adulthood. No individual in this group had less

than 3 police contacts (M=6.53; SD=3.89) and a handful of these offenders had actually

accumulated more than 10 police contacts despite the fact that their offending did not begin until

young adulthood.

The fourth group, representing 5.7% of the sample, can be best described as an adolescent-

peaked group (see D'Unger et al., 1998; Land, 2000) in that these individuals do not begin

offending until early adolescence but offend at relatively high rates during mid-adolescence

before peaking at age 15 and then virtually have desisted by age 23. No individual in this group

had less than 6 police contacts and more than half of the individuals in this group had

accumulated more than 10 police contacts (the majority of which were during mid-adolescence).

G5, representing the smallest percentage of the sample (1.9%), exhibited a high-rate chronic

pattern of offending that begins early on in life (prior to age 8) and their offending drastically

increases well into adolescence before reaching a peak in late adolescence. These individuals

still display high levels of offending throughout the remainder of the observation period or age

25. Individuals in this group averaged an alarming 32.48 (SD=10.30) police contacts and the









police contacts of this group ranged from a low of 22 to a high of 63 police contacts. This small

group of offenders managed to accumulate 812 police contacts during the ages of 6 to 25!

The mean posterior probabilities for group assignment indicated that, like the trajectory

analyses using the 1942 Racine continuous resident data, this model demonstrated the ability to

determine group membership with a high degree of probability (see Table 5-12). G1 had the

lowest mean probability (M=.84; median=.93), yet this was still way above the standard cutoff of

.70 (Nagin, 2005). The mean posterior probabilities were remarkably high for the low (G2) and

high-rate (G5) chronic offending groups indicating that these individuals' trajectories faced little

risk for mis-classification. According to the coefficients and their corresponding p-values

produced for this sex covariate, there was evidence that for all of the offender groups except the

late-onset offenders (i.e., adolescent-limited/peaked, low and high-rate chronic offenders) being

male significantly increased the likelihood of group membership into an offender group

compared with a non-offender group (see Table 5-13).

Table 5-14 presents the BIC model selection criteria for the 1949 continuous interviewed

residents and similar to the interviewed 1942 sample only three trajectory groups were produced.

This analysis failed to detect a late-onset or a low-rate chronic group as was the case in the prior

1949 continuous resident analysis. The average and predicted curves, as plotted in Figure 5-5,

indicated that the following three trajectory groups could be identified: (1) non-offenders; (2)

adolescent-peaked offenders; and (3) high-rate chronic offenders. G1 (75.1%) represented the

overwhelming majority of the sample and exhibiting practically no offending across the twenty

year time period. Comparatively, the second group (20.5%) did not begin offending until early

adolescence before reaching a peak in mid-adolescence and then desists by their early 20s.

Individuals in this group recorded an average of 4.32 (SD=2.81) police contacts with a range of 2









to 14 police contacts. G3, representing 4.3% of the sample, exhibited an early-onset of offending

(prior to age 8) and increased their offending patterns throughout mid and late-adolescence

before peaking at age 18. Nevertheless, these individuals still offended at relatively high rates by

age 25, although showing evidence of desistance in young adulthood. All of the individuals in

this group had accumulated at least 10 police contacts between the ages of 6 to 25 (M=23.21;

SD=12.95) and this small number of offenders were responsible for 557 police contacts prior to

reaching age 26.

Table 5-15 provides the mean posterior probabilities for group assignment and these

results show a strong ability for discriminating across trajectory groups with no mean posterior

probability less than .92 (G2). In fact, the mean posterior probabilities are remarkably high for

Gl (M=.97; median=.99) and G3 (M=.99; median=.99). In addition, consistent with all of the

prior trajectory analyses, being male significantly increased the likelihood for an individual

being assigned to either an adolescent-limited/peaked or high-rate chronic offending trajectory

when compared with assignment to a non-offending group (see Table 5-16).

Table 5-17 presents the BIC model selection criteria for the final trajectory analysis in this

study using the data from the 1949 Racine continuous interviewed residents I i// block data that

provides the trajectories and groups that will be of interest in the 1949 ANOVA and mlogit

analysis that follows. Similar to the prior 1949 continuous interviewed resident analysis and

consistent with the results of the comparable 1942 continuous interviewed residents I i// block

data analysis, the procedure indicated that a three-group model provided the best fit for these

data.

The average and predicted curves are displayed in Figure 5-6 below. The first group,

representing the majority of the sample (75.2%), exhibits almost no offending throughout the









entire twenty year observation period. In contrast, G2 (20.4%) resembles the adolescent-

limited/peaked group identified in the two previous trajectory analyses using the 1949 cohort

data in that individuals in this group did not begin offending until adolescence but did not reach a

peak in offending until 17 and then declined until appearing to desist by their early 20s.

Furthermore, these individuals averaged approximately 5 (SD=3.45) police contacts and they had

accumulated at least two police contacts between the ages of 6 and 25 with some of these group

members having amassed as high as 14 police contacts. G3 (3.7%) exhibits the high-rate chronic

offending patterns discussed previously in that these individuals begin offending early (prior to

age 8) and increase dramatically during adolescence before reaching a peak in young adulthood

(age 18). Yet, this group displays elevated levels of offending by age 25, but showed evidence

of desistance. No individual in this group had less than 14 police contacts and the average

number of police contacts for this group was 30.25 (SD=15.89). Overall, these handful of

offenders had accumulated 363 police contacts between the ages of 6 and 25.

Table 5-18 displays the mean posterior probabilities for group assignment and Table 5-19

illustrates the effect of sex (being male) on group membership. Like all of the previous analyses,

the mean posterior probabilities were noticeably high, ranging from .89 (G2) to .98 (G1) and as

high as .99 for G3. This was an indication that there was little ambiguity in assigning individual

patterns of offending to a group-based trajectory. Also, once again the results showed evidence

that being male significantly increased an individual's likelihood of being assigned to either an

adolescent-limited/peaked or high-rate chronic offending group compared with a non-offending

group.

With all of the trajectory analyses concluded, it is important to discuss these models within

the context of the relative stability/instability of identifying offender groups across the various









data sources (continuous residents, continuous interviewed residents, and continuous interviewed

residents in ilh block data). A graphical display of the offender groups across the data source and

by cohort are presented in Table 5-20. As can be seen there appears to be more similarities than

differences in the offense trajectories across the data sources and between the 1942 and 1949

cohorts. First, in both of the original 1942 and 1949 continuous resident samples with the largest

sample sizes (N=633 and N=1,297) produced the greatest number of groups. The four remaining

trajectory analyses with reduced samples were only able to identify three groups, which is

consistent with prior research that has indicated that the number of groups increases with the

sample size (D'Unger et al., 1998; Nagin & Tremblay, 2001; Sampson et al., 2004). Second, in

all of the analyses the non-offending group (G1) was always the most prevalent group in the

sample (ranging from 57.1% to 86.2%) and similarly, the high-rate chronic offending group was

always the smallest percentage of the offender groups (ranging from 1.9% to 9.8%). Third, the

1942 and 1949 original continuous resident samples also both identified a group that is atypical

to what Moffitt's (1993) taxonomy would expect or who have been classified as low-level

chronic offenders who offend early and maintain low non-zero rates throughout the twenty year

period of observation; however, this type of offender group has been observed elsewhere in

previous trajectory analyses (see D'Unger et al., 1998; 2002).

Regarding the differences, first, the largest sample in this study (1949 continuous resident

sample of 1,297 individuals) actually identified an additional atypical trajectory group or what

others have referred to as a late-onset group (see D'Unger et al., 1998; Sampson & Laub, 1993)

who do not begin offending until late adolescence/early adulthood, but offend at relatively high

rates in adulthood. Second, while all of the analyses displayed an adolescent-limited/peaked

group of offenders who was the largest group of offenders in four of the six analyses, in the 1949









trajectory analyses this group tended to resemble an adolescent-peaked group (D'Unger et al.,

1998; Land, 2000) in that they still exhibited low rates of offending into early adulthood before

desisting in the early 20s. The discussion in the next section now moves toward examining the

results from the ANOVAs in order to further investigate what macro-level and micro-level

factors are significant for distinguishing between offender groups.

Analysis of Variance

The results from the analysis-of-variance tests where the macro-level and micro-level

risk/protective factors are compared across groups and by cohort are presented in Table 5-21.

The group means of offending (number of police contacts) are also displayed in Table 5-21. The

trajectory models displayed in Figure 5-3 (1942 cohort) and Figure 5-6 (1949 cohort) are those

that all of the remaining analyses are based on. The first most obvious findings is that relatively

all of the risk factors are concentrated among the offender groups and the protective factors are

primarily found in the non-offender group (Gl). More specifically, the risk factors tend to be

heightened in the adolescent-limited/peaked groups (G2) and then further elevated in the high-

rate chronic groups (G3) (G3 for both the 1942 and 1949 cohorts will be referred to

interchangeably as high-rate or chronic offenders hereon). For the 1942 and 1949 cohorts, social

disorganization, sex (being male), race (being Black), living a low SES household, living with

natural parents, and perceived formal social control are distinguishing features of group

membership. Additionally, whether an individual had moved during their juvenile years was a

distinguishing feature for 1942 continuous residents, whereas stakes in conformity3 and family


3 In order to further explore the possible influence of marriage and employment on group-based offending
trajectories, the mean age at time of marriage and first full-time job were computed and compared across groups and
by cohort. Regarding the 1942 cohort, 83.0% of the non-offenders (G1), 70.2% of the adolescent-limited/peaked
(G2), and 86.9% of the chronic offenders (G3) reported being married by age 25 and of those that reported being
married, the mean age of marriage was 20.8 (G1), 21.6 (G2), and 20.7 (G3). In contrast, 97.6% of Gl, all of G2,
and 95.6% reported having their first full-time job by age 25 and the mean age of first full-time employment was
18.4 (G1), 18.4 (G2), and 18.0 (G3). When turning to the 1949 cohort, 72.7% of the non-offenders (G1), 63.6% of









structure (number of children in family) were also distinguishing features for group membership

among the 1949 continuous residents.

While the preliminary results from the f-test in the ANOVAs revealed the concentration of

risk factors among the offender groups, it was necessary to conduct Tukey's B post hoc tests in

order to determine which (if any) of the groups significantly differed from one another.

Concerning the 1942 cohort, individuals who were in either offender group lived in significantly

more socially disorganized communities and both the adolescent-limited/peaked groups (G2) and

the high-rate chronic groups (G3) had significantly more males and G3 was characterized by a

significantly greater number of Blacks. G1 also had a significantly greater number of individuals

reporting having grown up in a two natural parent household when compared with both G2 and

G3. In addition, all of the groups' mean number of police contacts significantly differed, or in

other words G3 had significantly more police contacts than G1 and G2, and G2 had significantly

more police contacts than G1.

When turning to the Tukey's B post hoc tests results for the 1949 cohort, while individuals

in G3 and G2 lived in more socially disorganized areas than G1 the groups means did not

significantly differ. Comparatively, there was a host of individual level factors that significantly

varied particularly when comparing G3 and G2 to G1. The offender groups had significantly

more males and Blacks in addition to having a significantly greater number of children in the

household and had significantly higher perceptions of formal social control (police presence)

when compared with G1 (although G2 and G3 did not significantly differ from each other). The

the adolescent-limited/peaked (G2), and only 33.3% of the chronic offenders (G3) experienced marriage by age 25
and of those that reported being married, the mean age of marriage was 21.0 (G1), 21.1 (G2), and 21.3 (G3).
Comparatively, 96.7% of G1, 95.5% of G2, and 91.6% reported having their first full-time job by age 25 and the
mean age of first full-time employment was 19.0 (G1), 18.8 (G2), 17.3 (G3). Thus it appeared that typically fewer
adolescent-limited/peaked offenders and fewer chronic offenders reported having experienced marriage and/or
employment by age 25, yet there was not consistent evidence that the mean age of having experienced these life
transitions varied greatly across groups and by cohort.









high-rate chronic offenders (G3) also had significantly lower household SES and less stakes in

conformity than both G1 and G2. Furthermore, G1 was comprised of significantly more

individuals that reported coming from a two natural parent household when compared with G2

and G3. Overall, similar to the 1942 ANOVA results, G3 had significantly more police contacts

than G1 and G2, and G2 had significantly more police contacts than G1. It is now important to

investigate if these same risk/protective factors that were shown to significantly distinguish

group membership in these bivariate analyses are also significant for predicting group

membership when estimated simultaneously, which now becomes the focus of the next section.

Multivariate Analysis

The following set of multinomial regression analyses are estimated in order to compare the

ability of the macro-level and micro-level factors to differentiate among the offender groups.

Each cohort-specific analysis (beginning with the 1942 cohort) will be divided into the following

three separate stages: (1) macro-level level; (2) micro-level level; and (3) the integrated model.

The first two regression analyses will assist in isolating the relative strength of prediction for the

different levels of analysis (macro-level and micro-level) and the latter will allow for an

assessment of the proposed integrated theoretical model for differentiating the non-offenders

from the adolescent-limited/peaked and high-rate chronic offenders.4

According to the results from the first macro-level-only model, using data from the 1942

Racine cohort (see Table 5-22) there was support for the effect of social disorganization in

differentiating the offender groups. Specifically, residing in a socially disorganized area nearly

doubled the odds of an individual being assigned to an adolescent-limited/peaked group and





4 Models were estimated using clustered robust standard errors in Stata 9.2.









more than doubled the odds of an individual being assigned to a high-rate chronic offending

group relative to a non-offender group.5

Comparatively, Table 5-23 presents the results from the micro-level-only model. First, the

findings showed a robust effect for sex on the odds of being assigned to either type of offender

group with being male nearly tripling the odds for being assigned to an adolescent-

limited/peaked offender group and increasing the odds by more than six times for an individual

being assigned to a high-rate offender group. Having been socialized in a natural two parent

household yielded a significant protective effect on the odds of being assigned to adolescent-

limited/peaked offender group by reducing the odds by 38%, and it produced a similar reduction

in the odds (41%) of being assigned to a high-rate chronic offender group relative to the non-

offender group. In addition, having moved during the juvenile years more than doubled the odds

of being assigned to a high-rate chronic offending trajectory. Although the majority of the other

individual level risk/protective factors were in the expected direction insofar as increasing or

decreasing the odds for being in a particular group, being Black, living in a low SES household,

stakes in conformity, number of children in the household, perceived delinquent peer influence,

and perceptions of formal social control did not achieve statistical significance for distinguishing

across trajectories.

Table 5-24 displays the multivariate results for the integrated theoretical model for

predicting offender group membership. The findings suggest that once the macro-level and

micro-level factors are estimated simultaneously, social disorganization is left only marginally

5 Percent non-White was originally included in the analyses as a control variable, but it failed to exert any significant
effects for predicting offender group membership. Models were also estimated using backwards deletion and the
results confirmed that percent non-White did not influence trajectory group membership or increase the model fit, so
it was dropped and not included in the analyses presented here. According to Sampson (1997) there has been little
support for this construct in relation to offending when included with neighborhood poverty and residential
instability and given the relatively high degree of homogeneity in these data, it is not surprising that this variable
was non-significant.









significant (p<.10) for increasing an individual's odds for being assigned to an adolescent-

limited/peaked offender group in relation to a non-offender group. In contrast, the effect of

social disorganization on the likelihood of being assigned to a chronic offender group was

completely mediated by the addition of the micro-level factors. Additionally, upon comparing

the pseudo R-square statistics, it appeared that the addition of social disorganization did

relatively little for providing a better model fit (McFadden's increased .23 to .25; Nagelkerke

increased .39 to .41). Overall, the same three micro-level variables that achieved statistical

significance in the micro-level only model discussed above (sex and natural two parent

household for both offender groups and having moved for the high-rate chronic only) still

remained statistically significant distinguishing factors for group assignment.

The results of the multivariate analysis for distinguishing across trajectories based on data

from the 1949 cohort are discussed below. Similar to the prior mlogit models presented above,

the first model illustrates the effect of social disorganization alone, the second demonstrates the

effect of only the micro-level factors, and the final model displays the findings from the

theoretically integrated model for distinguishing across offender groups.

Consistent with the results of the macro-level-only model for the 1942 cohort, social

disorganization statistically increased the odds for an individual being assigned to either an

adolescent-limited/peaked offender group or a high-rate chronic offender group in relation to a

non-offender group (albeit only marginally significant for the high-rate chronic offender group,

p=.07) for the 1949 cohort. More specifically, living in a socially disorganized area

approximately increased an individual's odds one and a half times for being assigned to either

offender group rather than to a non-offender group (see Table 5-25).









Table 5-26 reveals the results of the micro-level factors for predicting group membership

for the 1949 cohort. There were several similarities as well as differences for these individual

level factors for distinguishing across trajectories in the 1949 cohort analysis when compared

with the 1942 mlogit analysis discussed previously. The key similarity is that a rather robust

effect for sex was again found with being male more than doubling an individual's odds of being

assigned to an adolescent-limited/peaked group and increasing an individual's odds nearly two

and a half times for being assigned to a high-rate chronic offending group in relation to a non-

offender group.

Additionally, both family-level factors (family stability and family structure) were found to

be statistically significant distinguishing features of group assignment to an adolescent-

limited/peaked offender group with residing with natural parents resulting in a 46% reduction in

the odds and each additional child increasing an individual's odds 1.36 times for being assigned

to an adolescent-limited/peaked offender group. Being Black also marginally increased (p=.06)

an individual's odds of being assigned to an adolescent-limited/peaked offender group.

Comparatively, although these two family factors nearly achieved statistical significance for

increasing the odds of being assigned to a high-rate offender group [family stability (p=. 14);

family structure (p=. 11)], residing in a low SES household was the more important

distinguishing feature of group assignment to this particular group-based trajectory. An

individual's odds of being assigned to a high-rate chronic offender group in relation to a non-

offender group increased more than two and a half times if they reported having been socialized

in a low SES household. The remaining micro-level factors (stakes in conformity, differential

association, perceptions of formal social control, and having moved during the juvenile years)

were not statistically significant for distinguishing across trajectories.









The findings from the final mlogit model indicating the results of the theoretically

integrated model for predicting group membership using the 1949 cohort (see Table 5-27)

revealed similar results as to what was produced for the 1942 cohort. While social

disorganization remained marginally significant for the adolescent-limited/peaked offender

group in the 1942 analysis (p=.06), once the macro-level and micro-level factors were estimated

simultaneously for the 1949 cohort, the effect of social disorganization for distinguishing across

trajectories was no longer statistically significant for either offender group (although the odds

were still in the expected direction). Comparatively, the micro-level factors that were

statistically significant for predicting group membership in the previous micro-level-only model

still maintained their significance in the integrated model. More specifically, being male more

than doubled an individual's odds of being assigned to either an adolescent-limited/peaked or a

high-rate chronic offender group and family stability and structure were significant determinants

of adolescent-limited/peaked offender group membership. Additionally, having resided in a low

SES household more than doubled the odds of being assigned to a high-rate offender group.

Overall, upon inspecting the change in the pseudo-R-squares as indicators of model fit, there was

little gained from using a theoretically integrated model when compared with the micro-level-

only model (McFadden's increased from .19 to .20; Nagelkerke's increased from .31 to .32).

Conclusions

The results from the series of analytical procedures yielded several substantively important

findings. First, the six trajectory analyses (three for each cohort) provided evidence that non-

offender, adolescent-limited/peaked offender, and high-rate chronic offender groups were

recognized across all data sources (Racine continuous residents, Racine continuous interviewed

residents, and Racine continuous interviewed residents n i/h block data) and in both the 1942 and

1949 cohorts. Although the group percentages varied, there was support for hypothesis #1 in









that the semiparametric Poisson-based regression analyses produced these three distinct classes

of offenders and that (1) the non-offenders were the most prevalent of the groups; (2) the

adolescent-limited/peaked offenders were the most prevalent of the offending groups; and (3) the

high-rate offenders were the least prevalent group of offenders but were characterized by a

higher rate of mean offending across the observation period when compared with the adolescent-

limited/peaked offender group. However, there were a few groups that were recognized in the

larger sample trajectory analyses (low-level chronic and late-onsetters) who are not consistent

with Moffitt's taxonomy but have been discussed elsewhere (see D'Unger et al., 1998; Piquero,

2007; Sampson & Laub, 1993).

Support was also shown for the second hypothesis with sex being a significant covariate

for distinguishing group-based trajectories. Specifically, in all of the six trajectory analyses,

being male significantly increased the probability that an individual would be classified as either

an adolescent-limited/peaked or chronic offender with the non-offender group serving as the

reference group. The only exception was in the 1949 continuous resident sample (N=1,297).

While there were more males in the late-onset offender group for the cohort, being male did not

significantly increase the probability of assignment to this particular group-based trajectory.

Concerning hypothesis #3 that significant life transitions including marriage and

employment can affect desistance from crime, there was limited support for Sampson and Laub's

age-graded theory of informal social control. There was descriptive evidence at the bivariate

level suggesting that individuals in either offender group (adolescent-limited/peaked or high-rate

chronic offender group) scored lower on the stakes in conformity index when compared with

those individuals assigned to the non-offender group. The results from the ANOVA with the

1949 cohort also indicated that the stakes in conformity index was statistically associated with









group membership to a high-rate chronic offending trajectory with individuals in this group

having significantly lower mean scores on the stakes in conformity index when compared with

those in the adolescent-limited/peaked group and individuals in the non-offender group.

Qualified support was also shown for hypothesis #4 regarding the expectation that social

disorganization would be a significant distinguishing feature for group assignment. Being from a

socially disorganized area significantly increased the odds that an individual would be assigned

to either an adolescent-limited/peaked offender group or a high-rate offender group in relation to

a non-offender group in the 1942 and 1949 cohort mlogit analyses. The effect, however, was

only marginally significant (p=.07) for distinguishing high-rate chronic offenders in the 1949

analysis. The results indicated that social disorganization did not significantly distinguish the

adolescent-limited/peaked offenders from the chronic offenders.

Finally, there was little support provided for hypothesis #5 or the expectation that macro-

level and micro-level effects on the participation and frequency of offending will still be

observed net of the effects of each other. Although the results did indicate that once the multi-

level effects were considered simultaneously, the effect of social disorganization still maintained

a marginally significant effect for distinguishing the adolescent-limited/peaked offenders from

the non-offenders in theoretically integrated mlogit analysis of the 1942 cohort, the majority of

the effect appeared to be indirect. While the direction of the odds were in the expected direction

for both the 1942 and 1949 mlogits (with being from a socially disorganized area increasing the

odds of assignment to an offender group), an examination of the R-squares (McFadden's and

Nagelkerke's) showed that the addition of the social disorganization index contributed relatively

little for improving model fit. Once again, social disorganization was unable to significantly

differentiate the adolescent-limited/peaked group from the chronic offending group.









Chapter 6 will provide a further discussion of the results from the series of analytical

procedures and elaborate on the possible implications of the research findings. The chapter will

also highlight the limitations of the current study, offer suggestions as to how future research can

improve upon and extend the examination of integrated theoretical models, and close with a few

final thoughts.










Table 5-1. Descriptive Statistics for the Macro-level and Micro-level Measures by Cohort
1942 1949
Variables
Mean (SD) Mean (SD)
Independent Variables
Macro-level
Social Disorganization Index 0.00 (1.00) 0.00 (1.00)
Owner-Property Value 9374.30 (2895.29) 9196.40 (2951.55)
% Renter-Occupied Units 34.87 (20.83) 34.99 (21.43)
% Overcrowded 1.48 (3.83)* 1.89 (4.40)*

Micro-level
Male 0.47 (0.50) 0.52 (0.50)
Life Transitions
Stakes in Conformity Index 1.71 (0.45)* 1.60 (0.53)*
Employment 0.94 (0.22) 0.94 (0.23)
Marriage 0.77 (0.42)* 0.66 (0.47)*
Black 0.06 (0.24)* 0.11 (0.31)*
Low SES Household 3.61 (2.08)* 3.25 (1.91)*
Family Stability (Natural Parents) 0.84 (0.37)* 0.89 (0.31)*
Family Structure (No. of Children) 3.74 (2.21) 4.02 (2.17)
Differential association 0.22 (0.42) 0.25 (0.43)
Formal Social Control 1.24 (0.96) 1.32 (1.03)
Moved 0.61 (.0.49)* 0.49 (0.51)*

Police Contacts 1.29 (2.63)* 2.33 (6.71)*
Prevalence 37.9% 41.2%
Incidence 302* 754*


Note. The asterisk signs denotes significant mean differences between the 1942 and 1949 cohorts (p<.05). Listed
below the social disorganization and stakes in conformity indexes are the values of the measures that were used to
create each index prior to the data manipulations (principal components analysis for the social disorganization index
and the additive index for experiencing either none, 1, or 2 of the life transitions).


Table 5-2. Tabulated Bayesian Information Criteria (BIC) for 1942 Racine Continuous
Residents (N=633)
Optimal # of Groups BIC
Cubic ZIP Model
2 -2453.58
3 -2437.23
4 -2399.77
5 False Convergence









Table 5-3. Mean (Median) Posterior Probabilities for Group Assignments: 1942 Racine
Continuous Residents (N=633)
Group Prob Prob Prob Prob 25th 75th
(G1) (G2) (G3) (G4) Percentile Percentile
1 (n=463) .91(.99) .02(.00) .07(.01) .00(.00) .89 .99
2 (n=56) .02(.00) .84(.87) .12(.08) .03(.01) .71 .98
3 (n=82) .11(.03) .04(.00) .85(.93) .01(.00) .75 .98
4 (n=32) .00(.00) .04(.00) .05(.00) .91(.99) .91 .99
Note: Minimum and Maximum values presented for the group/prob(group) combination.


Table 5-4. Coefficient Estimates for Effect of Sex (being Male) on Trajectory Group
Membership: 1942 Racine Continuous Residents (N=633)
Groups coefficient SE p-value
Cubic ZIP Model
2 2.861 0.566 <0.001***
3 2.721 0.508 <0.001***
4 2.642 0.642 <0.001***
*p<.05 **p<.01 ***p<.001
a Group 1 (the lowest rate group) is the reference category of comparison.


Table 5-5. Tabulated Bayesian Information Criteria (BIC) for 1942 Racine Continuous
Interviewed Residents (N=333)
Optimal # of Groups BIC
Cubic ZIP Model
2 -1164.02
3 -1148.61
4 False Convergence









Table 5-6. Mean (Median) Posterior Probabilities for Group Assignments: 1942 Racine
Continuous Interviewed Residents (N=333)
Group Prob Prob Prob 25th 75t
(G1) (G2) (G3) Percentile Percentile
1 (n=287) .98(.99) .02(.00) .01(.06) .99 .99
2 (n=26) .06(.00) .89(.97) .06(.02) .77 .99
3 (n=20) .00(.00) .02(003) .98(.99) .99 .99
Note: Minimum and Maximum values presented for the group/prob(group) combination.


Table 5-7. Coefficient Estimates for Effect of Sex (being Male) on Trajectory Group
Membership: 1942 Racine Continuous Interviewed Residents (N=333)
Groups coefficient SE p-value
Cubic ZIP Model
2 2.873 0.711 <0.001***
3 2.123 0.638 <0.001***
*p<.05 **p<.01 ***p<.001
a Group 1 (the lowest rate group) is the reference category of comparison.


Table 5-8. Tabulated Bayesian Information Criteria (BIC) for 1942 Racine Continuous
Interviewed Residents i/lh Block Data (N=235)
Optimal # of Groups BIC
Cubic ZIP Model
2 -960.81
3 -940.76
4 False Convergence


Table 5-9. Mean (Median) Posterior Probabilities for Group Assignments: 1942 Racine
Continuous Interviewed Residents i/lh Block Data (N=235)
Group Prob Prob Prob 25th 75t
(G1) (G2) (G3) Percentile Percentile
1 (n=165) .89(.97) .11(.03) .00(.00) .78 .97
2 (n=47) .06(.00) .89(.97) .05(.01) .80 .99
3 (n=23) .00(.00) .09(.01) .91(.99) .85 .99
Note: Minimum and Maximum values presented for the group/prob(group) combination.









Table 5-10. Coefficient Estimates for Effect of Sex (being Male) on Trajectory Group
Membership: 1942 Racine Continuous Interviewed Residents / i/h Block Data
(N=235)
Groups coefficient SE p-value

Cubic ZIP Model
2 2.262 0.507 <0.001***
3 3.103 0.768 <0.001***

*p<.05 **p<.01 ***p<.001
a Group 1 (the lowest rate group) is the reference category of comparison.


Table 5-11. Tabulated Bayesian Information Criteria (BIC) for 1949 Racine Continuous
Residents (N=1,257)
Optimal # of Groups BIC
Cubic ZIP Model
2 -7123.17
3 -6851.59
4 -6769.89
5 -6694.99
6 False Convergence


Table 5-12. Mean (Median) Posterior Probabilities for Group Assignments: 1949 Racine
Continuous Residents (N=1,297)
Group Prob Prob Prob Prob Prob 25th 75th
(G1) (G2) (G3) (G4) (G5) Percentile Percentile
1 (n=740) .84(.93) .15(.07) .00(.00) .00(.00) .00(.00) .73 .93
2 (n=424) .00(.00) .97(.99) .02(.00) .02(.00) .00(.00) .98 .99
3 (n=34) .00(.00) .08(.01) .89(.97) .03(.00) .00(.00) .84 .99
4 (n=74) .00(.00) .04(.00) .01(.00) .93(.99) .02(.00) .91 .99
5 (n=25) .00(.00) .00(.00) .02(.00) .01(.00) .98(.99) .99 .99
Note: Minimum and Maximum values presented for the group/prob(group) combination.









Table 5-13. Coefficient Estimates for Effect of Sex (being Male) on Trajectory Group
Membership: 1949 Racine Continuous Residents (N=1,297)
Groups coefficient SE p-value
Cubic ZIP Model
2 1.272 0.132 <0.001***
3 0.508 0.353 0.150
4 2.700 0.432 <0.001***
5 2.714 0.741 <0.001***

*p<.05 **p<.01 ***p<.001
a Group 1 (the lowest rate group) is the reference category of comparison.


Table 5-14. Tabulated Bayesian Information Criteria (BIC) for 1949 Racine Continuous
Interviewed Residents (N=556)
Optimal # of Groups BIC
Cubic ZIP Model
2 -2979.90
3 -2848.07
4 False Convergence


Table 5-15. Mean (Median) Posterior Probabilities for Group Assignments: 1949 Racine
Continuous Interviewed Residents (N=556)
Group Prob Prob Prob 25th 75t
(GI) (G2) (G3) Percentile Percentile
1 (n=418) .97(.99) .03(.00) .00(.00) .98 .99
2 (n=114) .06(.01) .92(.99) .02(.00) .87 .99
3 (n=24) .00(.00) .01(.00) .99(.99) .99 .99
Note: Minimum and Maximum values presented for the group/prob(group) combination.


Table 5-16. Coefficient Estimates for Effect of Sex (being Male) on Trajectory Group
Membership: 1949 Racine Continuous Interviewed Residents (N=556)
Groups coefficient SE p-value
Cubic ZIP Model
2 1.712 0.272 <0.001***
3 2.774 0.738 <0.001***
*p<.05 **p<.01 ***p<.001
a Group 1 (the lowest rate group) is the reference category of comparison.









Table 5-17. Tabulated Bayesian Information Criteria (BIC) for 1949 Racine Continuous
Interviewed Residents i/ i/i Block Data (N=323)
Optimal # of Groups BIC
Cubic ZIP Model
2 -1816.65
3 -1747.74
4 False Convergence


Table 5-18. Mean (Median) Posterior Probabilities for Group Assignments: 1949 Racine
Continuous Interviewed Residents i/i/ Block Data (N=323)
Group Prob Prob Prob 25th 75th
(GI) (G2) (G3) Percentile Percentile
1 (n=245) .98(.99) .02(.00) .00(.00) .98 .99
2 (n=66) .11(.01) .89(.98) .01(.00) .82 .99
3 (n=12) .00(.00) .00(.00) .99(.99) .99 .99
Note: Minimum and Maximum values presented for the group/prob(group) combination.



Table 5-19. Coefficient Estimates for Effect of Sex (being Male) on Trajectory Group
Membership: 1949 Racine Continuous Interviewed Residents ii ilh Block Data
(N=323)
Groups coefficient SE p-value
Cubic ZIP Model
2 1.493 0.364 <0.001***
3 1.887 0.790 0.016*
*p<.05 **p<.01 ***p<.001
a Group 1 (the lowest rate group) is the reference category of comparison.










Table 5-20. Trajectory Groups by Sample and Cohort
Continuous-Resident Samples Sample Non-offenders Adolescent- Low-Level High-Rate Late-Onset
Size Limited/Peaked Chronic Chronic Offenders
(n) Offenders Offenders Offenders
1942 COHORT
Continuous-Residents 633 463 (73.1%) 56 (8.8%) 82 (13.0%) 32 (5.1%)

Continuous-Residents 333 287 (86.2%) 26 (7.8%) 20 (6.0%)
(w/ Interview Data)
Continuous-Residents 235 165 (70.2%) 47 (20.0%) 23 (9.8%)
(w/ Individual & Block Data)
1949 COHORT
Continuous-Residents 1297 740 (57.1%) 74 (5.7%) 424 (32.7%) 25 (1.9%) 34 (2.6%)
Continuous-Residents 556 418 (75.2%) 114(20.5%) -- 24 (4.3%)
(w/ Interview Data)
Continuous-Residents 323 245 (75.9%) 66 (20.4%) 12 (3.7%)
(w/ Individual & Block Data)











Table 5-21. Mean Covariate Levels by Trajectory Group Membership and Cohort
Variables 1942 1949
G1 G2 G3 G1 G2 G3
Mean Mean Mean F-test p-value Mean Mean Mean F-test p-value


Macro-level
Social Disorganization

Micro-level
Male
Black
Low SES Household
Stakes in Conformity
Family Stability (Natural
Parents)
Family Structure (No. of
Children)
Differential association
Formal Social Control
Moved

Number of Police Contacts


.41 .50 10.65 <.001***


33.79
3.84
5.99
2.00
6.49


3.69 3.57

.22 .24
1.14 1.43
.59 .60

.12 2.04


<.001***
.23
<.01**
.13
<.01**


1.46 .24

0.74 .84
3.56 .03*
2.47 .08+


.28 .29 4.38 .01*


.43
.02
2.99
1.62
.94

3.80

.24
1.21
.48


8.09 492.99 <.001***


.83
.33
5.58
1.25
.67


4.16 5.17

.29 .17
1.62 2.00
.53 .58


17.10
10.93
14.34
2.89
12.96


<.001***
<.001***
<.001***
.05*
<.001***


6.11 .01*

0.57 .57
7.12 <.001***
0.38 .68


.25 4.98 30.25 481.58 <.001***


p<.10 *p<.05 **p<.01 ***p<.001










Table 5-22. 1942 Racine Continuous Interviewed Residents i/ i/i Block Data: Odds Ratios and
Probabilities for Macro-Level Predictors of Offending Group Membership
Adolescent-Limited/Peaked High-Rate


Variables Offenders Chronic Offenders
Odds Prob. Odds Prob.
Macro-level
Social Disorganization 1.97 <.001*** 2.18 <.001***

Model Diagnostics
Chi-Square= 18.30***
-Log Likelihood= -176.95
McFadden's R2= .06
Nagelkerke= .11
p<.10 *p<.05 **p<.01 ***p<.001
Note. Reference category is group 1 (the nuin-oitider.').


Table 5-23. 1942 Racine Continuous Interviewed Residents i/ i/i Block Data: Odds Ratios and
Probabilities for Micro-Level Predictors of Offending Group Membership
Adolescent-Limited/Peaked High-Rate
Variables Offenders Chronic Offenders
Odds Prob. Odds Prob.
Micro-level
Male 2.99 <.001*** 6.12 <.001***
Black 1.01 .97 1.27 .30
Low SES Household 1.38 .13 1.10 .72
Stakes in Conformity 0.90 .57 0.96 .89
Family Stability (Natural 0.62 <.01** 0.59 .05*
Parents)
Family Structure (No. of 0.96 .80 1.37 .11
Children)
Differential association 1.01 .97 0.84 .57
Formal Social Control 1.13 .55 1.12 .71
Moved 1.11 .61 2.30 <.01**

Model Diagnostics
Chi-Square= 53.14***
-Log Likelihood= -143.38
McFadden's R2= .23
Nagelkerke= .39

p<.10 *p<.05 **p<.01 ***p<.001
Note. Reference category is group 1 (the nun-otiender.N).











Table 5-24. 1942 Racine Continuous Interviewed Residents / i/h Block Data: Odds Ratios and
Probabilities for Macro-level and Micro-level Predictors of Offending Group
Membership
Adolescent-Limited/Peaked High-Rate
Variables Offenders Chronic Offenders
Odds Prob. Odds Prob.
Macro-level
Social Disorganization 1.50 .06+ 1.36 .38

Micro-level
Male 2.87 <.001*** 5.79 <.001***
Black 1.01 .99 1.24 .44
Low SES Household 1.30 .27 1.04 .87
Stakes in Conformity 0.92 .65 0.98 .92
Family Stability (Natural 0.64 .01* 0.61 .06
Parents)
Family Structure (No. of 0.95 .76 1.33 .14
Children)
Differential association 1.01 .99 0.83 .57
Formal Social Control 1.06 .79 1.06 .85
Moved 1.03 .88 2.15 .03*

Model Diagnostics
Chi-Square= 55.47***
-Log Likelihood= -141.56
McFadden's R= .25
Nagelkerke R2= .41

p<.10 *p<.05 **p<.01 ***p<.001
Note. Reference category is group 1 (the nuo,-,tinenler.).











Table 5-25. 1949 Racine Continuous Interviewed Residents i/ i/i Block Data: Odds Ratios and
Probabilities for Macro-Level Predictors of Offending Group Membership
Adolescent-Limited/Peaked High-Rate


Variables Offenders Chronic Offenders
Odds Prob. Odds Prob.
Macro-level
Social Disorganization 1.50 <.01** 1.48 .07

Model Diagnostics
Chi-Square= 7.96*
-Log Likelihood= -207.61
McFadden's R2 .02
Nagelkerke R2 .04
p<.10 *p<.05 **p<.01 ***p<.001
Note. Reference category is group 1 (the nun-otiender.N).


Table 5-26. 1949 Racine Continuous Interviewed Residents i/ i/i Block Data: Odds Ratios and
Probabilities for Micro-Level Predictors of Offending Group Membership
Adolescent-Limited/Peaked High-Rate
Variables Offenders Chronic Offenders
Odds Prob. Odds Prob.
Micro-level
Male 2.31 <.001*** 2.40 .01*
Black 1.38 .06+ 1.20 .64
Low SES Household 1.26 .18 2.52 <.001***
Stakes in Conformity 1.09 .62 1.13 .50
Family Stability (Natural 0.64 <.001*** 0.67 .11
Parents)
Family Structure (No. of 1.36 .04* 1.61 .14
Children)
Differential association 1.25 .16 1.01 .99
Formal Social Control 1.14 .43 1.14 .64
Moved 0.89 .47 0.79 .51

Model Diagnostics
Chi-Square= 76.67***
-Log Likelihood= -169.86
McFadden's R2 .19
Nagelkerke R2= .31
p<.10 *p<.05 **p<.01 ***p<.001
Note. Reference category is group 1 (the nun-otiender.N).











Table 5-27. 1949 Racine Continuous Interviewed Residents n/ ith Block Data: Odds Ratios and
Probabilities for Macro-level and Micro-level Predictors of Offending Group
Membership
Adolescent-Limited/Peaked High-Rate
Variables Offenders Chronic Offenders
Odds Prob. Odds Prob.
Macro-level
Social Disorganization 1.19 .28 1.09 .54

Micro-level
Male 2.33 <.001*** 2.36 .02**
Black 1.30 .13 1.20 .64
Low SES Household 1.25 .21 2.53 <.001***
Stakes in Conformity 1.10 .60 0.79 .53
Family Stability (Natural 0.63 <.001*** 0.68 .15
Parents)
Family Structure (No. of 1.33 .05* 1.62 .13
Children)
Differential association 1.24 .19 1.01 .99
Formal Social Control 1.10 .58 1.17 .64
Moved 0.86 .37 0.78 .53

Model Diagnostics
Chi-Square= 79.26***
-Log Likelihood= -169.32
McFadden's R2 .20
Nagelkerke R= .32
p<.10 *p<.05 **p<.01 ***p<.001. Note. Reference category is group 1 (the noui-cllkder.N).









2
1.5
1
0.5
0


6 7 8 9 10 11 12 13 14 15 16 17 18 19 2021 22 23 24 25
6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25


Age


--G1 pred -- G2 pred


G3 pred


G4 pred


-* G1 amg


--G2 avg


- G3 avg G4 avg


Figure 5-1. 1942 Racine Continuous Residents Group-Based Trajectories (N=633)












0.8
0.6
0.4
0.2
0


6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25


Age


- G1 pred -- G2 pred


G3 pred


G1 avg -- G2 avg -- G3 avg


Figure 5-2. 1942 Racine Continuous Interviewed Residents Group-Based Trajectories (N=333)











t 1.5

o
0.5

*0
6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25

Age

-- G1 pred -i- G2 pred G3 pred G1 avg G2 avg -- G3 avg

Figure 5-3. 1942 Racine Continuous Interviewed Residents n i/h Block Data Group-Based Trajectories (N=235)










6
4
2
0


6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Age


-* G- pred -- G2 pred


G3 pred


G4 pred -- G5 pred


-- G1 ag


--G2 avg


- G3 avg


Figure 5-4. 1949 Racine Continuous Residents Group-Based Trajectories (N=1,297)


G4 amg


G5 ag











3
S2.5
S 2
L1.5

o 0

6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25

Age

-*- G1 pred -m- G2 pred G3 pred G1 avg G2 avg -- G3 avg

Figure 5-5. 1949 Racine Continuous Interviewed Residents Group-Based Trajectories (N=556)












4-
O




6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25

Age

-- G1 pred -i- G2 pred G3 pred G1 avg G2 avg -- G3 avg

Figure 5-6. 1942 Racine Continuous Interviewed Residents in itl Block Data Group-Based Trajectories (N=323)









CHAPTER 6
DISCUSSION AND CONCLUSION

Discussion

This research began by seeking to investigate the viability of integrated multi-level theory

for explaining offense trajectories. It was expected that residing in a socially disorganized area

early on in development would increase an individual's likelihood for being in an offender group

as opposed to a non-offender group. Particularly, living in a socially disorganized area should

increase the odds for exhibiting a chronic-offending trajectory. Second, it was anticipated that

certain individual and family-level factors would also elevate an individual's likelihood for being

an offender. Similar to the expectations regarding the influence of living in a socially

disorganized area, the belief was that the micro-level factors would significantly distinguish

adolescent-limited/peaked offenders and chronic offenders from the non-offenders.

Furthermore, in accordance with Sampson and Laub (1993), it is was assumed that individuals

who were more likely to report having experienced life transitions such as marriage and

employment would be more likely to be in the non-offender group and the adolescent-

limited/peaked group. Third, it was hypothesized that, once these factors were estimated

simultaneously (macro-level and micro-level together in the same model), they would each

contribute to predicting an individual's odds for being in an adolescent-limited/peaked offender

and chronic offender group relative to a non-offender group. In addition, sex (being male) was

expected to be significant for amplifying an individual's likelihood for being found in an

offender group.

Overall, the findings from this analysis yielded varying levels of support for the various

theoretical frameworks used. Support was shown Moffitt's (1993) taxonomy of offending in that

all of the trajectory analyses identified a non-offender, an adolescent-limited/peaked offender,









and a chronic offender group (with a declining trajectory in young adulthood). However, the

semi-parametric Poisson based models also discovered two groups that were inconsistent with

Moffitt's prediction when relying on police contact data from the largest samples of the 1942

(n=633) and 1949 (n=1,297) Racine continuous residents. A low-level chronic offending group

was identified in both cohorts who were characterized by an early onset of offending and

maintained a fairly low level of offending (yet a non-zero rate) throughout the twenty year

observation period (age 6-25). Comparatively, a late-onset offender group was discovered

among the 1949 cohort who had virtually no invovlement in offending until late in

adolescence/early adulthood but showed relatively high rates of offending throughout young

adulthood. In addition, the adolescent-limited/peaked group found in these analyses slightly

differed from the group as theoretically discussed by Moffitt (1993) in that they were classified

as adolescent-peaked and their offending patterns did not appear to desist until around age 21.

While these groups are atypical insofar as Moffitt's expectations, they have been discovered and

the source of discussion elsewhere (see D'Unger et al., 1998; Land, 2000; Piquero, 2007;

Sampson & Laub, 1993).

Also consistent with Moffitt and others (Rutter, 1983; Singer et al., 1968; Werry, 1986)

that have argued that life-course-persistent offending may be observed almost exclusively among

males and the prior research has shown that females typically show less involvement in crime in

either type of offender group (adolescent-limited offenders and life-course-persistent offenders)

(see Piquero, 2007), all of the trajectory results indicated that being male significantly increased

an individual's likelihood of being classified into any of the offender groups across all data

sources (continuous residents, continuous interviewed residents, and continuous interviewed

residents in i/h block data) and by cohort. The one exception was the late-onset offender group









identified in the 1949 continuous resident analysis (N=1,297) where although there were more

males in the group, sex was not found to be a significant covariate for distinguishing this

offender group from the non-offender group. The mlogit results also re-affirmed the importance

of sex for predicting group membership through the consistent and robust effect of being male on

increasing the odds by at least two times for an individual exhibiting either an adolescent-

limited/peaked or a chronic offending trajectory.

The results only revealed limited support for Sampson and Laub's age-graded theory of

informal social control and these findings were only evident at the bivariate level. In both cohort

analyses, the non-offender group had the highest mean score on the stakes in conformity index

when compared with either offender group, but only in the 1949 cohort analysis was the stakes in

conformity index significantly associated with group membership.

In line with the expectation of the influence of social disorganization, there was a minimal

amount of evidence that living in a socially disorganized area increased an individual's odds for

being in an offender group. The bivariate findings suggested that Racine continuous residents,

whose individual offending pattern could be classified into a adolescent-limited/peaked or

chronic offending group-based trajectory, tended to live in more socially disorganized areas than

the non-offenders. The multivariate results confirmed that social disorganization was a

significant predictor of group membership for both the 1942 and 1949 cohort. Although, living

in a socially disorganized area failed to distinguish the adolescent-limited/peaked offenders from

the chronic offenders.

Despite the fact that the bivariate and mlogit (macro-level-only) results indicated that

social disorganization was a distinguishing feature of group membership, the findings from the

mlogits, where both the macro-level and micro-level factors were estimated simultaneously,









failed to find much support for the usefulness of an integrated theoretical model for explaining

group-based offending trajectories. Once the micro-level factors (particularly sex, natural two

parent household, having moved, and low SES households depending on which cohort mlogit is

being referred to) were added to the model, social disorganization failed to exert a significant

independent effect on predicting group membership. Although, the various indicators of model

fit (chi-square, log likelihood, and the psuedo R-square statistics) showed that the integrated

model was a better fit, it appeared that the influence of social disorganization on increasing an

individual's likelihood for being in either an adolescent-limited/peaked or chronic offender

group tended to be indirect in that it operated through individual and family-level factors.

Overall, these findings tend to mirror those produced in prior multi-level studies that have shown

that the addition of neighborhood level processes typically only increase the variance explained

by 1% to 4% (see Cheong & Raudenbush, 2000; Duncan & Raudenbush, 1999; Elliot et al.,

1996; Simcha-Fagan & Schwartz, 1986), although other studies have begun to find more support

for school contextual effects ranging from a 6% to 7% increase in the variance explained

(Anderson, 2002; Welsh et al., 1999).

Limitations

As with all research, there are a series of data limitations that are worth mentioning. First,

the most apparent limitation relates to the era wherein these individuals were raised. These

individuals were either born during WWII (1942 cohort) or baby-boomers (i.e., 1949 cohort)

who grew up in the immediate post-WWII era. They were living their early years during the

1960s, which is a time characterized by large social change including the Civil Rights and

Women's Movement, Vietnam, riots, etc. Thus, it is likely that some of these findings may not

necessarily generalize to individuals who were socialized in different periods of time, say the

1980s or 1990s. Second, the area of Racine itself was a Midwestern industrial city with a









relatively small homogenous population of predominantly White residents, and it is not likely

that the results from this community context and demographic would apply to the larger more

metropolitan areas of today or in communities/neighborhoods with an ethnically diverse

population. Third, the substantial variety of the census-level measures that are typically used

currently in macro-level analyses were not available at the block level in 1950. Among the

measures that were accessible there was a rather noticeable amount of completely missing data

which precluded the ability to use mean substitution to retain a greater number of individuals for

the integrated theoretical models. Extending from this limitation, the generalizability of these

findings may not be applicable or reproduced if different conceptualizations or aggregations

were used to define the community/neighborhood such as cities, SMSAs, counties, police

districts, etc.

Fourth, the social disorganization index used in this study was limited in that it did not

include a measure for ethnic heterogeneity as discussed by Shaw and McKay (1942). Although

some research has shown that rates of delinquency are lower in ethnically homogenous areas (see

Lander, 1954), there is typically less support for the measure when compared with poverty and

residential instability (see Sampson, 1997). Additionally, Bursik (1988) and Sampson and

Groves (1989) have argued that various structural dimensions such as local friendship ties,

organizational density, and participation in local formal and informal voluntary organizations are

among the most important factors that affect delinquency rates. More recently, Sampson et al.

(1997) have argued that collective efficacy and the ability to form social capital matters more

than social disorganization, and Sampson (1997) has also indicated that one of the most

important structural dimensions for controlling neighborhood delinquency is the community's

ability to supervise and control teenage peer groups, particularly gangs. Sampson and Groves









(1989) also identified additional factors including family disruption, urbanization, and rapid

population change that can all impede the ability of a community to create informal networks of

social control for supervising teenage peer groups. Unfortunately, there were not any data

available to measure these important processes, so caution should be taken when interpreting

these results.

Similarly, the stakes in conformity index used to assess Sampson and Laub's (1993) age-

graded theory of informal social control was limited since it relied on a dichotomous index

ranging from 0 to 2. Laub et al. (1998) have argued that the quality or cohesiveness of the adult

social bonds is what is more important than mere participation in these life transitions. Others,

however, have argued that it is likely that being married and/or having a full-time job are

probably indicators that some level of attachment has occurred in that it seems doubtful that

these events would have happened at all if this was not the case (see Piquero et al., 2002).

It is also important to note that this research relied on official data of delinquency/crime as

measured through police contacts. Despite the known strengths and weaknesses of official and

self-report data when investigating criminal behavior (see Junger-Tas & Marshall, 1999), police

contacts may be the better measure of "true" officially-based offending because they are

typically closer in time to the actual criminal event than self-reported offending (see Hindelang

et al., 1981; Thornberry & Krohn, 2003).

Finally, there are some limitations of the trajectory modeling procedure as it specifically

pertains to this study. There was evidence provided from the sensitivity analysis conducted that

data loss did not equally affect all of the trajectory groups. For instance, it appeared that once

the multiple sources of data were merged together that the model only discovered three groups

and that the high-rate offending group (also referred to as chronic) actually showed evidence of









a sharp declining trajectory in early adulthood and even perhaps an indication of desistance by

age 25. Unfortunately, given the data limitations and the reliance on only official measures of

offending to model the trajectories, this research is unable to answer the question of whether or

not this is true evidence of desistance. Furthermore, there is also the possibility that these

individuals in the high-rate offending group which accumulated a rather large number of police

contacts over time had become incarcerated for some period of time by their young adulthood

years, thus not providing them an opportunity to offend. This is also an empirical question that

cannot be explored with this data. Lastly, once the individual and neighborhood level factors

were estimated in multivariate models, the results indicated that while these risk/protective

factors significantly distinguished the adolescent-limited/peaked offenders and the chronic-

offenders from the non-offenders, these same variables (with the exception of sex) were not able

to significantly distinguish the adolescent-limited/peaked offenders from the chronic-offenders.

Once again, this is likely the result of the data loss that appeared to differentially affect offender

groups.

Directions for Future Research

The findings from this research provided mixed results for the various theories including

Moffitt's developmental typology and the saliency of sex when distinguishing offender groups,

Sampson and Laub's age-graded theory of informal social control, and social disorganization

theory. This research did generate some preliminary findings that can direct future research on

integrated multi-level theoretical models.

Sex was found to be rather robust for significantly affecting group assignment when

estimated as a covariate in the trajectory analyses. It also exerted a consistent independent effect

for significantly increasing the odds for group assignment to either an adolescent-limited/peaked

offender group or a chronic offender group. Future research would benefit from estimating sex-









disaggregated trajectories and then estimating mlogit models to see whether or not the macro-

level and micro-level factors, such as those examined here, significantly differ across sex.

Unfortunately, there were not enough individuals in the analyses relied on in this study to permit

the estimation of sex-specific trajectories and make use of all of the risk/protective factors in the

mlogit models. Further studies employing considerably larger samples would permit this type of

analysis. It is worth noting that there are other theories that may account for sex differences as

well. One such example may be power-control theory (Hagan et al., 1985). It is possible that

parents and society socializes males and females differently and this may have an impact on sex-

specific pathways of offending. This possibility should be further explored.

Larger samples with multi-level data would also allow future researchers to examine these

effects on distinguishing between offender groups for groups that were uncovered in some of the

trajectory analyses that were not consistent with Moffitt's typology such as the low-level

chronic and the late-onset offenders. It may be the case that social disorganization matters more

so for influencing low-level chronic offenders with low yet relatively stable offending rates over

the life-course. Similarly, it seems reasonable to assume that the absence of the various adult

social bonds identified by Sampson and Laub (1993) or a low level of attachment to the

institutions of marriage and employment may serve as explanations for a late-onset pattern of

offending. Additionally, future studies investigating the chronic patterns of offending should

rely on data from longer observational periods particularly through middle and late adulthood to

see if these group-based offending patterns identified in this study continue beyond age 25.

This research relied solely on an official measure of crime to estimate offense trajectories

and more specifically, police contacts. It would be interesting if future studies would attempt to

reproduce these analyses using alternative official measures of offending such as arrest or









conviction records. Comparatively, it may also prove fruitful to examine the possible multi-level

effects on life-course patterns of offending based on self-reported criminal involvement.

Arguably, the best test would be to include some type of composite measure that represented a

combination of official and self-reported offending.

As far as the relatively weak found support for the macro-level effects, it would likely be

the case that future studies accounting for the processes identified by prior researchers (see

Bursik, 1988; Sampson, 1997; Sampson & Groves, 1989; Sampson et al., 1997) such as family

disruption (percentage of divorced or single-parent households), collective efficacy, social

capital, supervision of teenage peer groups, and local participation in voluntary organizations

may yield stronger support for macro-level influences on offending patterns and especially net of

the effects of micro-level factors. Future mutli-level studies should also make use of other

macro-level theoretical constructs to investigate their relative impact on trajectories of criminal

behavior including routine activities theory (see Cohen & Felson, 1979; Kennedy & Baron,

1993) and urban disadvantage perspectives (see Krivo & Peterson, 1996; Parker et al., 2005).

With attention to these aforementioned suggestions, it would seem that perhaps the ideal

research design for best examining the hypotheses highlighted in this research would be a

prospective longitudinal design which would identify individuals at birth in different social

contexts and follow them throughout their life-course and well into adulthood. Not only would

official and self-report data be gathered on the individuals at various periods over their life-

course, the optimal design would include both parent, teacher, and peer surveys throughout

different developmental periods. Additionally, surveys administered to community residents

over time would also allow for measuring the informal networks of social control, collective

efficacy, and social capital. Finally, if spousal surveys in addition to self-reports were also









administered in adulthood this would provide worthwhile information in order to gauge the level

of cohesiveness or attachment in the adult social bonds to better determine their effect in the

desistence process.

Nevetheless, it is obvious that collecting research in this magnitude, for a lengthy period of

time, involving a numerous amount of actors would require an enormous amount of monetary,

individual, political, and community support, yet there are early indications that research similar

to this magnitude (in a lesser degree) is possible and the preliminary findings leave hope for

future research in integrative mutli-level modeling (see related publications from the Project on

Human Development in Chicago Neighborhoods).

Implications

Taken together, the findings from this research are intended to be preminary and to offer a

starting point to guide and direct future research on theoretically integrated approaches. Given

the minimal significance of social disorganization in the baseline models, it is important that

future studies further investigate other social processes that have been recently discussed by

neighborhood-level researchers as perhaps being more important when estimating neighborhood

effects (for review, see Sampson et al., 2002) before any concerted effort can be made toward

imforming policy. Nevertheless, it is important here to at least highlight some of the possible

implications that research in this area may provide in the future.

Programs such as the Rochester Nurse-Home Visitation Program (see Olds et al., 1986) are

likely to provide the most benefit in targeting the cumulative effects of both neighborhood

disadvantage and factors particularly associated with chronic offending (i.e. low birth weight,

birth trauma). The Rochester program involved random assignment of expectant mothers in a

small semi-rural community to one of three conditions: (1) pre and post natal visits; (2) pre-natal

only visits; and (3) minimal standard services. The results indicated that the mothers who









received the home visits faired significantly better than those that received the minimal amount

of services and had higher levels of employment and fewer cases of child abuse and neglect. In

an outcome evaluation of a similar program, Seitz et al. (1985) also found that pre and post-natal

home visits to disadvantaged mothers reduced the impact of adverse risk factors and the children

whose mothers participated in the program were rated as less aggressive and disruptive than

children who were in the non-experimental group (see Provence and Naylor, 1983 for an

extended discussion of the Yale Child Welfare Research Program).

Programs targeting poor urban children during the preschool years have shown to have a

significant impact on reducing delinquent outcomes. Head Start is one of the most well-known

of these type of initiatives and subsequent evaluations (see Darlington et al., 1980 ; Lee et al.,

1990) have shown that the program was successful in producing short-term intellectual and

social gains during the early preschool years and that Head Start youth were more likely to

graduate high school. Other programs such as the Perry Preschool Project, which focused on

early childhood education (ages 3 to 5), parent meetings, and home visits, have yielded a host of

beneficial outcomes for program participants including better grades, higher standardized test

scores, higher graduation and employment rates, and lower delinquency and arrest rates (see

Berreuta-Clement et al., 1984; Schweinhart & Weikart, 1988). In addition, programs

implemented in the later elementary/early adolescent years specifically directed at improving

problem-solving and social competency among poor youth have been associated with decreased

levels of aggression and problem behaviors (see Weissberg et al., 1981; Hawkins et al., 1991;

Lochman et al., 1984; Sarason & Sarason, 1981; Hammond, 1991).

Family intervention programs facilitating the development of parental support networks

and increasing the link between parents and community services as well as improving parent-









child interactions have provided mixed results for reducing delinquent involvement (see Dumas,

1989; Johnson & Walker, 1987; Hawkins et al., 1991), but Guerra (1997:290) has argued that

these programs may not be as effective for urban minority disadvantaged youth because of

cultural differences and the multiple stressors that plague the residents in these areas. Peer-based

interventions like peer mentoring (Goldstein, 1978) and redirecting gang members toward more

prosocial community activities (see Klein, 1971) have also yielded varying results in influencing

an adolescent's invovlement in delinquency.

Residential relocation programs including the Moving to Opportunity for Fair Housing

Demonstration (or MTO) and the Gautreaux Program have been shown to be beneficial.

Gautreaux youth who moved to the suburbs substantially increased their educational attainment

and job opportunities (see Rosenbaum, 1995; Rosenbaum & Popkin, 1991) and recent MTO

evaluations have indicated that residential relocation was significantly related to decreased

involvement in violence, better overall health, and lower victimization rates (see Katz et al.,

2001; Ludwig et al., 2001). Leventhal and Brooks-Gunn (2003) have also highlighted the

significantly lower levels of depression, anxiety, and/or dependency problems among MTO

youth when compared with youth who remained in the disadvantaged areas. Other initiatives

that focus on increasing neighborhood supervision of teenage peer groups and promoting civic

participation are likely to provide noticeable benefits as well (see Sampson, 1997). Overall,

these programs reviewed above provide suggestions for how future research targeting

disadvantaged youth in disadvantaged areas can likely alleviate the conditions that plague

residents of these communities and decrease the prevalence of life-course-persistent chronic

patterns of offending.









Final Thoughts

Guerra (1997:300-301) perhaps best summarizes the direction for programmatic and policy

initiatives in the following statement presented below if we as a community have any hopes for

reducing the risk factors associated with both types of offending patterns identified by Moffitt,

but particularly for prevention/intervention with chronic offenders:

It is likely that the most effective interventions will address multiple causal factors, be

carried out in multiple contexts, and be extended in time over the course of development.

Families, peers, schools, and communities must work in concert to provide opportunities for

children to learn prosocial behavior and to create environments where prosocial behavior is

modeled and reinforced.

It is important for researchers, policymakers, and even individuals to be aware that all of

the abovementioned elements do not operate in isolation from one another. There are individual,

family, and community factors that all have influence on life-course pathways of development.

Sampson et al. (2002:474) also challenges neighborhood level researchers to move beyond mere

structural compositional measures of neighborhood effects and give greater attention toward

understanding "cultural, normative, and collective action perspectives that attach meaning to how

residents frame their commitment to places." Spatial dynamics are also important to

investigating the influence of neighborhoods on offending patterns. It is possible that while

youth may reside in a disadvantaged area, they may spend more time in an adjacent neighborhood

where there are stronger informal social ties and collective supervision of teenage peer groups.

Thus, while compositional measures of disadvantage may be related to crime rates, without

measuring the influence of spatial proximity to more advantaged areas the effect of

neighborhoods might be overestimated.









It is clear that community and individualistic perspectives have there own shortcomings

and limitations. Yet, it is also the case that perhaps by highlighting the strengths of each

approach and seeking an understanding of their interactive influence on life-course

developmental patterns of offending, future multi-level theoretical models can gain greater

support within the discipline. Although the results here failed to offer much evidence of the

saliency of this approach, research in this area is still an under-explored phenomenon. Future

studies that are able to better capture the complex nature of macro-level and micro-level

processes are needed before the usefulness of theoretical integration can be discounted.









APPENDIX
ZERO-ORDER CORRELATIONS BY COHORT


Table A-1. 1942 Cohort
1 2 3 4
1. Social Disorganization Index 1.0
2. Stakes in Conformity Index -.07 1.0
3. Low SES Household .30 .06 1.0
4. Family Structure .12 -.01 .19 1.0
Note. Significant effects (p<.05) are in italics.


Table A-2. 1949 Cohort
1 2 3 4
1. Social Disorganization Index 1.0
2. Stakes in Conformity Index -.07 1.0
3. Low SES Household .29 -.22 1.0
4. Family Structure .18 .10 .16 1.0
Note. Significant effects (p<.05) are in italics.









REFERENCES


Aber, J.L., Gephart, M.A., Brooks-Gunn, J., & Connell, J.P. (1997). Development in context:
Implications for studying neighborhood effects. In J. Brooks-Gunn, G. Duncan, & J.L.
Aber (eds.), Neighborhoodpoverty: Context and consequences for children. New York:
Sage Foundation.

Aguilar, B., Sroufe, L.A., Egeland, B., & Carlson, E. (2000). Distinguishing the early-onset-
persistent and adolescent-onset antisocial behavior types: From birth to 16 years.
Developmental and Psychopathology, 12, 109-132.

Akers, R.L. (1985). Deviant behavior: A social learning approach. Belmont, CA: Wadsworth.

Akers, R.L., Krohn, M.D., Lanza-Kaduce, L., & Radosevich, M. (1979). Social learning and
deviant behavior: A specific test of a general theory. American Sociological Review, 44,
635-655.

Akers, R.L. (1977). Deviant behavior: A social learning approach. 2nd edition. Belmont, CA:
Wadsworth.

Allison, P.D. (1999). Multiple regression: A primer. Thousand Oaks, CA: Sage Publications.

Anderson, E. (1990). Streetwise: race, class, and change in an urban community. Chicago:
University of Chicago Press.

Anderson, A.L. (2002). Individual and contextual influences on delinquency: The role of the
single-parent family. Journal of Criminal Justice 30, 575-87.

Agresti, A. (1990). Categorical data analysis. New York: John Wiley and Sons.

Bandura, A. (1982). The psychology of chance encounters and life paths. American
Psychologists, 37, 747-755.

Barrett, A.E., & White, H.R. (2002). Trajectories of gender role orientations in adolescence and
early adulthood: A prospective study of mental health effects of masculinity and
femininity. Journal ofHealth and Social Behavior, 43, 451-468.

Bartusch, D.R., Lynam, D.R., Moffitt, T.E., Silva, P.A. (1997). Is age important? Testing a
general versus a developmental theory of antisocial behavior. Criminology, 35, 13-48.

Bellair, P.E. (1997). Social interaction and community crime: Examining the importance of
neighbor networks. Criminology, 35: 677-703.

Berreuta-Clement, J.R., Schweinhart, L.J., Barnett, W.S., Epstein, A.S., Weikart, D.P. (1984).
Changed lives: The effects of the Perry Preschool Program ony ,inhi\ through age 19.
Ypsilanti, MI: High/Scope Press.

Berry, B.J., & Kasarda, J.D. (1977). Contemporary urban ecology. New York: Macmillan.









Blackwell, B.S., & Piquero, A.R. (2005). On the relationships between gender, power control,
self-control, and crime. Journal of Criminal Justice, 33, 1-17.

Blumstein, A., & Cohen, J. (1979). Estimation of individual crime rates from arrest records.
Journal of Criminal Law and Criminology, 70, 561-585.

Blumstein, A., Cohen, J., & Hsieh, P. (1982). The duration of adult criminal careers. Final
Report submitted to NIJ. Pittsburgh, PA: School of Urban and Public Affairs, Carnegie-
Mellon University.

Blumstein, A., Cohen, J., Roth, J.A., & Visher, C.A (eds.). (1986). Criminal careers and "career
criminals, Vol 1. Report of the Panel on Criminal Careers, National Research Council.
Washington, D.C.: National Academy of Sciences.

Blumstein, A. & Cohen, J. (1987). Characterizing criminal careers. Science, 237, 985-991.

Blumstein, A., Cohen, J., & Farrington, D.P. (1988). Criminal career research: Its value for
criminology. Criminology, 26, 1-35.

Braithwaite, J. (1979). Inequality, crime, andpublic policy. London: Routledge and Kegan Paul.

Brame, B., Nagin, D., & Tremblay, R. (2001). Developmental trajectories of physical aggression
from school entry to late adolescence. Journal of Child Psychology andPsychiatry, 42,
503-513.

Broidy, L.M., Nagin, D.S., Tremblay, R.E., Bates, J.E., Brame, R., Dodge, K.A., Fergusson,
D.M., Horwood, J.L., Loeber, R., Laird, R., Lynam, D.L., Moffitt, T.E., Petit, G.S., &
Vitaro, F. (2003). Developmental trajectories of childhood disruptive behaviors and
adolescent delinquency: a six-site, cross-national study. Developmental Psychology, 39,
222-245.

Bronfenbrenner, U., Moen, P., & Garbarino, J. (1984). Child, family, and community. In R.
Parke (ed.), Review of child developmental research: Vol. 7: The family. Chicago:
University of Chicago Press.

Brooks-Gunn, J., Duncan, G.J., Klebanov, P.K., & Sealand, N. (1993). Do neighborhoods
influence child and adolescent development? American Journal of Sociology, 99, 353-
395.

Bursik, R.J. (1984). Urban dynamics and ecological studies of delinquency. Social Forces, 63,
393-413.

Bursik, R.J. (1986). Ecological stability and dynamics of delinquency. In A.J. Reiss & M. Tonry
(eds.), Crime and Justice (vol. 66). Chicago: University of Chicago Press.

Bursik, R.J., & Grasmick, H.G. (1993). Neighborhoods and crime: The dimensions of effective
community control. New York: Lexington Books.









Bushway, S., Piquero, A.R., Broidy, L.M., Cauffman, E., & Mazerolle, P. (2001). Anempirical
framework for studying desistance as a process. Criminology, 39, 492-515.

Burgess, R., & Akers, R. (1966). A differential association-reinforcement theory of criminal
behavior. Social Problems, 14, 128-147.

Bushway, S.D., Thornberry, T.P., & Krohn, M.D. (2003). Desistance as a developmental
process: A comparison of static and dynamic approaches. Journal of Quantitative
Criminology, 19, 129-153.

Byrk, A.S., & Raudenbush, S.W. (1987). Application of hierarchical linear models to assessing
change. Psychology Bulletin, 101, 147-158

Byrk, A.S., & Raudenbush, S.W. (1992). Hierarchical Linear Models for social and behavioral
research: Application and data analysis methods. Newbury Park, CA: Sage Publications.

Caspi, A., Elder, G.H., & Bem, D.J. (1987). Moving against the world: Life-course patterns of
explosive children. Developmental Psychology, 23, 308-313.

Caspi, A, & Moffitt, T.E. (1993). The continuity of maladaptive behavior: From description to
understanding in the study of antisocial behavior. In D. Cicchetti & D. Cohen (eds.),
Manual in the development ofpsychopathology. New York: John Wiley & Sons.

Caspi, A., Moffitt, T.E., Stouthamer-Loeber, M., Krueger, R.F., & Schmutte, P.S. (1994). Are
some people crime-prone? Replications of the personality-crime relationship across
countries, genders, races, and methods. Criminology, 32, 163-195.

Catalono, R.F., & Hawkins, J.D. (1996). The social development model: A theory of antisocial
behavior. In D. Hawkins (ed.), Delinquency and crime: Current dihe ie. Cambridge:
Cambridge University Press.

Chaiken, J., & Chaiken, M. (1982). Varieties of criminal behavior. Report R-2814-NIJ. Santa
Monica, CA: Rand Corporation.

Chase-Lansdale, P.L., Gordon, R.A., Brooks-Gunn, J., & Klebanov, P.K. (1997). Neighborhood
and family influences on the intellectual and behavioral competence of preschool and
early school-age children. In J. Brooks-Gunn, G.J. Duncan, & J.L. Aber (eds.),
Neighborhood poverty: Vol. 1. Context and consequences for children. New York: Sage
Foundation.

Cheong, Y.F., & Raudenbush, S.W. (2000). Measurement and structural models for children's
problem behaviors. Psychological Methods, 5, 477-95.

Chung, I., Hill, K.G., Hawkins, J.D., Gilchrist, L.D., & Nagin, D.S. (2002). Childhood predictors
of offense trajectories. Journal ofResearch in Crime and Delinquency, 39, 60-90.

Cleckley, H. (1976). The mask of sanity (5th ed.). St. Louis, MO: Mosby.









Cohen, L.E., & Felson, M. (1979). Social changes and crime rate trends: A routine activities
approach. American Sociological Review, 44, 588-608.

Cohen, L.E., & Land, K.C. (1987). Age structure and crime: Symmetry versus asymmetry and
the projection of crime rates through the 1990s. American Sociological Review, 52, 170-
183.

Cohen, L.E., & Machalek, R. (1988). A general theory of expropriative crime. American Journal
ofSociology, 94, 465-501.

Coleman, J.S. (1988). Social capital in the creation of human capital. American Journal of
Sociology, 94, 95-120.

Coleman, J.S. (1990). Foundations of social theory. Cambridge, MA: Harvard University Press.

Cote, S., Zoccolillo, M., Tremblay, R.E., Nagin, D., & Vitaro, F. (2001). Predicting girls'
conduct disorder in adolescence from childhood trajectories of disruptive behaviors.
Journal of the American Academy of Child and Adolescent Psychiatry, 40, 678-684.

Cote, S., Tremblay, R.E., Nagin, D., Zoccolillo, M., Vitaro, F. (2002). The development of
impulsivity, fearfulness, and helpfulness during childhood: Patterns of consistency and
change in the trajectories of boys and girls. Journal of ChildPsychology & Psychiatry &
Allied Disciplines, 43, 609-618.

Cramer, H. (1946). Mathematical models of statistics. Princeton, NJ: Princeton University Press.

Crane, J. (1991). The epidemic theory of ghettos and neighborhood effects on dropping out and
teenage childbearing. American Journal of Sociology, 96, 1126-1159.

Daly, K. (1998). Gender, crime, and criminology. In M. Tonry (ed.), The handbook of crime and
punishment. New York: Oxford University Press.

Darlington, R.B., Royce, J.M., Snipper, A.S., Murray, H.W., & Lazar, I. (1980). Preschool
programs and later school competence of children from low-income families. Science,
208, 202-204.

De Li, S. (1999). Legal sanctions and youths' status achievement: A longitudinal study. Justice
Quarterly, 16, 377-401.

Donnellan, M.B., Ge, X., & Wenk, E. (2000). Cognitive abilities in adolescent-limited and life-
course-persistent criminal offenders. Journal ofAbnormal Psychology, 109, 396-402.

Dumas, J.E. (1989). Treating antisocial behavior in children: Child and family approaches.
Clinical Psychology Review, 9, 197-222.

Duncan, G.J. (1994). Families and neighbors as sources of disadvantage in the schooling
decisions of White and Black adolescents. American Journal ofEducation, 103, 20-53.









Duncan, G.J., Brooks-Gunn, J.P., & Klebanov, P.K. (1994). Economic deprivation and early-
childhood development. ChildDevelopment, 65, 296-318.

Duncan, G.J., & Raudenbush, S.W. (1999). Assessing the effect of context in studies of child and
youth development. Educational Psychologist, 34, 29-41.

D'Unger, A.V., Land, K.C., McCall, P.L, & Nagin, D.S. (1998). How many latent classes of
delinquent/criminal careers? Results from mixed Poisson regression analyses. American
Journal of Sociology, 103, 1593-1630.

D'Unger, A.V., Land, K.C., & McCall, P.L. (2002). Sex differences in age patterns of
delinquent/criminal careers: Results from Poisson latent class analyses of the
Philadelphia Cohort Study. Journal of Quantitative Criminology, 18, 349-375.

Eggleston, E.P., Laub, J., & Sampson, R.J. (2004). Methodological sensitivities to latent class
analysis of long-term criminal trajectories. Journal of Quantitative Criminology, 20, 1-
26.

Elder, G.H. (1975). Age differentiation and the life course. AnnualReview of Sociology, 1, 165-
190.

Elder, G.H. (1985). Perspectives on the life course. In Glen H. Elder, Jr. (ed.), Life course
dynamics: Trajectories and transitions, 1968-1980. Ithaca: Cornell University Press.

Elliott, D.S., Ageton, S.S., Huizinga, D., Knowles, B.A., & Canter, R.J. (1983). The prevalence
and incidence of delinquent behavior: 1976-1980. Boulder, CO: Behavioral Research
Institute.

Elliot, D.S., Huizinga, D., & Morse, B. (1986). Self-reported violent offending-A descriptive
analysis of juvenile violent offenders and their offending careers. Journal of
Interpersonal Violence, 1, 472-514.

Elliot, D.S. (1994). Serious violent offenders: Onset, developmental course, and termination-
The American Society of Criminology 1993 address. Criminology, 32, 1-21.

Elliott, D.S., Wilson, W.J., Huizinga, D., Sampson, R.J., Elliott, A. & Rankin, B. (1996). The
effects of neighborhood disadvantage on adolescent development. Journal of Research in
Crime andDelinquency, 33, 389-426.

Ensminger, M.E., Lamkin, R.P., & Jacobson, N. (1996). School leaving: A longitudinal
perspective including neighborhood effects. Child Development, 67, 2400-2416.

Entwisle, D.R., Alexander, K.L., & Olson, L.S. (1994). The gender gap in math: Its possible
origins in neighborhood effects. American Sociological Review, 59, 822-838.

Farrington, D.P. (1986a). Age and crime. In M. Tonry & N. Morris (eds.), Crime and Justice,
Vol. 7. Chicago, IL: University of Chicago Press.









Farrington, D.P. (1986b). Stepping stones to adult criminal careers. In D. Olweus, J. Block, & M.
Radke-Yarrow (eds.), Development of antisocial andprosocial behavior. New York:
Academic Press.

Farrington, D.P. (1995). The development of offending and antisocial behaviour from childhood:
Key findings from the Cambridge Study in Delinquent Development. Journal of
Child Psychology and Psychiatry, 360, 929-964.

Farrington, D.P. (1999). A criminological research agenda for the next millennium. International
Journal of Offender Therapy and Comparative Criminology, 43, 154-167.

Farrington, D.P. (2003). Developmental and life-course criminology: Key theoretical and
empirical issues-The 2002 Sutherland Address Award. Criminology, 41, 221-255.

Farrington, D.P., & West, D.J. (1990). "The Cambridge Study in delinquent development: A
prospective longitudinal survey of 411 males," in H. Kerner & G. Kaiser (eds.),
Criminality: Personality, behavior, and life history. Berlin: Springer-Verlag.

Farrington, D.P., Loeber, R., Elliott, D.S., Hawkins, J.D., Kandel, D.B., Klein, M.W., McCord,
J., Rowe, D.C., & Tremblay, R.E. (1990). Advancing knowledge about the onset of
delinquency and crime. In B.B. Lahey & A.E. Kazdin (eds.), Advances in clinical child
psychology. New York: Plenum Press.

Farrington, D.P., Sampson, R.J., & Wikstrom, P.H. (1994). Integrating individual and ecological
aspects on crime. Stockholm, Sweden: National Council for Crime Prevention.

Felson, R.B., & Haynie, D.L. (2002). Pubertal developmental, social factors, and delinquency
among adolescent boys. Criminology, 40, 967-988.

Fergusson, D.M., Horwood, L.J., Caspi, A., Moffitt, T.E., & Silva, P.E. (1996). The (artefactual)
transmission of reading disability: Psychometric lessons in the study of stability and
change in behavioral development. Developmental Psychology, 32, 132-140.

Fergusson, D.M., Horwood, L.J., & Nagin, D.S. (2000). Offending trajectories in a New Zealand
birth cohort. Criminology, 38, 525-551.

Fergusson, D.M., & Horwood, L.J. (2002). Male and female offending trajectories. Development
and Psychopathology, 14, 159-177.

Gephart, M.A. (1997). Neighborhoods and communities as contexts for development. In J.
Brooks-Gunn, G. Duncan, & J.L. Aber (eds.), Neighborhoodpoverty: Context and
consequences for children. New York: Sage Foundation.

Glueck, S., & Glueck, E. (1950). Unraveling juvenile delinquency. Cambridge, MA: Harvard
University Press.

Glueck, S., & Glueck, E. (1968). Delinquents and nondelinquents in perspective. Cambridge,
MA: Harvard University Press.









Goetting, A. (1994). The parenting crime connection. Journal ofPrimary Prevention, 14, 167-
184.

Goldstein, A.P., Sherman, M., Gershaw, N.J., Sprafkin, R.P., & Glick, B. (1978). Training
aggressive adolescents in prosocial behavior. Journal of Youth and Adolescence, 7, 73-
92.

Goldstein, H. (1995). Multilevel statistical models. 2nd edition. London, Edward Arnold.

Goldsmith, H.H., Bradshaw, D.L., & Reiser-Danner, L.A. (1984). Temperament as a potential
developmental influence on attachment. In J.V. Lerner & R.M. Lerner (eds.),
Temperament and social interaction during infancy and childhood. San Francisco: Jossey
Bass.

Gordon, R.A. (1968). Issues in multiple regression. American Journal of Sociology, 73, 592-616.

Gottfredson, M., & Hirschi, T. (1990). A general theory of crime. Palo Alto: Stanford University
Press.

Gottfredson, D.C, McNeil, R.J., & Gottfredson, D.G. (1991). Social area influences on
delinquency: A multi-level analysis. Journal ofResearch in Crime and Delinquency, 28,
197-226.

Granovetter, M. (1973). The strength of weak ties. American Journal of Sociology, 78,
1360-1380.

Grasmick, H., Tittle, C.R., Bursik, R.J., & Arneklev, B. (1993). Testing the core implications of
Gottfredson and Hirschi's general theory of crime. Journal ofResearch in Crime and
Delinquency, 30, 5-29.

Greenberg, D.F. (1985). Age, crime, and social explanation. American Journal of Sociology, 91,
1-21.

Greene, W.H. (1990). Econometric analysis. New York: Macmillan.

Griffiths, E., & Chavez, J.M. (2004). Communities, street guns and homicide trajectories in
Chicago, 1980-1995: Merging methods for examining homicide trends across space and
time. Criminology, 42, 941-978.

Guerra, N.G. (1997). Intervening to prevent childhood aggression in the inner city. In J. McCord
(ed.), Violence and childhood in the inner city. Cambridge, UK: University of Cambridge
Press.

Hagan, J. (1991). Destiny and drift: Subcultural preferences, status attainments, and the risks and
rewards of youth. American Sociological Review, 56, 567-582.









Hagan, J., Gillis, A.R., & Simpson, J. (1985). The class structure of gender and delinquency:
Toward a power-control theory of common delinquent behavior. American Journal of
Sociology, 90, 1151-1178.

Halpern-Felsher, B., Connell, J.P., Spencer, M.B., Aber, J.L., Duncan, G.J., Clifford, E.,

Crichlow, W., Usinger, P., & Cole, S.S. (1997). Neighborhood and family factors predicting
educational risk and attainment in African American and White children and adolescents.
In J. Brooks-Gunn, G.J. Duncan, & J.L. Aber (eds.), Neighborhoodpoverty: Vol. 1.
Context and consequencesfor children. New York: Sage Foundation.

Hamparian, D.M., Schuster, R., Dinitz, S., & Conrad, J. (1978). The violent few: A study of
dangerous juvenile offenders. Lexington, MA: Lexington Books.

Hamparian, D.M., Schuster, R., Davis, J., & White, J. (1985). The young criminal careers of the
violent few. Report to the U.S. Department of Justice, Office of Juvenile Justice and
Delinquency Prevention.

Harpur, J., & Hare, R.D. (1994). Assessment of psychopathology as a function of age. Journal of
Abnormal Psychology, 103, 604-609.

Hawkins, J.D., Von Cleve, E., & Catalano, R.F. (1991). Reducing early childhood aggression:
Results of a primary prevention program. Journal of the American Academy of Child and
Adolescent Psychiatry, 30, 208-217.

Hawkins, D. (ed.). (1995). Ethnicity, race, and crime: Perspectives across time and space.
Albany, NY: State University of New York Press.

Heilbron, D. (1994). Zero-altered and other regression models for count data with added zeros.
Biometrical Journal, 36, 531-547.

Heitgerd, J.L., & Bursik, R.J. (1987). Extra-community dynamics and the ecology of
delinquency. American Journal of Sociology, 92, 775-787.

Hindelang, M.J., Hirschi, T., & Weis, J.G. (1981). Measuring delinquency. Beverly Hills, CA:
Sage Publications.

Hipwell, A.E., Loeber, R., Stouthamer-Loeber, M., Keenan, K., White, H.R., &

Kroneman, L. (2002). Characteristics of girls with early-onset disruptive and antisocial
behaviour. Criminal Behaviour and Mental Health, 12, 99-118.

Hirschi, T., & Hindelang, M. J. (1977). Intelligence and delinquency: A revisionist review.
American Sociological Review, 42, 571-587.

Hirschi, T., & Gottfredson, M. (1983). Age and the explanation of crime. American Journal of
Sociology, 89, 552-584.









Hirschi, T., & Gottfredson, M. (1985). Age and crime, logic and scholarship: Comment on
Greenberg. The American Journal of Sociology, 91, 22-27.

Homey, J., Osgood, D.W., & Marshall, I.H. (1995). Criminal careers in the short-term: Intra-
individual variability in crime and its relation to local life circumstances. American
Sociological Review, 60, 655-673.

Huesmann, L.R., Eron, L.D., Lefkowitz, M.M., & Walder, L.O. (1984). Stability of aggression
over time and generations. Developmental Psychology, 20, 1120-1134.

Huff-Corzine, L., Corzine, J., & Moore, D.C. (1986). Southern exposure: Deciphering the
South's influence on homicide rates. Social Forces, 64, 906-924.

Jencks, C., & Mayer, S.E. (1990). The social consequences of growing up in a poor
neighborhood. In L. Lynn & M. McGeary (eds.), Inner-city poverty in the United States.
Washington: National Academy Press.

Jessor, R., Donovan, J.E., & Costa, F.M. (1991). Beyond adolescence: Problem behavior and
young adult development. New York: Cambridge University Press.

Johnson, D.L., & Walker, T. (1987). The primary prevention of behavior problems in Mexican-
American children. American Journal of Community Psychology, 15, 375-385.

Jones, B.L., Nagin, D.S., & Roeder, K. (2001). A SAS procedure based on mixture models for
estimating developmental trajectories. Sociological Methods andResearch, 29, 374-393.

Junger-Tas, J., & Marshall, I.H. (1999). The self-report methodology in crime research. In M.
Tonry (ed), Crime and Justice (vol. 25). Chicago, IL: University of Chicago Press.

Katz, L.F., Kling, J.R., & Liebman, J.B. (2001). Moving to opportunity in Boston: Early results
of a randomized mobility experiment. Quarterly Journal ofEconomics, 116, 607-654.

Kazemian, L., & Farrington, D.P. (2006). Exploring residual career length and residual number
of offenses for two generations of repeat offenders. Journal of Research in Crime and
Delinquency, 43, 89-113.

Kennedy, L.W., & Baron, S.W. (1993). Routine activities and a subculture ofviolence:A study
of violence on the street. Journal ofResearch in Crime and Delinquency, 20, 88-112.

Kim, J., & Mueller, C.W. (1978). Introduction to factor analysis: What it is and how to do it.
Thousand Oaks, CA: Sage Publications.

King, G. (1989). Event count models for international relations: Generalizations and
applications. International Studies Quarterly, 33, 123-147.









Klebanov, P.K., Brooks-Gunn, J., Chase-Landsdale, P., & Gordon, R.A. (1997). Are
neighborhood effects on young children mediated by features of the home environment?
In J. Brooks-Gunn, G.J. Duncan, & J. Aber (ed.), Neighborhood poverty: Context and
consequences for children. Sage: New York, NY.

Klein, M.W. (1971). Street gangs and street workers. Englewood Cliffs, NJ: Prentice Hall.

Kling, J.R., Liebman, J.B., & Katz, L.F. (2007). Experimental analysis of neighborhood effects.
Econometrica, 75, 83-119.

Kornhauser, R.R. (1978). Social sources of delinquency. Chicago: University of Chicago Press.

Kratzer, L., & Hodgins, S. (1999). A typology of offenders: A test of Moffitt's theory among
males and females from childhood to age 30. Criminal Behavior andMental Health, 9,
57-73.

Krivo, L.J., & Peterson, R.D. (1996). Extremely disadvantaged neighborhoods and urban crime.
Social Forces, 75, 619-650.

Kroneman, L., Loeber, R., & Hipwell, A.E. (2004). Is neighborhood context differently related
to externalizing problems and delinquency for girls compared with boys? Clinical Child
and Family Psychological Review, 7, 109-122.

Kupersmidt, J.B., Griesler, P.C., DeRozier, M.E., Patterson, C.J. & Davis, P.W. (1995).
Childhood aggression and peer relations in the context of family and neighborhood
factors. Child Development, 66, 360-375.

Lambert, D. (1992). Zero-inflated Poisson regression, with an application to defects in
manufacturing. Technometrics, 34, 1-14.

Land, K. (2000). Influence of neighborhood, peer, and family context: Trajectories of
delinquent/criminal offending across the life course, Final Report. Washington, DC:
National Institute of Justice.

Land, K.C., McCall, P.L., & Cohen, L.E. (1990). Structural covariates of homicide rates: Are
there any invariances across time and social space? American Journal of Sociology, 95,
922-963.

Land, K.C., McCall, P.L., & Nagin, D.S. (1996). A comparison of Poisson, Negative Binomial,
and Semiparametric Mixed Poisson Regression Models with empirical applications to
criminal careers research. Sociological Methods and Research, 24, 387-442.

Land, K.C., & Nagin, D.S. (1996). Micro-level models of criminal careers: A synthesis of The
criminal careers and life course approaches via Semiparametric Mixed Poisson
Regression Models, with empirical models. Journal of Quantitative Criminology, 12,
163-191.









Lander, B. (1954). Toward and understanding ofjuvenile delinquency. New York: Columbia
University Press.

Lash, T., Sigal, H., & Dudzinski, D. (1980). State of the child: New York City. New York:
Foundation for Child Development.

Laub, J.H., & Sampson, R.J. (1988). Unravelling families and delinquency: A reanalysis of the
Gluecks' data. Criminology, 26, 355-380.

Laub, J.H., & Sampson, R.J. (1993). Turning points in the life course: Why change matters to the
study of crime. Criminology, 31, 301-325.

Laub, J.H., Nagin, D.S., & Sampson, R.J. (1998). Trajectories of change in criminal offending:
Good marriages and the desistance process. American Sociological Review, 63, 225-238.

Laub, J.H., & Sampson, R.J. (2003). .\/hii e, beginnings, divergent lives: Delinquent boys to age
70. Cambridge, MA: Harvard University Press.

Lee, V.E., Brooks-Gunn, J., Schnur, E., & Liaw, F. (1990). Are Head Start effects sustained? A
longitudinal follow-up comparison of disadvantaged children attending Head, no
preschool, and other preschool programs. Child Development, 61, 495-507.

Lempers, J., Clark-Lempers, D., & Simons, R. (1989). Economic hardship, parenting and distress
in adolescence. Child Development, 60, 25-49.

Leventhal, T. & Brooks-Gunn, J. (2000). The neighborhoods they live in: Effects of
neighborhood residence on child and adolescent outcomes. Psychological Bulletin, 126,
309-337.

Leventhal, T., & Brooks-Gunn, J. (2003). Moving to opportunity: An experimental study of
neighborhood effects on mental health. American Journal of Public Health, 93, 1576-
1582.

Lindstrom, P. (1996). Family interaction, neighborhood context and deviant behaviour. Studies
in Crime and Crime Prevention, 5, 113-119.

Lochman, J.E., Burch, P.R., Curry, J.F., & Lampron, L.B. (1984). Treatment and generalization
effects of cognitive behavioral and goal setting interventions with aggressive boys.
Journal of Consulting and Clinical Psychology, 52, 915-916.

Loeber, R. (1982). The stability of antisocial child behavior: A review. Child Development, 53,
1431-1446.

Loeber, R., & Wikstrom, P. (1993). Individual pathways to crime in different types of
neighborhoods. In Integrating individual and ecological aspects of crime, edited by D.P.
Farrington, R.J. Sampson, & P. Wikstrom. Stockholm: Swedish National Council for
Crime Prevention.









Loeber, R., Burke, J.D., Lahey, B.B., Winters, A., & Zera, M. (2000). Oppositional defiant and
conduct disorder: A review of the past ten years, part 1. Journal of the American
Academy of Child and Adolescent Psychiatry, 39, 1468-1484.

Loftin, C., & Hill, R.R. (1974). Regional subculture and homicide: An examination of the Gastil-
Hackney thesis. American Sociological Review, 39, 714-724.

Loftin, C., & Parker, R.N. (1985). An errors-in-variable model of the effect of poverty on urban
homicide rates. Criminology, 23, 269-287.

Loehlin, J.C. (1989). Partitioning environmental and genetic contributions to behavioral
development. American Psychologist, 44, 1285-1292.

Long, J.S. (1997). Regression models for categorical and limited dependent variables.
Thousand Oaks, CA: Sage.

Ludwig, J., Duncan, G.J., & Hirschfield, P. (2001). Urban poverty and juvenile crime: Evidence
from a randomized housing-mobility experiment. Quarterly Journal ofEconomics, 116,
655-679.

Lynam, D.R., Moffitt, T.E., & Stouthamer-Loeber, M. (1993). Explaining the relationship
between IQ and delinquency: Class, race, test motivation, school failure, or self-control.
Journal ofAbnormal Psychology, 102, 187-196.

Lynam, D.R, Caspi A., Moffitt, T.E., Wikstrom, P.O., Loeber, R., & Novak, S. (2000). The
interaction between impulsivity and neighborhood context on offending: the effects of
impulsivity are stronger in poorer neighborhoods. Journal ofAbnormal Psychology, 109,
563-574.

Matsueda, R., & Anderson, K. (1998). The dynamics of delinquent peers and
delinquentbehavior. Criminology, 36, 269-308.

Maughan, B. (2005). Developmental trajectory modeling: A view from developmental
psychopathology. The ANNALS of the American Academy of Political and Social
Science, 602, 118-130.

Maume, M., Ousey, G., & Beaver, K. (2005). Cutting the grass: A reexamination of the link
between marital attachment, delinquent peers and desistance from marijuana use. Journal
of Quantitative Criminology, 21, 27-53.

Mazzerole, P, Brame, R., Piquero, A.R., & Dean, C. (2001). Onset age, persistence, and
offending versatility: Comparisons across gender. Criminology, 38, 1143-1172.

McArdle, J.J., & Epstein, D. (1987). Latent growth curves within developmental structural
equation models. ChildDevelopment, 58, 110-113.

McCord, W., & McCord, J. (1959). Origins of crime: A new evaluation of the Cambridge-
Somerville study. New York, NY: Cambridge University Press.









McDermott, S., & Nagin, D.S. (2001). Same or different? Comparing offender groups and
covariates over time. Sociological Methods and Research, 29, 282-318.

McDermott, S., & Nagin, D. (1998). Same or different over time? In S.O White (ed.), Handbook
ofLaw and Social Science: Youth and Justice. New York: Plenum Press.

Meredith, W., & Tisak, J. (1990). Latent curve analysis. Psychometrika, 54, 557-585.

Messner, S., Krohn, M., & Liska, A. (1989). Theoretical integration in the study of deviance and
crime: Problems and prospects. Albany, NY: State University of Albany Press.

Moffitt, T.E. (1990). Juvenile delinquency and attention-deficit disorder: Developmental
trajectories from age 3 to 15. ChildDevelopment, 61, 893-910.

Moffitt, T.E. (1993). Adolescent-limited and life-course-persistent antisocial behavior: A
developmental taxonomy. Psychological Review, 100, 674-701.

Moffitt, T.E. (1994). Natural histories of delinquency. In E.G.M. Weitekamp & Hans-Jurgen
Kerner (eds.), Cross-national longitudinal research on human development and criminal
behavior. Dordrecht, the Netherlands: Kluwer Academic.

Moffitt, T.E. (1997a). Neuropsychology, antisocial behavior, and neighborhood context. In J.
McCord (ed.), Violence and childhood in the inner city. Cambridge, UK: University of
Cambridge Press.

Moffitt, T.E. (1997b). Adolescence-limited and life-course-persistent offending: A
complementary pair of developmental theories. In T.P Thornberry (ed.), Developmental
ithew ie\ of crime and delinquency: Advances in criminological theory, Volume Seven.
New Brunswick, NJ: Transaction Publishers.

Moffitt, T.E. (2003). Life-course-persistent and adolescent-limited antisocial behavior: A ten-
year research review and a research agenda. In B.B. Lahey, T.E. Moffitt, & A. Caspi
(eds.), Causes of conduct disorder and delinquency. New York: Guilford Press.

Moffitt, T.E. (2005). Life-course-persistent versus adolescence-limited antisocial behavior. In D.
Cicchetti & D. Cohen (eds.), Developmental psychopathology (2nd ed.). New York:
Wiley.

Moffitt, T.E., & Henry, B. (1989). Neuropsychological assessment of executive functions in self-
reported delinquents. Development and Psychopathology, 1, 105-118.

Moffitt, T.E., & Henry, B. (1991). Neuropsychological studies of juvenile delinquency and
violence: A review. In J. Milner (ed.), The neuropsychology of aggression Norwell, MA:
Kluwer Academic.

Moffitt, T.E., Lynam, D.R., & Silva, P.A. (1994). Neuropsychological tests predicting persistent
male delinquency. Criminology, 32, 277-300.









Moffitt, T.E., & Caspi, A. (2001). Childhood predictors differentiate life-course persistent and
adolescent-limited antisocial pathways among males and females. Developmental and
Psychopathology, 13, 355-375.

Moffitt, T.E., Caspi, A, Rutter, M., & Silva, P.A. (2001). Sex differences in antisocial behavior:
Conduct disorder, delinquency, and violence in the Dunedin Longitudinal Study.
Cambridge, UK: Cambridge University Press.

Nagin, D.S. (1999). Analyzing developmental trajectories: A semiparametric, group-based
approach. Psychological Methods, 4, 139-157.

Nagin, D.S. (2004). Response to "Methodological sensitivities to latent class analysis of long-
term criminal career trajectories." Journal of Quantitative Criminology, 20, 27-35.

Nagin, D.S. (2005). Group-based modeling of development. Cambridge, MA: University of
Harvard Press.

Nagin, D.S., & Paternoster, R. (1991). On the relationship of past to future delinquency.
Criminology, 29, 163-189.

Nagin, D.S., & Farrington, D.P. (1992). The onset and persistence of offending. Criminology, 30,
301-523.

Nagin, D.S., & Land, K.C. (1993). Age, criminal Careers, and population heterogeneity:
Specification and estimation of a nonparametric, mixed Poisson model. Criminology, 31,
327-362.

Nagin, D.S., Farrington, D.P., & Moffitt, T.E. (1995). Life-Course trajectories of different types
of offenders. Criminology, 33, 111-140.

Nagin, D.S., & Waldfogel, J. (1998). The effect of conviction on income through the life cycle.
International Review ofLaw and Economics, 18, 25-40.

Nagin, D.S., & Tremblay, R.E. (2001). Analyzing developmental trajectories of distinct but
related behaviors: A group-based method. Psychological Methods, 6, 18-34.

Nagin, D.S., & Tremblay, R.E. (2005a). Developmental trajectory groups: Fact or a useful
statistical fiction. Criminology, 43, 873-904.

Nagin, D.S., & Tremblay, R.E. (2005b). From seduction to passion: A response to Sampson and
Laub. Criminology, 43, 915-918.

Nagin, D.S., & Tremblay, R.E. (2005c). What has been learned from group-based trajectory
modeling? Examples from physical aggression and other problem behaviors. The
ANNALS of the American Academy of Political and Social Science, 602, 82-117.









Nagin, D.S., & Tremblay, R.E. (2005d). Further reflections on modeling and analyzing
developmental trajectories: A response to Maughan and Raudenbush. The ANNALS of the
American Academy of Political and Social Science, 602, 145-154.

O'Brien, R.M., Stockard, J., & Isaacson, L. (1999). The enduring effects of cohort characteristics
on age-specific homicide rates, 1960-1995. The American Journal ofSociology, 104,
1061-1095.

Olds, D.L., Henderson, C.R., Tatelbaum, R., & Chamberlin, R. (1986). Improving the delivery of
prenatal care and outcomes of pregnancy: A randomized trial of nurse-home visitation.
Pediatrics, 77, 16-28.

Orr, L., Feins, J.D., Jacob, R., Beecroft, E., Sanbonmatsu, L., Katz, L.F., Liebman, J.B., & Kling,
J.R. (2003). Moving to Opportunity: Interim impact evaluations. Washington, DC: U.S.
Department of Housing and Urban Development, Office of Policy Development and
Research.

Pagani, L.S., Tremblay, R.E., Nagin, D.S., Zoccolillo, M., Vitaro, F., & McDuff, P. (2004). Risk
factor models for adolescent verbal and physical aggression toward mothers.
International Journal ofBehavioral Development, 28, 528-537.

Park, R.E., Burgess, E.W. (1925). Introduction to the science of sociology. Chicago: University
of Chicago Press.

Parker, R.N., & Smith, M.D. (1980). Deterrence, poverty, and type of homicide. American
Journal ofHomicide, 85, 614-624.

Parker, K.F., Stults, B.J., & Rice, S.K. (2005). Racial Threat, concentrated disadvantage and
social control: Considering the macro-level sources of variation in arrests. Criminology,
43, 1111-1134.

Paternoster, R., & Brame, R. (1997). Multiple routes to delinquency? A test of developmental
and general theories of crime. Criminology, 35, 49-84.

Paternoster, R., Dean, C.W., Piquero, A.R., Mazerolle, P., & Brame, R. (1997). Generality,
continuity, and change in offending. Journal of Quantitative Criminology, 13, 231-266.

Petersilia, J., Greenwood, P.W., & Lavin, M. (1977). Criminal careers of habitual felons. Report
R-2144-DOJ. Santa Monica, CA: Rand Corporation.

Peterson, M.A., & Braiker, H.B. (1980). Doing crime: A survey of California prison inmates.
Report R-2200-DOJ. Santa Monica, CA: Rand Corporation.

Peeples, F., & Loeber, R. (1994). Do individual factors and neighborhood context explain ethnic
differences in juvenile delinquency? Journal of Quantitative Criminology, 10, 141-157.









Piquero, A.R. (2007). Taking stock of developmental trajectories of criminal activity over the
life course. In A. Liberman (ed.), Longitudinal research on crime and delinquency. New
York: Springer.

Piquero, A.R., & Brezina, T. (2001). Testing Moffitt's account of adolescence-limited
delinquency. Criminology, 39, 353-370.

Piquero, A.R., Blumstein, A., Brame, R., Haapanen, R., Mulvey, E.P., & Nagin, D.S. (2001).
Assessing the impact of exposure time and incapacitation on longitudinal trajectories of
criminal offending. Journal ofAdolescent Research, 16, 54-74.

Piquero, A.R., & Buka, S. (2002). Linking juvenile and adult patterns of criminal activity in the
Providence Cohort of the National Collaborative Perinatal Project. Journal of Criminal
Justice, 30, 259-272.

Piquero, A.R., Brame, R., Mazzerole, P., & Haapanen, R. (2002). Crime in emerging adulthood.
Criminology, 40, 137-169.

Piquero, A.R., Farrington, D.P., & Blumstein, A. (2003). The criminal career paradigm. Crime
and Justice, 30, 359-506.

Piquero, A.R., & White, N. (2003). On the relationship between cognitive abilities and life-
course-persistent offending among a sample of African Americans: A longitudinal test of
Moffitt's hypothesis. Journal of Criminal Justice, 31, 399-409.

Piquero, A.R., & Moffitt, T.E. (2004). Life-course-persistent offending. In J.R. Adler (ed.),
Forensic Psychology: Concepts, debates, and practices. Portland, OR: Willan Publishing.

Piquero, A.R., Brame, R., & Lynam, D. (2004). Studying career length through early adulthood
among serious offenders. Crime & Delinquency, 50, 412-435.

Piquero, A.R., Brame, R., & Moffitt, T.E. (2005). Extending the study of continuity and change:
Gender differences in the linkage between adolescent and adult offending. Journal of
Quantitative Criminology, 21, 219-243.

Piquero, A.R., Brezina, T., & Turner, M.G. (2005). Testing Moffitt's account of delinquency
abstention. Journal ofResearch in Crime andDelinquency, 42, 27-54.

Piquero, A.R., Moffitt, T.E., & Lawton, B. (2005). Race and Crime: The contribution of
individual, familial, and neighborhood level risk factors to life-course-persistent
offending. In D. Hawkins & K. Kempf-Leonard (eds.), Our children, your children:
Confronting race and ethnicity in the American juvenile justice system. Chicago, IL:
University of Chicago Press.

Plomin, R. (1990). The role of inheritance in behavior. Science, 248, 183-188.

Plomin, R., Chipuer, H.M., & Loehlin, J.C. (1990). Behavioral genetics and personality. In L.A.
Pervin (ed.), Handbook ofpersonality theory and research. New York: Guilford Press.









Pratt, T.C., Turner, M., & Piquero, A.R. (2005). Parental socialization and community context: A
longitudinal analysis of the structural sources of low self-control. Journal of Research in
Crime andDelinquency, 41, 219-243.

Pratt, T.C., & Cullen, F.T. (2005). Assessing macro-level predictors of theories of crime: A
meta-analysis. In M. Tonry (ed.), Crime and Justice (vol. 32). Chicago: University of
Chicago Press.

Provence, S., & Naylor, A. (1983). Wo king ni ith disadvantaged parents and children: Scientific
issues and practice. New Haven, CT: Yale University Press.

Raudenbush, S.W. (2005). How do we study "what happens next"? The ANNALS of the
American Academy of Political and Social Science, 602, 131-144.

Reiss, A.J., & Lewis, A.L. (1961). The distribution of juvenile delinquency in the social class
structure. American Sociological Review, 26, 720-732.

Reiss, A.J. (1986). Why are communities important in understanding crime? In A.J. Reiss & M.
Tonry (eds.), Communities and Crime. Chicago: University of Chicago Press.

Reiss, A.J. (1989). Ending criminal careers. Prepared for the final report of the "Desistance-
Persistance Working Group" of the Program on Human Development and Criminal
Behavior. MacArthur Foundation and National Institute of Justice.

Roncek, D.W. (1981). Dangerous places: Crime and residential environment. Social Forces, 60,
74-96.

Rosenbaum, J.E. (1995). Changing the geography of opportunity by expanding residential
choice: Lessons from the Gautreaux Program. Housing Policy Debate, 6, 231-269.

Rosenbaum, J.E., & Popkin, S. (1991). Employment and earnings of low-income Blacks who
move to middle-class suburbs. In C. Jencks & P. Peterson (eds.), The urban underclass.
Washington, DC: Brookings Institution.

Rowe, D.C., Vazsonyi, A.T., & Flannery, D.J. (1995). Sex differences in crime: Do means and
within-sex variation have similar causes? Journal ofResearch in Crime and Delinquency,
32, 84-100.

Rutter, M. (1983). Developmental neuropsychiatry. New York, NY: Guilford.

Rutter, M. (1989). Pathways from childhood to adult life. Journal of Child Psychology and
Psychiatry, 30, 23-51.

Sampson, R.J. (1987). Urban black violence: The effect of male joblessness and family
disruption. American Journal of Sociology, 93, 348-382.

Sampson, R.J. (1995). The community. In J.Q. Wilson & J. Petersilia (eds.), Crime. San
Francisco, CA: ICS, Inc









Sampson, R.J. (1997). The embeddedness of child and adolescent development: A
neighborhood-level perspective on urban violence. In J. McCord (ed.), Violence and
childhood in the inner city. Cambridge, UK: University of Cambridge Press.

Sampson, R.J. (2002). Transcending tradition: New directions in community research, Chicago
style. Criminology, 40, 213-230.

Sampson, R.J., & Groves, W.B. (1989). Community structure and crime: Testing social-
disorganization theory. American Journal of Sociology, 94, 774-802.

Sampson, R.J., Raudenbush, S.W., & Earls, F. (1997). Neighborhoods and violent crime: A
multilevel study of collective efficacy. Science, 277, 918-924.

Sampson, R.J., Laub, J.H. (1990). Crime and deviance over the life course: The salience of adult
social bonds. American Sociological Review, 55, 609-627.

Sampson, R.J., & Laub, J.H. (1992). Crime and deviance in the life course. Annual Journal of
Sociology, 18, 63-84.

Sampson, R.J., & Laub, J.H. (1993). Crime in the making: Puathii uy and turning points through
life. Cambridge, MA: Harvard University Press.

Sampson, R.J., & Laub, J.H. (1995). Understanding variability in lives through time:
Contributions of life-course criminology. Studies on Crime and Crime Prevention, 4,
143-158.

Sampson, R.J., & Laub, J.H. (1997). A life-course theory of cumulative disadvantage and the
stability of delinquency. In T.P. Thornberry (ed.), Developmental dile, ile, of crime and
delinquency (pp. 131-161). New Brunswick, NJ: Transaction.

Sampson, R.J., Morenoff, J.D., & Gannon-Rowley, T. (2002). Assessing "neighborhood effects":
Social processes and new directions in research. Annual Review of Sociology, 28, 443-
478.

Sampson, R.J., & Laub, J.H. (2003). Life-course desisters? Trajectories of crime among
delinquent boys followed to age 70. Criminology, 41, 555-592.

Sampson, R.J., & Raudenbush, S.W. (2004). Seeing disorder: Neighborhood stigma and the
social construction of "broken windows." Social Psychological Quarterly, 67, 319-342.

Sampson, R.J., Laub, J.H., & Eggleston, E.P. (2004). On the robustness and validity of groups:
Response to Daniel Nagin. Journal of Quantitative Criminology, 20, 37-42.

Sampson, R.J., & Laub, J.H. (2005a). Seductions of method: Rejoinder to Nagin and Tremblay's
"Developmental Trajectory Groups: Fact or Fiction." Criminology, 43, 905-914.









Sampson, R.J., & Laub, J.H. (2005b). When prediction fails: From crime-prone boys to
heterogeneity in adulthood. The ANNALS of the American Academy of Political and
Social Science, 602, 73-79.

Sampson, R.J., & Laub, J.H. (2005c). A life-course view of the development of crime. The
ANNALS of the American Academy of Political and Social Science, 602, 12-45.

Sarason, I.G., & Sarason, B.R. (1981). Teaching cognitive and social skills to high school
students. Journal of Clinical and Consulting Psychology, 49, 908-918.

Schreck, C.J., Fisher, B., & Miller, J.M. (2004). The social context of delinquent peer networks
and violent victimization. Justice Quarterly, 21, 23-48.

Schuerman, L., & Kobrin, S. (1986). Community careers in crime. In A.J. Reiss & M. Tonry
(eds.), Crime and justice: A review of research. Chicago: University of Chicago Press.

Schweinhart, L.J., & Weikart, D.P. (1988). The High/Scope Perry preschool program. In R.H.
Price, E.L. Cohen, R.P. Lorion, & J. Ramos-McKay (eds.), 14 ounces ofprevention: A
casebookfor practicioners (pp. 53-65). Washington, DC: American Psychological
Association.

Seitz, V., Rosenbaum, L.K., & Apfel, N.H. (1985). Effects of family support intervention: A ten-
year follow up. ChildDevelopment, 56, 376-391.

Shannon, L.W. (1988). Criminal career continuity: Its social context. New York: Human
Sciences Press.

Shannon, L.W. (1991). Changing patterns of delinquency and crime: A longitudinal study in
Racine. Boulder: Westview Press.

Shannon, L.W. (1994). Juvenile delinquency and adult crime, 1948-1977 [Racine, Wisconsin]:
City ecological data [Computer file]. 2nd ICPSR version. Iowa City, IA: University of
Iowa, Iowa Urban Community Research Center [producer], 1980. Ann Arbor, MI:
Interuniversity Consortium for Political and Social Research [distributor], 1994.

Shavit, Y., & Rattner, A. (1988). Age, crime, and the early life course. The American Journal of
Sociology, 93, 1457-1470.

Shaw, C.D., & McKay, H.D. (1942). Juvenile delinquency and urban areas. Chicago: University
of Chicago Press.

Short, J.F. (1990). Gangs, neighborhoods, and youth crime. Criminal Justice Research Bulletin,
5, 1-11.

Silverthorn, P., & Frick, P.J. (1999). Developmental pathways to antisocial behavior: The
delayed-onset pathway in girls. Development andPsychopathology, 11, 101-126.









Silverthorn, P., Frick, P.J., & Reynolds, R. (2001). Timing of onset and correlates of severe
conduct problems in adjudicated girls and boys. Journal ofPsychopathology and
Behavioral Assessment, 23, 171-181.

Simcha-Fagan, O., & Schwartz, J.E. (1986). Neighborhood and delinquency: An assessment of
contextual effects. Criminology, 24, 667-703.

Singer, J.E., Westphal, M., & Niswander, K.R. (1968). Sex differences in the incidence of
neonatal abnormalities and abnormal performance in early childhood. Child
Development, 39, 103-112.

Skogan, W.G. (1990). Disorder and decline. New York: Free Press.

Snyder, H. (1988). Court careers ofjuvenile offenders. Washington, DC: U.S. Department of
Justice, Office of Juvenile Justice and Delinquency Prevention.

Snyder, J., & Patterson, G. (1987). Family interaction and delinquent behavior. In H. Quay (ed.),
Handbook ofjuvenile delinquency. New York: Wiley.

Steffensmeier, D.J., Allan, E.A., Harer, M.D., & Streifel, C. (1989). Age and the distribution of
crime. The American Journal of Sociology, 94, 803-831.

Steffensmeier, D.J., Streifel, C., & Shihadeh, E.S. (1992). Cohort size and arrest rates over the
life course: The Easterlin Hypothesis revisited. American Sociological Review, 57, 306-
314.

Sutherland, E.H. (1947). Principles of criminology. 4th ed. Philadelphia: Lippincott. Sutherland,
E.H. (1939). Principles of criminology. 3rd edition. Philadelphia: Lippincott.

Taylor, R.B. (1995). The impact of crime on communities. Annals of the American Academy of
Political and Social Science, 539, 28-45.

Taylor, R.B., & Covington, J. (1988). Neighborhood changes in ecology and violence.
Criminology, 26, 553-589.

Thornberry, T.P., & Krohn, M.D. (2003). The self-report method for measuring delinquency and
crime. In D. Duffee (ed.), Criminal justice 2000, volume 4, measurement and analysis of
crime. Washington, DC: National Institute of Justice.

Tibbetts, S.G., & Piquero, A.R. (1999). The influence of gender, low birth weight and
disadvantaged environment in predicting early-onset of offending: A test of Moffitt's
interactional hypothesis. Criminology, 37, 843-878.

Tittle, C.R. (1988). Two empirical regularities (maybe) in search of explanation: Commentary on
the age-crime debate. Criminology, 26, 75-86.

Tonry, M., Ohlin, L.E., & Farrington, D.P. (1991). Human development and criminal behavior.
New York: Springer-Verlag.









Tracy, P.E., Wolfgang, M.E., & Figlio, R.M. (1990). Delinquency in two birth cohorts. New
York: Plenum.

Uggen, C. (2000). Work as a turning point in the life course of criminals: A duration model of
age, employment, and recidivism. American Sociological Review, 65, 529-546.

Visher, C.A., & Roth, J.A. (1986). Participation in criminal careers. In A. Blumstein, J. Cohen,
J.A. Roth, & C.A. Visher (eds.), Criminal careers and "career criminals," Volume 1.
Washington, DC: National Academy Press.

Warr, M. (1998). Life-course transitions and desistance from crime. Criminology, 36, 183-216.

Warr, M. (1993). Age, peers, and delinquency. Criminology, 31, 17-40.

Warr, M., & Stafford, M. (1991). The influence of delinquent peers: What they think or what
they do? Criminology, 29, 851-866.

Weissberg, R.P., Gesten, E.L., Rapkin, B.D., Cowen, E.L., Davidson, E., de Apodaca, R.F., &
McKim, B.J. (1981). The evaluation of a social problem-solving training program for
suburban and inner-city third grade children. Journal of Clinical and Consulting
Psychology, 49, 251-261.

Weisburd, D., Bushway, S., Lum, C., &Yang, S. (2004). Trajectories of crime at places: A
longitudinal study of street segments in the city of Seattle. Criminology, 42, 283-322.

Wellman, B., & Leighton, B. (1979). Networks, neighborhoods, and communities: Approaches
to the study of the community question. Urban Affairs Quarterly, 14, 363-390.

Welsh, W., Greene, J. & Jenkins, P. (1999). School disorder: The influence of individual,
institutional, and community factors. Criminology, 37, 73-115.

Werry, J.S. (1986). Biological factors. In H.C. Quay & J.S. Werry (eds.), Py-i/hlip,,th,1,/gi1ul
disorders of childhood. New York, NY: Wiley.

West, D.J., & Farrington, D.P. (1973). Who becomes delinquent? London: Heinemann.

West, D.J., & Farrington, D.P. (1977). The delinquent way of life. London: Heinemann.

White, N., & Piquero, A.R. (2004). A preliminary empirical test of Silverthorn and Frick's
delayed-onset pathway in girls using an urban, African-American, US-based sample.
Criminal Behaviour and Mental Health, 14, 291-309.

Wikstrom, P.H. (1991a). Urban crime, criminals, and victims. New York: Springer-Verlag.

Wikstrom. P.H. (1991b). Housing tenure, social class and offending. Criminal Behaviour, 1, 69-
89.









Wikstrom, P.H., & Loeber, R. (2000). Do disadvantaged neighborhoods cause well-adjusted
children to become adolescent delinquents? A study of male juvenile serious offending,
individual risk and protective factors, and neighborhood context. Criminology, 38, 1109-
1142.

Wilson, J.Q., & Kelling, G.L. (March, 1982). Broken windows: The police and neighborhood
safety. Atlantic Monthly, 29-38.

Wilson, J.Q., & Hernstein, R. (1985). Crime and human nature. New York: Simon & Schuster.

Wilson, W.J. (1987). The truly disadvantaged: The inner city, the underclass, and public policy.
Chicago: University of Chicago Press.

Wirth, L. (1938). Urbanism as a way of life. American Journal of Sociology, 44, 1-24.

Wolfgang, M.E., Figlio, R.M., & Sellin, T. (1972). Delinquency in a birth cohort. Chicago:
University of Chicago Press.

Wolfgang, M., Thomberry, T.P., & Figlio, R.M. (1987). From boy to man, from delinquency to
crime. Chicago: University of Chicago Press.

Zorn, C.J. (1998). An analytic and empirical assessment of zero-inflated and hurdle Poisson
specifications. Sociological Methods and Research, 26, 368-400.









BIOGRAPHICAL SKETCH

Wesley Glenn Jennings received his bachelor's degree in psychology from the University

of South Carolina in 2002. After completing his undergraduate education he continued on to

receive his master's degree in criminal justice in 2004 also from the University of South

Carolina. Upon finishing his master's degree he moved down to the University of Florida to

pursue his doctorate in criminology. His research interests specifically include offense

trajectories, sex offenders, and multi-level modeling.





PAGE 1

1 TRAJECTORIES OF TWO RACINE BI RTH COHORTS: A THEORETICALLY INTEGRATED MODEL FOR EXPLAINING OFFENDING By WESLEY G. JENNINGS A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2007

PAGE 2

2 2007 Wesley G. Jennings

PAGE 3

3 ACKNOWLEDGMENTS I would like to provide a brief acknowledgement to all of those who have helped me along the way to achieve this ultimate goal. First and foremost, I would like to thank my family for all of their continued love and support throughout my many, many years of schooling. They have been a source of strength and financial support. I would also like to sin cerely thank my wife for her willingness to understand how long this proc ess may take and working hard to support us financially throughout my graduate education. Lastly, I would lik e to offer an extreme amount of gratitude to all of the f aculty at the Department of Cr iminology, Law, and Society, and specifically to my committee me mbers who were steadfast in their support throughout my entire graduate student career. Alex Piquero, Lonn Lanza-Kaduce, Brian Stul ts, Karen Parker, and Chuck Peek played an extremely integral role in my success in completing this dissertation and have provided me with excellent advice, criticisms, and thoughts th at have certainly helped me mature as an academic professional. For this I will always be grateful and I hope to continue collaborating with all of these individuals for a long, long time.

PAGE 4

4 TABLE OF CONTENTS page ACKNOWLEDGMENTS...............................................................................................................3 LIST OF TABLES................................................................................................................. ..........6 LIST OF FIGURES................................................................................................................ .........8 ABSTRACT....................................................................................................................... ..............9 CHAPTER 1 INTRODUCTION..................................................................................................................11 Introduction................................................................................................................... ..........11 Background..................................................................................................................... ........11 Present Study.................................................................................................................. ........15 Layout of Dissertation......................................................................................................... ...16 2 THEORETICAL PERSPECTIVES........................................................................................18 Introduction................................................................................................................... ..........18 What is a Community?...........................................................................................................20 Social Disorganization.....................................................................................................21 Age-Crime Relationship a nd Criminal Careers...............................................................26 Developmental Criminol ogy/Life-Course Theory..........................................................33 Sampson and Laubs Age-Graded Theory of Informal Social Control...........................36 Moffitts Taxonomy........................................................................................................40 3 REVIEW OF THE LITERATURE........................................................................................45 Introduction................................................................................................................... ..........45 Impact of Social Disorganiza tion at the Individual Level......................................................45 Prevalence and Incidence of Chronic Offending.............................................................50 The Effect of Life Transitions on Offending...................................................................54 The Saliency of Moffitts Taxonomy across Sex............................................................58 Prior Applications of GroupBased Models of Offending..............................................61 Integrated Theoretical Model for Explaining Offending.................................................65 Hypotheses and Expectations..........................................................................................68 4 DATA AND METHODS.......................................................................................................70 Introduction................................................................................................................... ..........70 Data Sources................................................................................................................... ........71 Independent Variables.....................................................................................................73 Macro-level independent variables..........................................................................73

PAGE 5

5 Micro-level indepe ndent variables...........................................................................74 Dependent Variables.......................................................................................................81 Study Limitations.............................................................................................................. ......82 Methodology.................................................................................................................... .......85 Group-Based Modeling...................................................................................................85 A Seductive or Useful Statistical Method?.....................................................................90 Analytic Procedure..........................................................................................................96 Trajectory estimation................................................................................................96 Multinomial logistic regression................................................................................97 Study Strengths..............................................................................................................101 5 RESULTS........................................................................................................................ .....103 Introduction................................................................................................................... ........103 Descriptive Statistics......................................................................................................... ...103 Independent Variables...................................................................................................103 Dependent Variable.......................................................................................................105 Trajectory Analysis............................................................................................................ ...105 Analysis of Variance........................................................................................................... ..116 Multivariate Analysis.......................................................................................................... ..118 Conclusions.................................................................................................................... .......122 6 DISCUSSION AND CONCLUSION..................................................................................144 Discussion..................................................................................................................... ........144 Limitations.................................................................................................................... ........147 Directions for Future Research.............................................................................................150 Implications................................................................................................................... .......153 Final Thoughts................................................................................................................. .....156 APPENDIX: ZERO-ORDER CORRELATIONS BY COHORT...............................................158 REFERENCES..................................................................................................................... .......159 BIOGRAPHICAL SKETCH.......................................................................................................181

PAGE 6

6 LIST OF TABLES Table page 4-1 Resident Samples by Cohort, Size, Data Sources, and Offender Groups........................102 4-2 Observational Periods of Data Sources by Cohort..........................................................102 4-3 Social Disorganization Inde x Factor Loadings by Cohort...............................................102 5-1 Descriptive Statistics for the Macrolevel and Micro-level Measures by Cohort...........126 5-2 Tabulated Bayesian Information Crit eria (BIC) for 1942 Racine Continuous Residents...................................................................................................................... ....126 5-3 Mean (Median) Posterior Probabili ties for Group Assignments: 1942 Racine Continuous Residents.......................................................................................................127 5-4 Coefficient Estimates for Effect of Sex (being Male) on Trajectory Group Membership: 1942 Racine Continuous Residents...........................................................127 5-5 Tabulated Bayesian Information Crit eria (BIC) for 1942 Racine Continuous Interviewed Residents......................................................................................................127 5-6 Mean (Median) Posterior Probabili ties for Group Assignments: 1942 Racine Continuous Interviewed Residents...................................................................................128 5-7 Coefficient Estimates for Effect of Sex (being Male) on Trajectory Group Membership: 1942 Racine Continuous Interviewed Residents.......................................128 5-8 Tabulated Bayesian Information Crit eria (BIC) for 1942 Racine Continuous Interviewed Residents with Block Data ...........................................................................128 5-9 Mean (Median) Posterior Probabili ties for Group Assignments: 1942 Racine Continuous Interviewed Residents with Block Data ........................................................128 5-10 Coefficient Estimates for Effect of Sex (being Male) on Trajectory Group Membership: 1942 Racine Continuous Interviewed Residents with Block Data ............129 5-11 Tabulated Bayesian Information Crit eria (BIC) for 1949 Racine Continuous Residents...................................................................................................................... ....129 5-12 Mean (Median) Posterior Probabili ties for Group Assignments: 1949 Racine Continuous Residents.......................................................................................................129 5-13 Coefficient Estimates for Effect of Sex (being Male) on Trajectory Group Membership: 1949 Racine Continuous Residents...........................................................130

PAGE 7

7 5-14 Tabulated Bayesian Information Crit eria (BIC) for 1949 Racine Continuous Interviewed Residents......................................................................................................130 5-15 Mean (Median) Posterior Probabili ties for Group Assignments: 1949 Racine Continuous Interviewed Residents...................................................................................130 5-16 Coefficient Estimates for Effect of Sex (being Male) on Trajectory Group Membership: 1949 Racine Continuous Interviewed Residents.......................................130 5-17 Tabulated Bayesian Information Crit eria (BIC) for 1949 Racine Continuous Interviewed Residents with Block Data ...........................................................................131 5-18 Mean (Median) Posterior Probabili ties for Group Assignments: 1949 Racine Continuous Interviewed Residents with Block Data ........................................................131 5-19 Coefficient Estimates for Effect of Sex (being Male) on Trajectory Group Membership: 1949 Racine Continuous Interviewed Residents with Block Data ............131 5-20 Trajectory Groups by Sample and Cohort.......................................................................132 5-21 Mean Covariate Levels by Traj ectory Group Membership and Cohort..........................133 5-22 1942 Racine Continuous Interviewed Residents with Block Data : Odds Ratios and Probabilities for Macro-Level Predic tors of Offending Group Membership..................134 5-23 1942 Racine Continuous Interviewed Residents with Block Data : Odds Ratios and Probabilities for Micro-Level Predic tors of Offending Group Membership...................134 5-24 1942 Racine Continuous Interviewed Residents with Block Data : Odds Ratios and Probabilities for Macro-level and Micr o-level Predictors of Offending Group Membership.....................................................................................................................135 5-25 1949 Racine Continuous Interviewed Residents with Block Data : Odds Ratios and Probabilities for Macro-Level Predic tors of Offending Group Membership..................136 5-26 1949 Racine Continuous Interviewed Residents with Block Data : Odds Ratios and Probabilities for Micro-Level Predic tors of Offending Group Membership...................136 5-27 1949 Racine Continuous Interviewed Residents with Block Data : Odds Ratios and Probabilities for Macro-level and Micr o-level Predictors of Offending Group Membership.....................................................................................................................137 A-1 1942 Cohort................................................................................................................ .....158 A-2 1949 Cohort................................................................................................................ .....158

PAGE 8

8 LIST OF FIGURES Figure page 3-1 Integrated Theoretical Model: Social Di sorganization, An Age-Graded Theory of Informal Social Control and Mo ffitts Developmental Taxonomy...................................69 5-1 1942 Racine Continuous Reside nts Group-Based Trajectories.......................................138 5-2 1942 Racine Continuous Interviewed Residents Group-Based Trajectories...................139 5-3 1942 Racine Continuous Interviewed Residents with Block Data Group-Based Trajectories................................................................................................................... ...140 5-4 1949 Racine Continuous Reside nts Group-Based Trajectories.......................................141 5-5 1949 Racine Continuous Interviewed Residents Group-Based Trajectories...................142 5-6 1942 Racine Continuous Interviewed Residents with Block Data Group-Based Trajectories................................................................................................................... ...143

PAGE 9

9 Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy TRAJECTORIES OF TWO RACINE BI RTH COHORTS: A THEORETICALLY INTEGRATED MODEL FOR EXPLAINING OFFENDING By Wesley G. Jennings December 2007 Chair: Lonn Lanza-Kaduce Cochair: Alex Piquero Major: Criminology, Law, and Society Macro-level research evidencing the relationship between communities and crime dates back to the early ecological work of Park and Burgess and later further developed by Shaw and McKay into what is commonly referred to as soci al disorganization theor y. This research has since shown that poverty, ethnic heterogeneity, and residential mobility lead to the inability for members of the community to develop mechanisms of informal social control, which in turn leads to higher ra tes of crime. In contrast, micro-level theories such as Sa mpson and Laubs theory argues that individual pathways of offending can be altered by life tr ansitions such as marri age and employment. Comparatively, Moffitts taxonomy asserts that ther e are two theoretically distinct groups of offenders, where one group displays an adolesce nt-limited type of offending and the other group are considered to be life-course-persistent or chronic offenders. Moffitt has also argued that the life-course-persistent type is predominantly a male phenomenon. This study will provide one of th e first attempts to explore th e possibility of developing a theoretically integrated model for explaining gr oup offending using trajectory analysis and relying on data from two Racine Birth Cohorts. Re sults from this study relying on official data

PAGE 10

10 yielded varying levels of suppor t for all of the theoretical frameworks mentioned above. Specifically, all of the trajectory models discove red groups consistent with Moffitts typology, as well as uncovering groups not necessarily consis tent with her typology. Also consistent with Moffitt, sex was shown to be a significant cova riate for predicting group assignment. Bivariate support was found for Sampson and Laubs age-graded th eory of informal social control. Lastly, multivariate support was revealed for the impact of living in a socially disorganized area. Yet, once mlogit models were estimated incorporating the individual and neighborhood effects there was evidence of mediation. It appeared that in dividual and family variables at times completely or at least partially mediated th e effect of social disorganizatio n as a distinguishing feature of group membership. Implications for the current fi ndings and directions for future research are also discussed.

PAGE 11

11 CHAPTER 1 INTRODUCTION Introduction There exists a large body of research that has identified the association between the variation in the rate of offenders and community structural dynamics, most notably in the social disorganization literature (Bursik & Grasmick, 1993). Similarly, there has been a long line of research that has linked indivi dual-level factors to involvement in offending (Farrington, 1995). Yet, few efforts combine these levels of analysis On this score, Samp son (1997:32) provided an accurate summary of the current state of research in integrated approaches to studying offending in noting that few studies have successfully de monstrated a unified approach to individual and neighborhood-level dimensions of crime. The primary focus of this dissert ation is to present empirical da ta that will shed light on the question of the extent to which macro-level, micr o-level, or some combination of factors best relates to offending over the life course. In orde r to answers these questio ns, this study relies on extensive longitudinal data from two birth c ohorts of Racine, Wisconsin residents (Shannon, 1988, 1991) and utilizes a novel methodology for identif ying latent classes of offenders. After classifying the Racine residents into offender gr oups, a series of multinom ial regression models will allow for an assessment of the relative strength macro-level and micro-level factors for increasing and/or decreasing the likelihood of belonging to one tr ajectory group as opposed to another. Background While a handful of studies have compared ra tes of offending with individual/family SES and area (or school) SES and have found that th e highest rates of of fending generally were observed in the lowest SES areas among the indi viduals with the lowest SES (Braithwaite, 1979;

PAGE 12

12 Reiss & Rhodes, 1961; Wikstrom, 1 991a), only a few studies have i nvestigated the influence of other individual and community covariates on o ffending. Lindstrom (1996) found that the level of family interaction was significantly associated with juvenile offendi ng and this relationship was the most pronounced in more socially di sadvantaged neighborhoods. Kupersmidt et al. (1995) showed that juveniles living in singleparent families in low SES neighborhoods were more likely to be aggressive than their c ounterparts in high SES neighborhoods. Relying on a comparison of two different groups of P ittsburgh neighborhoods (i.e., underclass and nonunderclass), Peeples and Loeber ( 1994) indicated that the relati onship between race and juvenile offending was only apparent in the underclass neighborhoods. Even less research has targeted the integr ation of individual and neighborhood effects on offending under the auspices of the criminal career paradigm (Blumstein et al., 1986) and recent developmental frameworks. Wikstrom (1991b) conduc ted one of the first st udies to address this deficiency in the existing research and found that individuals that resided in the most socially disadvantaged areas had the highe st levels of both participati on and frequency of offending. Loeber and Wikstrom (1993) also showed that no t only did juveniles offend at higher rates (i.e., greater participation and frequency) in low SE S neighborhoods when compared with those from neighborhoods with higher SES a si gnificantly larger proportion of the offenders in the low SES neighborhoods became serious offenders. In a more recent study, Wikstrom and Loeber (2000) examined whether the interaction of risk and/or protective factors and community cont ext had an influence on early and late-onset offending among three samples of boys in 90 Pittsburgh neighborhoods. They found that neighborhood socioeconomic context did not have a direct effect on the prevalence of offending for boys who had a high number of risk factors, but for those who had balanced risk-protective

PAGE 13

13 factor scores, neighborhood contex t mattered. In addition, they i ndicated that individual riskprotective scores had a greater im pact on the rate of early-onset offending (before age 13) than did neighborhood context, but th at a direct neighborhood effect was observed for late-onset offending (after age 12). Lynam et al. (2000) found that the effects of impulsivity on juvenile offending was stronger in poor neighborhoods; however, livi ng in a disadvantaged neighborhood did not heighten nonimpulsive boys risk for juvenile of fending when compared with residing in more advantaged neighborhoods. Lynam et al. (2000) also noted that the direct neighborhood effects on individual juvenile delinquency were relativ ely weak, whereas the neighborhood effects were stronger when using self-reported measures of the quality of the neighborhood obtained directly from the inner-city youth. In a more recent study, Piquero, Moffitt, a nd Lawton (2005) investigated the possible interaction effects of individual, familial, and neighborhood-level factors on life-coursepersistent offending. Their data were derived from the National Collabo rative Perinatal Project (Baltimore site) and consisted of information fr om the original NCPP study that included early childhood and census-tract measures along with addi tionally including follow-up interviews with the children in the 1990s. Piquero et al. ( 2005) failed to find direct effects for either neighborhood or individual risk for life-course-persistent offendi ng; however, they did find that individual, family, and neighbor hood-level effects predicted li fe-course-persistent offending patterns among African-Americans in disadvantage d neighborhoods. In a ddition, while Piquero et al. (2005:230) concluded that their analysis demonstrated si milarities across race in lifecourse-persistent patterns of offending, ther e were marked differences in the neighborhood contexts where the Whites and African-Americans lived.

PAGE 14

14 With attention to the relative arbitrariness of how prior research has defined offender groups, more recent research applying Nagin a nd Lands (1993) group-based modeling technique has begun to identify latent classes of offenders th at exhibit distinct pathways of offending (for review see Piquero, 2007; see also Nagin, 2005). The consistent finding across these studies has been the identification of three main type s of offending groups: (1) non-offenders; (2) adolescent-limited offenders; and (3) chronic or high-rate offe nders. DUnger, Land, McCall, and Nagin (1988) were also ab le to discover similar offendi ng types across a variety of longitudinal data sources. It is important to emphasize here that the data relied on in this study is official data not self-report; therefore it is likel y that the non-offender group will be much larger than it would be if the trajectories were estimated using self-report data instead. Additionally, DUnger, Land, a nd McCall (2002) found evidence of relatively comparable latent offending classes using data from the Second Philadelphia Birth Cohort (Tracy, Wolfgang, & Figlio, 1990); however, the ad olescent-limited group they unc overed was characterized as adolescent-peaked because of their elevated levels of of fending in adolescence. This adolescentpeaked grouped has also been noted by other res earchers (for review, se e Piquero, 2007). Yet, other groups have also emerged su ch as low-level chronics or la te-onsetters (see DUnger et al., 1998, 2002). Prior applications of this methodology have al so found similar offender groups in male and female trajectories (Cote et al., 2001, 2002; DUnger et al., 2002; Fergusson & Horwood, 2002; Land, 2000; Piquero, Brame, & Moffitt, 2005) and ot her research has identified factors other than sex such as low socioeconomic status (McDermott & Nagin, 1998) and peer and school influences (Chung et al., 2001) as being strongly related to the likelihood of be ing classified as a chronic offender. Land (2000) has also shown th at sex was a significant predictor of chronic

PAGE 15

15 offending patterns, although its effect was comple tely mediated by the inclusion of neighborhood context. Present Study This current study contributes to the literature in following ways First, by drawing from the present advancements in the criminological lite rature regarding the application of trajectory analysis for classifying offenders (Nagin & Land, 1993; Nagin, 2005), this study seeks to provide another replication of th e latent class groups that have been previously identified by prior researchers (for review, see Piquero, 2007) such as the non-offenders, adolescentlimited/peaked offenders and chronic offenders across two Racine Birth Cohorts. Second, this study seeks to investigate whether sex is a disti nguishing feature of groupbased trajectories as has been discussed in prior applications of this group-based modeling techni que (see Cote et al., 2001, 2002; DUnger et al., 2002; Fergusson & Horwood, 2002; Land, 2000; Piquero et al., 2005). Third, this research will include an explora tion into the impact that social disorganization has on individual offending patterns. Four th, guided by two prominent developmental theoretical frameworks, specifica lly Sampson and Laubs (1993) ag e-graded theory of informal social control and Moffitts (1993) taxonomy of adolescent-limited and life-course-persistent offenders, this study seeks to explore the impact of key life transitions (marriage, employment) on desistance from offending in adulthood and assess the viability of Moffitts taxonomy. Fourth, the saliency of Moffitts (1994; see also Moffitt et al., 1994) claims that males are more likely to be life-course-persisten t offenders than females will al so be explored. Finally, this study seeks to address the rather limited amount of research integrati ng macro-level and micro-

PAGE 16

16 level factors that may simultaneously affect an individuals pathways of offending using data from two of Shannons (1988, 1991) Racine Birth Cohorts.1 Layout of Dissertation This dissertation will consist of six chapters. Chapter 2 will begin with a discussion of the various theoretical frameworks th at provide the basis for the st udy. More specifically, it will begin with a brief description of how scholar s have defined a community over time followed by an outline of the origins and theoretical tenets of social disorganization theory. Next will be an explanation of the relationship between ag e and crime and how research on the age-crime curve complements the emergence of the criminal career paradigm. Similarly, the narrative will continue with a discussion of how the criminal career paradigm had a dire ct contribution to the development of what has become known as deve lopmental/life-course cr iminology. This chapter will conclude with a presentation of Sampson and Laubs (1993) age-graded theory of informal social control (see also Laub & Sampson, 1993) and Moffitts (1993) taxonomy of offending, the latter providing the theoretical rationale for hypothesizing why diffe rent classes of offenders may exist and the former describing certain life transi tions that may affect an individuals offending trajectory. Chapter 3 will continue by offering an in-depth exploration into the prior extant literature that can be found examining each of the theoreti cal frameworks mentioned above. Most notably, this chapter will illustrate the links between so cial disorganization and individual patterns of offending, along with addressing the prior rese arch on chronic offending, crime over the lifecourse (i.e., Sampson and Laubs age-graded theo ry of informal social control), and offer a 1 This study relies on data from the 1942 and 1949 Racine Birth Cohorts, as opposed to all three of the cohorts (1942, 1949, 1955), because there is not any self-reported interview data available for the 1955 cohort precluding its use for investigating the particular individual-level factors addressed in this research. According to Shannon (1988, 1991), lack of continued federal funding prevented him from interviewing the most recent 1955 cohort.

PAGE 17

17 description of studies that have supported Moffitts (1993 ) proposed taxonomy and also demonstrated evidence that life-course-persi stent offending may be disproportionately found among males. This chapter concludes with a pr esentation of the findings from several recent applications of Nagin and Lands (1993) met hodological technique for identifying group-based offending trajectories. Chapter 4 will provide a systematic explanat ion of the source of the data and their limitations for the current rese arch along with an in-depth description of the 1942 and 1949 Racine Birth Cohorts used in the analysis. Th e sample and data information will be followed by a comprehensive identification of all of the meas ures incorporated in this study including the macro-level and micro-level fact ors and the two dependent measur es (number of police contacts and offending group membership). The final section of this chapter will expl ain the stages of the analysis as far as the ordering of the analytic procedures and the model specifications and will conclude with a brief discussion of the strengths of the study. Chapter 5 will present the findings from all of the statistical analyses including the results from (1) the group-based trajectory estimations a nd model selection criter ia; (2) a discussion of the importance of sex in predicting group assign ment; (3) description of factors that are associated with group assignment; and (4) an ex amination of the impact of macro-level and micro-level variables (both se parately and simultaneously) fo r significantly distinguishing between trajectory groups. The fi nal chapter (Chapter 6) will be gin with a further discussion on the findings as well as limitations of this resear ch. This section will be followed by an offering of some suggestions and guidance for future research on mutli-level modeling and theoretical integration and will conclude with a few final thoughts.

PAGE 18

18 CHAPTER 2 THEORETICAL PERSPECTIVES Introduction The origins of macro-level res earch, specifically social disorg anization, can be traced back to the early work in the first half of the twentieth century by Park and Burgess (1925) and the later expansions by Shaw and McKay (1942). These researchers and others that followed in their footsteps have formed what is commonly refe rred to today as the Chicago School. These researchers, most notably Shaw and McKay, we re influenced by the d eep-rooted sociological tradition of studying the eco logy of crime and focused on explaining why crime rates not individual criminality was higher in certain areas of the co mmunity. This tradition of studying and theorizing about crime in the aggregate persisted for some time until there became an increased interest in studying i ndividual-level factors that a ffect offending, which was strongly influenced by the increased use of survey me thodology in sociology and criminology. This early individual-level research sparked some of the most influential criminological theories that exist today including social learning theory (Burgess & Akers, 1966; Akers et al., 1979) and social control/bonding theory (Hirschi, 1969). By the late 1970s and early 1980s individual-leve l research had evolved in the sociological and criminological literature and led to what is still considered one of the most robust relationships in the social scienc es, i.e., the age-crime relationshi p, or the notion that crime rises to a peak in middle to late adolescence follo wed by a precipitous decl ine in early adulthood. Hirschi and Gottfredson (1983) argued that there is no need for any theory to account for the relationship between age and crime because it is invariant across culture, time, and space. Although, there has been a relatively strong amount of debate over this early a ssertion (which will be returned to later), th is along with evidence from Wolfgang et al.s (1972) research on

PAGE 19

19 chronic offending fueled the development of the criminal career paradigm. This perspective involves a series of dimensions that will be addressed below, but most importantly, this framework highlighted the similarities and di fferences among offenders and their criminal careers (see Blumstein et al., 1986). Coinciding with this new research interest was the birth of developmental criminology or life-course criminology. This theoretical pers pective focuses on how involvement in crime changes over the life course. It hi ghlights factors such as life tran sitions or turning points that can alter an individuals o ffending trajectory (Elder, 1985). In addition, Sampson and Laub (1993) formulated an age-graded th eory of informal social contro l that recognizes that different types of individuals have different offending ca reers and that there ar e age-dependent or agegraded processes that influence, shape, or al ter ones pathway of offending. In comparison, one of the most recognized developmental theories of crime (other than Sampson and Laubs agegraded theory of informal social control) is that of Moffitts (1993) taxonomy of adolescentlimited and life-course-persistent offenders. Moffitt argues that there are three theoreti cally and qualitatively distinct groups of individuals including: (1 ) the non-offenders; (2) the adolesce nt-limited offenders; and (3) the life-course-persistent offenders. Furthermore, she also argues that the patterns of offending for the two latter groups (i.e., the groups who are actually involved in o ffending) are driven by different causal processes, have different ages of onset, and are particular ly distinct in their frequency of offending and career lengths. All of the theoretical issues touched upon above will be the focus for the remainder of this chapter, but it is first important to briefly visit the debate among scholars (sociologists and criminologists alike) on what is a community or neighborhood.

PAGE 20

20 This definitional disagreement over the appropriat e unit of analysis has been at the heart of macro-level research for some time. What is a Community? Community theory gained increased attention w ithin the realms of sociology in response to the ecological work of Park and his colleagues at the University of Chicago in the 1920s and 1930s. Their study of the dynamics of urban life in Chicago supported th e idea that variables such as population size, density, and heterogene ity interact to produce communities that can be characterized as urban. Following a brief decl ine in community-based sociological research in the 1950s and 1960s, the 1970s and 1980s hosted a gene ral resurgence in atte ntion devoted to the community/neighborhood debate. This movement was fueled by research that relied on using local community areas (Heitgerd & Bursik, 198 7; Bursik, 1984, 1986), census tracts or some aggregation thereof (Schuerman & Kobr in, 1986; Shannon, 1988, 1991; Taylor & Covington, 1988), police districts (Shannon, 1988, 1991), and ev en electoral wards (Sampson & Groves, 1989) as the units of analysis. The consistenc y among the findings that were produced based on these varying aggregations and definitions of community/neighborhood wa s that there existed significant differences in local politics, economic s, and lifestyles between communities and that the world is no longer comprised of small isolated groups of huma n beings scattered over a vast territory (Wirth, 1938:2). An extension to the community/neighborhood definitional debate and its historical origins was more recently provided in Burs ik and Grasmicks (1993) classic work, Neighborhoods and Crime: The Dimensions of Effective Community Control The authors began with the contention that although consensus had ye t to be obtained regarding the definition of neighborhoods, there were several ch aracteristics that the majority of sociologists would indicate that delineated a neighborhood includ ing: (1) a small physical area wi thin a larger area; (2) some

PAGE 21

21 sort of common interest in the preservation of the inhabited area; and (3) a collective identity that tends to be constant over tim e (Bursik & Grasmick, 1993:6). Bursik and Grasmick (1993) further argued that social control may exist within communities in different forms or what they id entify as private, parochial, and/or public networks. However, they did contend that all communities tend to operate with the same underlying desire for informal and formal associa tions for eliciting social control in order to minimize crime (bearing in mind that particular crim es are likely to be differentially viewed as unconventional as well as differentially enforced across communities). Although, the community/neighborhood debate may be unresol ved, but there appears to be a movement toward fleshing out the strengths of its definiti onal attributes and providing a more encompassing characterization. This concept of community and the process of identifying particular factors that are associated or lead to crime were the basis for wh at has emerged as of one of the classic theories of crime, or what is commonly known in the disc ipline as social disorgan ization theory. This next section presents the historic al roots of the theory and lays out the core assumptions behind why crime should be higher in socially disorganized communities. Social Disorganization Early macro-level research inve stigating the importance of ecological characteristics for explaining crime rates has its origins from the Ch icago School and the work of Park and Burgess (1925) and Shaw and McKay (1942). The work of the latter scholars focused on attempting to offer a conceptual framework to explain the st ability of aggregate cr ime/delinquency rates over time in the city of Chicago despite the dynamic ch ange in the racial and ethnic composition of its neighborhoods (Shaw & McKay, 1942). Shaw and Mc Kay argued that the majority of the crime tended to be concentrated in pockets of the city that Park and Burgess (19 25) characterized as the

PAGE 22

22 zones of transition that surround th e central business district. Th e negative inverse relationship that Shaw and McKay observed between the ne ighborhoods spatial distance from the central business district and their crime rates provide d the framework for their theory of social disorganization. More specifically, their theory emphasizes the concentration of high levels of population turnover, racial and/ or ethnic heterogeneity, and poverty in the neighborhoods that exhibit the highest rate s of crime/delinquency. At the fundamental level, social disorganizati on is assumed to be the causal factor between neighborhood characteristics and o ffending patterns at the aggreg ate level because it emphasizes the importance of the ability of a communitys re sidents to realize shared goals and values and regulate the conduct of its citizen s (i.e., informal social contro l) (Berry & Kasarda, 1977; Kornhauser, 1978; Sampson & Groves, 1989; Sampson, Raudenbush, & Earls, 1997). High population turnover (i.e., residentia l instability) as identified by Shaw and McKay (1942), is seen as a structural antecedent that may impede the ab ility of communities to organize effectively and to self-regulate residents behavior because of th e constant out-migration that characterizes highcrime communities. This high rate of reside ntial mobility is believe d to produce a social atmosphere where many residents are uninterested in communities they hope to leave at the first opportunity (Kornhauser, 1978:78). This situation leads to th e potential failure of the communitys residents to forge interpersonal ties or even weak ties which some scholars have argued are the most important (see Bellair, 1997 ; Granovetter, 1973), because of the prevailing opinion that either they or their neighbors will soon be leaving the area. Racial and/or ethnic heter ogeneity is the second structur al characteristic recognized by Shaw and McKay (1942) that may perhaps hinder the ability of communities to develop mechanisms of informal social control. It is also causally linked to the high rates of population

PAGE 23

23 turnover mentioned above. This connection is prim arily due to the fact that in communities that are characterized by rapid compositional change, it becomes difficult to promote a unified front against the invasion of potentially thr eatening groups (Bursik & Grasmick, 1993:33). Furthermore, heterogeneity can inhibit comm unication, especially in communities that are plagued by language barriers. This will more th an likely make it increasingly difficult for the residents to solve or even express their co mmon problems and goals (Kornhauser, 1978:75). Shaw and McKay (1942) also indicated that high rates of poverty/low socioeconomic status are characteristic of socially disorgan ized communities. According to Shaw and McKay, this situation was attrib utable to the constant out-migration to the suburbs of residents that had the ability to move, and the area that the former residents left behind stil l tended to be plagued by poverty and increased crime/delinque ncy rates, regardless of which racial/ethnic group settled into the area. These individuals who lack the opportunity and/ or resources to escape this socially disorganized area (i.e ., the inner city) have been referred to by Wilson (1987) as the truly disadvantaged because they have to live in areas of extreme concentrated poverty and with a high density of subsidized housing. This dens ity of subsidized and predominantly renteroccupied housing has also been associated with cr ime rates (see Schuerman & Kobrin, 1986). Schuerman and Kobrin (1986) have indicated that although little is known about how neighborhoods evolve into high-cr ime areas, prior research has evidenced distinct stages of neighborhood development that typify the recipro cal effects of crime a nd community decline. Emerging is the first stage where neighborhood deterioration precedes the high-crime rates, but once communities move through the transitional stage and into the last stage (i.e., enduring phase), then the high-crime ra tes precede the further deterioration of the neighborhood. A host of factors has been at tributed to neighborhood de terioration including: movement from single to

PAGE 24

24 multiple family dwellings, residential mobilit y, broken homes, rise in young adult population, minority group relations, female participation in the work force, and even advanced education. In addition, according to Taylor (1995), prior scholars have contended that community decline is primarily the result of residentia l mobility or at least the reside nts desire to move from their current area. This resident turnover has been connected to many adverse neighborhood outcomes including weaker attachments among residents, lower sa tisfaction with the neighborhood, less local involvement, a nd decreased housing values. Although not explicitly referenced by Sh aw and McKay in the formulation of their theory, the relative inability of researchers to perhaps appropriately interpret Shaw and McKays underlying assumptions and/or dise ntangle the reciprocal effect s of social disorganization meaning that the factors identified above may both influence crime rates and foster social disorganizationled to a shift in the discipline in favor of indi vidual-level theories of crime causation (Land, 2000). However, several researcher s have made recent attempts to revitalize the work of Shaw and McKay (1942) and better define the causal pathwa ys that exist between neighborhood characteristics and offending (see Bursik & Grasmick, 1993; Sampson & Groves, 1989). Bursik and Grasmick (1993) put forth an atte mpt to define their systemic theory of neighborhood organization for explaining crime. At the fundamental level, systemic theory is predicated upon the ability of neighborhoods to control themselves and their surrounding environment in order to minimize their likeli hood of victimization a nd reduce crime. The authors argue that formal and informal mechanisms of social control are equally important if the neighborhood is to be successful in reducing crime.

PAGE 25

25 The private, parochial, and public spheres are regarded as the three specific elements that are important in eliciting th e necessary controls on the behaviors of the neighborhoods members. The private level of control is expressed through the strong intimate informal networks that exist within the communities th at are able to reduce crime by threatening or withdrawing social support from those members who engage in inappropriate or non-normative behavior. The parochial level of control is directly tied to the ability of so cial institutions such as the school and church to influence behavior th rough the shared sentimen ts of the neighbors for reducing victimization risk. The public level of control is perhaps the most difficult to obtain and emphasizes the ability of the neighborhood to mobilize government resources at the local level and acquiring crime controllin g agents (i.e., police) as an ex ternal resource to reduce the crime in their neighborhood. These three mechan isms are believed to manifest themselves through neighborhood-level in teractions over time; however, the greatest impediment for the formation of the ties is residential instabil ity, which presents a d ilemma being that those communities who need these three mechanisms the most will be the least capable of developing them (Bursik & Grasmick, 1993). Sampson and Groves (1989) also emphasized the importance of family disruption when examining the effect of social disorganization on crim e. They argued that marital and family disruption (i.e., divorce and singl e-parent households) may decrea se the ability for community residents to develop informal soci al controls. This assumption is based on the belief that twoparent households provide stabili ty and increased supervision fo r their children. Two parents also allow for more supervision over their person al property. Sampson (1987) has shown rather large effects for family disruption on rates of juvenile crime. Although, Sampson and Groves (1989) demonstrated that social disorganizatio n (as measured by sparse friendship networks,

PAGE 26

26 unsupervised teenage peer groups, and low organi zational participation) mediated a rather significant portion of the direct effect of family disruption on crime/delinquency rates as well as mediating the effects of other structural characte ristics such as classical social disorganization measures (low SES, residential mobility, and ethnic heterogeneity). In sum, the social disorganization fram ework provides a backdrop for explaining why crime rates may be higher in communities that are characterized by hi gh population turnover, ethnic/racial heterogeneity, and re sidential instability, but just how social disorganization can also be expected to have an affect on indivi dual outcomes will become clearer in the next Chapter (Ch. 3). The next section (along with th e remaining sections in Chapter 2) illustrates a complete shift in theoretical framework by not focusing on explanations of crime in the aggregate but by presenting a series of criminol ogical paradigms and micro-level developmental theories that have been offered to e xplain individual variation in crime. Age-Crime Relationship and Criminal Careers Perhaps the most commonly accepted fact about crime is the association between age and crime (Cohen & Land, 1987; Hirschi & Gottf redson, 1983; Hirschi & Gottfredson, 1985; Farrington, 1986a; Greenberg, 1985; Obrien, St ockard, & Isaacson, 1999; Sampson & Laub, 1990, 1992; Shavit & Rattner, 1988; Steffensme ier, Allan, Harer, & Streifel, 1989; Steffensmeier, Streifel, & Shihadeh, 1992). The progr ession of what has been referred to in the literature as the age-crime curve is that the crime rate, as measured by arrest, typically increases from the minimum age of criminal re sponsibility and peaks in the teenage years followed by a decline in early adulthood (Farri ngton, 1986a). Although th is general statement accurately characterizes the age-cr ime curve, in order to fully understand the complexity the agecrime relationship it is necessary to revisit Hi rschi and Gottfredsons ( 1983) original article discussing this phenomenon entitled Age and the Explanation of Crime

PAGE 27

27 Hirschi and Gottfredson (1983:54), in response to what they considered to be the misapplication of the age-crime distribution by schol ars in criticizing social theories of crime and supporting longitudinal methods over cross-sectional designs, assert that the effect of age is invariant across time, place, social and cultural conditions, and crime type. They indicated that rates of delinquency have been comparable over the last 150 years and that the age distribution of delinquency in England and Wales in the 1940s, Argentina in the 1960s, and the United States in the 1980s are virtually indi stinguishable. They further argued that although non-White offenders may start their delinque ncy earlier and offend at a highe r rate than Whites and that while boys offend at higher rates th an girls for particular offenses the age-crime curve is still invariant across race and sex (1983:556). In addition, they provided evidence against the differences in the age-distributions of property and person offenses. They illustrated that while official data may reveal diffe rent age-crime curves for pers on and property offenses, this difference is not observed in self -report data; therefore, the effects of age are indeed invariant across crime type. Hirschi and Gottfredsons (1983) early claims about the invariance of the effects of age across historical and cultural cond itions and crime type were s oon met by criticism. Farrington (1986a) specified that the main limitation of inte rpreting the types of age-crime curves used by Hirschi and Gottfredson (1983) is th at there is probable ambiguity in differentiating between age and period effects. Period eff ects are significant factors that ar e time-period specific, such as eras of high unemployment, riots, and/or civil unrest, and these influenc es are only related to changes in the crime rates of those cohorts that are aligned with that particul ar era in history. In contrast, age effects are universal changes that occur with age (i.e., natural maturation) that lead to physical deterioration for all cohort s during all periods of observation.

PAGE 28

28 Farrington (1986a) further argued that there is al so the issue of disent angling cohort effects when interpreting the age-crime re lationship. Cohort effects are in fluences that are due to an offenders membership in a specific cohort, su ch as belonging to th e baby-boomer generation. Taken together, all of these matters present prob lems because an individuals actual age is always measured in number of years and typically equals the current year (period) subtracted by birth year (cohort) (Farrington, 1986a). Furthermore, according to Farrington (1986a) the inherent characteristic of the age-crime curve that receives the least am ount of attention is that it reflects variations in the prevalence (how many individuals are involved in crime) rather than the incidence (how many crimes are committed) of offending (see also Blumstein et al., 1988). While age-crime curves for individuals are typically unimoda l (with only one peak in the teenage years), Farrington (1986a) argues that these curves do not mirror the aggreg ate curve that was referred to by Hirschi and Gottfredson (1983). Additionally, in dividual crime rates have been shown to be constant during a criminal career, thus providing evidence that ar rest rates do not always decrease with age for all offenders (Blumstein & Cohen, 1979). Farrington (1986a) also showed that aggregate age-crime curv es can be comparable, but it is their measures of distribution and central te ndency that vary across time, place, sex, and offense type. Nevertheless, despite the conflic ting arguments discussed above particularly between Hirschi and Gottfredson (1983) and Farr ington (1986a), there is virtually no definitive evidence refuting the str ong relationship between age and crime. Given the robustness of this relationship, recent attention has drawn research ers to focus on how criminal behavior changes throughout the various stages of life (Sam pson & Laub, 1990, 1992) and this approach has

PAGE 29

29 become characterized as the criminal career paradigm (Blu mstein, Cohen, Roth, & Visher, 1986). Blumstein et al. (1986) pointed out that the bulk of prior research in criminology was devoted to investigating crime using aggregate crim e rates and/or crimes pe r capita in the general population. This aggregate crime rate is the product of the particip ation rate of offending and the frequency of offending among the general populati on involved in crime, which according to Blumstein et al. makes studying individual criminalit y difficult. Hence, in an effort to better enable researchers to study diffe rential offending patterns, the crim inal career paradigm is unique because it divides the aggregate crime rate into two modules and has its own set of terminology and dimensions. The first module of the aggregate crime rate is participation in offe nding, or prevalence. Criminal career research distinct ly categorizes indivi duals into those who are criminally active and those who are not (Blumstein & Cohen, 1987). Participation m easurement is an expression of both the offenders involvement in crime and the range of the criminal acts committed during the observation period (i.e. serious or serious and non-serious offenses ). Regardless of the range of specified offenses or the particular observa tion period, active offenders always include new offenders (i.e., first offense occurs during obser vational period) and offenders who began their offending prior to the observati on period but remain criminally active during the period of study (Blumstein et al., 1986:18). Therefore, according to Piquero, Farrington, and Blumstein (2003:246), the longer the aver age duration of offending, the greater the c ontribution of persisters to measured participation in successive observation periods. According to Blumstein et al (1986), national estimates of participation in offending can range anywhere from a low of 25% to almost 45% depending on the operationalization of what

PAGE 30

30 constitutes an offense. Visher and Roth (1986) conducted one of the most comprehensive attempts to capture a true estimate of the participation rate of offending across the United States and Great Britain. They found that across both countries, th e average particip ation rate for non-traffic related offenses was approximately 30%. However, the rate fluctuated based on which type of criteria was employed, with the participation rate being the lowest when a stringent official measure was used such as conviction (West & Farrington, 1973, 1977) and the participation rate tended to be the highest when a more comprehensive official definition of offending was utilized such as police contact (Shannon, 1988, 1991). The second partition of the aggregate crime rate is frequency of offending. This refers to the average annual crime rate contributed by a pa rticular group of offe nders. Participation illustrates the pervasiveness of offending in th e general population, whereas the frequency refers to the intensity of th e individual offenders cr iminal activity (Blumste in et al., 1988). The frequency of offending is referred to as lambda ( ), in order to discriminate it from the aggregate rate used in the Uniform Crime Reports (U CR) (Blumstein & Cohen, 1987), and regarding lambda, most studies have found relative stabil ity over age, (Blumstein et al., 1986; Chaiken & Chaiken, 1982), sex (Blumstein et al., 1986), and crime type (P eterson & Braiker, 1980). The length of an individuals criminal caree r (duration) and the seriousness of offending are also important concepts in the criminal car eer paradigm (Blumstein et al., 1986:13). The duration of a criminal career indicates the total length of time of a criminals offending (Blumstein & Cohen, 1988). It represents the time that elapsed from the offenders initial offense to the time of his/her last offense (Blu mstein et al., 1986:13), or in other words the interval between an initiation a nd a termination in offending (Pique ro et al., 2003). Research has shown that there is a significant amount of variat ion in criminal careers, where many offenders

PAGE 31

31 criminal careers are brief and end in their teenage years and others continue to commit crimes up to and beyond middle age. Piquero, Brame, and Lynam (2004) have argued that the relative inattention of research regarding the duration of criminal careers is due to the difficulty in determining the true end of a criminal career. This difficulty has been addr essed by Blumstein et al (1986:18) where they stated the importance of distingu ishing total career length from residual career length, wherein the later refers to the estimated time remaining in an offenders criminal career given the specific observational period. Attention to this element is what Piquero et al. ( 2003) referred to as subsequently calculating estimates of career termin ation probabilities after an offender's arrest. Numerous studies (for review, see Piquero et al., 2003) have attempted to measure offender career lengths in actual number of years produci ng varying results ranging from a low of five years (Blumstein, Cohen, & Hsieh, 1982) to a relative high of 17.27 years (Piquero, Brame, & Lynam, 2004). Most recently, Kazemian and Farrington (2006) found evidence of a general decline in residual career length and offending frequency over time based on data from the Cambridge Study in Delinquent Development. The decline in residual career lengt h was observed in the original South London male sample and among th eir sons, although offending frequency did not decline as a result of conviction for the fathers. In addition, age of onset was associated with both residual career length and offending frequency, but the rela tionship was noticeably weaker among the fathers. The seriousness dimension of criminal career re search reflects both the number of offenses committed, as well as the particular patterns of switching (i.e., vers atility) among offenses (Blumstein et al., 1986:13). There are numerous t ypes of offenses in which criminals can engage

PAGE 32

32 in across the course of their career, as is there are wide variation among offenders in their levels of specialization and seriousness of offending. Offenders have been classified as either specialists (concentrating in only one offens e type or closely correlated offenses) or generalists (active in numerous types of re lated and non-related offe nses) and the types of offenses can also fluctuate, either escalating or de-escalating in seriousness (Blumstein et al., 1986:18). In a summary provided by Piquero et al. (2003:2 50) the authors note that most research has provided support for the generality of offending th roughout the duration of a criminal career and that specialization might occur more frequently among adult offenders versus juvenile offenders. In addition, there is some evidence in favor of offense versatility, although this switching tends to occur across clusters of logi cally related crime types (i.e., violence, property, and other). Also, minimal research has addressed the issue of co-offending patterns an d found that juveniles tend to commit crimes in groups, especially rega rding burglary and robbery, and males tended to co-offend with other males. Overall, criminal career research, comprised of all its relative dimensions, attempts to address questions of if how and why different factors influence the beginning of criminal activity, the continued participati on and prevalence of offending, and factors related to desistance or termination of criminal activity. This crimin al career paradigm also directly coincided with the movement toward developmental and life-cour se perspectives for explaining crime; however, the life-course approach differs from the traditio nal approach to criminal career research because it recognizes that there may be di fferent types of offenders that may have different types of careers (Farrington, 1999). The next section is de voted toward describing life-course theory and

PAGE 33

33 presents both sides of the debate regarding the persistent-heterogeneity versus state-dependent explanations of criminal behavior. Developmental Criminology/Life-Course Theory The life-course perspective can be defined as a series of pathways throughout the developmental process where there are varying ag e-dependent expectations and available options in an individuals decision-making process. Th ese decisions along with th e natural course of life events are seen to shape the trajectory of an i ndividuals life and can be influenced by transitions and turning points (Elder, 1985). Trajectories are viewed as pa thways of development over the life span such as work life, marriage, schooling, or involvement in crime. They are seen as longterm behavioral patterns that are marked by a sequence of life events and transitions. These transitions are then described as distinct life even ts that occur in shorter intervals that may alter an individuals behavioral trajec tory (such as first job, first ma rriage, first child, etc.) (Elder, 1985:31-32). Sampson and Laub (1990) extended this fram ework by focusing on the strong association between childhood events and adulthood experiences, a nd the idea that transiti ons can redirect an individuals life-course behavioral pathway. They indicated that life-course analysis focus on the length, timing, and the order of significant lif e events (including crime) and their effect on future social development (Sampson & Laub, 1992). Attention to the imp act of life events is critical because prior research has shown that committing a crime has a considerable behavioral influence on the likelihood of an individual comm itting crime in the future (Nagin & Paternoster, 1991). Loeber (1982) indicated that children who initially exhibi ted high rates of antisocial behavior were more likely to continue this behavi or than those who initia lly showed slower rates of antisocial behavior. Rely ing on data from 411 South L ondon males, Farrington (1995:956),

PAGE 34

34 concluded that offending may increase or decrease over time, but the worst offenders at one age still tend to be the worst at another age. This statement is essentially the basis of Gottfredson and Hirschis persistent-heterogeneity interpre tation (or static appro ach) of the relationship between past and future offending where they state that: The best predictor of crime is prior crim inal behavior [and] research shows that differences between people in the likelihood that they will commit criminal acts persist over timeand these differences appear early and remain stable over much of the life course (1990:107-108). Gottfredson and Hirschi (1990) expand upon this viewpoint in their di scussions regarding their theory of low self-control. According to Gottfredson and Hirsch i (1990: 97), the main underlying cause of low self-control is ineffective child-re aring. They argue th at it is the primary caregiver's (i.e., the parents) role to teach the child self-control through monitoring their behavior, recognizing deviant beha vior when it occurs, and eff ectively punishing the observed deviant behavior. These practices are believed to foster the child's development of self-control and an ability to delay gratificati on. The formulation of this core trait will also better enable the child to become more sensitive to the interests and desires of others, become more independent and willing to restrain their beha viors, and have a lesser propensity for using force or violence as a mechanism for resolving disputes or achieving their ends. Gottfredson and Hirschi (1990:90-91) further ar gue that the six primary elements that characterize individuals who exhibit low self-contr ol are impulsivity, insensitivity, physically as opposed to mentally oriented, short-sightedness, risk-takers, and displaying a low tolerance for frustration. Although these six co mponents are known to be distinct personality and behavioral traits, these various dimensions are all believed to be estab lished early on in childhood and coalesce within individuals to manifest a single construct that remains persistent throughout the life course. The fact that these elements fuse together to generate a unitary construct, what

PAGE 35

35 Gottfredson and Hirschi refer to as criminal pr opensity, does not necessarily mean a crime will occur. The theorists indicate that this propens ity will only cause an individual to engage in a crime or an analogous behavior once it converges w ith an opportunity to co mmit a criminal act. However, this temporal convergence of propensity and opportunity is not meant to argue the fact that it occurs without the indi vidual's knowledge of some social legal, or natural sanctions associated with their behavior. It is meant to indicate that individual's with low self-control will tend to behave criminally given the opportunity b ecause they typically perceive fewer negative consequences and have less to lose when comp ared with individuals with high self-control (Gottfredson & Hirs chi, 1990). In contrast to Gottfredson and Hirsch is (1990) viewpoint, a state-dependence interpretation (or dynamic approach) has also been presented to account for the positive association between past and future criminal be havior. This perspec tive asserts that prior participation in criminal activity reduces internal inhibitions and external constraints to future criminal activity and intensifies the prior offende rs motivation to commit future crime. Or in other words, the prior commission of a crime cause s a subsequent increase in the probability of an individual engaging in a future crime (Nagin & Paternoster, 1991; Nagin & Farrington, 1992). The state-dependence hypothesis al so recognizes the impact of life events on criminal behavior (Piquero, Brame, Mazerolle, & Haapanen, 2002). In support of this interpretation, Sampson and Laub contend that: Lives are often unpredictable and dynamic; e xogenous changes are ever present. Many changes in life result from chance or random events in individual liveswhile other changes in life directions stem from macro-level exogenous shocks (1995:50). Despite the existing points of contention betw een the two rival explanations, the results from Piquero et al. (2002) ba sed on data from 524 males re leased from California Youth Authority (CYA) institutions indi cated that criminal offending a ppears to involve a combination

PAGE 36

36 of the persistent-heterogeneity and state-dependence explanations. This finding led Piquero et al. to conclude that theoretical models must account for the relative contribution of both explanations in attempting to describe the ove rall progression of criminal behavior. Other research has also argued in favor of taking th e theoretical middle gr ound when studying crime (Paternoster & Brame, 1997; Paternoster et al., 1997). Additionally, previous resear ch has found evidence for the relationship between the stability of offending and antisocial behavior, but indicate that the trajectory of offending will differ across ages and vary over the life course (Nagin, Farrington, & Moffitt, 1995). Taking this into account, Sampson and Laub (1993) proposed a theoretical model that addresses the impact that life events and transitions can have on an individuals behavioral pathway and the following section provides a larger discus sion of their theory as contained within a developmental lifecourse perspective. Sampson and Laubs Age-Graded Theo ry of Informal Social Control Sampson and Laubs (1993) age-graded theory in formal social control can be viewed as a merging of the ideas of state dependence (Nagin & Paternoster, 1991) and cumulative continuity, wherein cumulative continuity assumes that adve rse life circumstances accumulate and exert a more significant effect as their frequency incr eases over time (Caspi & Moffitt, 1993; Moffitt, 1993) in order to create a model for explaining crim e over the life course. Th ey argue that social capital (investment or stakes in conformity) and turning points (employment, marriage) are important factors that may challenge traditional dete rministic theories of crime that assume that early childhood experiences lay the groundwork for later development (Bandura, 1982). According to Sampson and Laub (1993:301), tw o books rocked the criminological field and renewed researchers interest in childhood: A General Theory of Crime and Crime and Human Nature In Gottfredson and Hirschis (1990) work, A General Theory of Crime they

PAGE 37

37 argue that ineffective child re aring (including the pa rents failure to effectively monitor, recognize and punish their childs un wanted behaviors) in the early years of a childs life (prior to age 8) fosters the development of low self-cont rol. This latent trait is then viewed as a relatively stable phenomenon that increases th e likelihood of involvement in criminal and analogous behavior throughout an individuals life. Comparativ ely, in Wilson and Hernsteins (1985) book, Crime and Human Nature these scholars state that cr ime could be best explained by individual differences in impulsivity and temp erament, along with an interaction of these qualities with familial factors (s ee also Grasmick et al., 1993; Nagin & Paternoster, 1991). Fueled by this renewed interest in childhood and drawing from social control theory (Hirschi, 1969) and Elders life-course a pproach (1975, 1985), Sampson and Laub (1993:303) began to focus on age-graded life transitions and argue that institutions of formal and informal social control matter and vary over the life course. They contend that age-graded informal social control is essential for promoti ng interpersonal bonds that link i ndividuals to the larger social institutions in which they live (i.e., work, fa mily, school). They also acknowledge that both continuity and within-individual changes do occur over time and that specific life transitions and other factors related to development may interv ene in ones pathway of offending (i.e., also referred to as the stepping-stone approach by Farrington, 1986b). Sampson and Laub (1993) further developed thei r theoretical framewor k with attention to what they characterize as a series of building blocks. The first building block focuses on the intervening role of informal family and school social bonds. Their second building block focuses on the continuity of antisocial behavior th at begins early on in a childs development and extends throughout adulthood. They also recognize th at particular life even ts and social ties in adulthood can alter an individuals behavioral trajectory and this viewpoint is central to their

PAGE 38

38 third building block, i.e., the im portance of adult social bonds. They propose that social bonds developed in adulthood (employment, marriage) can explain involvement/lack of involvement in crime regardless of any underlyi ng individual criminal propensit y. In addition, they emphasize that it is not only the mere expos ure of individuals to these bonds the quality and intensity of these bonds is critical. Drawing on the work of Elder (1975, 1985), Sampson and Laub also hypothesize that different adaptations to key life events can affect an individuals trajectory, and crucial turning points such as divorce or death of a love one can redirect life trajectories. Two additional important concepts in Sampson and Laubs age-graded theory of informal social control include the proces ses of self-selection a nd cumulative continuit y. Self-selection is the classic birds of a feathe r flock together argument, and is based on the belief that individuals with low self-contro l resulting from ineffective parenting systematically sort themselves into criminogenic environments. For example, delinquent youth will tend to gravitate toward delinquent peers in adolescenc e, and similarly, will find delinquent spouses in adulthood and continue their offending beha vior (Gottfredson & Hirschi, 1990). In comparison, cumulative continuity (Moffitt, 1993) integrated with the concept of statedependence (Nagin & Paternoster, 1991) refers to the idea that involvement in delinquency has an incremental effect on the inte rpersonal social bonds formed in adulthood (i.e., labor force attachment, marital cohesion). For instance, an arrest and subseque nt incarceration in adolescence may lead the youth toward dropping ou t of school. This practice would logically affect their future job prospects and thus result in the failure of the individual to develop strong adulthood bonds to the labor force, which inev itably increasing their lik elihood of involvement in crime (Tittle, 1988).

PAGE 39

39 Furthermore, Sampson and Laub (1993) view change as occurring along a continuum where change can be shaped by a shift in societal roles and environments and this individual or environmental change can lead to the devel opment of social capital (Coleman, 1988, 1990; Nagin & Paternoster, 1992). Sampson and Laub ( 1993) further argue that the lack of social capital is one of the central antecedents of w eak social bonds. Similarl y, individuals who lack social capital are expected to be closed off from the informal networks among individuals in the social system, which also has a negative influenc e on their future life prospects (Coleman, 1990). However, Sampson and Laub (1993) also caution that just getting married or employed is insufficient for manufacturing social capital. Th ese social ties are only seen as important in cases where they impose obligations and restraints on individuals th at lead them to refrain from criminal involvement. According to Sampson and Laub (1993), this process offers an explanation for not only career de sistance but late-onset criminal ity because regardless of an individuals early developmental hi story, if they are able to form social capital or investment in societal institutions, then they will always be less delinquent/cri minal. Additionally, they argue that social capital can be reciprocal in nature and that there are racial ethnic, and structural variations in social capital (s ee Anderson, 1990; Short, 1990). In short, according to Sampson and Laub, their life-cour se theory seeks to: Return development to where it probably s hould have been all along, conceived as the constant interaction between individuals and their environment, coupled with purposeful human agency and random developmental noise (2005:40). Sampson and Laubs (1993) age-graded theory of informal social control soon sparked researchers interests; however, in the same year developmental psychologist Terrie Moffitt (1993) unveiled a theoretical typology that claims that there are differe nt types of offending trajectories that exist in the social world and her taxonomy classifies offenders as either adolescent-limited or life-course-persistent. Th e next section provides a discussion of Moffitts

PAGE 40

40 proposed taxonomy of offenders and highlights th e features that dis tinguish an adolescentlimited offender from a life-course-persistent offe nder. Also discussed is Moffitts (1994; see also Moffitt et al., 1994) more recent claims as to why life-course-persistent offending may be limited almost exclusively to males. Moffitts Taxonomy Moffitt (1993) identifies the adolescent-limited offenders as normative and believes they include individuals who are generally well socialized and adjusted, do not begin their offending early in life, and are simply acting out a pa ssing phase in the developmental process of adolescence. These individuals only offend because of the process that she identifies as social mimicry, whereby the youth achieve status a nd power in their adoles cent social world by mirroring the antisocial behaviors of others (m ost notably the behavior of their life-coursepersistent counterparts). Th erefore, through direct observation and reinforcement, the adolescent-limited offenders begi n to engage in similar behavi ors in order to assert their individuality and emerging adulth ood status. However, alt hough their rates of offending may converge with that of their life-course-persist ent counterparts only brie fly in mid-adolescence (ages 14 to 17), the adolescent-limited offenders (by definition) desist from offending soon after this developmental time period. These adolescen t-limited offenders are also typified by inconsistent, erratic, and brief offending careers. Moffitt (1993) further argues that they may shoplift and smoke some pot, but at th e same time follow school rules. In contrast, Moffitt (1993) identifies the li fe-course-persistent group as offenders who begin their criminality/behavior problems early in childhood and exhibi t an ongoing pattern of behavioral problems throughout their life. Thes e offenders begin biting and hitting at age 4, shoplifting and truancy at age 10, selling drugs and stealing cars at age 16, robbery and rape at age 22, and fraud and child abuse at age 30 (Mo ffitt, 1993:679). Moffitt argues that there is no

PAGE 41

41 reason to believe that these offenders ever extr aordinarily abandon their an tisocial tendencies and assume a prosocial identity later in life. A lthough prior research has shown that while offending may decline in middle to late adulthood among crim inal psychopaths, their antisocial personality traits (Cleckley, 1976) have b een shown to persist until ag e 69 (Harpur & Hare, 1994). Farrington et al. (1990) ha ve also indicated that early-onset offending is st rongly correlated with the seriousness of offending and the length of an individuals offending career. Furthermore, compared with their adolescen t-limited counterparts, life-course-persistent offenders problematic behavior is not the result of social mimicr y but it is believed to result from neurodevelopmental impairments that lead to a lack of self-control. This inability to form self-control (see also Gottfredson & Hirschi, 1990) affects their socialization and development and thus makes these individuals vulnerable to committing crime. Moffitt (1993) has also attributed this groups persistent offending to an interaction of liv ing in a poor social environment and neurological problems duri ng early fetal development that inhibit tempermental, cognitive, and behavioral developmen t. A series of studies have shown that one of the most robust predictors of antisocial beha vior is neurological deficiencies (see Moffitt, 1990; Moffitt & Henry, 1991; Hirschi & Hindelang, 1977). Moffitt (1993) continues to argue that a subs tantial portion of the antisocial behavior exhibited by these life-course-persistent offe nders can be attributed to heritability (or intergenerational transmission of antisocial beha vior) (Heusmann et al., 1984). Prior research has shown that there is a strong consistency between parents and thei r children in their temperament and personality traits (Plomin, Ch ipuer, & Loehlin, 1990) and in their cognitive abilities and intelligence scores (Plomin, 1990; Loehlin, 1989). Loeber et al. (2000) indicated that at-risk children are often exposed to adverse home environments and neighborhood

PAGE 42

42 conditions because of their parents vulnerability to a host of problems incl uding their inability to deal with difficult children because they lack sufficient coping resources (see also Synder & Patterson, 1987). Furthermore, Moffitt (1997) has found that these lif e-course-persistent offenders were always more likel y to be involved in crime regardless of the status of the neighborhood in which they resided. Moffitt continues by asserting that age will exert a direct influence on the strength of the association between an individual s underlying criminal propensity and their offending patterns. According to Moffitt, it is expected that crimin al propensity should be strongly associated with antisocial behavior in early ch ildhood, yet the strength of the association should be greatly reduced in adolescence because the adolescent-l imited offenders exhibit the highest prevalence rates during this time period and these offende rs lack a criminal propensity. The strong correlation between criminal propensity and crim e should again resurface in adulthood because the life-course-persistent offenders become the predominant group once again in this time period (see also Piquero & Brezina, 2001).1 Guided by prior research, Moffitt and colleagues have offered some theoretical refinements to the two offender typologies to explain di fferential rates of offending across sex. The adolescent-limited type exemplifies female offe nding patterns (Moffitt, 1994) and life-coursepersistent offending pattern is almost exclusively a male phenomenon because (1) 1 It is important to note here that there is a strong degr ee of overlap between Moffitts life-course-persistent offender and Gottfredson and Hirschis (1990) charact erization of an offender with low se lf-control. The main similarity is found in their definitions of propensity, which is believed to be influenced by impulsivity, hyperactivity, and low verbal ability (see also Bartusch et al., 1997). Similarly, in line with Moffitt et al.s (1994) sex expectations for lifecourse-persistent offending, Gottfredson and Hirschi (1990:148-153) argue that male and female differences in opportunity or supervision can explain the disproportiona te rates of offending across sex. Blackwell and Piquero (2005) found support for the relationship between low self-control and offending among males and females, but indicated that there were sex di fferences in their responses toward their pare nts' control. Therefore, it is likely that Gottfredson and Hirschi would argue that sex differences in self-control can explain the concentration of life-coursepersistent offending patterns among males.

PAGE 43

43 neuropsychological disorders are more preval ent among males (Rutter, 1983; Singer et al., 1968); (2) childhood-onset conduct problems are ve ry rare among girls (Werry, 1986); and (3) female delinquent behavior tends to be more erratic than male delinquent behavior (Moffitt et al., 1994:283). While large scale self-report studies versus th ose based on official data have shown that virtually every adolescent commits some illegal behavior (see Elliot et al., 1983), according to Moffitt (1993) there is yet another group of indivi duals who are characterized as abstainers or non-offenders. Similarly, Moffitt (1993) provides some theoretical speculations as to why these individuals fail to become involved in crime a nd delinquency. Moffitt argues that perhaps these youth experience late-onset puberty and therefore do not sense the maturity gap or perceive the social assets associated with crime (1993:689). Prior research has indicated that females that began menstruating after the age of 15 did not show any involvement in delinquency. The abstainers may also lack the opportuni ties to engage in delinquency. Adolescent delinquency rates have been shown to be higher in urban versus rural areas (Skogan, 1990), and Sampson and Groves (1989:789) have also commented that one of the str ongest correlates of violence is the presence of unsupervised groups of teenagers. In another study, Farrington and Wests (1990) research with their London male cohort showed that adolescents who resided in criminogenic environments but did not engage in delinquency could be characterized as nervous and withdrawn and also tended to entirely lack friends or only have a few friends. Most recently, Piquero, Turner, a nd Brezina (2005) provided one of the first empirical tests investigating Moffitts (1993) ab stainer hypothesis and found that there did exist a relatively small group of adolescents (a sligh tly greater proportion of females) that could be classified as abstainers. Piquero et al. (2005) indicated that teacher attach ment was inversely related to

PAGE 44

44 association with delinquent peers for the females and that less associati on with delinquent peers was related to being an abstainer among the males. However, Piquero et al. also noted evidence that was contrary to Moffitts expectations. Mo re specifically, they did not find that that the abstainers were more likely to be sad and depr essed or completely excluded from all peer networks. Piquero et al. (2005) showed that th e abstainers were still relatively active in the dating scene and were not typically loners, but ar gued that they were just adolescents who were involved in a more prosocial network of peers. Overall, Moffitts taxonomy provides a solid theoretical framework to explain not only differential offending patterns of individuals, but also offers sp eculations as to why some youth never become involved in delinquency. Furtherm ore, her offender typology can also account for sex differences in offending. This concludes the chapter on theoretical fr ameworks. It attended to the assumptions underlying social disorganization theory, the age-crime curve and the criminal career paradigm, the recent theoretical advancements of an age-graded theory of in formal social control (Sampson & Laub, 1993), and the existence of adolescent-limited and lifecourse-persistent offenders (Moffitt, 1993) including the re cent theoretical extensions to account for sex differences in offending. The next chapter cont ributes a thorough discussion of what the prior literature has discovered regarding these perspectives.

PAGE 45

45 CHAPTER 3 REVIEW OF THE LITERATURE Introduction The theoretical perspectives pr esented above have been the fo cus of a considerable amount of empirical testing. While, soci al disorganization theory has r eceived a substa ntial amount of attention for explaining crime rates at the macrolevel (for review, see Bursik & Grasmick, 1993; Sampson & Groves, 1989). The theory (particu larly surrounding neighborhood SES) has also been applied to accounting for individual variatio n in offending. This line of research will be discussed in more depth below. It will be followed by an examinati on of the studies (most notably that of Wolfgang et al ., 1972) that have identified the existence of a relatively small proportion of chronic high-rate offenders (bet ween 2% and 14% depending on which study is being referred to) who are responsi ble for at least half of the ove rall crime and even a larger percentage of the violent offenses. The next se ctions will present the research evidence on the impact that life transitions such as marriag e and employment can ha ve on an individuals pathway of offending and the literature that has provided support for Moffitts (1993) taxonomy as an explanation of sex differences in offending. Finally, this chapter will conclude with an illustration of the prior applica tions of Nagin and Lands (19 93) methodological technique for identifying groups of individuals with similar o ffending trajectories as well as providing several hypotheses that will be test ed in this current study. Impact of Social Disorganizat ion at the Individual Level In general, community-level scholars (e.g. Bu rsik & Grasmick, 1993) have associated the geographic concentration of soci oeconomic disadvantage to the so cial disorganization of the neighborhoods, and these disorgan ized neighborhoods have been ar gued to have adverse effects on the individuals and families that reside within those areas. In contrast, other scholars contend

PAGE 46

46 that the concentrated effects of poverty cannot be fully explained w ithout including control measures for both familial and individual-level characteristics (see Gephart, 1997:14). Although, while social disorganization theo ry (as originally formulated) sought to explain crime at the aggregate level, there are empirically-based reas ons to expect that social disorganization may also have an effect on individua l behavior. Simcha-Fagan and Sc hwartz took a step toward this possibility: An aggregate-level theory cannot be expected to provide a comprehe nsive explanation of individual delinquency or crim inality, it should clarify thos e contextual effects which, independently of or in interaction with th e adolescents predispositi on, affect individual behavior (1986:688). Jencks and Mayer (1990) provided a review of the current extant knowledge on the effects that social compositional measures (e.g., poverty) have on a series of developmental outcomes including: educational attainment, cognitive sk ills, criminal activity, and economic success. They indicated that the research is generally mi xed at best regarding the influence of average SES (in the aggregate) on the likelihood of an i ndividuals chances for planning to go to college, attending college, and/or gradua ting from college, especially for Whites. Research has shown that youth from high SES neighborhoods are more likely to have higher educational attainment compared with their counterparts who come fr om lower SES neighborhoods. In addition, Black mens chances of competing in the job market and finding well-paid jobs were greatly reduced when they resided in a heavily concentrated Black or welfare depende nt neighborhood (see also Aber et al., 1997). Klebanov et al. (1997) provided an extension of the resear ch on the influences of neighborhood effects on crime by specifically testi ng whether or not these effects are mediated and/or moderated (or interact with) by family processes such as maternal warmth, cognitive stimulation in the home, maternal coping style, av ailability of social s upport, and the physical

PAGE 47

47 environment of the home. They found that the positive effects of living in a high SES neighborhood on academic achievement, higher verbal ability scores, and less behavior problems and the negative effects of living in heterogenou s neighborhoods on verbal ability scores were mediated by the quality of the home learning envir onment. The effect of home environment was found across all age groups, measures, and res ource-rich and resource -poor neighborhoods. They also found support for the interaction of th e home environment and family characteristics in each neighborhood type in ways related to criminal outcomes. Similarly, additional research by Brooks-Gunn et al. (1993) showed that neighbor hoods had an effect on childhood IQ and the likelihood of becoming a teenage parent and/or dropping out of school. Thus, it appears that neighborhood level effects may be both direct an d indirect, but perhaps a substantial proportion of their effects are indirect in that they operate through i ndividual and family factors. Simcha-Fagan and Schwartz (1986:671) argue that structural char acteristics have an impact on the organizational networks among me mbers of the community, and these networks can subsequently affect an individuals abil ity to form social bonds to the community. Individuals who rent property as opposed to ow ning their own property are less likely to have a stake in the community in which they live, and thus may be less inclined to participate in neighborhood organizations that provide a mechanis m of informal social control. Similarly, individuals who rent their propert y may be less likely to care a bout the maintenance or up-keep of their dwelling because they do not intend to be there long. This lack of investment could lead to neighborhoods incivilities (W ilson & Kelling, 1982), which have been linked a breakdown in informal social control and crime (Sampson, 1995) Dwelling units that are over-crowded such as public housing units are also likely to lead to a sense of anonymity among the residents. The failure of individuals to ge t out of their units and develop informal social ties with their

PAGE 48

48 neighbors may be due to their exposure to viol ence and fear of crime (Roncek, 1981; Sampson, 1995). Sampson and Groves (1989) ha ve also shown that high rate s of residentia l instability were associated with juvenile delinquency and crime. Gephart (1997) reviewed th e developmental impact of community effects on adverse individual outcomes. In early childhood, children who reside in areas of extreme disadvantage face a slightly higher risk for having a low bi rth weight and suffering from child abuse. Comparatively, additional analyses have shown that the presence of affluent neighbors was associated with higher IQs and the presence of low-income neighbors was related to a greater likelihood of behavioral probl ems in early childhood. Furthermore, research has also found that the influence of a neighborhoods socioeconomic status on individual outcomes in early chil dhood and middle childhood were different across races. At age 3, the benefit of affluent nei ghbors (versus middle SES neighbors) on childrens IQ scores was more salient among White child ren than African-American children (BrooksGunn et al., 1993). Similar findi ngs were also found among children at ages 5 and 6 for IQ scores, verbal ability, and read ing achievement scores (Chase-L ansdale et al., 1997; Duncan et al., 1994). In addition, all of thes e studies revealed that the pr esence of low SES neighbors was positively associated with early childhood behavior problems. Gephart (1997) also indicated that research investigati ng neighborhood effects during late childhood and early adolescence found that commun ity impoverishment, the burden of childcare, and residential instability were significant predictors of va rious forms of delinquency such as violent crime, drug trafficking, juvenile de linquency, and teen pregna ncy (see Gephart, 1997). Two additional studies have also shown that neighborhood SES had a stronger effect for male adolescents than female adolescents (Entwisl e et al., 1994; HalpernFelsher et al., 1997).

PAGE 49

49 Research has also shown that living in a high SES neighborhood increased older adolescents chances of gra duating high school and attending college (Brooks-Gunn et al., 1993; Duncan, 1994; Ensminger et al., 1996; Halpern-Fe lsher et al., 1997). All of these studies suggested that the protective e ffect of living in a higher SES neighborhood was more salient for males. Crane (1991) has also provided evidence that males had the worst outcomes when living among low SES neighbors. In addition, a series of studies reviewed earlier have all consistently shown that living in a low SES neighborhood was positively associated with higher rates of criminal and delinquent behavi or (Braithwaite, 1979; Lindstrom, 1996; Loeber & Wikstrom, 1993; Lynam et al., 2000; Peeples & Loeber, 199 4; Piquero et al., 2005; Reiss & Rhodes, 1961; Sampson & Groves, 1989; Wikstrom, 1991a, 1991b; Wikstrom & Loeber, 2000). The U.S. Department of Housing and Urban Development (HUD) initiated what is known as the Moving to Opportunity for Fair Housi ng Demonstration (or MTO) in 1994 motivated by the success of the Gautreaux Program that has been in operation in Chicago for almost three decades. Rosenbaum (1995; see also Rosenbaum & Popkin, 1991) presented quasi-experimental evidence that Gautreaux youth w ho moved to the suburbs greatly increased their educational attainment and job opportunities when compared with those that remained in the poor urban areas. Intrigued by these findings and the idea that residential mobility may benefit individuals from disadvantaged communities, the MTO program became operational in five selected sites (Baltimore, Boston, Chicago, Los Angeles, and New York City). The MTO program involved randomly providing families vouchers allowing some families with children in high poverty areas to move to less poor areas with the h ope that this relocation would increase their educational and employment opportunities. Eligibility for the program was determined by the families that lived in Sect ion 8 public housing units with at least a 40%

PAGE 50

50 poverty rate according to 1990 census tract data. Prior MTO evaluations have indicated that residential relocation significan tly contributed to a reduction in violent arrests for males and better overall health among the children that moved from the public housing units (Katz et al., 2001; Ludwig et al., 2001). Katz et al. (2001) also noted signif icant reductions in criminal victimization among individua ls that moved out of the Section 8 housing. Leventhal and Brooks-Gunn (2003) noted that the MTO program had profound effects on mental health outcomes for the individuals and fa milies that relocated. They reported that the parents who moved reported less distress than did those that remained a nd the boys that moved (especially those between ages 8 and 13) repor ted significantly less depr ession, anxiety, and/or dependency problems when compared with their counterparts who were still in public housing. Most recently, Kling et al. (2007) noted the substa ntial mental health and education benefits for the adults and females, but commented on the ad verse effects associated with the males who received vouchers (for an exte nded review of MTO impact evalua tions, see also Orr et al., 2003). Overall, there is reason to believe that stru ctural factors includi ng those highlighted in social disorganization theory can have an eff ect on individuals outcomes and perhaps influence their pathway of offending. It is likely that living in areas character ized by social disorganization will increase an individuals likelihood of bei ng a chronic offender. This concept of chronic offending is the topic of the discussion presented in the following section. Prevalence and Incidence of Chronic Offending Relying on data from the well-known Philade lphia Birth Cohort Study, Wolfgang, Figlio, and Sellin (1972) provided some of the earliest ev idence of the existence of chronic offenders. Wolfgang et al. found that there were a group of delinquents that repres ented only 6% of the cohort and 18% of the cohort's delinquents but we re responsible for comm itting roughly half of all the offenses and about two-thir ds of all of the violent offenses In a ten-year follow-up study

PAGE 51

51 of 10% of the original Philade lphia birth cohort, Wolfgang, T hornberry, and Figlio (1987) found that these chronics, originally de fined as having had five or more police contacts as juveniles, had increased the seriousness of their offending into adulthood. Tracy, Wolfgang, and Figlio (1990) found even higher rates of chronic offe nding among members of the Second Philadelphia Birth Cohort Study. They found that the chronics represented 7% of the cohort and 23% of the offenders, yet were responsible for committing 61% of all the offenses including 60% of homicides, 75% of forcible rapes, 73% of robbe ries, and 65% of the aggravated assaults. In a similar study, Snyder (1988) examined a cohort of Phoenix and Utah youths and found that the chronics, defined as havi ng had five or more referrals to juvenile court, represented only 5% of the cohort. However, li ke the original Philadelphia Birt h Cohort, they were responsible for 51% of all the offenses committed includi ng 61% of murders, 64% of rapes, 67% of robberies, 61% of aggravated assaults, and 66% of burglaries. Sha nnon (1988) studied three youth cohorts in Racine, Wisconsin and found a sli ghtly higher percentage of chronics (8% to 14%) than the previously mentioned studies, bu t still found that they were involved in an overwhelming majority (75%) of all of the felony arrests. Hamparian, Schuster, Davis, and White ( 1985) in a follow-up study of the original Columbus Cohort (Hamparian, Schuster, Din itz, & Conrad, 1978) found that 60% of the previously identified chronics (r epresenting only 2% of the cohor t) had been arrested at least once by their early-twenties for an adult felony offe nse. These offenders were also more likely to have had an earlier onset of delinquency, more frequent an d serious offending, and displayed continuity in offending from thei r youth through their mid-twenties. The high frequency offending of chronics that ha s been illustrated when examining official arrest data is also evident in self-report survey s. Results from the National Youth Survey (NYS)

PAGE 52

52 have shown that about 5% of juveniles at each age level from twelve to seventeen years old can be classified as serious, chronic offenders. Thes e offenders are characterized as having engaged in three or more aggravated as saults, sexual assaults, gang fight s, and/or strong arm robberies within the previous year. In addition, on average these chronics committed 132 delinquent offenses annually and 8% of these offenses are considered serious and violent. Perhaps more alarming is the fact that 84% of the offenders iden tified as chronics in the survey had no official record of delinquency and this disparity between official measures and self-reported measures of offending cannot be underestimated (Elliot, 1994; Elliot, Huizinga, & Morse, 1986). Further support for the existen ce of chronic offenders was also provided in a series of RAND Inmate studies conducted during the late 1970s and earl y 1980s. Petersilia, Greenwood, and Lavin (1977) conducted the firs t of these studies and focused on the criminal careers of 49 inmates in a California medium-security prison. All of these offenders were currently serving time for armed robbery and were characterized as re peat offenders because they all had served at least one prior prison term. Petersilia et al. (1977) gathered information on the frequency of offending from structured interviews with the o ffenders and their official criminal records. Overall, the 49 inmates self-reported committi ng 10,500 serious crimes even though the majority of the offenders had spent the better part of their criminal careers behind bars. Peterson and Braiker (1980) carried out the second Rand Inmate study where they administered self-report questionn aires to 624 incarcerated male fe lons in five of Californias state prisons asking the inmates to report what ty pes of offenses they committed during the three years prior to their current imprisonment. Pete rson and Braiker found that the majority of the inmates did not specialize in offending and that most of the felons who committed one crime type at a high frequency also re ported significant criminal activity in other types of crime. On

PAGE 53

53 average, these chronic offenders represented about 8% of the inmates but self-reported committing over 60 crimes per year. Chaiken and Chaiken (1982) conducted the thir d Rand Inmate study investigating high-rate offending, wherein they analyzed the criminal hi stories and survey results from 2,190 prison and jail inmates in California, Texas, and Michig an. Chaiken and Chaiken identified what they referred to as a small group of violent pred ators or omni felons and these individuals reported committing robberies, assaults, and drug deals at disproportionately higher rates than the other inmates. In addition, they tended to commit burglaries, thef ts, and other property crimes at higher rates than even the offenders who self-reported to spec ialize in those specific crime types. Overall, all of the studies described above have shown strong empirical support for chronic offenders and these offenders represen t only 2% to 14% of the offending population (depending on which study is being referred to), yet they are respons ible for at least half of all the offenses committed (Farrington, 1999). The fact that a sizeable fraction of the crime problem can be attributed to a relatively small number of identifiable chronic offenders provides support for preventative measures and early identifica tion (see Piquero, Farringto n, & Blumstein, 2003). However, in contrast to the robust finding of chronic offending among various samples and time periods and the evidence that these individual s tend to be criminally active for long periods of time, Sampson and Laub (1993) argue that ke y life transitions such as marriage and/or employment can have an impact on an individual s pathway of offending, even if they are/were chronic offenders. The following section addres ses the empirical evidence in support of the effect that these various life events can have on behavioral outcomes, most notably desistance from crime.

PAGE 54

54 The Effect of Life Tran sitions on Offending Sampson and Laubs (1993) age-graded theory of informal social control takes the position that life key life transitio ns (including delinquency) can affect adult outcomes. Similarly, Caspi and Moffitt (1993) have referred to delinquency (particularly serious) as a knifing off of future opportunities, or in other words delinquency in ad olescence is expected to impede the formation of conventional social bonds that will c ontinue to have an affect in adulthood Relying on the Gluecks data (1950, 1968) Sampson and Laub (1993) presented preliminary support for their theory, whereby in dividuals who exhibited the strongest ties to work and family had the lowest rates of delinquenc y. In contrast, incarcer ation in the adolescent years was inversely related to job stability and positively associated with continued criminal involvement. Nagin and Waldfogel (1998) have also noted the advers e effects that an adolescent conviction has on future job opportunities amon g West and Farrington s (1977) London male cohort. Additional research has shown that invo lvement in delinquency results in the worst adult outcomes for lower class boys, particularly amon g lower class boys with a high frequency of police contacts (Hagan, 1991). Jess or et al. (1991) have also i ndicated that delinquency is not a major impediment for the formation of prosocia l adult bonds for middle class adolescents. Based on the results of Sampson and Laubs (1993) original test of their age-graded theory of informal social control, support was found th at delinquency was associated with crime in adulthood as well as being significan tly related to a series of adverse life events including being AWOL from the military, economic dependence, and having had martial problems. Contrary to what Gottfredson and Hirschi (1990) would pred ict, they also found significant independent effects for marital attachment and job stability on involvement in adult crime net of the effects of criminal propensity.

PAGE 55

55 Sampson and Laub (1993:317-320) c oncluded by offering six key issues to guide future research testing their theory. First, there is some element of luck, randomness, or chance in an individuals life course (also referred to by Rutter, 1989 as adve ntitious happenings). Second, there is individual variation in the effects that life transitions have on be havior. Third, macrolevel opportunity structures for marriage and labor market activity are important (i.e., segmentation of the labor market, ethnic encl aves). Fourth, criminal propensity cannot completely explain adult crime because life transitions can also affect opportunities and motivations to commit crime (see also Cohen & M achalek, 1988). Fifth, the effect of cumulative continuity is strongest among those in disadv antaged racial and econom ic conditions. Sixth, historical context matters. With attention to some of these suggestions, Horney et al. (1995) f ound clear evidence that meaningful change in life circumstances was associated with desistance from crime based on their interviews with 658 offenders in a Nebraska Correctional Facility. They also estimated hierarchical linear models that allowed them to control for criminal propensity and still found that the offenders reported reduced involvement in criminal activity when they were living with their wife. Overall, their results suggested that life transitions can alte r behavioral patterns of offending at particular periods of time, especially in the short-term. Laub et al. (1998) estimated trajectories with the Gluecks da ta and found that the timing and qua lity of marriages were what was important and that the inhibiting effects of marriage on crime were gradual and grew in intensity over time. Similar to Horney et al (1995), these findings were still observed while controlling for criminal propensity and a host of other factors. More recently, Piquero et al. (2002) developed a stakes in conformity index where they combined the life circumstances of employment and marriage to gauge the possible cumulative

PAGE 56

56 effects of local life circumstances on pathwa ys of offending among serious offenders from the California Youth Authority. Based on their results, Piquero et al. concluded that it is likely the case the these mechanisms of informal social control (or life transiti ons) can have a stronger effect on instrumental as opposed to interperso nal crimes because commission of the former tends to involve a relative degree of planning. Laub and Sampson (2003) estimated models that took into acc ount within-individual change and found that when controlling for age, individuals who were currently married had lower criminal propensity scores when compared w ith their scores when they were not married. They also found these differences in propensity existed in times where the individuals were employed or in active military service compared with periods when they were not. In another study testing Sampson and Laubs th eory, Warr (1998) provided an analysis of two waves of longitudinal data from the Nationa l Youth Survey of 1,725 individuals. Counter to Sampson and Laubs (1993) position, Warr hypothe sized that the link between life transitions and crime might be better explained by the effect that these transitions have on altering the relationships with delinquent peers (Akers et al., 1979, 1985; Sutherland, 1947). Warrs results did reveal that marriage was strongl y associated with desistance from crime, but he indicated that for many of these individuals, getting married was a key life transition that signaled a severing of heavy involvement with their delinquent peers a nd instead re-focused th eir attention to their spouse and the desire to have children. These findings led Warr to conclude that marriage not only reduces individuals motivation for continuing criminal behavior by re -directing their goals, it also greatly minimizes their opportunities for offending by limiting the amount of time that they have to spend with their former delinquent peers. In addition, Warr caution ed that there still exis ted the possibility of a

PAGE 57

57 selection effect in that individua ls who do in fact get married ar e perhaps qualitatively different from those who do not get married, and that i ndividuals who are not i nvolved with delinquent peers and/or engaged in crime may appear to be more suitable marriage prospects. Maume, Ousey, and Beaver (2005) have si nce reexamined Warr s (1998) hypotheses and conclusions using data from the National Y outh Survey, but improve upon Warrs earlier study by specifically investigating the e ffects of marital attachment on desistance from marijuana use and control for the possible selectivity bias in the relationship between marital attachment, delinquent peer associations, and desistance. In line with Samp son and Laubs theory, Maume et al. found that marriage was important to the desistan ce, particularly for those marriages that were characterized by high levels of attachment. C ontrary to Warrs (1998) findings, marriage still maintained significant directs effects for affecting desistance net of the effect of delinquent peer association (measured as the change in the number of delinquent associates). Uggen (2000) investigated the impact that employment (i.e., a turning point) has on an offenders recidivism relying on data from over 3,000 individuals from nine U.S. cities, which were originally part of the 1977 National Supported Work Dem onstration Project that targeted the underclass and the ghetto poor. The individu als who participated in this project were randomly assigned to either of two treatment groups (construction work or service jobs) or to a control group where no job placement was given. Th e work and arrests histories were collected for all of these individuals every nine months for three years. The results from the event history analysis yielded a variety of noteworthy findings First, although the program failed to reduce crime overall across the entire sample, its imp act was definitely age-graded. Job treatment significantly reduced recidivism among the older participants (p articularly aged 27 or older) whereas the job treatment had little to no effect on crim e among the young adults (late

PAGE 58

58 teens/early twenties). Therefore, it appeared that this life transition (employment) affected desistance from crime only for a certain age group of individuals (older adults) providing support for Sampson and Laubs (1993) theory. In a recent review of the state of their theory and in response to some noted criticisms (see Moffitt, 2005), Sampson and Laub (2005c:18) refined their views toward the influence of life transitions on adult offending by arguing that involvement in institutions such as marriage and work provide both short-term and l ong-benefits. In the short-term, these life transitions can alter an individuals motivation for crime by re-f ocusing their energy and attention toward conventional outlets (spending time and caring for their spouse/family or their job) and can produce the long-term benefit of fo stering their development of pros ocial adult bonds or stakes in conformity. Sampson and Laub (2005c:36) continued by argu ing that marriage (or employment) should not be viewed as a one particular point in ti me, but viewed as a dynamic and cumulative causal factor that affects desistance from crime over tim e (see also Laub et al., 1998). Given the above presented evidence in support for Sampson and La ubs age-graded theory of informal social control, it is important also to visit the re search providing support for Moffitts taxonomy of offending for explaining sex variat ions in offending and this becomes the focus of the following section. The Saliency of Moffitts Taxonomy across Sex According to Moffitt et al. (2001), females typi cally have fewer risk factors than males; therefore, they are less likely to be early-onset and/or life-c ourse-persistent offenders. In contrast, the sex similarities in adolescent-limite d offending are due to th e process of social mimicry where there is an equal-opportunity for offending among males and females in the adolescent years (Moffitt et al., 2001). Moffitt et al. (2001) have also suggested that while

PAGE 59

59 neurodevelopmental disorders affect the life-cou rse pathways of offending similarly for males and females, males experience early cognitive deficits far more frequently than females. Their results showed that significant sex differences existed in almost all cognitive areas including hyperactivity, impulsivity, and temperament. Tibbetts and Piquero (1999) conducted one of the earliest tests of whether Moffitts taxonomy applies to males and females and report ed that they failed to find enough females who exhibited the life-course-persisten t pattern of offending in order to provide sufficient statistical power in a Philadelphia cohort. Subsequently, Fer gusson et al. (2000) were able to show that a single model could explain antiso cial behavior across sex, but the ratio of involvement in earlyonset offending and late-onset offending was 4:1 and 2:1 for males and females respectively. In contrast, Aguilar et al. (2000) found that the prevalence of early-onset offending was similar across sex. In another study, Moffitt and Caspi (2001) f ound that the sex ratio for life-coursepersistent offending was 10:1, whereas the sex ra tio was virtually negligible for adolescentlimited offending (1.5:1). Similarly, Kratzer an d Hodgins (1999) found a rather large sex ratio for life-course-persistent offending (15:1) and adolescent-limited offending (4:1) among a Swedish cohort of 13,000 individuals. Mazerolle et al. (2000) reported that early-onset offending was associated with persistent offendi ng and a lack of specialization among members of a Philadelphia cohort (n= 3,655). In addition, in a longitudinal study of 820 girls, Cote et al. (2001) did observe an antisocial group charac terized by persistent and frequent offending, although this group only made up 1.4% of the entire sample of females. Comparatively, Silverthorn and Frick (1999) have argued anti social girls tend to share many of the correlates of chil dhood-onset offending and adulthood adjustment problems that the

PAGE 60

60 childhood-onset boys express. The difference is that antisocial girls typically demonstrate a delayed-onset pathway (do not begin their o ffending until adolescence). They propose that this third developmental pathway is unique to females and while the cognitive and neurological deficits and the dysfunctional family environment is present in these gi rls childhood, they do not produce overt antisocial behavi or until adolescence. Silverthorn and Frick (2001) provided a prelimin ary test of their delayed-onset offender typology among 72 incarcerated youth. They found that there were similar percentages of early and late-onset offenders among the males, but that the females were more likely to show the delayed-onset (late-onset) patterns of offending and these girls disp layed similar risk factors as the childhood-onset boys. Subsequently, two addi tional tests found that early-onset and delayedonset offending were similar among boys and girls a nd that the early-onset girls shared the same risk factors as the early-onset boys (see Fergusson & Horwood, 2001; Moffitt & Caspi, 2001). More recently, White and Piquero (2 004) provided another test of sex differences/similarities in early versus delaye d-onset offending, except us ing a large sample of African-American participants from the Philadelphia portion of the National Collaborative Perinatal Project. White and Piquero (2004) discovere d that females were just as likely as males to display early-onset patterns of offending, but comparatively, the late-onset females shared many risk factors with the early-onset males. Noting all of the research presented above, according to recent review, Piquero and Moff itt (2005) argued that females do show less involvement in criminal activities at both leve ls (adolescent-limited of fenders and life-coursepersistent offenders) when compared with males. Taken together, all of the prev iously reviewed research has indicated that (1) macro-level factors can affect individual be havior, (2) there exists a sma ll group of chronic offenders who

PAGE 61

61 commit disproportionate amounts of crime; (3) li fe transitions can al ter individual offending pathways; and (4) Moffitts taxonomy can be used to explain sex differences in offending, particularly the findings that males are more likely to be found among the life-course-persistent offender group than the females. Noting this ev idence, the following section directs attention toward examining the stability of offender groups across recent prior applications of Nagin and Lands group-based modeling techni que, along with a discussion of the factors that have been found to be associated with offending group membership. Prior Applications of GroupBased Models of Offending Similar to the theoretical developments in criminal career and developmental research, modern innovations in met hodological approaches to studying group-based modeling, specifically the technique devel oped by Nagin and Land (1993) (i.e ., what has been referred to by many names including trajectory analysis, finite mixture models, or semiparametric mixed Poisson-based regression models). The tec hnique permits disaggregating aggregate offending patterns into group-based o ffending trajectories of i ndividuals that exhibit distinct pathways over time (Nagin & Land, 1993; Land, McCall, & Na gin, 1996; Land & Nagin, 1996; Nagin, 2005). This method for modeling multiple pathways of offending provides a mechanism to identify and distinguish between certain types of behavioral trajectories. It also enables the use of regression models such as multinomial regression in order to examine which covariates are best able to predict which offending group that an individual is assigned. Nagin and Land (1993) provided the first application of th is methodology based on data from the well known South London cohort of 411 males (West & Farrington, 1973, 1977). Relying on convictions as their measure of offe nding, they found evidence of four groups of offenders including a non-offender group, an adolescent-limited group, and two groups of chronic offenders with different degrees of high-rate offending.

PAGE 62

62 McDermott and Nagin (1998) found support for the existence of three offender groups based an analysis of male respondents from th e National Youth Survey, which is a well-known self-report survey. The first group was classi fied as the non-offenders. The second group was referred to as adolescent-peaked. They tended to resemble the third or chronic offending group in adolescence but then began to desist at th e age of 18 until there was no observed offending at age 24. Additionally, trajectory analyses based on two distinctly di fferent types of offenders (Gluecks Boston delinquents vs. California Yo uth Authority parolees ) from different time periods and contexts found that there seemed to be 4 to 6 trajectory groups that all began to decline by adulthood (see Laub et al., 1998; Piquer o et al., 2002). Laub et al. (1998) and Piquero et al. (2002) also noted that marriage or stakes-in-conformity (that included marriage) was related to lower trajectories. A more comprehensive attempt to identify th e consistency of latent offending groups was performed by DUnger et al. (1998) using da ta from the London cohort (West & Farrington, 1973, 1977), the Second Philadelphia Birth Cohort (W olfgang, Figlio, & Se llin, 1972; Tracy et al., 1990) and the three Racine Birth Cohorts (Shannon, 1988, 1991). They found support for the considerable similarity of offending groups across the data sources, with the constant finding of non-offending, adolescent-limited or peaked offending, and chronic offending groups. Specifically concerning the Racine Birth Cohorts, DUnger et al. s results revealed the similar four classes of offenders noted above for the 1942, 1949 and 1955 cohorts, but discovered an additional fifth group for the 1942 and 1955 cohorts with a late-onset pattern of offending. DUnger et al. (1998) illustrate d that this group show ed relatively stable rates of offending throughout their twenties. Broidy et al. (2003) also found three to four trajectory groups when

PAGE 63

63 examining childhood physical aggression and later involvement in delinquency/crime using data with various demographic groups (males/femal es) and six different study sites represented including: Child Development Project; Christ church Health and Development Study, Dunedin Multidisciplinary Health and Human Deve lopment Study, Montreal Longitudinal Study, Pittsburgh Youth Study and the Quebec Provincial Study. More recent research has also observed th ese similar offending gr oups disaggregated by sex (Cote et al., 2001, 2002; DUnger et al ., 2002; Fergusson & Horwood, 2002; Land, 2000; Piquero et al., 2005), although some studies were unable to identify a chronic offending group among females (DUnger et al., 2002; Piquero et al., 2005). DUnger et al.s (2002) analysis provided five offender groups for the males and only three for the females. Furthermore, they indicated that while the age patt erns of offending tended to be similar across sex, the frequency of offending was higher among the males. In contrast, Cote et al. (2001) discovered a life-course-per sistent group of offenders among the females, but noted that this group only comprised 1.4% of the sample of nearly 900 females. In a more recent study, Cote et al. (2002) found that the broad patterns of development (as far as groups) were similar across sex, but males were more likely to exhibit higher impulsivity trajectories when compared with the females. Relying on data from 896 members of the New Zealand Dunedin Birth Cohort, Fergusson and Horwood (2002) indicated that while identical offending trajectories were found for the males and females, the females were more likely to be of low risk and found predominantly among th e adolescent-limited gr oup, whereas the males were more likely to be in the chronic offendi ng groups. Most recently, Piquero et al. (2005) discovered three classes of of fending (low, medium, and high-rate) among males and only two for the female models (low and medium-rate) us ing data from the Dunedin birth cohort. In

PAGE 64

64 summary, Hipwell et al. (2002) concluded that the main difference between boys and girls behavioral patterns of offending is that only a very small proportion of girls follow the lifecourse-persistent trajectory. Chung et al. (2001) also discovere d another trajectory group that they termed desisters that illustrated an atypical trajectory in that they were characterized by early-onset offending (before age 13) but instead of showing the steady rise in mid-adolescence in to young adulthood, the groups trajectory declined and appeared to de sist by age 21. According to Chung et al. (2001), further analyses suggested that it was not individual and familial f actors that could be attributed to their desistance but rather the influence of less exposure to peer neighborhood, and school factors typical associated with offending. They also indicated that aggressiveness, antisocial peers, and drug availability were consistent predic tors of initial levels of offending measured at age 13 and that individual factors including ag gression, depression, and anxiety significantly distinguished the late-onset offenders from the non-offending group. Piquero (2007) provided a comprehensive revi ew on the studies to date applying Nagins and Lands (1993) group-based modeling technique and concluded that th ere existed relative consistency in the number of trajectory groups and the shape of the curves, but there was variability across the studies depending on the length of follo w-up and the ages that were represented in the analysis. A dditionally, in line with Moffitt these studies have shown evidence of adolescent-limited and life-course-persistent trajectories across dive rse data sets, cohorts, countries and using various measures for offending including self-reports, teacher ratings, parent ratings, and police records based on either contacts, arrests, or convictions. However, these applications have also uncovered other traj ectories not consistent with Moffitts (1993) predictions such as the low-level chronics and late-onset type.

PAGE 65

65 The next section provides an elaboration of the theoretically integrated model using macrolevel and micro-level factors for explaining group-based offending tr ajectories. Specifically, the discussion focuses on how the social disorganiz ation framework can be incorporated with Samspon and Laubs age-graded theory of inform al social control and Moffitts developmental taxonomy. Integrated Theoretical Mode l for Explaining Offending Typically, criminological and sociological research focuses on either examining community level or aggregate rates of delinquency/crime or investigating the differences between delinquents/criminals and nondelinquents/cr iminals. Although, in recent years scholars have begun to recognize the importance of multi-leve l research and integrated theoretical models (see Reiss, 1986; Moffitt, 1997a; Sampson, 1997; Fa rrington, Sampson, & Lauritsen, 1994). In an attempt to answer these calls for research in this area, the proposed model in this study emphasizes the importance of three particular th eoretical perspectives (social disorganization, age-graded theory of informal social cont rol, and Moffitts taxonomy) for explaining groupbased offending trajectories. The proposed mode l is further elaborated on below and the conceptual figure (along with addi tional variables included in the an alyses) is displayed in Figure 3-1. In the 1920s Shaw and McKay reported that the areas that were characterized by the highest rates of crime and delinque ncy also tended to display comp aratively high rates of infant mortality, low birth weight, child abuse, along with other adverse factors for early childhood development. According to Sampson (1997:43) the consistent finding of moderate-strong continuity of early onset offending behavior and later criminal i nvolvement and the longstanding covariation of delinquency rates and social prob lems clearly emphasizes that there is a connection between healthful de velopment of young children and community structure. The

PAGE 66

66 quality of health care services has also been shown to vary across communities with the most socially disorganized areas ha ving the poorest services (Lash, Si gal, & Dudzinski, 1980). Similarly, Moffitt (1997a) and others (see Bronfenbrenner, Moen, & Garbarino, 1984) have argued nueropsychologically vulnerable ch ildren may be disproportionately found in poor areas given the association between brain de ficiencies and povert y. Moffitt (1997a) also highlights the concentration of low birth wei ght, poor infant nutrition, and the incidence of exposure to toxic agents as be ing greater in poor neighborhoods. Furthermore, children with early developmental/neurological deficits are not typically born into protective and supportive environments with in-tact families, but rather into families with pare nts who socialize their children in a criminogenic environment where both pa rent and child are vuln erable to the adverse neighborhood conditions (Moffitt, 1997a). These at-r isk or difficult children are then likely to receive negative reactions from their parents in efforts to cont rol their problem behavior in early development and this negative reaction is likely to be con tinued throughout adolescence among peer groups at school and indivi duals in the community (see Go ldsmith, Bradshaw, and ReiserDanner, 1986) whereby further exacerbating the de velopment/continuity of antisocial behavior. Caspi et al. (1987:308) summarizes this intera ction as one in which the child acts; the environment reacts; and the child reacts back in mutually interl ocking evocative interaction. Thus, in the proposed theoretical model, living in a socially disorgan ized area should impact group-based offending trajectories by increasing the likelihood of being assigned to a chronic offender group. Moffitt (1994) has also suggested that chroni c offending can perhaps be regarded as an exclusively male phenomenon because of the dispr oportionate concentration of risk factors such as neurodevelopmental disorders and early be havior problems among males. Additionally,

PAGE 67

67 Moffitt and others (see Moffitt et al., 1994) have argued that female delinquency tends to be more erratic when compared with male delinqu ency. More recently, Piquero and Moffitt (2005) have contended that males do tend to show more involvement in crime/delinquency regardless of offender group, i.e., adolescent-limited or life-cou rse-persistent. Consequently, it is expected that a greater proportion of males will exhibit both types of offense trajectories and more specifically, being male will be a significantly distinguishing feature of group-based patterns of offending. In contrast, Sampson and Laubs age-graded theo ry of informal social control argues that social bonds or stakes in conformity in adulthood affect desistance from crime. It is expected that individuals who exhibit a virtual non-offe nding trajectory will also display the greatest amount of stakes in conformity since they are likely to be al ready bonded as youth and adolescence since they show no involvement in delinquency. Similarly, this assumed early attachment to institutions such as family and school is likely to continue into adulthood in that they will also be more likely to be married a nd have full-time jobs which will further restrict their involvement in crime. Second, while soci al disorganization is li kely to affect group assignment to an offender group, it is believed that the experience of life transitions such as marriage and employment will increase the likelihood that individuals will express an adolescent-limited/peaked offens e trajectory compared with a chronic offending pattern since these social bonds are theorized to have an eff ect on criminal career desistance. Or in other words, while these groups may have relatively similar patterns of offe nding during adolescence, these life transitions will be related to desist ance among the adolescent-limited/peaked offender group.

PAGE 68

68 In short, recent theoretical and methodologica l advancements in cr iminological research permits me to examine the possible generalizabi lity of formerly discussed offender groups (nonoffenders, adolescent-limited/peaked offenders, a nd chronic offenders) and exploring a potential integration of macro-level and micro-level theories for explai ning offending. These issues along with a description of the methodology and analytic procedures that wi ll be utilized in this current study will be the focal point of the following chapte r (Chapter 4) and the re sults chapter (Chapter 5) will involve a systematic exploration into the several key hypotheses mentioned below. Hypotheses and Expectations Hypothesis #1: According to Moffitts (1993) taxonomy a nd prior applications of trajectory analysis (for review, se e Piquero, 2007), it is e xpected that the semiparametric Poisson-based regression analyses will produce th ree distinct classes of offend ers including: 1) non-offenders; 2) adolescent-limited; an d 3) chronic offenders. Hypothesis #2: In accordance with the theoretical ex tensions provided by Moffitt (1994) and Moffitt et al. (1994), it is expected that sex wi th be a significant covariate for distinguishing group-based trajectories or in other words, males will be more likely to be in offending groups than females. Hypothesis #3: In line with Sampson and Laubs age-graded theory of informal social control, it is hypothesized that individuals who self-report havi ng experienced these key life events (or stakes in conformity) will be less likely to be chronic offenders. Hypothesis #4: It is expected that indi viduals whose neighborhood of origin is characterized by poverty/low neighborhood socioeconomic stat us and residential instability (i.e., social disorganization ) will be more likely to be chronic offenders than adolescent-limited offenders or non-offenders. Hypothesis #5: Macro-level and micro-level effects on the participation and frequency of offending will still be observed net of the effects of each otheror in other words, both macrolevel and micro-level effects will be significantly related to latent class membership. However, the possibility of mediation will also be explored.

PAGE 69

69 Figure 3-1. Integrated Theoretical Model: Social Disorganization, An Age-Graded Theory of Informal Social Control and Moffitts Developmental Taxonomy Social Disorganization Poverty Residential Instability Overcrowding Non-offenders AdolescentLimited/Peaked Offenders High Stakes in Conformity Male Low SES Household Black Moved # of Children Formal Social Control Natural Parents Chronic Offenders Differential Association

PAGE 70

70 CHAPTER 4 DATA AND METHODS Introduction The former two chapters (Chapters 2 and 3) provided the theoretical framework for explaining individual variations in offending and the latter chap ter expressed how macro-level and micro-level factors could affect an individu als pathway of offending and why males can be expected to be disproportiona tely found in chronic offending groups. This information along with the closing section of Chapter 3 that di scussed prior applications of Nagin and Lands (1993) group-based methodology all bear equal importa nce in framing the outline of this current chapter. This chapter begins with a description of the data and sample for which all the analyses of this current study are based. This is followe d by a section highlighti ng the macro-level and micro-level independent variables incorporated in this study to empirically test Sampson and Laubs (1993) age-graded theory of informal social control and Moffitts (1993) taxonomy of adolescent-limited and life-course -persistent offending. The macrolevel measures used in this study were developed with attention to those factors emphasized as having relatively strong effect sizes according to a recent meta analysis by Pratt and Cullen (2005) and the micro-level measures were theoretically-deriv ed and based on prior empirica l research. After the section describing the measures, a discussion of the sp ecific group-based modeling technique (Nagin & Land, 1993) used in this current st udy to identify latent classes of offenders is presented. Next, there is also a discussion devoted to ward fleshing out the recent debate on the usefulness of the technique that has occurred be tween the originators of the methodological tool and some of its most vocal critics. This chapte r will conclude with a description of the software and statistical procedures that will be used in or der to estimate the group trajectories and then to

PAGE 71

71 assess the relative strength of macro-level and micro-levels variables for predicting the likelihood of group membership. Data Sources The data used for these analyses were drawn from the first two birth cohorts collected by Shannon (1988, 1991) in Racine, Wisconsin. Racine was a Midwestern indus trial city with an estimated population of 89,000 residents according to the 1960 U.S. Census. Shannon initially gathered detailed information on three births co horts, the first being 1942, followed by a second cohort in 1949, and the third from the year of 1955. Overall, official police contact data was collected on all of th e individuals from each of the three cohorts from birth up until the end of the follow-up period (i.e., 1974) which yielded sample sizes of 1,352 for the 1942 cohort, 2,099 for the 1949 cohort and 2,676 for the 1955 cohort. In order to create a uniform baseline age of offending across cohorts, this current study only included information on police contacts afte r the age of six. Mo re specifically, Shannon was particularly interested in a dolescents that were defined as c ontinuous residents of Racine through their eighteenth birthdays (regardless of their mobility within the citys limits). This continuous-residence data (which provi des the data for the focus of this study) generated sub-samples of 633 for the 1942 cohor t, 1,297 for the 1949 cohort, and 2,159 from the 1955 cohort. Of these sub-samples, Shannon (1988, 1991) selected random samples for the purposes of collecting in-person interview data for the first tw o cohorts that began on June 1, 1976 (when the 1942 cohort was age 34 and the 1949 cohort was 27 years of age).1 Group-based offending trajectories will be estimated sepa rately for the 1942 and 1949 continuous-resident subsamples presented below. A later section in this chapter entitled Trajectory Estimation 1 According to Shannon (1988, 1991), the lack of continued federal funding restricted him from collecting interview data for the 1955 cohort.

PAGE 72

72 provides a more thorough discussion of the vari ous models and a tabular display can also be found below in Table 4-1. In addition, Table 42 illustrates the different years of observation that can be attributed to each of the data sour ces along with the correspon ding ages of the Racine continuous-residents by cohort. Thus, the final continuous-residence based sa mples previously alluded to involves taking the police contact and interview-level data separately for the 1942 and 1949 residents and merging this individual-level data with the limited available neighborhood-level data based on a unique Census block identifier for each individua l. Once the individu al and neighborhood-level data were merged and sample frequencies we re conducted, it was observed that a number of individuals were missing information on all of the neighborhood-level vari ables of interest. Therefore, in order to create an accurate base line and minimize missing data, we chose only to include the cases that had existi ng 1950 Census block information.2 This selection criterion left a total sample of 235 and 323 individuals fo r the 1942 and 1949 continuous-residents samples respectively. These samples were predominantly White residents (92% and 85%), followed by Black (6% and 11%), and the approximately remaining 3% of each sample were Chicanos. There was an almost even split as far as the sex composition between the males (47% and 53%) and the females (53% and 48%). 2 In order to incorporate change into the models (i.e., such as computing change scores for the Census measures by using 1960 data as well) would have resulted in the loss of more than 60 additional individuals which would have affected the stability of the trajectory models (see Nagin & Tremblay, 2005; Sampson & Laub, 2004). Preliminary comparisons indicated that the 1950 and 1960 block data were relatively similar; therefore, the 1950 Census measures were used with the intention to preserve any additional sample loss and to represent macro-level-level characteristics that existed closer to the early developmen tal years of the Racine residents that Moffitt (1993) has indicated as a crucial time period (see also Gottfredson & Hirschi, 1990). Prior research has shown the adverse effect of living in a low SES neighborhood on factors asso ciated with life-course-persi stent patterns of offending, especially in early childhood (see Brooks-Gunn et al., 1993; Chase-Lansdale et al., 1997; Duncan et al., 1994). The original principal investigator Shannon (1994:7) also indicat ed that if a researcher was interested in the ecological characteristics of which the cohort memb ers spent their early childhood/juvenile years, then the 1950 Census data should be used.

PAGE 73

73 Independent Variables Macro-level independent variables Social Disorganization. Based on the factors highlighted by Shaw and McKay (1942) that are believed to contribute to soci al disorganization and associated with crime (as defined as the breakdown of the formal and informal social networks for regulating neighborhood resident behavior), measures of neighborhood socioeconomi c status (SES) and residential mobility were included.3 Neighborhood SES was measured directly by the average housing values of the residential properties on each block. Prior research has shown the positive effects of living in a high SES neighborhood has on academic achievement, verbal ability scores, and behavioral problems (see Klebanov et al., 1997). Resident ial mobility was indirectly measured by the percent of housing units on a bloc k that were renter-occupied co mpared with those that were owner-occupied.4 Simcha-Fagan and Schwartz (1986: 671) have argued that structural characteristics can affect an individuals ab ility to form social bonds to the community. Individuals who rent as opposed to owning their pr operty are less likely to develop the necessary ties (see Bellair, 1997; Granovetter, 1973) that are important for providing a mechanism of informal social control. Finally, percent overcrowding was Census-based measure indicating the percent of rooms within hous ing units on each block with more than one person per room. Wilson (1987) has also attributed housing dens ity to concentrated poverty, which has been 3 Because of the sparse amount of U.S. Census data availa ble in 1950 and relative homogeneity of Racine, this study was unable to include a measure for ethnic heterogeneity in order to provide a complete measure of social disorganization incorporating all of the processes that lead to crime. 4 The original measure of social disorganization used by Land (2000) was a scal e created by Shannon (1988, 1991, 1994) where a rank of 1-26 was assigned to an individual to represent their neighborhood SES based on the average Census block characteristics of the area in which th ey lived. This measure was actually referred to by Shannon as a natural area code with no reference to social disorganization and given that the measures used (average dollar value of rent, average dollar value of owned property, percent lacking plumbing, percent overcrowding, and percent renter-occupied) did not load on one construct it is likely that Shannons (1988, 1991, 1994) index and Lands (2000) later use of this index is not an entirely accurate proxy measure of social disorganization.

PAGE 74

74 posited as a precursor of social disorganization. Dwelling units that are over-crowded also may lead to sense of anonymity between the residents because of their failure to interact with their neighbors (see Roncek, 1981; Sampson, 1995). The results from principal components analysis confirmed that these three variables were measuring a similar construct (i.e., social disorg anization) across the two cohorts as presented below in Table 4-3; therefore, the corresponding component loadings were utilized in order to construct the appropriate indexes.5 This procedure allows for pres erving the integrity of the data by removing the potential of a partialing fallacy (Gordon, 1968) and reduci ng the potential for multicollinearity which is common in macro-le vel data (Land, McCall, & Cohen, 1990). Thus, positive z-scores were indicative of more socially disorganized areas Micro-level independent variables Stakes in Conformity. Sampson and Laub (2005c) argue that involvement in institutions such as marriage and work provide both short-te rm and long-benefits for individuals. These life transitions can change an individuals motivatio n for crime by re-focusing their attention toward spending time with their spouse and/or investing tim e in their job. These events are believed to foster an individuals development of prosocia l adult bonds. Prior research has shown support for the impact that these transitions can ha ve on desistance from crime (Sampson & Laub, 1993; Warr, 1998) and other research has shown that the e ffect of these life transitions are age-graded (see Uggen, 2000). Therefore, in accordance with Sampson and Laubs (1993) theory and prior research (see Piquero et al., 2002) a stakes in conformity inde x was created for the two local life circumstances 5 This procedure involves multiplying the raw variable scores by the standardized weights obtained as a result of the principal components analyses and then summing these values together to produce a single measure (see Kim & Mueller, 1985).

PAGE 75

75 of marriage and employment. Each of these tran sitions are dummy-coded with 1 indicating if the individual self-reported havi ng experienced the particular life transition by age 25 and 0 indicating if the individual faile d to report having experienced th e specific life transition by age 25.6 Therefore, the index ranged from 0 to 2, w ith 2 indicating that th e individual reported having experienced both of these life circumstances.7 Sex. Sex has long been recognized as one of the strongest predictors of involvement in delinquency and violent crime where crime rates have consistently shown higher prevalence rates of offending for males when compared with females (Rowe et al., 1995). Daly (1998) has also noted that self-repor t studies have repeatedly indicated th at males report a higher prevalence (involvement) and incidence (number of acts) of offending. Wilson and Hernstein (1985) have commented on the strength of the relationship be tween sex and crime over time and cultures, and according to Gottfredson and Hirschi (1990), an ywhere and everywhere males tend to commit more crime than females (with the exception, check fraud, shoplifting, and prostitution). Additionally, Moffitt and some of her colleagues have suggested that life-course-persistent offending may be observed almost exclusiv ely among males because neuropsychological disorders and childhood-onset condu ct problems are more preval ent among males (Rutter, 1983; Singer et al., 1968; Werry, 1986). Female delinque ncy also tends to be more erratic when 6 There were self-reported measures that existed in the da ta for yes/no responses to these life events; however, in order to ensure that the actual life transition occurred with in the observation period of the data available for both cohorts, the experience of these lif e transitions after age 25 for the 1942 cohort were not included. 7 It is important to note up front that this measure is limited in that Sampson and Laub have argued that mere participation in life transitions may not be as important as the quality of the attachment to the various social institutions (i.e., marriage, work, etc.). Laub et al. (1998) also indicate that the inhibiting effects of life transitions are cumulative over time and it takes a while for individu als to express the necessa ry level of cohesion or attachment to affect their behavioral pathway of offending, or affect their desistance from crim e. In the end, according to Piquero et al. (2002) it is likely that just being married or having a job is some kind of indication that individuals have developed some affective ties to their spou ses and/or jobs or else they would not be likely to be involved in the life circumstance to begin with.

PAGE 76

76 compared with male delinquency (Moffitt et al., 1994). Prior research has shown that females typically show less involvemen t in crime in both offender groups (adolescent-limited offenders and life-course-persistent offenders) when compared with males. Recent applications of Nagin and Lands (1993) trajectory analysis have al so shown distinct offender groups among females (see Cote et al., 2001; DUnge r et al., 2002; Fergusson & Horwood, 2002; Land, 2000; Piquero et al, 2005). Provided with th is theoretical framework and building on the prior research mentioned above, a dichotomous variable re presenting the sex of the 1942 and 1949 cohort members was included with males coded as 1 an d females coded as 0 as a covariate in the trajectory models as well as in the multi nomial logit models presented in Chapter 5. Black. Similar to sex, race has also been a long-standing correlate of crime (see Hawkins, 1995) and Moffitt (1994) has argued that the hi gher prevalence of both adolescentlimited and life-course-persistent offenders among Blacks may account for the racial variation in crime rates because the theoretical origins of the latter patter n of offending disproportionately affect Blacks because of the strong association between race and poverty. Prior research has shown a higher prevalence of lif e-course-persistent offenders among Blacks (Piquero & Buka, 2002), and other researchers have argued that the risk factors associated with life-coursepersistent offending typically tend to be disp roportionately concentr ated among Blacks (see Piquero & Moffitt, 2004; Piquero et al., 2005). Furthermore, macro-level research has indica ted that Blacks were more likely than Whites and Hispanics to live in poor neighborhoods and be involved in more serious forms of criminality. Prior research has also shown that Blacks who reside in these areas are disproportionately affected by th e burden of child-care and resi dential instability and these factors were strong predictors of delinquency an d violence (for review, see Gephart, 1997).

PAGE 77

77 Additional research utilizing trajectory analysis has also noted the increas ed likelihood of Blacks in chronic offending groups (see Land, 2000); therefor e, it was necessary to include race. The original nominal level variab le indicating the race of the 1942 and 1949 cohort members was recoded into a dummy variable w ith Blacks being coded as 1. Low SES Household. This prestige-based measure will be included in order to provide an individual-level SES complement to the macr o-level measures of neighborhood SES. Prior research has shown that i ndividuals who are from low SES households in low SES neighborhoods tend to have the hi ghest rates of offending and comparatively those from the high SES households in high SES communities have the lowest offending rates (Braithwaite, 1979; Reiss & Rhodes, 1961; Wikstrom, 1991a). Usin g trajectory analysis, McDermott and Nagin (1998; see also Land, 2000) have also shown that low SES is an important factor for predicting group membership, more specifically the likelihood of being classified as a chronic offender. Additionally, Brooks-Gunn and Leve nthal (2000) have noted the importance of including this family-level measure when estimating neighborhood effects. This variable was measured as a scale i ndicating the prestige of the head of the households occupation in the family in which the individual grew up. It ranged from a low of one to a high of nine with the following occ upational types represente d: (1) professional, technical, managerial; (2) clerical and sales; (3) craftsman, foreman; (4) operative; (5) maintenance and service; (6) priv ate household labor; (7) industrial labor; (8) farm, agricultural labor; and (9) unemployed. Consistent with the direction of the macro-level social disorganization index, higher valu es indicated worse situations (i.e., growing up in lower SES households).

PAGE 78

78 Family Stability. This measure was used to represen t the stability of the family dynamic in which the cohort member was socialized. Most research has shown an established relationship between family stability and delin quency/crime (Glueck & Glueck, 1950; McCord & McCord, 1959; Goetting, 1994) and crime rates (Sampson, 1987). Studies have found that adolescents living with both bi ological parents have lower rates of offending when compared with those who reside in broken homes or recons tituted families. According to Gottfredson and Hirschi (1990:104), single-parentheaded households (usually by a woman) may be unable to provide the same amount of supervision because of the time constraints imposed on them and thus may be more likely to have negative res ponses to their childrens unwanted behaviors. Furthermore, this variable along with the followi ng measure listed below (family structure) have been identified as key family-level variable s when estimating neighborhood effects (see BrooksGunn & Leventhal, 2000) Originally, this variable was based on a scale created by Shannon (1988, 1991) which combined responses to several items regarding family composition ranging from 1 to 20, where according to Shannon (1988, 1991), 1 represented the most stable family dynamic (two parents present during the entire childhood /juvenile period) and 20 the leas t stable family dynamic (lived only with grandparents, never with parents). For the purposes of this stud y and due to the little variability among the categories, the scale was collapsed into a dichotomous variable with 1 indicating the presence of two parents during the entire ch ildhood/juvenile period and 0 indicating any other type of less stable family dynamic. Family Structure. In order to account for the possi ble differential opportunities of parental supervision affecting individuals offending pathways (Gottfredson & Hirschi, 1990) along with having a complementary individual meas ure of household density to the macro-level

PAGE 79

79 household density measure (percent overcrowding) a second measure of family dynamics was included to represent family structure. This wa s a measure of the number of children within each of the cohort members family (including the cohort member).8 Differential Association. Differential association theory (Sutherland, 1939, 1947) has long stressed criminal and deviant behavior is learned through exposure to and acceptance of definitions favorable to crime. Burgess and Ak ers (1966), and in later reformulations by Akers (1977), expanded on this idea by specifying the pr ocess in which learning occurs. Favorable definitions of crime and deviance are learned in a social setting where the imitation of criminal and deviant behavior also occurs Delinquency, like any behavior, ge nerates either beneficial or adverse consequences positive or negative reinforcem ent. If delinquent or criminal behavior is reinforced either socially or nonsocially, the behavior is likely repeated in the presence of cues to which it has been paired. Or in other words, Akers (1998:50) argues th at the probability that individuals will engage in deviant behavior is in creased when they differentially associate with others who commit the behavior and express favorable definitions toward it. Thus, based on research that has produced strong evidence for wh at is now termed the perceived delinquent peer effect (Felson & Haynie, 2002; Matsueda & Anderson, 1998; Schr eck, Fisher, & Miller, 2003; Warr, 1993; Warr & Stafford, 1991), a measur e was included with 1 indicating that the 8 Overall, there was very little missing data at the individual-level (less than 1% for any of the variables used in the analyses). The only two variables that had missing data were family stability and family structure. In order to correct for the missing values, th e method of group specific means were computed. In order to do this procedure the sample was sorted into five categories: 1) White males; 2) White females; 3) Black males; 4) Black females; and 5) Chicano males (there were not any Chicano females in the sample). For the few cases (n=4) where a missing value was recorded, the sex/race specif ic group means were imputed.

PAGE 80

80 individual self-reported peer influence as th e strongest contributing factor toward their delinquent/criminal behavior in junior high and high school.9 Formal Social Control. This measure was a scale of the individuals perception of the level of police patrol in their neighborhood during their junior high and high school years.10 This indicator of institutionalized formal social control ranged from 0 repres enting not at all or no police patrol, 1 for lightly patrolled, 2 for moderately pa trolled, to a high of 3 which indicated that the individual perceived their neighborhood as be ing heavily patrolled during their early to late adolescent years.11 Moved. In order to provide a contextual measure of residential mobility to complement the compositional measure included in the social disorganization index, th is variable was coded as 1 if the individual had moved within Racine during th eir juvenile years. It is likely that moving, particularly during the adolescent years, is likely to have an adverse effect on an 9 It is worth mentioning here that this is certainly a less than ideal measure of differentia l association and more direct measure would have been prefer able; however, due to the limitations of data this measure was included as a control variable. 10 It is important to note that this measure is a weak measur e of formal social control an d that direct measures such as prior conviction or incarceration are optimal; however, there was only two items that allowed the individuals to report having received jail time as a result from their polic e contact and these questions were only in reference to their first and second police contact. Because only 1 to 4 in dividuals across the cohorts actually reported this as a response I made the determination to use the police patrol measure because there was a sufficient amount of variation in responses to this latter survey item. Additionally, this measure is also limited due to the fact that it was asking individuals to retrospectively answer this question based on their perceptions of how they felt as a middle and high-school juvenile (which is at least 10 years prior and for the 1949 cohort and 20 years prior to the date of the interviews in 1974 for the 1942 cohort) and it is also likely that those individuals who actually came into contact with the police may have had a bias toward reporting higher levels of neighborhood patrol. Future studies would benefit from incorporating more direct measures of formal social control such as those discussed by Sampson and Laub (1997) and De Li (1999), which relied on convictions and on court appearances and convictions, respectively. 11 Shannon (1988, 1991, 1994) originally developed a scale to assess formal social control, but this measure had a series of flaws, with the most serious being that the scal e was constructed by summing self-reported responses of the cohort members perceptions of the (1) level of police patrol when they were in junior high and high school (1=heavily, 2=moderately, 3=lightly; 4=not at all); (2) their and a couple of their friends views toward the police in junior and senior high, (1=positive, 2=negative, 3=indifferent) ; and (3) to report whether or not they got into trouble with the police with a couple of their friends (yes/no). Even more oddly, this scale was constructed by coding an individuals response as 1 if they had either negative or indifferent views toward the police, adding 2 points if the individual reported that they and their friends got in trouble with the police, and adding points if the individual also responded that they perceived police patrol was either heavy or moderate during their juvenile years.

PAGE 81

81 individuals pattern of offending by disrupting the ability to form /maintain prosocial bonds with peers and their schools, which ar e important factors for inhibiti ng delinquency (see Hirschi, 1969; Akers, 1985).12 Dependent Variables Delinquent/Criminal Involvement. The variable used to measure offending was the number of official police contacts for each individual from age 6 to age 25.13 The police contact measure was limited to those instances in which an individual came into contact with a Racine police officer specifically for a crime that he/she was suspected to have committed or did in fact commit. In addition, the official police contacts were also restrict ed to contacts that had true crime codes with a felony/misdemeanor distincti on as well as contacts for status offenses to measure offending (i.e., excluding tr affic offenses, and city ordinance violations) because (1) they are a more accurate official measure of t rue offending when compared with arrests and convictions; (2) they are closer in time to the actual criminal ac t than self-reported offending (see Hindelang et al., 1981; Thornberry & Krohn, 2003); a nd (3) the inclusion of status offenses increases the potential for identifying a dolescent-limited offenders (see Moffitt, 1993).14 Offending Group Membership. The results of the semiparametric mixed Poisson-based regression models will guide th e identification of the number of offending groups that exist within these data and the shapes of their offending trajectories. The identified latent classes of offenders will operate as the dependent vari able for the estimation of the multinomial 12 It is important to note that this measure is limited in that is was impossible to discern from the available data if the individual moved from a social disorganized area to a more or less disorganized area; therefore, it is important to interpret the findings with caution and keep in mind that the expectation is that moving is seen as having an adverse effect on offending patterns within this study. 13 This police contact data refers to the years that are consistently available for the 1942 and 1949 continuousresidents or up to and including age 25. 14 Prior trajectory analyses with Shannons (1988, 1991) Ra cine cohorts relied solely on felonies and misdemeanors (DUnger et al., 1998) or everything including tra ffic offenses and suspicion contacts (Land, 2000).

PAGE 82

82 (polytomous) regression models (Models 1 and 2) with the non-offender group omitted in order to serve as the reference category. The followi ng sections describe the group-based modeling technique designed by Nagin and Land (1993) that wi ll be utilized in this study for identifying the offender groups along with a discussion of th e existing debate that has been ongoing related to the usefulness of the methodology. Study Limitations There are a series of limitations of the data source itself and th e measures used in this study that are worth noting. First, the timeliness of th e data is an issue sin ce the cohorts themselves were born in the 1940s, socialized as children and/or adolescents in the 1950s and 1960s. Therefore, it is likely that these samples may not be generalizable to other cohorts from more recent time periods because they were born during WWII (e.g., 1942 cohort) or are from the baby-boomer generation (e.g., 1949 cohort) who grew up in the post-WWII era. Second, Racine, WI was a Midwestern industrial city with an estimated populati on of 89,000 residents, and it is not likely that the re sults from this study would apply to the larger metro politan urban cities of today (i.e., cities with populat ions greater than 100,000). Third, particularly during the time of data collection, Racine was a relatively homogenous city with limited numbers of minor ities, and according to Shannon (1988), although the overall population did rise to 95,000 persons by 1970, this was followed by a decline to 86,000 persons by 1980. He attributed this popula tion decline to a growing number of young people in need or work and increased female part icipation in the work force, but this was met with a decreased number of individuals leaving for retirement. Taken together, the results discovered here may not necessarily apply to ar eas with larger minority representation or dissimilar population shifts.

PAGE 83

83 Noting the limitations of the cohort members a nd the location, there is also a series of limitations with the measures used in this analysis First and foremost, the data available in the 1950 Census at the block level is very limited or in other words, there was not the typical long list of measures accessible to macro-level resear chers today for this research, and there also was a noticeable amount of missing 1950 census data for the 1942 and 1949 continuous-residents. With this is mind, the social disorganization me asure used here is relatively weak in the sense that it is only measuring two (neighborhood SES and residential mobility) of the three elements that Shaw and McKay (1942) highlight ed as important factors associated with the breakdown in informal social control and crime rates. Once again, Racine was a relatively homogeneous area, so it is likely that racial dissimilarity would be difficult to directly observe at the block level. Additionally, Sampson et al. (1997) have argued that what matters more than social disorganization is collective efficacy or wh at happens at the individual level as far as developing community ties and informal social cont rol. There were not an available individuallevel data that would allow for an investigati on of this hypothesis, so caution should be noted that these individual-level processes are not accounted for in this study. In addition, as footnoted earlier the stakes in conformity index used to assess Sampson and Laubs (1993) age-graded informal social contro l theory is limited in that it is a dichotomous index with a short range (0 to 2). More re cently, Laub et al. (1998) have argued that mere participation in local life circumstances may not be as important as the quality of the attachment and that the inhibiting effects of life transitions on desistance are cumulative and gradual over time. Therefore, while this measure does not captu re the level of attachment or cohesiveness of the relationship, it is likely that just being married or having a full-time job is indicates that some degree of attachment has already been formed or else the life transition would probably not have

PAGE 84

84 occurred in the first place (see Piquero et al., 200 2). The differential association measure is also noticeably weak. Regarding the dependent variable (police contacts), there ar e also well known strengths and weaknesses from using either official data (p olice contact, arrest, co nviction) or self-report data to measure criminal behavior. Under-reporti ng is one of the principal limitations of official statistics because the majority of police activity is a direct result of the wi llingness of citizens to report crimes (see Hindelang et al ., 1981; Thornberry & Krohn, 2003). However, there are also extensive disparities in victim reporting based on various factor s including: requirements for insurance companies, confidence or lack of confidence in police de partments, etc. Furthermore, there are many offenders who will never come to the attention of law enforcement, such as shoplifters, burglars, or white-c ollar criminals (Junger-T as & Marshall, 1999). This result might be due to the fact that more skillful and shrewd offenders are less likely of being arrested and/or may reflect the differences in pol ice discretion and priorities acro ss regions/agencies (Blumstein & Cohen, 1987). One key advantage of self-report measures is that they often offer more comprehensive information than official data. They pr ovide background and soci odemographic data on individual offenders, as well as personal characteristics. Research has indicated th at self-reports are effective in answering questi ons of causality, motivation, cha nge, and stability in criminal behavior (Junger-Tas & Marshall, 1999:296). Self-reports also tend to include both a broader range of deviant behavior and less serious types of crimes than official police records indicate (Blumstein, Cohen, Roth & Visher, 1986:32). Some examples of the types of crimes that are likely to be found in self-reports but not in offi cial data are prostitution, drug abuse, public order crimes, and drunk driving.

PAGE 85

85 Notwithstanding the mentioned benefits, there are several limitations to the self-report method. Under-reporting and poor recall of crimin al activity are disadvantages because selfreports tend to be retrospective in nature (as was the case with the interviews in these data). In addition, some respondents may even inflate their actual level of criminal activity. Cross-study comparisons of self-reports may also be hindere d due to differences in the wording of the questions, differences in the administration of the survey, or differences in time periods (Blumstein et al., 1986:33). Taki ng into account the noted limitati ons of police records and selfreported measures of offending, this study chose to measure police contacts because they are a (1) more accurate official measure of true offending than arrests and convictions, and (2) typically closer in time to the actual criminal act than self-reported offending which sometimes asks individuals to recall their pa st behavior in their lifetime, last twelve months, last six months, etc. (see Hindelang et al., 1981; Thornberry & Krohn, 2003). Within the context of all of the above mentioned limitations surrounding the data, time period, samples, and the location along with the measurement shortcomings, the following section highlights the group-ba sed methodology that will be used to estimate the group trajectories. Methodology Group-Based Modeling Longitudinal research in its most basic form (as applied to the study of crime), involves the identification of a cohort at birth and following th at group of individuals throughout their life, or at least long enough to capture the termination of the majority of criminal careers (Blumstein et al., 1986:104). The longitudinal method allows for a specified cohorts level of involvement in crime to be traced as the cohort, itself, matures (Greenberg, 1985). This knowledge is crucial for detecting causal factors or sequen ces that associate individual char acteristics or certain events to

PAGE 86

86 consequential offending (Blumstein et al., 1988). Longitudinal data also permits the use of complex behavioral models because they allo w one to examine the effects of missing or unobserved variables that may be correlated w ith crime (i.e., also known as the unobserved heterogeneity problem). It is often the case that even among people who have similar characteristics (race, age, sex) there is a s ubstantial amount of heterogeneity. Thus it is important to utilize longitudina l models or run the risk of using misspecified models and producing inconsistent paramete r estimates (see Nagin, 2005). According to Nagin (2005:1-2), longitudinal data also provide researchers with the basis to study developmental trajectories. Most of the common methods for studying developmental trajectories, however, are those th at are calculated to account for individual variability, which masks the possibility of observing meaningful su bgroups within certain samples and populations. Furthermore, this over-reliance on traditional longitudinal approaches has led to much ad-hoc theorizing and classifying individuals on the basi s of what could be random variation. Hence important developmentally distinct groups could be missed or misclassified. The particular method used in this study to model the group-based offending trajectories was originally developed by Nagin and Land (1 993) and is known by several different names including: group-based modeling, fi nite mixture models; trajectory analysis, and semiparametric Poisson-based regression analysis. It is impor tant to note that this technique shares some similarities with the traditional group-based mo deling approaches such as hierarchical linear modeling (Byrk & Raudenbush, 1987, 1992; Goldst ein, 1995) and latent curve analysis (McArdle & Epstein, 1987; Me redith & Tisak, 1990). Still, there are distinct differences. For in stance, both hierarchical modeling and latent curve analysis assume continuous distribution fu nctions. Trajectory analysis does not assume

PAGE 87

87 this continuous distribution, but ra ther assumes that there exist di stinct clusters of group-based trajectories that exhibit differe nt etiological processes, and the methodology approximates the continuous distribution by identify ing points of support (i.e., dis tinct groups or trajectories) (Nagin, 2005). These trajectory models are also based on a maximum likelihood (ML) function, which of course permits them to display the qualities of ML parame ter estimatesthey are consistent and asymptotically normally distri buted (Cramer, 1946; Greene, 1990). Given the assumptions and statistical prope rties, this method assumes that individual differences in behavioral trajectories can be described as pol ynomial functions of age or time. The following formula presents the probability fo r belonging to a particular group: J P (Yi) = j Pj (Yi) j where P(Yi) is the unconditional probability of observing individual i s longitudinal sequence of police contacts, Yi, over time. It is equivalent to the sum across J groups of the probability of Yi given i s membership in group j weighted by the probability of membership in group j (see Nagin, 2005:25). There are various forms and derivatives of th e trajectory model, but the two most common parametric forms that are likely to apply to th e data used for this study are the censored normal model (CNORM) and the zero-inflated Poisson m odel (ZIP). The CNORM model is appropriate for data where the outcome variable of interest is normally distributed bu t there exists clustering of the data at either the minimum or maximum (which typically occurs with offending data). This model can also be applied to data wit hout censoring by assigning the scale maximum and minimum values to values that are respectively sm aller and larger than any value observed in the data. The equations listed below correspond to the statis tical derivations of the cases when (1)

PAGE 88

88 the data are censored at the scale minimum; (2 ) normally distributed between the scale minimum and maximum; and (3) the data are censored at the scale maximum: pj ( yit = S min) = ( S min j xit) / pj ( yit) = 1 / ( yit j xit / ) for S min yit S max, and pj ( yit = S max) = 1 ( S max j xit) / where it is assumed that yit is normally distributed with a mean of j Xit and conditional on age ( t ) with a standard deviation of and represents the cumulative di stribution of a normal random variable. In addition, S min represents the scale maximum and S max represents the scale minimum (Nagin, 2005:30). In contrast, the parametric form of the trajectory model that is needed for modeling countbased data such as the number of police contacts is the ZIP vers ion, which was first presented in Nagin and Lands original work and is based on the generalization of what Lambert (1992) called the zero-inflated Poisson (ZIP) distributi on. Count-based models have been more frequently utilized in the social sciences in an effort to correct for th e inefficient and biased estimates that are produced when linear regres sion is employed for modeling data where the dependent variable is measured as a count or frequency of an event (for review, see Long, 1997). Although the Poisson regression model is the most basic form of the count-based models, it is often not the most suitable model due to its prim ary assumption of mean variance equivalence, or equidispersion. It is more like ly the case that, in practice a nd especially in criminological research, the distribution of the outcome variable is overdispersed, (i.e., the variance exceeds the mean); therefore, a more appropriate mode ling procedure is recommended (Long, 1997:218).

PAGE 89

89 The ZIP model allows for the assumption of a dual-state-system. A dual-state-system signifies two situations: one in a zero-contact state where the assumption is that a group of individuals exists that can be regarded as police-c ontact free, and the s econd state or non-zero contact state where the expected incidences of police contacts are assumed to follow some known distribution, such as the Poisson (Long, 1997; Lambert, 1992). Several other studies have also noted the practicality of the ZIP model wh en using count data with excess zeros (King, 1989; Land, McCall, & Nagi n, 1996; Zorn, 1998). The ZIP model takes the following form where the first term is the probability of no police contacts and the second term is the probability of the combined events being active and the probability of zero police contacts given jt > 0 (Nagin, 2005:35). j yit pj ( yit) = (1 t ) ( jt e jit / yit!) ( yit = 1, 2,.), Once the parametric form of the trajectory model is established it is necessary to determine the functional form of the groups trajectories and there are three possibil ities including a linear, quadratic, and cubic form (or age, age-squared, and age-cubed). Depending on the ages and the length of the observation period, it is unlikely that individual o ffending will have a linear form. Similarly, although the quadratic model is more often used in research (see Bushway et al., 2003:140), it only permits one major directional cha nge in the trajectory, and it is likely that cohort members will and have undergone periods of intermittency in their involvement with the police. Intermittency periods refer to those point s in time (i.e., years) where no official police contacts were recorded for any pa rticular cohort members. Theref ore, it is likely that the cubic model will demonstrate the best functional form of the trajectories. The CNORM and the ZIP equations that allow for all three of these possible functional forms are presented below: yit = 0 j + 1 jAgeit 1 + 2 jAgeit 2 + 3 jAgeit 3 + it (CNORM),

PAGE 90

90 ln ( t j ) = 0 j + 1 jAgeit 1 + 2 jAgeit 2 + 3 jAgeit 3 + it (ZIP), where Age1 it, Age2 it, and Age3 it are the cohort members ( i s) age, age-squared, and agecubed respectively, t represents time, and it is a disturbance parameter that is assumed to be normally distributed and the s are the parameters that determine the functional form of the trajectory (Nagin, 2005:28-33). Finally, the exact number of groups is chosen based on an examination of the Bayesian Information Criteria (BIC) in orde r to maximize model fit. This process is necessary because conventional likelihood ratio tests cannot be used to assess whether the addition of another group adds any explanatory power to the model (see DUnger et al., 1998). In the end, the exact determination of the number of groups is based on the ability of the model to find a relative convergence using all of the available informati on and by the utility of the groups for indicating distinct and homogenous trajector ies, while also having a suffici ent number of persons within each group to allow for further sta tistical analyses (Bus hway et al., 2003:140; see also Brame et al., 2001). The BIC values can be relied on to determine the numb er of groups in both nested and unnested models (see DUnger et al., 1998; Nagin, 1999), and the BIC values are estimated based on the following equation, where L is the maximum likelihood, n is the sample size, and k is the number of parameters (Nagin, 2005:64): BIC = -2 log ( L ) + log ( n ) k A Seductive or Useful Statistical Method? More than thirty years has past since Hirsch i (1969:53) initially pr ovided the word of caution to criminologists when using offender typol ogies by asserting that it begs the question of causal homogeneity by focusing exclusively on the question of behavioral homogeneity. Criminologists have come a long way since Hirschis earlier remarks regarding offender

PAGE 91

91 typologies; however, this does not mean that Nagi ns and Lands latest methodological technique has been insulated from theoretical scrutiny. The group-based modeling technique has been the source of recent criticisms and counterpoints between its origin ators and two prominent lifecourse theorists, Sampson and Laub. This debate began with the findings from an earlier study by Sampson and Laub (2003) questioning the stability of offendi ng patterns over the life course a nd more specifically with an article published in the Journal of Quantitative Criminology entitled Methodological sensitivities to latent cl ass analysis of long-term criminal tr ajectories. The authors (Eggleston, Laub, & Sampson, 2004) argued that the enthusiasm for the use of Nagin and Lands (1993) semiparametric group-based models has not b een equally supplemented with robustness and sensitivity analyses that are ofte n inherent in offending data. They proceed by summarizing the obvious advant ages of these models (see also Nagin, 2005) that include (1) the fact that they are desi gned for longitudinal data ; (2) they lead to the ability to distinguish groups of offenders that ar e homogenous within their trajectory but distinct from other trajectories; (3) th ey provide an objective system atic criterion for establishing offending groups by using the Bayesian Informa tion Criterion (BIC) rath er than presupposing subjective sub-groups of offenders; (4) posterior probabilities are used to assess the precision of group assignment; and (5) this method accommodates missing data, whereby offenders that have unavailable information for several time periods ar e not excluded from the analysis (Eggleston et al., 2004:2). Relying on 500 juvenile delinquent s data from the archives of the classic Gluecks Unraveling Juvenile Delinquency research (1950, 1968), Egglest on et al. (2004:23) concluded that the length of follow-up can affect the trajecto ry patterns observed and argue that data that

PAGE 92

92 focuses on say age 6 to 15 versus 10 to 32 may pr oduce trajectories with different curves (i.e., shapes), peak ages of offending, and possibly membership to different offending groups. Additionally, they argue that research using trajectory an alysis needs to attend to both incarceration and mortality information because these factors can also affect offending trajectories (see Piquero et al., 2001 for a similar discussion on the importance of accounting for incarceration data). This article discussed above prompted an initial response by Nagin (2004) and a subsequent response by the Sampson, Laub, and Eggl eston (2004) that were both printed in the same issue of the Journal of Quantitative Criminology Nagins (2004) initia l response indicated that he had no criticisms regarding the anal ysis conducted by Eggles ton et al. (2004) and typically agreed with their fi ndings; however, he did note some counterpoints to a few of their conclusions. Specifically, Nagin (2004:28) argu ed that the criticisms offe red by Eggleston et al. aimed at the trajectory model are limitations that plague all longitudinal data (i .e., projecting beyond the range of the data and missing data resulting from incarceration and/or deat h). Other researchers have also commented on the idea of career termination or false desistance resulting from death (see Reiss, 1989) or based on data that does not track the individua ls throughout their entire life course (see Bushway et al., 2001; Piquero et al., 2003). Nagin concluded by reminding researchers that the groups that result based on the tr ajectory models must be understood as an approximation and this warning was already previously presen ted in Nagin and Lands (1993) original work. Sampson et al. (2004) reanalyzed the Gluecks data in response to Nagins comments and subsequently provided several more suggestions for future resear chers relying on the trajectory

PAGE 93

93 method. They also felt that the majority of thei r criticisms should also be heeded by any future criminological research incorporating longitudinal data. First, using random samples of varying sizes, they found that the number of groups only began to stabilize with samples of over 200. Other research has indicated that the number of identified groups is partly a function of sample size (see Nagin & Tremblay, 2001) a nd that the number of groups te nds to be consistent with samples above this range (s ee DUnger et al., 1998). In their final remarks, Sampson et al. (2004:41) indicate that the larg er problem with the use of trajectory analysis and inferences ma de from longitudinal data with short follow-up periods concerns the causal theory for using groups. They state that future research should resort to grounding all methods in strong theory and be wary of models that provide evidence of an existence of groups just because th ey are statistically discovered or what is referred to as the reification of groups (see also Laub & Sampson, 1993; Sampson & Laub, 1993). As if this above exchange was not enough to settle the debate and prompt future longitudinal researchers to be mindful to theo ry and consider the estimation of groups in trajectory analysis as approximate, this dialogu e was continued most recently in a 2005 issue in the journal of Criminology Nagin and Tremblay (2005a:873) set forth with an attempt to debunk three commonly held misconceptions regardi ng trajectory analysis that originate from researchers failure to understa nd that the method results in an approximation of groups including: (1) the individuals actually belong to a trajectory group; (2) the number of trajectory groups is immutable; and (3) th e trajectories of group members fo llow the group-level trajectory in lock step. Regarding the first misconception, Nagin and Tr emblay (2005a:882) state that trajectory groups are referred to as is they are real/t rue entities only for the purpose of literary

PAGE 94

94 convenience in that it facilita tes the ability to clearly discu ss groupings of individuals that display similar offending patterns. Since the di scovering of trajectory groups is a statistical approximation, they are not meant to be interpre ted as factual representations of reality (a common limitation of all statistica l models). Similarly, this fals e understanding of the fact that the method results in a statistical approximation of groups also has fostered the development of the second misconception, i.e., that the number of groups in the sample is immutable. According to Nagin and Tremblay (2005a: 888), the number of groups and the shapes of their offending curves are not a fixed reality e ither. Instead, the number of groups and their respective trajectories ar e influenced by the quality of the long itudinal data that are relied on for the analysis. In this sense, the quality of the da ta refers to the sample size and the number of observation periods, and obviously larger samples and longer follow-up periods are always the most optimal. The research presented above al so showed that the number of groups and the shapes of their trajectories were relatively stable in samples that were larger than 200 cases (DUnger et al., 1998; Sampson et al., 2004). Nagin and Tremblay (2005a:892) concluded by offering a couple of comments concerning the final misconception (that indi viduals will follow their offendi ng groups trajectory in lock step). First, the shape of the groups trajecto ry within which the individual is assigned is intended to reflect a behavioral process over time versus possible individual variation across this identified pattern of offending. There will always be individual variation, just as there is variation in offending if a rese archer relies on other units of analysis (households, cities, counties, etc). It is highly unlikely that any one particular member of any these aggregations will exactly display the average offending rates of th e group. In addition, ther e are also statistical

PAGE 95

95 reasons why individuals within a group trajectory will deviate from the process captured over the entire period of observation, most not ably involving sampling variation. Notwithstanding this further attempt by Nagin and Tremblay to defend the methodology and to caution its users about the limitations of trajectory analysis this article was inevitably met with a rejoinder by Sampson and Laub (2005a). In a departure from the previous dialogue of the debate as depicted in the Journal of Quantitative Criminology this article had a much more forthright tone. Sampson and Laub (2005a: 905) basically conclude d by suggesting that trajectory analysis is just anothe r sexy statistic al tool much like path and factor analysis long ago, followed by event history analysis in the 1980s, and most recently preceded by LISREL and HLM in the 1990s. Although, they did argue that th e methodology has some uses, they believe that the larger problem is criminologys obsessi on with statistical tools without attention to theoretical underlying processes that influence behavior. In a last attempt by Nagin and Tremblay to provide a response to Sampson and Laubs criti ques, Nagin and Tremblay (2005b) themselves were surprised by the tone of the latest rejoin der, and equally cautioned criminologists to be wary of seductive theories as well as se ductive methodologies (espec ially when the critics are promoting a life-course theory th at they themselves developed). In another work, Sampson and Laub (2005b:41) did concede that they were not arguing that grouping techniques such as Nagins did no t have any place in criminology, they were just trying to warn criminologists about the mistake of the reification of gr oups as a result of the application of the method, which Nagin (2005) hi mself also mentioned as a possible outcome. They argued that this mistake is comparable to researchers making errors in interpreting

PAGE 96

96 regression results in causal versus associational terms (for furthe r discussions also see Sampson & Laub, 2005b, 2005c; Maughan, 2005; Nagin & Tr emblay, 2005c, 2005d; Raudenbush, 2005). With these criticisms and limitations in mind, this chapter concludes with identifying the software that will be used to estimate th e group-based offending trajectories. The final discussion is on the use of multinomial regression for assessing the relative strength of macrolevel versus micro-level predictors fo r predicting latent class membership. Analytic Procedure Trajectory estimation PROC TRAJ is the name of the computer soft ware that will be relied on in order to estimate the offending trajectories for the 1942 and 1949 Racine continuo us-resident samples by cohort and data source including sex as covariate. This is a special procedure that is obtainable as a macro-level function in SAS (Jones et al., 2001) and is made available for free use by the National Consortium on Violence Research. Alth ough a relatively new procedure, PROC TRAJ has become a widely used method for investigat ing trajectories of development (Piquero, 2007) and has been applied to a variet y of different research questi ons including criminal career desistance (Bushway et al., 2003), offense trajec tories (Chung et al., 200 2; Fergusson et al., 2000; Piquero et al., 2003), sex di fferences in mental health (B arrett & White, 2002), age-crime trajectories (McDermott & Nagin, 2002), and aggre ssion toward mothers (Pagani et al., 2004). While there have been two previous studies th at have used PROC TRAJ with macro-level data, they relied solely on aggregate crime rate s in Chicago census tracts (Griffiths & Chavez, 2004) or neighborhood street segments in Seattle (Weisburd et al., 2004). Therefore, this is possibly the first attempt to date (see Land, 2000 fo r a possible exception) that has utilized this method to estimate offending traj ectories and then estimate mode ls incorporating macro-level and micro-level characteristics for predicting group membership.

PAGE 97

97 The initial trajectory models will be es timated separately for the 1942 and 1949 Racine continuous-residents including sex as a covariate in the trajecto ry estimation (Models 1 and 2).15 This will be followed by two models estimated si milarly by cohort with sex as a covariate but only using the random subsamples of intervie wed Racine continuous-residents from the 1942 and 1949 cohorts (Models 3 and 4). These models will allow a preliminar y examination of the stability of offender groups across the cohorts and by data source. In other words, they offer a type of sensitivity analysis for the stability of group identification th at Nagin and Tremblay (2005a) have argued has rarely been done in the tr ajectory literature (for exceptions see DUnger et al., 1998; Sampson et al., 2004). Finally, because the purpose of this study is to examine the importance of macro-level and micro-level factors in pr edicting longitudinal trajector ies of offending, the 1942 and 1949 subsamples will be merged respectively with the sparse 1950 Census block data and an additional model will be estimated for each cohort with sex as a covariate (Models 5 and 6). These final trajectory models with the 1942 and the 1949 continuous-resident subsamples (merged with the interview and block data) are the basis for a ssessing the possible independent effects of macro-level a nd micro-level measures for predicting latent class membership (see also Table 4-1 presented above that graphically disp lays the trajectory analyses that will be estimated). Multinomial logistic regression After identifying the latent offender classes and establishing their re spective participation and frequency rates of offending, multinomial logi stic regression models will be estimated in 15 Ideally, race-specific models would also be estimated al so; however, due to the hom ogeneity of Racine in the 1940s and the small numbers of non-White s in the continuous-resident samples, race will remain a covariate in the multinomial regression models estimated in the following chapter.

PAGE 98

98 order to evaluate th e relationships among the macro-leve l and micro-level measures for predicting offender group membership within the c ontext of a multivariate model with attention to the hypotheses presented at the end of chapte r 3. This multivariate procedure will be used because the multinomial logit model is able to deal with dependent variables in which there is no hierarchical order to the responses and can handle both continuous and discrete independent variables, similar to ordinary regression models (Agresti, 1990). While de pendent variables that are interval scales are completely appropriate for the common multiple regression technique known as ordinary least squares (O LS) and although this same technique is often used in cases where the dependent variable has ordinal outcome s that range from one to four (i.e., strongly disagree, disagree, agree, and str ongly agree), this latter applica tion is typically not correct. There are a variety of assumptions associat ed with ordinary least squares regression including the assumptions of linearity, mean independence, homoscedasticity, uncorrelated disturbance (error) terms, and the assumption of normality (see Allison, 1999); however, the most important assumption that relates to the inappropriate decision to use OLS with categorical dependent variables is the assu mption of linearity. If OLS is applied to an ordinal outcome variable, then there is the expectation that an increase or decrease of one unit on the scale is interpreted as having the same influence no ma tter where you begin on the scale. Similarly, there are cases when the dependent variable out comes have no order at all in that they are nominal scales (Allison, 1999). Although it is theoretically possible to assume some inherent degree of order to outcomes such as being a non-offender, adolescent-limited offender, and a chronic offender (i.e., 1, 2, 3) these offending groups are based on nominal typologies that are theoretically and methodologically driven. There is also the poten tial outcome of where more than these three

PAGE 99

99 groups can be discovered and there is no logica l way to order these groups, for instance earlyonset adolescent-limited offenders, low-low chro nic offenders, late-ons et offenders, etc. Therefore, in this case the most appropriate modeling technique for the form of the dependent variable (i.e., group membership) in this study is what is referred to as the multinomial logit model. Fundamentally, the multinomial logit model (MNLM) can be referred to as an extension of the typical binary logit model ( 0, 1) where binary logits are si multaneously estimated for all of the potential comparisons across the available ca tegorical outcomes (offender groups in this case) (Long, 1997:149). For example, its results ar e similar to estimating binary logit models hypothetically comparing group 1 (non-offende rs) against group 2 (adolescent-limited offenders), group 3 (chronic offenders) against group 1 (non-offenders), and then finally group 2 (adolescent-limited offenders) against group 3 (chr onic offenders). This obviously can become much more work and present difficulty for expl aining the results across these models as the number of groups increase. The multinomial lo git model can be expressed in terms of the probability of observing outcome m given x (assuming y is the dependent variable with J nominal outcomes, i.e., adolescent-limited offende r, chronic offender, etc.). The MNLM as a probability model takes the fo llowing form (Long, 1997:152): Pr ( yi = m | xi) = exp ( xim) / j J =1 exp ( xij). Similarly, the MNLM can also be presented in the form of an odds model, where the odds of outcome m versus outcome n given x is illustrated as follows (Long, 1997:154): m | n ( xi) = exp ( xi m | n), and, after several derivations, the odds of an outcome m (being an adolescent-limited offender) versus outcome n (being a non-offender) are expected to change by a factor of exp ( k m | n

PAGE 100

100 multiplied by ), controlling for the eff ects of the other variables in the model (Long, 1997:169). This odds model allows for a clean er interpretation. For example, given this formula we could now hypothetically say that the odd s of being an adolescent-limite d offender relative to being a non-offender are 2.9 times greater for males, holding all other variables constant. For all of the following multinomial analyses, la tent class membership will be treated as the outcome variable with dichotomous variab les representing each offender group and the nonoffender group (group 1) will be left out as the reference group. The results of the multinomial logit models that provide the best model fit for these data will be asse ssed using the pseudo-R2 statistic which is bounded between the values of zero and one, with va lues closer to one indicating better model fit (Long, 1997). The multi nomial analyses will be estimated separately by cohort for the 1942 and 1949 Racine continuous-re sidents (total=558 indi viduals) who had all of the available information (interview and bl ock data) necessary to assess a theoretically integrated model for explaining gr oup-based offending patterns. The progression of the models will be as fo llows: Model 1 will be a baseline macro-level model where the effects of the social disorg anization index (neighborhood SES and residential mobility) for predicting group membership will be determined. Model 2 will be estimated similarly to Model 1, although this model will on ly include the individual-level independent variables (life transitions, race, and low SES hous ehold, and sex). Finally, Model 3 will present the full model where both the macro-level a nd micro-level measures with be estimated simultaneously, which would allow for an explor ation into the possible mediation effects of individual and family processes on the relationshi p between living in a social disorganized area and group membership. The results of all of thes e analyses will be discussed and elaborated on in Chapter 5.

PAGE 101

101 Study Strengths Notwithstanding the limitations noted previous ly in this chapter, there are several important strengths of this study that are worth mentioning. First, this study answers the recent calls among criminologists regardi ng moving toward theoretical in tegration of macro-level and micro-level levels of analysis (Sampson et al., 1997; Messner et al., 1989), specifically focusing on social disorganization theory (Shaw & McKay, 1942), Sampson and Laubs age-graded theory of informal social control (1993), a nd Moffitts (1993) offender taxonomy. Furthermore, it approaches this process by employing a relati vely new methodological t ool that has gained a substantial recognition in the field as an appropr iate modeling technique for identifying distinct behavioral patterns (for re view, see Piquero, 2007). Second, as an addition to the trajectory literat ure, this study provides an exploration into the importance of sex for distingu ishing group-based trajectories. Third, prior studies that have estimated trajectories with the Shannons (1988, 1991) Racine cohorts have relied on only felonies and misdemeanors as the measure of offending (see DUnger et al., 1998); therefore, this research also includes status offenses in the hopes to better capture adolescent-limited offending patterns. This study also improves by Lands (2000) study by estimating a series of trajectory models as a sensitivity analysis to discover if similar offe nding trajectories are produced by cohort and data source. This is es pecially of concern on ce the Racine continuousresident subsamples are merged with the interv iew level data, and subsequently, once the even fewer of the individuals are merged with the limited amount of available census data characterizing the community in which they were socialized. Finally, th is study is perhaps the first one (see Land, 2000 for a possible exception) to use trajectory analysis along with multilevel regression models to predict the likelihood of exhibiting a distinct pattern of offending over the life-course.

PAGE 102

102 Table 4-1. Resident Samples by Cohort, Size, Data Sources, and Offender Groups. Continuous-Resident Samples Sample Size (n) 1942 COHORT Continuous-Residents 633 Continuous-Resident s (w/ Interview Data) 333 Continuous-Residents (w/ I ndividual & Block Data) 235 1949 COHORT Continuous-Residents 1297 Continuous-Resident s (w/ Interview Data) 556 Continuous-Residents (w/ I ndividual & Block Data) 323 Table 4-2. Observational Peri ods of Data Sources by Cohort Data Sources Years 1942 Cohort 1949 Cohort Police Contact Data 1942 Cohort 1949 Cohort 1948-1967 1955-1974 6 to 25 years of age 6 to 25 years of age 6 to 25 years of age 6 to 25 years of age Interview Data 1976 34 years old 27 years old Census Data 1950 8 years old 1 year old Table 4-3. Social Disorganization Index Factor Loadings by Cohort 1950 Census Measures 1942 1949 Owner-Property Value -.774 -.818 % Renter-Occupied Units .794 .794 % Overcrowded .612 .665 Eigenvalue % Variance Explained N 1.61 53.47 235 1.75 58.04 323

PAGE 103

103 CHAPTER 5 RESULTS Introduction This chapter begins with a description of the summary statistics and mean differences across the cohorts for all of th e independent and dependent meas ures. Following this discussion is a presentation of the groupbased trajectories for all of the samples of the 1942 and 1949 Racine continuous residents (full sample, sample with interview data, and sample with interview and block data). Included in the explanation of the group-based trajecto ry results is also a discussion of the effect of sex (being male) as a covariate for increasing the likelihood of being in an offender group as opposed to a non-offend er group. This chapter concludes with an ANOVA presenting the mean covariate levels ac ross groups for all variables used in the multivariate analysis by cohort. This is followed by an illustration of the multivariate results and a description of the impact of macro-level and micro-level f actors for distinguishing between group-based trajectories. The multivariate anal ysis concludes with a presentation of the theoretically integrated model fo r predicting group membership. Descriptive Statistics Table 5-1 below presents the descriptive statis tics that correspond to all of the macro-level and micro-level variables by cohort described previ ously that will be used in the mlogit analysis that follows and the zero-order correlations among the contin uous measures can be found in Appendix A. Independent Variables As can be seen in Table 5-1 there was a vari ety of measures that significantly differed between the 1942 and 1949 Racine continuous resi dents that precluded the possibility of combining the cohorts in further analyses. Regard ing the measures that we re combined to create

PAGE 104

104 the social disorganization index, there was evid ence that while both cohorts were characterized by comparable measures of mean owne r-property value (1942=$9,374; 1949=$9,196) and roughly equivalent percentages of rente r-occupied units (1942=34.87%; 1949=34.99%), there was an indication that the 1949 c ohort (M=1.89; SD=3.83) had a sign ificantly greater percent of overcrowded units on average when compared with the 1942 cohort (M=1.48; SD=4.40). Turning to the micro-level measures, there were fairly comparable sex percentages between the cohorts with 47% of the 1942 samp le and 52% of the 1949 sample being male. While equivalent percentages of 1942 and 1949 R acine continuous reside nts reported having been employed prior to age 25 (94%), there wa s a significant mean di fference in reporting whether or not they had also been married pr ior to age 25 with more of the 1942 residents (1942=77%; 1949=66%) reporting havi ng experienced marriage. Once these measures were combined into the additive stakes in conform ity index (range 0-2 life transitions), the 1942 cohort reported having significan tly higher stakes in confor mity (M=1.71; SD=0.45) when compared with the 1949 cohort (M=1.60; SD =0.63). Comparatively, the 1949 cohort was comprised of significantly more Bl acks (11% versus 6%) and were more likely to have reported having moved at some point in their juvenile years (61% versus 49%). In addition, the 1942 cohort reported having been socialized in significantly lower SES households (1942=3.61; 1949=3.25) and were less likely to have grown up in a natural two-parent household (84% versus 89%) when compared with the 1949 cohort. Al though, the 1949 cohort repor ted slightly greater numbers of children in their families, perceived more succeptible to differential association, and had higher perceptions of formal social contro l in their neighborhoods during their juvenile years, these differences were not significantly different.

PAGE 105

105 Dependent Variable The prevalence and incidence of police co ntacts for the 1942 and 1949 Racine continuous residents can also be found in Table 5-1. While the cohorts had relatively comparable prevalence estimates of having come into co ntact with the police with 37.9% of the 1942 residents and 41.2% of the 1949 residents having an officially recorded police contact, the 1949 cohort had significantly greater m ean numbers of police contacts than did the 1942 cohort. On average, the 1949 residents had 2.33 police contac ts (SD=6.71) whereas the 1942 residents had a 1.29 average number of police contacts (S D=2.63). Additionally, the 1949 cohort had accumulated a significantly greater total num ber of police contacts by age 25 (754 police contacts) when compared with the 302 accumulated police contacts for the 1942 cohort. Trajectory Analysis In order to aid in the substantive interpreta tion of the nature of the offending trajectories identified and estimated for the 1942 and 1949 Raci ne continuous residents, the predicted and average curves for each of the trajectories implied by the model are plotted in their corresponding figures presented below. The tr ajectory results are described in order of progression and by cohort, thus the first model is that which was produced using the full sample of 1942 continuous residents, the second model disp lays the results from a model relying on data for the 1942 continuous interviewed residents, and the third model illustrates the trajectories of the 1942 continuous interviewed residents with block data This same process is then repeated for the 1949 Racine continuous residents (total =six models; 1942=3 models; 1949=3 models). Each model is discussed in term s of the various model selecti on criteria (Nagin, 2005), or more specifically: (1) the increase in the BIC value (as previously described in Chapter 4); (2) the posterior probabilities for group assignment; a nd (3) the models ability to demonstrate a

PAGE 106

106 theoretically distinct and subs tantively meaningful number of groups. Finally, the models are described with attention to examining the e ffect of sex (being male) on group membership. The model estimations began by fitting th e 1942 Racine continuous residents police contact data to a two group model and then co mparing the BIC values from a three group model to the two group model. The three-group model pr oved a better fit and th e analysis continued by adding groups and observing the improvement in the BIC values. The trajectory procedure ultimately arrived at a four-group solution beca use despite several attempts the model was unable to converge with a five-g roup model (see Table 5-2). The average and the predicted curves are di splayed below in Figure 5-1 and these plots illustrate that the predicted (smoothed) curves closely followed the true mean offending curves. The first group (G1) represents the over whelming majority of the sample or 73.1% of all of the 1942 Racine continuous residents. Individu als in this trajectory group exhibit virtually no offending (as measured by official data) across the entire twenty years of observation (i.e., ages 6-25). The second group includes 8.8% of the sample and individua ls in this group appear to demonstrate the type of offending that woul d characterize adolescent-limited offenders as described by Moffitt (1993) with their offending beginning in adolescence, reaching a peak by age 16 before declining and desisting prior to young adulthood. No i ndividual in this group accumulated more than 10 police contacts by age 25 and the average number of police contacts for this group was 3.91 (SD=1.98). The third gr oup, representing 13% of the sample, exhibit a low-level of chronic offending beginning in early adolescence (age 12) and continue to illustrate a non-zero rate of offending into adulthood and are still offending at age 25; however, no individual in this group accumu lated more than seven police c ontacts during the twenty year period. Group 4 (5.1%) presents a sharp contrast to the other groups in that individuals in this

PAGE 107

107 group begin offending relatively earl y in life (some even as early as age 7) and they maintain a rather high amount of police c ontacts throughout the remainder of the observation period but does decline in young adulthood. Although, it does appe ar that these individuals peak in late adolescence they still exhibit high levels of offending in early a dulthood. No individual in this group accumulated less than 5 police contacts and a handful of offenders in this group had even amassed more than 20 police contacts before re aching age 26. Overall, this small group of offenders was responsible fo r 352 total police co ntacts between the ages of 6 and 25. When turning to an examination of the mean posterior probabilities for group assignment (see Table 5-3), the mean assignment probabiliti es for each group are noticeably high suggesting that there was little ambiguity in the model when determining group assignments. For example, for G1, the mean posterior probability was .91, and the next highest was G3 with a mean probability of .07 of being assigned to G1; for group 2, the mean posterior probability was .84 (median=.87); for group 3 it was .85 (median= .93); and for group 4 it was .91 (median=.99).1 The results of this trajectory analysis provided evidence that sex was a highly significant covariate for determining group a ssignment (see Table 5-4). With G1 serving as the reference group (i.e., the non-offender group), be ing male significantly increase d an individuals likelihood for being assigned to any of the three offender groups (i.e., adolescentlimited/peaked, low-level chronic, and high-le vel chronic). Table 5-5 illustrates the BIC cr iteria for assisting in model selection for the trajectory analysis using the reduced random sample of in terviewed Racine continuous residents and the average and predicted group-based offending traj ectories are displayed in Figure 5-2. These 1 All average posterior probab ilities are above the standard cut-off of .70 (Nagin, 2005).While th ere is always the possibility of mis-classification of individuals to particular trajectories none of the posterior probabilities in these data were below .80 thus indicating a high degree of discrimination between groups.

PAGE 108

108 trajectories were similar to those produced in th e previous analysis, alth ough only three distinct groups were identified in this particular analysis.2 The first group represented the overwhelming ma jority of the Racine interviewed residents with 86.2% of the sample being classified as non-offenders with virtua lly no offending during the observation period. Comparativ ely, G2 represented 7.8% of th e sample and individuals in this group appeared to display an offending patt ern consistent with Moffitts (1993) adolescentlimited offender type wherein these individuals began offending in adolescence before peaking in mid-adolescence and showed evidence of desist ance near age 19 with no individual in this group having any officially recorded police cont acts after age 20. Indivi duals in this group had an average of 4.50 (SD=2.49) police contacts and no individual in this group accumulated more than 12 police contacts prior to reaching age 26. The third group, representing 6% of the sample, is recognizable by their early onset of offending (p rior to age 10) and rela tively high levels of offending throughout the rest of the observation pe riod. Although there is evidence that this groups rate of offending peaked in late adolescen ce/early adulthood they still maintained relatively high levels of offending through age 25. On average, individuals in this group had 7.85 (SD=3.42) police contacts and a handful of these offenders accumulated more than 10 police contacts between the ages of 6 and 25. Furt hermore, this small group of chronic offenders was responsible for a tota l of 157 police contacts. Similar to the results of the previous traject ory analysis, the posterior probabilities for group assignment were remarkably high providing evidence that the individual trajectories in these data could be assigned to a group-based trajectory with a relatively high degree of 2 A four group model was not able to fit these data, thus the low-level chronic offending group was not identified in this stage of the analysis.

PAGE 109

109 probability (see Table 5-6). Only in G2 did the mean posterior probabilities fall slightly below .90 but still remained well above the .70 cut-off (Nagin, 2005). Consistent with the previous model, the results in Table 5-7 indicated that being male signifi cantly increased an individuals likelihood of being assigned to either offende r group (adolescent-limited/peaked or chronic) when compared with the non-offender group. Table 5-8 presents the BIC model selection cr iteria for the trajectory analysis using the data from the 1942 Racine continuous interviewed residents with block data that provides the trajectories and groups that will be of interest in the 1942 ANOVA and mlogit analysis that follows. A three-group model once ag ain provided the best fit for these data and the average and predicted offending curves are plotted below in Figure 5-3. G1, representing 70.2% of the sample, are virt ually a non-offending group with little to no police contacts during the ages of 6 to 25. The second group, representing 20% of the sample is similar to the adolescent-limited offender type proposed by Moffitt (1993) with the individuals in this group beginning to offend in adolescence before peaking around the age of 16 before declining and showing evidence of having desist ed by 21. No individual in this group had accumulated more than 5 total police contacts during the twen ty years of follow-up. Also consistent with the earlier analyses, the third gro up exhibited an early onset of offending (prior to age 10), relatively high levels of offending throu ghout adolescence before reaching a peak at age 16; however, these individuals continued to mainta in comparatively high rates of offending into adulthood and through the age of 25, although the trajec tory was declining. No individual in this group had accumulated less than 4 police contacts and a handful of these individuals had as many as more than ten police contacts prior to reaching age 26 (M=8.09; SD=3.30). This small number of Racine resident s amassed 186 police contacts be tween the ages of 6 and 25.

PAGE 110

110 Table 5-9 presents the posterior probabilities for group assignment and consistent with the two prior trajectory analyses us ing residents from the 1942 cohort, there was relatively little ambiguity when the model was determining group assignment with mean posterior probabilities as high as .91 (median=.99) for G3 and still co mparatively high for G1 (M=.89; median=.97) and G2 (M=.89; median=.97). Additionally, the results indicated that sex (being male) significantly increased an individuals like lihood of belonging to either an adolescent-limited/peaked or a chronic offending group when compared with membership in a non-offending group (see Table 5-10). The next set of trajectory analyses i nvolves data from the 1949 Racine continuous residents following the same progression of mode ls that were estimated for the 1942 cohort: (1) continuous resident s; (2) continuous interviewed residents; and (3) continuous interviewed residents with block data The results of the BIC model selection criteria for the 1949 continuous resident sample is pr esented below in Table 5-11. This was the only sample in which a five-group model provided the best fit for these data, but this is not su rprising given that prior researchers have shown evidence that the number of groups typically increase with sample size (DUnger et al., 1998; Nagin & Tremblay, 2001; Samp son et al., 2004) and this was in fact the largest sample in this study (N=1,297). The average and predicted curves for the 1949 c ontinuous resident trajectory analysis can be found in Figure 5-4 and consistent with th e analyses produced using data from the 1942 continuous residents, the pred icted offending curves tended to closely follow the true offending curves. The first group, representing 57.1% of the sample, is purely a non-offender group as none of these individuals had any reco rded police contacts thro ughout the entire twenty years of observation. The second group, represen ting 32.7% of the sample, exhibited very low

PAGE 111

111 rates of offending over the entire observation peri od. There was a tendency for these individuals to peak at age 16 and then decline, but they still showed non-zero rate s of offending through age 25. On average, individuals in this group ha d 1.98 (SD=1.24) police cont acts and no individual in this group had more than 7 police contacts. G3, representing 2.6% of the sample, is an atypical group that is no t identified in any of the other anal yses presented in this study although this offender group has been the point of discussi on in prior research (s ee DUnger et al., 1998; Sampson and Laub, 1993). Individuals in this gr oup are unique in that they had virtually no officially recorded police cont acts until they reached adulthood (age 18) and their rates of offending increased rather noticeably in young adu lthood. No individual in this group had less than 3 police contacts (M=6.53; SD=3.89) and a handful of these offenders had actually accumulated more than 10 police contacts despite the fact that their offending did not begin until young adulthood. The fourth group, representing 5.7% of the sample, can be best described as an adolescentpeaked group (see DUnger et al., 1998; Land, 20 00) in that these individuals do not begin offending until early adolescence but offend at relatively high rates during mid-adolescence before peaking at age 15 and then virtually have desisted by age 23. No individual in this group had less than 6 police contacts and more than half of the individuals in this group had accumulated more than 10 police contacts (the majo rity of which were dur ing mid-adolescence). G5, representing the smallest percentage of th e sample (1.9%), exhibited a high-rate chronic pattern of offending that begins early on in life (prior to age 8) and their offending drastically increases well into adolescence before reaching a p eak in late adolescence. These individuals still display high levels of offending throughout the remainder of the observation period or age 25. Individuals in this group averaged an alarming 32.48 (SD=10.30) police contacts and the

PAGE 112

112 police contacts of this group ranged from a low of 22 to a high of 63 police contacts. This small group of offenders managed to accumulate 812 poli ce contacts during the ages of 6 to 25! The mean posterior probabilities for group assi gnment indicated that like the trajectory analyses using the 1942 Racine continuous resident data, this model demons trated the ability to determine group membership with a high degree of probability (see Table 5-12). G1 had the lowest mean probability (M=.84; median=.93), yet th is was still way above the standard cutoff of .70 (Nagin, 2005). The mean posterior probabilitie s were remarkably high for the low (G2) and high-rate (G5) chronic offending groups indicating th at these individuals trajectories faced little risk for mis-classification. According to the coefficients and their corresponding p-values produced for this sex covariate, there was eviden ce that for all of the offender groups except the late-onset offenders (i.e., adol escent-limited/peaked, low and hi gh-rate chronic offenders) being male significantly increased the likelihood of group membership into an offender group compared with a non-offender group (see Table 5-13). Table 5-14 presents the BIC model se lection criteria for the 1949 continuous interviewed residents and similar to the interviewed 1942 samp le only three trajectory groups were produced. This analysis failed to detect a late-onset or a low-rate chronic group as wa s the case in the prior 1949 continuous resident analysis. The average a nd predicted curves, as plotted in Figure 5-5, indicated that the following three trajectory gr oups could be identified: (1) non-offenders; (2) adolescent-peaked offenders; and (3) high-rate ch ronic offenders. G1 (75.1%) represented the overwhelming majority of the sample and exhi biting practically no offending across the twenty year time period. Comparatively, the second group (20.5%) did not be gin offending until early adolescence before reaching a peak in mid-adol escence and then desists by their early 20s. Individuals in this group recorded an average of 4.32 (SD=2.81) pol ice contacts with a range of 2

PAGE 113

113 to 14 police contacts. G3, repres enting 4.3% of the sample, exhibi ted an early-ons et of offending (prior to age 8) and increased their offendi ng patterns throughout mi d and late-adolescence before peaking at age 18. Neve rtheless, these individuals still offended at relatively high rates by age 25, although showing evidence of desistance in young adulthood. All of the individuals in this group had accumulated at least 10 police co ntacts between the ages of 6 to 25 (M=23.21; SD=12.95) and this small number of offenders were responsible for 557 police contacts prior to reaching age 26. Table 5-15 provides the mean posterior pr obabilities for group assignment and these results show a strong ability for discriminating across trajectory groups with no mean posterior probability less than .92 (G2). In fact, the mean posterior probabilities are remarkably high for G1 (M=.97; median=.99) and G3 (M=.99; median= .99). In addition, consis tent with all of the prior trajectory analyses, being male significan tly increased the likelihood for an individual being assigned to either an a dolescent-limited/peaked or high -rate chronic offending trajectory when compared with assignment to a non-offending group (see Table 5-16). Table 5-17 presents the BIC model selection criteria for the final trajectory analysis in this study using the data from the 1949 Racine continuous interviewed residents with block data that provides the trajectories and gr oups that will be of intere st in the 1949 ANOVA and mlogit analysis that follows. Similar to the prior 1949 continuous interviewed resident analysis and consistent with the results of the comparable 1942 continuous interviewed residents with block data analysis, the procedure indicated that a thre e-group model provided the best fit for these data. The average and predicted curves are displa yed in Figure 5-6 below. The first group, representing the majority of the sample (75.2%), exhibits al most no offending throughout the

PAGE 114

114 entire twenty year observation period. In contrast, G2 (20.4%) resembles the adolescentlimited/peaked group identified in the two previo us trajectory analyses using the 1949 cohort data in that individuals in th is group did not begin offending until adolescence but did not reach a peak in offending until 17 and then declined until appearing to desist by their early 20s. Furthermore, these individuals averaged approxi mately 5 (SD=3.45) police contacts and they had accumulated at least two police c ontacts between the ages of 6 and 25 with some of these group members having amassed as high as 14 police contact s. G3 (3.7%) exhibits the high-rate chronic offending patterns discussed previously in that th ese individuals begin o ffending early (prior to age 8) and increase dramatically during adoles cence before reaching a peak in young adulthood (age 18). Yet, this group displays elevated levels of offending by age 25, but showed evidence of desistance. No individual in this group had less than 14 po lice contacts and the average number of police contacts for this group was 30.25 (SD=15.89). Overall, these handful of offenders had accumulated 363 police contac ts between the ages of 6 and 25. Table 5-18 displays the mean posterior proba bilities for group assignment and Table 5-19 illustrates the effect of sex (being male) on group membership. Like all of the previous analyses, the mean posterior probabilities were noticeably high, ranging from .89 (G2) to .98 (G1) and as high as .99 for G3. This was an indication that there was little ambiguity in assigning individual patterns of offending to a group-ba sed trajectory. Also, once again the results showed evidence that being male significantly in creased an individuals likelihood of being assigned to either an adolescent-limited/peaked or high-rate chroni c offending group compared with a non-offending group. With all of the trajectory anal yses concluded, it is important to discuss these models within the context of the rela tive stability/instability of identifying offender groups across the various

PAGE 115

115 data sources (continuous residents, continuous interviewed residents, and continuous interviewed residents with block data ). A graphical display of the offe nder groups across the data source and by cohort are presented in Table 5-20. As can be s een there appears to be more similarities than differences in the offense trajectories acro ss the data sources a nd between the 1942 and 1949 cohorts. First, in both of the original 1942 and 1949 continuous resident samples with the largest sample sizes (N=633 and N=1,297) produced the great est number of groups. The four remaining trajectory analyses with reduced samples were only able to identify three groups, which is consistent with prior research that has indica ted that the number of groups increases with the sample size (DUnger et al., 1998; Nagin & Trem blay, 2001; Sampson et al., 2004). Second, in all of the analyses the non-o ffending group (G1) was always the most prevalent group in the sample (ranging from 57.1% to 86.2%) and sim ilarly, the high-rate chronic offending group was always the smallest percentage of the offende r groups (ranging from 1.9% to 9.8%). Third, the 1942 and 1949 original continuous resident samples al so both identified a group that is atypical to what Moffitts (1993) taxonomy would expect or who have b een classified as low-level chronic offenders who offend early and maintain low non-zero rates throu ghout the twenty year period of observation; however, this type of offender group has been observed elsewhere in previous trajectory analyses (s ee DUnger et al., 1998; 2002). Regarding the differences, first, the largest sa mple in this study (1949 continuous resident sample of 1,297 individuals) actu ally identified an additional at ypical trajectory group or what others have referred to as a late-onset group (see DUnger et al., 1998; Sampson & Laub, 1993) who do not begin offending until late adolescenc e/early adulthood, but offe nd at relatively high rates in adulthood. Second, while all of the anal yses displayed an adol escent-limited/peaked group of offenders who was the larg est group of offenders in four of the six analyses, in the 1949

PAGE 116

116 trajectory analyses this group te nded to resemble an adolesce nt-peaked group (DUnger et al., 1998; Land, 2000) in that they stil l exhibited low rates of offendi ng into early adulthood before desisting in the early 20s. The discussion in th e next section now moves toward examining the results from the ANOVAs in order to further in vestigate what macro-level and micro-level factors are significant for disti nguishing between offender groups. Analysis of Variance The results from the analysis-of-variance te sts where the macro-level and micro-level risk/protective factors are compared across group s and by cohort are presented in Table 5-21. The group means of offending (number of police cont acts) are also displayed in Table 5-21. The trajectory models displayed in Figure 5-3 (1942 cohort) and Figure 5-6 (1949 cohort) are those that all of the remaining analys es are based on. The first most obvious findings is that relatively all of the risk factors are concentrated among the offender groups and the protective factors are primarily found in the non-offender group (G1). Mo re specifically, the ri sk factors tend to be heightened in the adolescent-limite d/peaked groups (G2) and then further elevated in the highrate chronic groups (G3) (G3 for both th e 1942 and 1949 cohorts will be referred to interchangeably as high-rate or chronic offende rs hereon). For the 1942 and 1949 cohorts, social disorganization, sex (being male), race (being Black), living a low SES household, living with natural parents, and perceived formal social control are distinguishing features of group membership. Additionally, whether an individua l had moved during their juvenile years was a distinguishing feature for 1942 continuous re sidents, whereas stakes in conformity3 and family 3 In order to further explore the possible influence of marriage and employment on group-based offending trajectories, the mean age at time of marriage and first full-time job were computed and compared across groups and by cohort. Regarding the 1942 cohort, 83.0% of the non-offenders (G1), 70.2% of the adolescent-limited/peaked (G2), and 86.9% of the chronic offenders (G3) reported being married by age 25 and of those that reported being married, the mean age of marriage was 20 .8 (G1), 21.6 (G2), and 20.7 (G3). In contrast, 97.6% of G1, all of G2, and 95.6% reported having their first full-time job by age 25 and the mean age of first full-time employment was 18.4 (G1), 18.4 (G2), and 18.0 (G3). When turning to the 1949 cohort, 72.7% of the non-offenders (G1), 63.6% of

PAGE 117

117 structure (number of children in family) were also distinguishing features for group membership among the 1949 continuous residents. While the preliminary results from the f-test in the ANOVAs revealed the concentration of risk factors among the offender groups, it was nece ssary to conduct Tukeys B post hoc tests in order to determine which (if any) of the gr oups significantly differed from one another. Concerning the 1942 cohort, individuals who were in either offender group lived in significantly more socially disorganized communities and both the adolescent-limited/peaked groups (G2) and the high-rate chronic groups (G 3) had significantly more males and G3 was characterized by a significantly greater number of Black s. G1 also had a significantly greater number of individuals reporting having grown up in a two natural parent household when compared with both G2 and G3. In addition, all of the groups mean number of police contacts signif icantly differed, or in other words G3 had significantly more police cont acts than G1 and G2, and G2 had significantly more police contacts than G1. When turning to the Tukeys B post hoc tests results for the 1949 cohort, while individuals in G3 and G2 lived in more socially disorgan ized areas than G1 the groups means did not significantly differ. Comparativel y, there was a host of individual level factors that significantly varied particularly when comparing G3 and G2 to G1. The offender groups had significantly more males and Blacks in addition to having a si gnificantly greater number of children in the household and had significantly higher perceptions of formal social control (police presence) when compared with G1 (although G2 and G3 did not significantly differ from each other). The the adolescent-limited/peaked (G2), an d only 33.3% of the chronic offenders (G3) experienced ma rriage by age 25 and of those that reported being married, the mean age of marriage was 21.0 (G1), 21.1 (G2), and 21.3 (G3). Comparatively, 96.7% of G1, 95.5% of G2, and 91.6% reported having their first full-time job by age 25 and the mean age of first full-time employment wa s 19.0 (G1), 18.8 (G2), 17.3 (G3). Thus it appeared that typically fewer adolescent-limited/peaked offenders and fewer chronic offenders reported having experienced marriage and/or employment by age 25, yet there was not consistent eviden ce that the mean age of having experienced these life transitions varied greatly across groups and by cohort.

PAGE 118

118 high-rate chronic offenders (G3) also had signi ficantly lower household SES and less stakes in conformity than both G1 and G2. Furthermor e, G1 was comprised of significantly more individuals that reported coming from a two natural parent hous ehold when compared with G2 and G3. Overall, similar to the 1942 ANOVA resu lts, G3 had significantly more police contacts than G1 and G2, and G2 had significantly more po lice contacts than G1. It is now important to investigate if these same risk/protective factor s that were shown to significantly distinguish group membership in these bivariate analyses are also significan t for predicting group membership when estimated simultaneously, which now becomes the focus of the next section. Multivariate Analysis The following set of multinomial regression analys es are estimated in order to compare the ability of the macro-level and micro-level fact ors to differentiate among the offender groups. Each cohort-specific analysis (beginning with th e 1942 cohort) will be divided into the following three separate stages: (1) macro-level level; (2) micro-level level; and (3) the integrated model. The first two regression analyses will assist in is olating the relative strength of prediction for the different levels of analysis (macro-level and micro-level) and the latter will allow for an assessment of the proposed integrated theoretic al model for differentiating the non-offenders from the adolescent-limited/peaked and high-rate chronic offenders.4 According to the results from the first macr o-level-only model, using data from the 1942 Racine cohort (see Table 5-22) there was support for the effect of soci al disorganization in differentiating the offender groups. Specifically, resi ding in a socially disorganized area nearly doubled the odds of an individua l being assigned to an adol escent-limited/peaked group and 4 Models were estimated using clustered robust standard errors in Stata 9.2.

PAGE 119

119 more than doubled the odds of an individual being assigned to a high -rate chronic offending group relative to a non-offender group.5 Comparatively, Table 5-23 presents the results fro m the micro-level-only model. First, the findings showed a robust effect for sex on the odds of being assigned to ei ther type of offender group with being male nearly tripling the odds for being assigned to an adolescentlimited/peaked offender group and increasing the odds by more than six times for an individual being assigned to a high-rate offender group. Ha ving been socialized in a natural two parent household yielded a significant protective effect on the odds of being assigned to adolescentlimited/peaked offender group by reducing the od ds by 38%, and it produced a similar reduction in the odds (41%) of being assigned to a high -rate chronic offender group relative to the nonoffender group. In addition, having moved during th e juvenile years more than doubled the odds of being assigned to a high-rate chronic offendi ng trajectory. Although the majority of the other individual level risk/pro tective factors were in the expected direction insofar as increasing or decreasing the odds for being in a particular group, being Black, living in a low SES household, stakes in conformity, number of children in the household, percei ved delinquent peer influence, and perceptions of formal social control did not achieve statistic al significance for distinguishing across trajectories. Table 5-24 displays the multivariate results for the integrated theoretical model for predicting offender group membership. The findi ngs suggest that once the macro-level and micro-level factors are estimated simultaneously, so cial disorganization is left only marginally 5 Percent non-White was originally included in the analyses as a control variable, but it failed to exert any significant effects for predicting offender group membership. Models were also estimated using backwards deletion and the results confirmed that percent non-White did not influence trajectory group membership or increase the model fit, so it was dropped and not included in the analyses presented here. According to Sampson (1997) there has been little support for this construct in relation to offending wh en included with neighborhood poverty and residential instability and given the relatively high degree of homogeneity in these data, it is not surprising that this variable was non-significant.

PAGE 120

120 significant (p<.10) for increasi ng an individuals odds for bei ng assigned to an adolescentlimited/peaked offender group in relation to a non-offender group. In contrast, the effect of social disorganization on th e likelihood of being assigned to a chronic offender group was completely mediated by the addition of the mi cro-level factors. Additionally, upon comparing the pseudo R-square statistics, it appeared that the addition of social disorganization did relatively little for providing a better model fit (McFaddens in creased .23 to .25; Nagelkerke increased .39 to .41). Overall, the same three micro-level variables that achieved statistical significance in the micro-level only model discussed above (s ex and natural two parent household for both offender groups and having moved for the highrate chronics only) still remained statistically significant dis tinguishing factors for group assignment. The results of the multivariate analysis for di stinguishing across trajectories based on data from the 1949 cohort are discussed below. Simila r to the prior mlogit models presented above, the first model illustrates the e ffect of social disorganization alone, the second demonstrates the effect of only the micro-level factors, and th e final model displays the findings from the theoretically integrated model for distinguishing across offender groups. Consistent with the results of the macrolevel-only model for the 1942 cohort, social disorganization statistica lly increased the odds for an indivi dual being assigne d to either an adolescent-limited/peaked offender group or a hi gh-rate chronic offender group in relation to a non-offender group (albeit only ma rginally significant for the high-rate chronic offender group, p=.07) for the 1949 cohort. More specificall y, living in a socially disorganized area approximately increased an indi viduals odds one and a half time s for being assigned to either offender group rather than to a nonoffender group (see Table 5-25).

PAGE 121

121 Table 5-26 reveals the results of the microlevel factors for predicting group membership for the 1949 cohort. There were several similari ties as well as differences for these indivdual level factors for distinguishing across trajecto ries in the 1949 cohort an alysis when compared with the 1942 mlogit analysis discussed previousl y. The key similarity is that a rather robust effect for sex was again found with being male mo re than doubling an i ndividuals odds of being assigned to an adolescent-limite d/peaked group and increasing an individuals odds nearly two and a half times for being assigned to a highrate chronic offending group in relation to a nonoffender group. Additionally, both family-level factors (family st ability and family structure) were found to be statistically significant di stinguishing features of group assignment to an adolescentlimited/peaked offender group with residing with na tural parents resulting in a 46% reduction in the odds and each additional child increasing an individuals odds 1.36 times for being assigned to an adolescent-limited/peaked offender group. Be ing Black also marginally increased (p=.06) an individuals odds of being assigned to an adolescent-limited/peaked offender group. Comparatively, although th ese two family factors nearly ach ieved statistical significance for increasing the odds of being assigned to a high -rate offender group [family stability (p=.14); family structure (p=.11)], residing in a low SES household was the more important distinguishing feature of group assignment to this particular groupbased trajectory. An individuals odds of being assi gned to a high-rate chronic o ffender group in relation to a nonoffender group increased more than two and a half times if they reported having been socialized in a low SES household. The remaining micro-le vel factors (stakes in conformity, differential association, perceptions of formal social control, a nd having moved during the juvenile years) were not statistically significant fo r distinguishing across trajectories.

PAGE 122

122 The findings from the final mlogit model in dicating the results of the theoretically integrated model for predicting group memb ership using the 1949 cohort (see Table 5-27) revealed similar results as to what was produced for the 1942 cohort. While social disorganization remained marginally signifi cant for the adolescent-limited/peaked offender group in the 1942 analysis (p=.06), once the macrolevel and micro-level factors were estimated simultaneously for the 1949 cohort, the effect of social disorganization for distinguishing across trajectories was no longer statis tically significant for either offender group (although the odds were still in the expected direction). Comp aratively, the micro-level factors that were statistically significant for pred icting group membership in the previous micro-level-only model still maintained their significance in the integrated model. More specifically, being male more than doubled an individuals odds of being assigned to either an adolesce nt-limited/peaked or a high-rate chronic offender group and family stabi lity and structure were significant determinants of adolescent-limited/peaked offender group memb ership. Additionally, having resided in a low SES household more than doubled the odds of be ing assigned to a high-rate offender group. Overall, upon inspecting the change in the pseudo-R-squares as indi cators of model fit, there was little gained from using a theoretically integrat ed model when compared with the micro-levelonly model (McFaddens increased from .19 to .20; Nagelkerkes increased from .31 to .32). Conclusions The results from the series of analytical pro cedures yielded several substantively important findings. First, the six trajectory analyses (three for each cohort) provided evidence that nonoffender, adolescent-limited/peaked offender, and high-rate chronic offender groups were recognized across all data sources (Racine continuous residents, Racine continuous interviewed residents, and Racine continuous interviewed residents with block data ) and in both the 1942 and 1949 cohorts. Although the group percentages va ried, there was support for hypothesis #1 in

PAGE 123

123 that the semiparametric Poisson-based regression analyses produced these three distinct classes of offenders and that (1) the non-offenders we re the most prevalent of the groups; (2) the adolescent-limited/peaked offenders were the most prevalent of the offending groups; and (3) the high-rate offenders were the least prevalent gr oup of offenders but we re characterized by a higher rate of mean offending across the observa tion period when compared with the adolescentlimited/peaked offender group. However, there were a few groups that were recognized in the larger sample trajectory analyses (low-level ch ronics and late-onsetters ) who are not consistent with Moffitts taxonomy but have been discusse d elsewhere (see DUnge r et al., 1998; Piquero, 2007; Sampson & Laub, 1993). Support was also shown for the second hypothesi s with sex being a significant covariate for distinguishing group-based trajectories. Specif ically, in all of the six trajectory analyses, being male significantly increased th e probability that an individual would be classified as either an adolescent-limited/peaked or chronic offender with the no n-offender group serving as the reference group. The only exception was in the 1949 continuous resident sample (N=1,297). While there were more males in the late-onset offender group for the cohort, being male did not significantly increase the probability of assignment to this particular group-based trajectory. Concerning hypothesis #3 that significant li fe transitions including marriage and employment can affect desistance from crime, there was limited support for Sampson and Laubs age-graded theory of informal social control. There was descriptive evidence at the bivariate level suggesting that individuals in either offender group (adolesc ent-limited/peaked or high-rate chronic offender group) scored lower on the stakes in conformity index when compared with those individuals assigned to the non-offende r group. The results from the ANOVA with the 1949 cohort also indicated that the stakes in conf ormity index was statistically associated with

PAGE 124

124 group membership to a high-rate chronic offendi ng trajectory with indi viduals in this group having significantly lower mean scores on the stak es in conformity index when compared with those in the adolescent-limited/peaked group and individuals in th e non-offender group. Qualified support was also shown for hypothesis #4 regarding the expectation that social disorganization would be a signif icant distinguishing feature for group assignment. Being from a socially disorganized area signi ficantly increased the odds that an individual would be assigned to either an adolescent-limited/peaked offender group or a high-rate offender group in relation to a non-offender group in the 1942 and 1949 cohort ml ogit analyses. The effect, however, was only marginally significant (p=.07) for disti nguishing high-rate chronic offenders in the 1949 analysis. The results indicated that social diso rganization did not signifi cantly distinguish the adolescent-limited/peaked offenders from the chronic offenders. Finally, there was little support provided for hypothesis #5 or the expectation that macrolevel and micro-level effects on the participa tion and frequency of offending will still be observed net of the effects of each other. A lthough the results did i ndicate that once the multilevel effects were considered simu ltaneously, the effect of social disorganization still maintained a marginally significant effect for distinguishing the adolescent -limited/peaked offenders from the non-offenders in theoretically integrated mlog it analysis of the 1942 cohort, the majority of the effect appeared to be indir ect. While the direction of the odds were in the expected direction for both the 1942 and 1949 mlogits (with being from a socially disorganized area increasing the odds of assignment to an offender group), an examination of the R-squares (McFaddens and Nagelkerkes) showed that the a ddition of the social disorganizat ion index contributed relatively little for improving model fit. Once again, social disorganization was unable to significantly differentiate the adolescent-limited/peaked group from the chronic offending group.

PAGE 125

125 Chapter 6 will provide a further discussion of the results from the series of analytical procedures and elaborate on the possible implicati ons of the research findings. The chapter will also highlight the limitations of th e current study, offer suggestions as to how future research can improve upon and extend the examination of integr ated theoretical models and close with a few final thoughts.

PAGE 126

126 Table 5-1. Descriptive Statistics for the Ma cro-level and Micro-level Measures by Cohort Variables 1942 Mean (SD) 1949 Mean (SD) Independent Variables Macro-level Social Disorganization Index 0.00 (1.00) 0.00 (1.00) Owner-Property Value 9374.30 (2895.29) 9196.40 (2951.55) % Renter-Occupied Units 34.87 (20.83) 34.99 (21.43) % Overcrowded 1.48 (3.83)* 1.89 (4.40)* Micro-level Male 0.47 (0.50) 0.52 (0.50) Life Transitions Stakes in Conformity Index 1.71 (0.45)* 1.60 (0.53)* Employment 0.94 (0.22) 0.94 (0.23) Marriage 0.77 (0.42)* 0.66 (0.47)* Black 0.06 (0.24)* 0.11 (0.31)* Low SES Household 3.61 (2.08)* 3.25 (1.91)* Family Stability (Natural Parents) 0.84 (0.37)* 0.89 (0.31)* Family Structure (No. of Children) 3.74 (2.21) 4.02 (2.17) Differential association 0.22 (0.42) 0.25 (0.43) Formal Social Control 1.24 (0.96) 1.32 (1.03) Moved 0.61 (.0.49)* 0.49 (0.51)* Police Contacts 1.29 (2.63)* 2.33 (6.71)* Prevalence 37.9% 41.2% Incidence 302* 754* Note. The asterisk signs denotes si gnificant mean differences between the 1942 and 1949 cohorts (p<.05). Listed below the social disorganization and stak es in conformity indexes are the values of the measures that were used to create each index prior to the data mani pulations (principal components analysis for the social disorganization index and the additive index for experiencing either none, 1, or 2 of the life transitions). Table 5-2. Tabulated Bayesian Information Criteria (BIC) for 1942 Racine Continuous Residents (N=633) Optimal # of Groups BIC Cubic ZIP Model 2 -2453.58 3 -2437.23 4 -2399.77 5 False Convergence

PAGE 127

127 Table 5-3. Mean (Median) Posterior Proba bilities for Group Assignments: 1942 Racine Continuous Residents (N=633) Group Prob (G1) Prob (G2) Prob (G3) Prob (G4) 25th Percentile 75th Percentile 1 (n=463) .91(.99) .02(.00) .07(.01) .00(.00) .89 .99 2 (n=56) .02(.00) .84(.87) .12(.08) .03(.01) .71 .98 3 (n=82) .11(.03) .04(.00) .85(.93) .01(.00) .75 .98 4 (n=32) .00(.00) .04(.00) .05(.00) .91(.99) .91 .99 Note: Minimum and Maximum values presented for the group/prob(group) combination. Table 5-4. Coefficient Estimates for Eff ect of Sex (being Male) on Trajectory Group Membership: 1942 Racine Continuous Residents (N=633) Groupsa coefficient SE p-value Cubic ZIP Model 2 2.861 0.566 <0.001*** 3 2.721 0.508 <0.001*** 4 2.642 0.642 <0.001*** *p<.05 **p<.01 ***p<.001 a Group 1 (the lowest rate group) is the reference category of comparison. Table 5-5. Tabulated Bayesian Information Criteria (BIC) for 1942 Racine Continuous Interviewed Residents (N=333) Optimal # of Groups BIC Cubic ZIP Model 2 -1164.02 3 -1148.61 4 False Convergence

PAGE 128

128 Table 5-6. Mean (Median) Posterior Proba bilities for Group Assignments: 1942 Racine Continuous Interviewed Residents (N=333) Group Prob (G1) Prob (G2) Prob (G3) 25th Percentile 75th Percentile 1 (n=287) .98(.99) .02(.00) .01(.06) .99 .99 2 (n=26) .06(.00) .89(.97) .06(.02) .77 .99 3 (n=20) .00(.00) .02(003) .98(.99) .99 .99 Note: Minimum and Maximum values presented for the group/prob(group) combination. Table 5-7. Coefficient Estimates for Eff ect of Sex (being Male) on Trajectory Group Membership: 1942 Racine Continuous Interviewed Residents (N=333) Groupsa coefficient SE p-value Cubic ZIP Model 2 2.873 0.711 <0.001*** 3 2.123 0.638 <0.001*** *p<.05 **p<.01 ***p<.001 a Group 1 (the lowest rate group) is the reference category of comparison. Table 5-8. Tabulated Bayesian Information Criteria (BIC) for 1942 Racine Continuous Interviewed Residents with Block Data (N=235) Optimal # of Groups BIC Cubic ZIP Model 2 -960.81 3 -940.76 4 False Convergence Table 5-9. Mean (Median) Posterior Proba bilities for Group Assignments: 1942 Racine Continuous Interviewed Residents with Block Data (N=235) Group Prob (G1) Prob (G2) Prob (G3) 25th Percentile 75th Percentile 1 (n=165) .89(.97) .11(.03) .00(.00) .78 .97 2 (n=47) .06(.00) .89(.97) .05(.01) .80 .99 3 (n=23) .00(.00) .09(.01) .91(.99) .85 .99 Note: Minimum and Maximum values presented for the group/prob(group) combination.

PAGE 129

129 Table 5-10. Coefficient Estimates for Eff ect of Sex (being Male) on Trajectory Group Membership: 1942 Racine Continuous Interviewed Residents with Block Data (N=235) Groupsa coefficient SE p-value Cubic ZIP Model 2 2.262 0.507 <0.001*** 3 3.103 0.768 <0.001*** *p<.05 **p<.01 ***p<.001 a Group 1 (the lowest rate group) is the reference category of comparison. Table 5-11. Tabulated Bayesian Information Criteria (BIC) for 1949 Racine Continuous Residents (N=1,257) Optimal # of Groups BIC Cubic ZIP Model 2 -7123.17 3 -6851.59 4 -6769.89 5 -6694.99 6 False Convergence Table 5-12. Mean (Median) Posterior Proba bilities for Group Assignments: 1949 Racine Continuous Residents (N=1,297) Group Prob (G1) Prob (G2) Prob (G3) Prob (G4) Prob (G5) 25th Percentile 75th Percentile 1 (n=740) .84(.93) .15(.07) .00(.00) .00(.00) .00(.00) .73 .93 2 (n=424) .00(.00) .97(.99) .02(.00) .02(.00) .00(.00) .98 .99 3 (n=34) .00(.00) .08(.01) .89(.97) .03(.00) .00(.00) .84 .99 4 (n=74) .00(.00) .04(.00) .01(.00) .93(.99) .02(.00) .91 .99 5 (n=25) .00(.00) .00(.00) .02(.00) .01(.00) .98(.99) .99 .99 Note: Minimum and Maximum values presented for the group/prob(group) combination.

PAGE 130

130 Table 5-13. Coefficient Estimates for Eff ect of Sex (being Male) on Trajectory Group Membership: 1949 Racine Continuous Residents (N=1,297) Groupsa coefficient SE p-value Cubic ZIP Model 2 1.272 0.132 <0.001*** 3 0.508 0.353 0.150 4 2.700 0.432 <0.001*** 5 2.714 0.741 <0.001*** *p<.05 **p<.01 ***p<.001 a Group 1 (the lowest rate group) is the reference category of comparison. Table 5-14. Tabulated Bayesian Information Criteria (BIC) for 1949 Racine Continuous Interviewed Residents (N=556) Optimal # of Groups BIC Cubic ZIP Model 2 -2979.90 3 -2848.07 4 False Convergence Table 5-15. Mean (Median) Posterior Proba bilities for Group Assignments: 1949 Racine Continuous Interviewed Residents (N=556) Group Prob (G1) Prob (G2) Prob (G3) 25th Percentile 75th Percentile 1 (n=418) .97(.99) .03(.00) .00(.00) .98 .99 2 (n=114) .06(.01) .92(.99) .02(.00) .87 .99 3 (n=24) .00(.00) .01(.00) .99(.99) .99 .99 Note: Minimum and Maximum values presented for the group/prob(group) combination. Table 5-16. Coefficient Estimates for Eff ect of Sex (being Male) on Trajectory Group Membership: 1949 Racine Continuous Interviewed Residents (N=556) Groupsa coefficient SE p-value Cubic ZIP Model 2 1.712 0.272 <0.001*** 3 2.774 0.738 <0.001*** *p<.05 **p<.01 ***p<.001 a Group 1 (the lowest rate group) is the reference category of comparison.

PAGE 131

131 Table 5-17. Tabulated Bayesian Information Criteria (BIC) for 1949 Racine Continuous Interviewed Residents with Block Data (N=323) Optimal # of Groups BIC Cubic ZIP Model 2 -1816.65 3 -1747.74 4 False Convergence Table 5-18. Mean (Median) Posterior Proba bilities for Group Assignments: 1949 Racine Continuous Interviewed Residents with Block Data (N=323) Group Prob (G1) Prob (G2) Prob (G3) 25th Percentile 75th Percentile 1 (n=245) .98(.99) .02(.00) .00(.00) .98 .99 2 (n=66) .11(.01) .89(.98) .01(.00) .82 .99 3 (n=12) .00(.00) .00(.00) .99(.99) .99 .99 Note: Minimum and Maximum values presented for the group/prob(group) combination. Table 5-19. Coefficient Estimates for Eff ect of Sex (being Male) on Trajectory Group Membership: 1949 Racine Continuous Interviewed Residents with Block Data (N=323) Groupsa coefficient SE p-value Cubic ZIP Model 2 1.493 0.364 <0.001*** 3 1.887 0.790 0.016* *p<.05 **p<.01 ***p<.001 a Group 1 (the lowest rate group) is the reference category of comparison.

PAGE 132

132Table 5-20. Trajectory Gr oups by Sample and Cohort Continuous-Resident Samples Sample Size (n) Non-offenders AdolescentLimited/Peaked Offenders Low-Level Chronic Offenders High-Rate Chronic Offenders Late-Onset Offenders 1942 COHORT Continuous-Residents 633 463 (73.1%) 56 (8.8%) 82 (13.0%) 32 (5.1%) -Continuous-Residents (w/ Interview Data) 333 287 (86.2%) 26 (7.8%) -20 (6.0%) -Continuous-Residents (w/ Individual & Block Data) 235 165 (70.2%) 47 (20.0%) -23 (9.8%) -1949 COHORT Continuous-Residents 1297 740 (57.1%) 74 (5.7%) 424 (32.7%) 25 (1.9%) 34 (2.6%) Continuous-Residents (w/ Interview Data) 556 418 (75.2%) 114 (20.5%) -24 (4.3%) -Continuous-Residents (w/ Individual & Block Data) 323 245 (75.9%) 66 (20.4%) -12 (3.7%) -

PAGE 133

133Table 5-21. Mean Covariate Levels by Trajectory Group Membership and Cohort Variables 1942 G1 G2 G3 Mean Mean Mean F-test p-value 1949 G1 G2 G3 Mean Mean Mean F-test p-value Macro-level Social Disorganization -.18 .41 .50 10.65 <.001*** -.09 .28 .29 4.38 .01* Micro-level Male .31 .79 .91 33.79 <.001*** .43 .79 .83 17 .10 <.001*** Black .04 .08 .17 3.84 .23 .02 .23 .33 10.93 <.001*** Low SES Household 3.30 4.29 4.35 5.99 <.01** 2.99 3.76 5.58 14.34 <.001*** Stakes in Conformity 1.76 1.62 1.65 2.00 .13 1.62 1.59 1.2 5 2.89 .05* Family Stability (Natural Parents) .90 .72 .70 6.49 <.01** .94 .76 .67 12.96 <.001** Family Structure (No. of Children) 3.69 3.57 4.48 1.46 .24 3.80 4.16 5.17 6.11 .01* Differential association .22 .24 .17 0.74 .84 .24 .29 .17 0.57 .57 Formal Social Control 1.14 1.43 1.61 3.56 .03* 1.21 1.62 2 .00 7.12 <.001*** Moved .59 .60 .83 2.47 .08+ .48 .53 .58 0.38 .68 Number of Police Contacts .12 2.04 8.09 492.99 <.001*** .25 4.98 30.25 481.58 <.001*** +p<.10 *p<.05 **p<.01 ***p<.001

PAGE 134

134 Table 5-22. 1942 Racine Continuous Interviewed Residents with Block Data : Odds Ratios and Probabilities for Macro-Level Predic tors of Offending Group Membership Variables Adolescent-Limited/Peaked Offenders High-Rate Chronic Offenders Odds Prob. Odds Prob. Macro-level Social Disorganization 1.97 <.001*** 2.18 <.001*** Model Diagnostics Chi-Square= 18.30*** -Log Likelihood= -176.95 McFaddens R2= .06 Nagelkerke= .11 +p<.10 *p<.05 **p<.01 ***p<.001 Note. Reference category is group 1 (the non-offenders ). Table 5-23. 1942 Racine Continuous Interviewed Residents with Block Data : Odds Ratios and Probabilities for Micro-Level Predic tors of Offending Group Membership Variables Adolescent-Limited/Peaked Offenders High-Rate Chronic Offenders Odds Prob. Odds Prob. Micro-level Male 2.99 <.001*** 6.12 <.001*** Black 1.01 .97 1.27 .30 Low SES Household 1.38 .13 1.10 .72 Stakes in Conformity 0.90 .57 0.96 .89 Family Stability (Natural Parents) 0.62 <.01** 0.59 .05* Family Structure (No. of Children) 0.96 .80 1.37 .11 Differential association 1.01 .97 0.84 .57 Formal Social Control 1.13 .55 1.12 .71 Moved 1.11 .61 2.30 <.01** Model Diagnostics Chi-Square= 53.14*** -Log Likelihood= -143.38 McFaddens R2= .23 Nagelkerke= .39 +p<.10 *p<.05 **p<.01 ***p<.001 Note. Reference category is group 1 (the non-offenders ).

PAGE 135

135 Table 5-24. 1942 Racine Continuous Interviewed Residents with Block Data : Odds Ratios and Probabilities for Macro-level and Micr o-level Predictors of Offending Group Membership Variables Adolescent-Limited/Peaked Offenders High-Rate Chronic Offenders Odds Prob. Odds Prob. Macro-level Social Disorgan ization 1.50 .06+ 1.36 .38 Micro-level Male 2.87 <.001*** 5.79 <.001*** Black 1.01 .99 1.24 .44 Low SES Household 1.30 .27 1.04 .87 Stakes in Conformity 0.92 .65 0.98 .92 Family Stability (Natural Parents) 0.64 .01* 0.61 .06+ Family Structure (No. of Children) 0.95 .76 1.33 .14 Differential association 1.01 .99 0.83 .57 Formal Social Control 1.06 .79 1.06 .85 Moved 1.03 .88 2.15 .03* Model Diagnostics Chi-Square= 55.47*** -Log Likelihood= -141.56 McFaddens R2= .25 Nagelkerke R2= .41 +p<.10 *p<.05 **p<.01 ***p<.001 Note. Reference category is group 1 (the non-offenders ).

PAGE 136

136 Table 5-25. 1949 Racine Continuous Interviewed Residents with Block Data : Odds Ratios and Probabilities for Macro-Level Predic tors of Offending Group Membership Variables Adolescent-Limited/Peaked Offenders High-Rate Chronic Offenders Odds Prob. Odds Prob. Macro-level Social Disorganization 1.50 <.01** 1.48 .07+ Model Diagnostics Chi-Square= 7.96* -Log Likelihood= -207.61 McFaddens R2= .02 Nagelkerke R2= .04 +p<.10 *p<.05 **p<.01 ***p<.001 Note. Reference category is group 1 (the non-offenders ). Table 5-26. 1949 Racine Continuous Interviewed Residents with Block Data : Odds Ratios and Probabilities for Micro-Level Predic tors of Offending Group Membership Variables Adolescent-Limited/Peaked Offenders High-Rate Chronic Offenders Odds Prob. Odds Prob. Micro-level Male 2.31 <.001*** 2.40 .01* Black 1.38 .06+ 1.20 .64 Low SES Household 1.26 .18 2.52 <.001*** Stakes in Conformity 1.09 .62 1.13 .50 Family Stability (Natural Parents) 0.64 <.001*** 0.67 .11 Family Structure (No. of Children) 1.36 .04* 1.61 .14 Differential association 1.25 .16 1.01 .99 Formal Social Control 1.14 .43 1.14 .64 Moved 0.89 .47 0.79 .51 Model Diagnostics Chi-Square= 76.67*** -Log Likelihood= -169.86 McFaddens R2= .19 Nagelkerke R2= .31 +p<.10 *p<.05 **p<.01 ***p<.001 Note. Reference category is group 1 (the non-offenders ).

PAGE 137

137 Table 5-27. 1949 Racine Continuous Interviewed Residents with Block Data : Odds Ratios and Probabilities for Macro-level and Micr o-level Predictors of Offending Group Membership Variables Adolescent-Limited/Peaked Offenders High-Rate Chronic Offenders Odds Prob. Odds Prob. Macro-level Social Disorganization 1.19 .28 1.09 .54 Micro-level Male 2.33 <.001*** 2.36 .02** Black 1.30 .13 1.20 .64 Low SES Household 1.25 .21 2.53 <.001*** Stakes in Conformity 1.10 .60 0.79 .53 Family Stability (Natural Parents) 0.63 <.001*** 0.68 .15 Family Structure (No. of Children) 1.33 .05* 1.62 .13 Differential association 1.24 .19 1.01 .99 Formal Social Control 1.10 .58 1.17 .64 Moved 0.86 .37 0.78 .53 Model Diagnostics Chi-Square= 79.26*** -Log Likelihood= -169.32 McFaddens R2= .20 Nagelkerke R2= .32 +p<.10 *p<.05 **p<.01 ***p<.001. Note. Reference category is group 1 (the non-offenders ).

PAGE 138

138 0 0.5 1 1.5 2 678910111213141516171819202122232425 AgePolice Contacts G1 pred G2 pred G3 pred G4 pred G1 avg G2 avg G3 avg G4 avg Figure 5-1. 1942 Racine Con tinuous Residents Group-Base d Trajectories (N=633)

PAGE 139

139 0 0.2 0.4 0.6 0.8 1 678910111213141516171819202122232425 AgePolice Contacts G1 pred G2 pred G3 pred G1 avg G2 avg G3 avg Figure 5-2. 1942 Racine Continuous Interviewed Residents Group-Based Tr ajectories (N=333)

PAGE 140

140 0 0.5 1 1.5 678910111213141516171819202122232425 AgePolice Contacts G1 pred G2 pred G3 pred G1 avg G2 avg G3 avg Figure 5-3. 1942 Racine Continuous Interviewed Residents with Block Data Group-Based Trajectories (N=235)

PAGE 141

141 0 2 4 6 678910111213141516171819202122232425 AgePolice Contacts G1 pred G2 pred G3 pred G4 pred G5 pred G1 avg G2 avg G3 avg G4 avg G5 avg Figure 5-4. 1949 Racine Con tinuous Residents Group-Base d Trajectories (N=1,297)

PAGE 142

142 0 0.5 1 1.5 2 2.5 3 678910111213141516171819202122232425 AgePolice Contacts G1 pred G2 pred G3 pred G1 avg G2 avg G3 avg Figure 5-5. 1949 Racine Continuous Interviewed Residents Group-Based Tr ajectories (N=556)

PAGE 143

143 0 2 4 6 678910111213141516171819202122232425 AgePolice Contacts G1 pred G2 pred G3 pred G1 avg G2 avg G3 avg Figure 5-6. 1942 Racine Continuous Interviewed Residents with Block Data Group-Based Trajectories (N=323)

PAGE 144

144 CHAPTER 6 DISCUSSION AND CONCLUSION Discussion This research began by seeking to investigate th e viability of integrated multi-level theory for explaining offense trajectories. It was expect ed that residing in a socially disorganized area early on in development would increase an indi viduals likelihood for being in an offender group as opposed to a non-offender group. Particularly, liv ing in a socially disorganized area should increase the odds for exhibiting a chronic-offending trajectory. Second, it was anticipated that certain individual and family-level factors would also elevate an individuals likelihood for being an offender. Similar to the expectations re garding the influence of living in a socially disorganized area, the belief was that the micro-level factor s would significantly distinguish adolescent-limited/peaked offenders and ch ronic offenders from the non-offenders. Furthermore, in accordance with Sampson and La ub (1993), it is was assumed that individuals who were more likely to repor t having experienced life tran sitions such as marriage and employment would be more li kely to be in the non-offe nder group and the adolescentlimited/peaked group. Third, it was hypothesize d that, once these factors were estimated simultaneously (macro-level and micro-level toge ther in the same model), they would each contribute to predicting an indivi duals odds for being in an a dolescent-limited/peaked offender and chronic offender group relative to a non-offe nder group. In addition, sex (being male) was expected to be significant for amplifying an individuals likelihood for being found in an offender group. Overall, the findings from this analysis yielded varying leve ls of support for the various theoretical frameworks used. Support was shown Moffitts (1993) taxonomy of offending in that all of the trajectory analyses identified a non-offender, an adol escent-limited/peaked offender,

PAGE 145

145 and a chronic offender group (with a declining tr ajectory in young adulthood). However, the semi-parametric Poisson based models also disc overed two groups that we re inconsistent with Moffitts prediction when relying on police cont act data from the largest samples of the 1942 (n=633) and 1949 (n=1,297) Racine continuous residents. A lo w-level chronic offending group was identified in both cohorts who were char acterized by an early onset of offending and maintained a fairly low level of offending (y et a non-zero rate) throughout the twenty year observation period (age 6-25). Comparatively, a late-onset offender group was discovered among the 1949 cohort who had virtually no invovlement in offending until late in adolescence/early adulthood but showed re latively high rates of offending throughout young adulthood. In addition, the adoles cent-limited/peaked group found in these analyses slightly differed from the group as theoretically discussed by Moffitt (1993) in that they were classified as adolescent-peaked and their offending patterns did not app ear to desist until around age 21. While these groups are atypical inso far as Moffitts expectations, they have been discovered and the source of discussion elsewhere (see D Unger et al., 1998; La nd, 2000; Piquero, 2007; Sampson & Laub, 1993). Also consistent with Moffitt and others (Rutter, 1983; Singer et al., 1968; Werry, 1986) that have argued that life-course-persistent o ffending may be observed almost exclusively among males and the prior research has shown that fema les typically show less i nvolvement in crime in either type of offender group (adolescent-limited offenders and life-course-persistent offenders) (see Piquero, 2007), all of the traj ectory results indicated that be ing male significantly increased an individuals likelihood of bei ng classified into any of the offender groups across all data sources (continuous re sidents, continuous interviewed residents, and continuous interviewed residents with block data ) and by cohort. The one exception was the late-onset offender group

PAGE 146

146 identified in the 1949 continuous resident anal ysis (N=1,297) where although there were more males in the group, sex was not found to be a si gnificant covariate for distinguishing this offender group from the non-offender group. The mlogit results also re-affirmed the importance of sex for predicting group membership through the consistent and robust effect of being male on increasing the odds by at least two times for an individual exhibiting either an adolescentlimited/peaked or a chro nic offending trajectory. The results only revealed limited support for Sampson and Laubs age-graded theory of informal social control and these findings were on ly evident at the bivariat e level. In both cohort analyses, the non-offender group ha d the highest mean score on the stakes in conformity index when compared with either offender group, but only in the 1949 cohort analys is was the stakes in conformity index significantly asso ciated with group membership. In line with the expectation of the influence of social disorganization, there was a minimal amount of evidence that living in a socially disorganized area in creased an individuals odds for being in an offender group. The bivariate findings suggested that Racine conti nuous residents, whose individual offending pattern could be clas sified into a adolesce nt-limited/peaked or chronic offending group-based trajec tory, tended to live in more soci ally disorganized areas than the non-offenders. The multivariate results c onfirmed that social disorganization was a significant predictor of group membership fo r both the 1942 and 1949 cohort. Although, living in a socially disorganized area failed to disti nguish the adolescent-limited/ peaked offenders from the chronic offenders. Despite the fact that the bivariate and mlog it (macro-level-only) results indicated that social disorganization was a di stinguishing feature of group memb ership, the findings from the mlogits, where both the macro-level and microlevel factors were estimated simultaneously,

PAGE 147

147 failed to find much support for the usefulness of an integrated theoretic al model for explaining group-based offending trajectories. Once the micr o-level factors (particu larly sex, natural two parent household, having moved, and low SES ho useholds depending on which cohort mlogit is being referred to) were added to the model, soci al disorganization failed to exert a significant independent effect on predicting group membership. Although, the various indicators of model fit (chi-square, log likelihood, and the psuedo R-square statistics) showed that th e integrated model was a better fit, it appeared that the infl uence of social disorgan ization on increasing an individuals likelihood for being in either an adolescent-limited/peaked or chronic offender group tended to be indirect in that it operated through individual and family-level factors. Overall, these findings tend to mirror those produced in prior multi-level studies that have shown that the addition of neighborhood le vel processes typically only in crease the variance explained by 1% to 4% (see Cheong & Raudenbush, 2000; Duncan & Raudenbush, 1999; Elliot et al., 1996; Simcha-Fagan & Schwartz, 1986), although ot her studies have begun to find more support for school contextual effects ranging from a 6% to 7% increase in the variance explained (Anderson, 2002; Welsh et al., 1999). Limitations As with all research, there are a series of data limitations that are worth mentioning. First, the most apparent limitation relates to the era wherein these individuals were raised. These individuals were either bor n during WWII (1942 cohort) or baby-boomers (i.e., 1949 cohort) who grew up in the immediate pos t-WWII era. They were living their early years during the 1960s, which is a time characterized by large social change includi ng the Civil Rights and Womens Movement, Vietnam, riots, etc. Thus, it is likely that some of these findings may not necessarily generalize to individu als who were socialized in diffe rent periods of time, say the 1980s or 1990s. Second, the area of Racine itself was a Midwestern industrial city with a

PAGE 148

148 relatively small homogenous population of predomin antly White residents, and it is not likely that the results from this community context a nd demographic would appl y to the larger more metropolitan areas of today or in communitie s/neighborhoods with an ethnically diverse population. Third, the substantial va riety of the census-level meas ures that are typically used currently in macro-level analyses were not av ailable at the block le vel in 1950. Among the measures that were accessible there was a rath er noticeable amount of completely missing data which precluded the ability to use mean substituti on to retain a greater number of individuals for the integrated theoretic al models. Extending from this lim itation, the generali zability of these findings may not be applicable or reproduced if different conceptualizat ions or aggregations were used to define the community/neighborhood such as cities, SMSAs, counties, police districts, etc. Fourth, the social disorganization index used in this study was limited in that it did not include a measure for ethnic heterogeneity as discussed by Shaw and McKay (1942). Although some research has shown that rates of delinquenc y are lower in ethnically homogenous areas (see Lander, 1954), there is typically less support for the measure wh en compared with poverty and residential instability (see Sampson, 1997). Additionally, Bursik (1988) and Sampson and Groves (1989) have argued that various structural dimensions such as local friendship ties, organizational density, and particip ation in local formal and informal voluntary organizations are among the most important factors that affect de linquency rates. More recently, Sampson et al. (1997) have argued that collective efficacy and the ability to form social capital matters more than social disorganization, and Sampson (1997) has also indicated th at one of the most important structural dimensions for control ling neighborhood delinquency is the communitys ability to supervise and control teenage peer groups, particularly gangs. Sampson and Groves

PAGE 149

149 (1989) also identified additional factors includi ng family disruption, urbanization, and rapid population change that can all impe de the ability of a community to create informal networks of social control for supervising teenage peer gr oups. Unfortunately, th ere were not any data available to measure these important processes, so caution should be taken when interpreting these results. Similarly, the stakes in conf ormity index used to assess Sampson and Laubs (1993) agegraded theory of informal social control wa s limited since it relied on a dichotomous index ranging from 0 to 2. Laub et al. (1 998) have argued that the quality or cohesiveness of the adult social bonds is what is more important than mere participation in these life transitions. Others, however, have argued that it is likely that being married a nd/or having a full-time job are probably indicators that some le vel of attachment has occurred in that it seems doubtful that these events would have happened at all if this was not the cas e (see Piquero et al., 2002). It is also important to note th at this research relied on offici al data of delinquency/crime as measured through police contacts. Despite the kn own strengths and weaknesses of official and self-report data when investigating criminal be havior (see Junger-Tas & Marshall, 1999), police contacts may be the better measure of tru e officially-based offending because they are typically closer in time to the actual criminal event than self-reported offending (see Hindelang et al., 1981; Thornberry & Krohn, 2003). Finally, there are some limitations of the traj ectory modeling procedur e as it specifically pertains to this study. There was evidence provided from the sensitivity analysis conducted that data loss did not equally affect all of the trajectory groups. Fo r instance, it appeared that once the multiple sources of data were merged toge ther that the model only discovered three groups and that the high-rate offending grou p (also referred to as chronics ) actually showed evidence of

PAGE 150

150 a sharp declining trajectory in early adulthood and even perhap s an indication of desistance by age 25. Unfortunately, given the data limitations and the reliance on only official measures of offending to model the trajectories this research is unable to an swer the question of whether or not this is true evidence of de sistance. Furthermore, there is also the possibility that these individuals in the high-rate offending group whic h accumulated a rather large number of police contacts over time had become incarcerated for some period of time by their young adulthood years, thus not providing them an opportunity to offend. This is also an empirical question that cannot be explored with this data. Lastly, once the individual and ne ighborhood level factors were estimated in multivariate models, the results indicated that while these risk/protective factors significantly distinguished the adoles cent-limited/peaked offenders and the chronicoffenders from the non-offenders, these same variab les (with the ex ception of sex) were not able to significantly distinguish the adolescent-limited/ peaked offenders from the chronic-offenders. Once again, this is likely the result of the data lo ss that appeared to diffe rentially affect offender groups. Directions for Future Research The findings from this research provided mixe d results for the various theories including Moffitts developmental typology and the saliency of sex when distinguishing offender groups, Sampson and Laubs age-graded theory of informal social control, and social disorganization theory. This research did generate some prelimin ary findings that can dire ct future research on integrated multi-level theoretical models. Sex was found to be rather robust for si gnificantly affecting group assignment when estimated as a covariate in the trajectory analyses. It also exerted a consis tent independent effect for significantly increasing the odds for group assignment to either an adolescent-limited/peaked offender group or a chronic offender group. Future research would benefit from estimating sex-

PAGE 151

151 disaggregated trajectories and then estimating ml ogit models to see whether or not the macrolevel and micro-level factors, such as those examined here, significantly differ across sex. Unfortunately, there were not enoug h individuals in the analyses re lied on in this study to permit the estimation of sex-specific trajectories and make use of all of the ri sk/protective factors in the mlogit models. Further studies employing considerab ly larger samples would permit this type of analysis. It is worth noting that there are other theories that may account for sex differences as well. One such example may be power-control theo ry (Hagan et al., 1985). It is possible that parents and society socializes males and females differently and this may have an impact on sexspecific pathways of offending. This po ssibility should be further explored. Larger samples with multi-level data would also allow future researchers to examine these effects on distinguishing between offender groups fo r groups that were unc overed in some of the trajectory analyses that were not consistent with Moffitts typology su ch as the low-level chronics and the late-onset offende rs. It may be the case that soci al disorganization matters more so for influencing low-level chronic offenders with low yet relatively stable offending rates over the life-course. Similarly, it seems reasonable to assume that the absence of the various adult social bonds identified by Sampson and Laub ( 1993) or a low level of attachment to the institutions of marriage and employment may serv e as explanations for a late-onset pattern of offending. Additionally, future studies investig ating the chronic patterns of offending should rely on data from longer observational periods pa rticularly through middle and late adulthood to see if these group-based offendi ng patterns identified in th is study continue beyond age 25. This research relied solely on an official m easure of crime to estimate offense trajectories and more specifically, police contacts It would be intere sting if future studies would attempt to reproduce these analyses using al ternative official measures of offending such as arrest or

PAGE 152

152 conviction records. Comparatively, it may also prove fruitful to examine the possible multi-level effects on life-course patterns of offending ba sed on self-reported criminal involvement. Arguably, the best test would be to include some type of compos ite measure that represented a combination of official and self-reported offending. As far as the relatively weak found support for the macro-level effects, it would likely be the case that future studies accounting for the processes identified by prior researchers (see Bursik, 1988; Sampson, 1997; Sampson & Groves, 1989; Sampson et al., 1997) such as family disruption (percentage of divorced or single-pa rent households), collective efficacy, social capital, supervision of teenage peer groups, and local participation in voluntary organizations may yield stronger support for macro-level influen ces on offending patterns and especially net of the effects of micro-level factors. Future mu tli-level studies should also make use of other macro-level theoretical constructs to investigate their relative impact on trajectories of criminal behavior including routine activities theo ry (see Cohen & Felson, 1979; Kennedy & Baron, 1993) and urban disadvantage perspectives (see Krivo & Peterson, 1996; Parker et al., 2005). With attention to these aforementioned suggess tions, it would seem that perhaps the ideal research design for best examining the hypothese s highlighted in this research would be a prospective longitudinal design which would identif y individuals at birth in different social contexts and follow them throughout their life-co urse and well into adulthood. Not only would official and self-report data be gathered on the individuals at various periods over their lifecourse, the optimal design would include both parent, teacher, and peer surveys throughout different developmental periods. Additionally, surveys administered to community residents over time would also allow for m easuring the informal networks of social control, collective efficacy, and social capital. Fina lly, if spousal surveys in add ition to self-reports were also

PAGE 153

153 adminstered in adulthood this would provide wort hwhile information in order to gauge the level of cohesiveness or attachment in the adult social bonds to better determine their effect in the desistence process. Nevetheless, it is obvious that collecting resear ch in this magnitude, for a lengthy period of time, involving a numerous amount of actors woul d require an enormous amount of monetary, individual, political, and community support, yet there are early indi cations that research similar to this magnitude (in a lesser degree) is possi ble and the preliminary findings leave hope for future research in integrative mutli-level mode ling (see related publications from the Project on Human Development in Chicago Neighborhoods). Implications Taken together, the findings from this research are intended to be preminary and to offer a starting point to guide and direct future research on theoretically integrated approaches. Given the minimal significance of social disorganization in the baseline models, it is important that future studies further investigate other social processes that have been recently discussed by neighborhood-level researchers as perhaps bei ng more important when estimating neighborhood effects (for review, see Sampson et al., 2002) befo re any concerted effort can be made toward imforming policy. Nevertheless, it is important here to at least highlight some of the possible implications that research in this area may provide in the future. Programs such as the Rochester Nurse-Home Visitation Program (see Olds et al., 1986) are likely to provide the most benefit in target ing the cumulative eff ects of both neighborhood disadvantage and factors particul arly associated with chronic offending (i.e. low birth weight, birth trauma). The Rochester program involved random assignement of expectant mothers in a small semi-rural community to one of three conditi ons: (1) pre and post natal visits; (2) pre-natal only visits; and (3) minimal standard services. The results indicated that the mothers who

PAGE 154

154 received the home visits faired significantly better than those that received the minimal amount of services and had higher levels of employment and fewer cases of child abuse and neglect. In an outcome evaluation of a similar program, Seitz et al. (1985) also found that pre and post-natal home visits to disadvantaged mo thers reduced the impact of adve rse risk factors and the children whose mothers participated in the program were rated as less aggressi ve and disruptive than children who were in the nonexperimental group (see Provence and Naylor, 1983 for an extended discussion of the Yale Ch ild Welfare Research Program). Programs targeting poor urban children during the preschool years ha ve shown to have a significant impact on reducing delin quent outcomes. Head Start is one of the most well-known of these type of initiatives and subsequent eval uations (see Darlington et al., 1980 ; Lee et al., 1990) have shown that the program was successful in producing short-term intellectual and social gains during the early preschool years a nd that Head Start youth were more likely to graduate high school. Other programs such as the Perry Preschool Project, which focused on early childhood education (ages 3 to 5), parent meetings, and home visits, have yielded a host of beneficial outcomes for program participants in cluding better grades, higher standardized test scores, higher graduation and employment rates, and lower delinquency and arrest rates (see Berreuta-Clement et al., 1984; Schweinhart & Weikart, 1988). In addition, programs implemented in the later elementary/early adol escent years specifically directed at improving problem-solving and social competency among poor youth have been associated with decreased levels of aggression and probl em behaviors (see Weissberg et al., 1981; Hawkins et al., 1991; Lochman et al., 1984; Sarason & Sarason, 1981; Hammond, 1991). Family intervention programs facilitating th e development of parental support networks and increasing the link between parents and comm unity services as well as improving parent-

PAGE 155

155 child interactions have provided mixed results for reducing delinquent involvement (see Dumas, 1989; Johnson & Walker, 1987; Hawkins et al., 1991) but Guerra (1997:290) has argued that these programs may not be as effective for urban minority disadvantaged youth because of cultural differences and the multiple stressors that plague the residents in these areas. Peer-based interventions like peer mentori ng (Goldstein, 1978) and redirec ting gang members toward more prosocial community activ ities (see Klein, 1971) have also yiel ded varying results in influencing an adolescents invovleme nt in delinquency. Residential relocation programs including the Moving to Opportunity for Fair Housing Demonstration (or MTO) and the Gautreaux Prog ram have been shown to be beneficial. Gautreaux youth who moved to the suburbs substa ntially increased their educational attainment and job opportunities (see Rosenbaum, 1995; Rosenbaum & Popkin, 1991) and recent MTO evaluations have indicated that residential relocation was significantly related to decreased involvement in violence, better overall health, and lower victimiz ation rates (see Katz et al., 2001; Ludwig et al., 2001). Leventhal and Br ooks-Gunn (2003) have also highlighted the significantly lower levels of depression, anxiety, and/or dependency problems among MTO youth when compared with youth who remained in the disadvantaged areas. Other initiatives that focus on increasing neighborhood supervisio n of teenage peer groups and promoting civic participation are likely to provide noticeable benefits as well (see Sa mpson, 1997). Overall, these programs reviewed above provide sugge stions for how future research targeting disadvantaged youth in disadvant aged areas can likely alleviat e the conditions that plague residents of these communities and decrease th e prevalence of life-course-persistent chronic patterns of offending.

PAGE 156

156 Final Thoughts Guerra (1997:300-301) perhaps best summarizes the direction for programmatic and policy initiatives in the following statement presented below if we as a community have any hopes for reducing the risk factors associated with both types of offending patterns identified by Moffitt, but particularly for prevention/inte rvention with chronic offenders: It is likely that the most effective interventions will address multiple causal factors, be carried out in multiple contexts, and be extende d in time over the course of development. Families, peers, schools, and communities must work in concert to provide opportunities for children to learn prosocial beha vior and to create environments where prosocial behavior is modeled and reinforced. It is important for researchers, policymakers, and even individuals to be aware that all of the abovementioned elements do not operate in isol ation from one another. There are individual, family, and community factors that all have infl uence on life-course pathways of development. Sampson et al. (2002:474) also challenges neig hborhood level researcher s to move beyond mere structural compositional measures of neighborhood effects and gi ve greater attention toward understanding "cultura l, normative, and collective action persp ectives that attach meaning to how residents frame their committment to places." Spatial dynamics are also important to investigating the influence of neighborhoods on offending patterns. It is possible that while youth may reside in a disdvantaged area, they may spend more time in an adjacent neighborhood where there are stronger informal social ties and collective supervision of teenage peer groups. Thus, while compositional measures of disadvant age may be related to crime rates, without measuring the influence of spatial proximity to more advantaged areas the effect of neighborhoods might be overestimated.

PAGE 157

157 It is clear that community and individualistic perspectives have there own shortcomings and limitations. Yet, it is also the case that perhaps by highlighting the strengths of each approach and seeking an unde rstanding of their interac tive influence on life-course developmental patterns of offending, future multilevel theoretical models can gain greater support within the discipline. Al though the results here failed to offer much evidence of the saliency of this approach, research in this ar ea is still an under-explor ed phenomenon. Future studies that are able to bette r capture the complex nature of macro-level and micro-level processes are needed before the usefulness of theoretical integration can be discounted.

PAGE 158

158 APPENDIX ZERO-ORDER CORRELATIONS BY COHORT Table A-1. 1942 Cohort 1 2 3 4 1. Social Disorganization Index 1.0 2. Stakes in Conformity Index -.07 1.0 3. Low SES Household 30 .06 1.0 4. Family Structure .12 -.01 19 1.0 Note. Significant effects (p<.05) are in italics. Table A-2. 1949 Cohort 1 2 3 4 1. Social Disorganization Index 1.0 2. Stakes in Conformity Index -.07 1.0 3. Low SES Household .29 -.22 1.0 4. Family Structure .18 .10 .16 1.0 Note. Significant effects (p<.05) are in italics.

PAGE 159

159 REFERENCES Aber, J.L., Gephart, M.A., Brooks-Gunn, J., & Conne ll, J.P. (1997). Development in context: Implications for studying neighborhood effect s. In J. Brooks-Gunn, G. Duncan, & J.L. Aber (eds.), Neighborhood poverty: Context and c onsequences for children. New York: Sage Foundation. Aguilar, B., Sroufe, L.A., Egeland, B., & Carl son, E. (2000). Distinguishing the early-onsetpersistent and adolescent-onset antisocial behavior types: From birth to 16 years. Developmental and Psychopathology, 12, 109-132. Akers, R.L. (1985). Deviant behavior: A social learning approach Belmont, CA: Wadsworth. Akers, R.L., Krohn, M.D., Lanza-Kaduce, L., & Radosevich, M. (1979). Social learning and deviant behavior: A specific test of a general theory. American Sociological Review, 44, 635-655. Akers, R.L. (1977). Deviant behavior: A social learning approach 2nd edition. Belmont, CA: Wadsworth. Allison, P.D. (1999). Multiple regression: A primer Thousand Oaks, CA: Sage Publications. Anderson, E. (1990). Streetwise: race, class, and change in an urban community Chicago: University of Chicago Press. Anderson, A.L. (2002). Individual an d contextual influences on delinquency: The role of the single-parent family. Journal of Criminal Justice 30 575. Agresti, A. (1990). Categorical data analysis New York: John Wiley and Sons. Bandura, A. (1982). The psychology of chance encounters and life paths. American Psychologists, 37, 747-755. Barrett, A.E., & White, H.R. (2002). Trajectories of gender role orientations in adolescence and early adulthood: A prospectiv e study of mental health effects of masculinity and femininity. Journal of Health and Social Behavior, 43, 451-468. Bartusch, D.R., Lynam, D.R., Moffitt, T.E., Silv a, P.A. (1997). Is age important? Testing a general versus a developmental theory of antisocial behavior. Criminology, 35, 13-48. Bellair, P.E. (1997). Social interaction and co mmunity crime: Examining the importance of neighbor networks. Criminology, 35 : 677-703. Berreuta-Clement, J.R., Schweinhart, L.J., Barne tt, W.S., Epstein, A.S., Weikart, D.P. (1984). Changed lives: The effects of the Perry Pr eschool Program on youths through age 19 Ypsilanti, MI: High/Scope Press. Berry, B.J., & Kasarda, J.D. (1977). Contemporary urban ecology. New York: Macmillan.

PAGE 160

160 Blackwell, B.S., & Piquero, A.R. (2005). On the relationships between gender, power control, self-control, and crime. Journal of Criminal Justice, 33, 1-17. Blumstein, A., & Cohen, J. (1979). Estimation of individual crime rates from arrest records. Journal of Criminal Law and Criminology, 70, 561-585. Blumstein, A., Cohen, J., & Hsieh, P. (1982). The duration of adult criminal careers. Final Report submitted to NIJ. Pittsburgh, PA: School of Urban and Public Affairs, CarnegieMellon University. Blumstein, A., Cohen, J., Roth, J. A., & Visher, C.A (eds.). (1986). Criminal careers and career criminals, Vol 1. Report of the Panel on Criminal Ca reers, National Research Council. Washington, D.C.: National Academy of Sciences. Blumstein, A. & Cohen, J. (1987). Ch aracterizing criminal careers. Science, 237 985-991. Blumstein, A., Cohen, J., & Farrington, D.P. (1988) Criminal career re search: Its value for criminology. Criminology, 26 1-35. Braithwaite, J. (1979). Inequality, crime, and public policy London: Routledge and Kegan Paul. Brame, B., Nagin, D., & Tremblay, R. (2001). Deve lopmental trajectories of physical aggression from school entry to late adolescence. Journal of Child Psychol ogy and Psychiatry, 42, 503. Broidy, L.M., Nagin, D.S., Tremblay, R.E., Bate s, J.E., Brame, R., Dodge, K.A., Fergusson, D.M., Horwood, J.L., Loeber, R., Laird, R., Lynam, D.L., Moffitt, T.E., Petit, G.S., & Vitaro, F. (2003). Developmental trajector ies of childhood disruptive behaviors and adolescent delinquency: a si x-site, cross-national study. Developmental Psychology, 39, 222-245. Bronfenbrenner, U., Moen, P., & Garbarino, J. (1984). Child, family, and community. In R. Parke (ed.), Review of child developmental research: Vol. 7: The family Chicago: University of Chicago Press. Brooks-Gunn, J., Duncan, G.J., Klebanov, P.K ., & Sealand, N. (1993). Do neighborhoods influence child and adolescent development? American Journal of Sociology, 99, 353395. Bursik, R.J. (1984). Urban dynamics a nd ecological studies of delinquency. Social Forces, 63, 393-413. Bursik, R.J. (1986). Ecological stability and dynami cs of delinquency. In A.J. Reiss & M. Tonry (eds.), Crime and Justice (vol. 66) Chicago: University of Chicago Press. Bursik, R.J., & Grasmick, H.G. (1993). Neighborhoods and crime: The di mensions of effective community control. New York: Lexington Books.

PAGE 161

161 Bushway, S., Piquero, A.R., Broidy, L.M., Cauffm an, E., & Mazerolle, P. (2001). Anempirical framework for studying desistance as a process. Criminology, 39, 492. Burgess, R., & Akers, R. (1966). A differential association-reinforcement theory of criminal behavior. Social Problems, 14, 128-147. Bushway, S.D., Thornberry, T.P., & Krohn, M.D. (2003). Desistance as a developmental process: A comparison of st atic and dynamic approaches. Journal of Quantitative Criminology, 19, 129-153. Byrk, A.S., & Raudenbush, S.W. (1987). Application of hierarchical linear models to assessing change. Psychology Bulletin, 101, 147-158 Byrk, A.S., & Raudenbush, S.W. (1992). Hierarchical Linear Models for social and behavioral research: Application a nd data analysis methods. Newbury Park, CA: Sage Publications. Caspi, A., Elder, G.H., & Bem, D.J. (1987). Mo ving against the world: Life-course patterns of explosive children. Developmental Psychology, 23, 308-313. Caspi, A, & Moffitt, T.E. (1993). The continuity of maladaptive behavior: From description to understanding in the study of antisocial beha vior. In D. Cicchetti & D. Cohen (eds.), Manual in the development of psychopathology. New York: John Wiley & Sons. Caspi, A., Moffitt, T.E., Stouthamer-Loeber, M., Krueger, R.F., & Schmutte, P.S. (1994). Are some people crime-prone? Replications of the personality-crime relationship across countries, genders, races, and methods. Criminology, 32, 163-195. Catalono, R.F., & Hawkins, J.D. (1996). The soci al development model: A theory of antisocial behavior. In D. Hawkins (ed.), Delinquency and crime: Current theories. Cambridge: Cambridge University Press. Chaiken, J., & Chaiken, M. (1982). Varieties of criminal behavior. Report R-2814-NIJ. Santa Monica, CA: Rand Corporation. Chase-Lansdale, P.L., Gordon, R.A., Brooks -Gunn, J., & Klebanov, P.K. (1997). Neighborhood and family influences on the intellectual and behavioral competence of preschool and early school-age children. In J. Brooks-G unn, G.J. Duncan, & J.L. Aber (eds.), Neighborhood poverty: Vol. 1. Context and consequences for children. New York: Sage Foundation. Cheong, Y.F., & Raudenbush, S.W. (2000). Measurem ent and structural models for childrens problem behaviors. Psychological Methods, 5 477. Chung, I., Hill, K.G., Hawkins, J.D., Gilchrist, L.D., & Nagin, D.S. (2002). Childhood predictors of offense trajectories. Journal of Research in Crime and Delinquency, 39, 60-90. Cleckley, H. (1976). The mask of sanity (5th ed.). St. Louis, MO: Mosby.

PAGE 162

162 Cohen, L.E., & Felson, M. (1979). Social changes and crime rate trends: A routine activities approach. American Sociological Review, 44 588-608. Cohen, L.E., & Land, K.C. (1987). Age structure and crime: Symmetry versus asymmetry and the projection of crime rates through the 1990s. American Sociological Review, 52 170183. Cohen, L.E., & Machalek, R. (1988). A ge neral theory of e xpropriative crime. American Journal of Sociology, 94, 465-501. Coleman, J.S. (1988). Soci al capital in the crea tion of human capital. American Journal of Sociology, 94 95-120. Coleman, J.S. (1990). Foundations of social theory Cambridge, MA: Harvard University Press. Cote, S., Zoccolillo, M., Tremblay, R.E., Nagi n, D., & Vitaro, F. (2001). Predicting girls' conduct disorder in adolescence from childhood trajectories of disruptive behaviors. Journal of the American Academy of Child and Adolescent Psychiatry, 40, 678-684. Cote, S., Tremblay, R.E., Nagin, D., Zoccolill o, M., Vitaro, F. (2002). The development of impulsivity, fearfulness, and helpfulness during childhood: Patterns of consistency and change in the trajectories of boys and girls. Journal of Child Psyc hology & Psychiatry & Allied Disciplines, 43, 609-618. Cramer, H. (1946). Mathematical models of statistics Princeton, NJ: Princeton University Press. Crane, J. (1991). The epidemic theory of ghettos and neighborhood eff ects on dropping out and teenage childbearing. American Journal of Sociology, 96, 1126-1159. Daly, K. (1998). Gender, crime, and criminology. In M. Tonry (ed.), The handbook of crime and punishment New York: Oxford University Press. Darlington, R.B., Royce, J.M., Snipper, A.S ., Murray, H.W., & Lazar, I. (1980). Preschool programs and later school competence of children from low-income families. Science, 208 202-204. De Li, S. (1999). Legal sanctions and youths status achievement: A longitudinal study. Justice Quarterly, 16 377-401. Donnellan, M.B., Ge, X., & Wenk, E. (2000). Cogn itive abilities in adolescent-limited and lifecourse-persistent criminal offenders. Journal of Abnormal Psychology, 109, 396-402. Dumas, J.E. (1989). Treating antis ocial behavior in children: Child and family approaches. Clinical Psychology Review, 9 197-222. Duncan, G.J. (1994). Families and neighbors as sources of disadvantage in the schooling decisions of White and Black adolescents. American Journal of Education, 103, 20-53.

PAGE 163

163 Duncan, G.J., Brooks-Gunn, J.P., & Klebanov, P. K. (1994). Economic deprivation and earlychildhood development. Child Development, 65, 296-318. Duncan, G.J., & Raudenbush, S.W. (1999). Assessing th e effect of context in studies of child and youth development. Educational Psychologist, 34 29. D'Unger, A.V., Land, K.C., McCall, P.L, & Na gin, D.S. (1998). How many latent classes of delinquent/criminal careers? Results from mixed Poisson regression analyses. American Journal of Sociology, 103, 1593-1630. D'Unger, A.V., Land, K.C., & McCall, P.L. ( 2002). Sex differences in age patterns of delinquent/criminal careers: Results from Poisson latent class analyses of the Philadelphia Cohort Study. Journal of Quantitative Criminology, 18, 349-375. Eggleston, E.P., Laub, J., & Sampson, R.J. (2004). Methodological sensitiviti es to latent class analysis of long-term criminal trajectories. Journal of Quantitative Criminology, 20, 126. Elder, G.H. (1975). Age differentiation and the life course. Annual Review of Sociology, 1, 165190. Elder, G.H. (1985). Perspectives on the lif e course. In Glen H. Elder, Jr. (ed.), Life course dynamics: Trajectories and transitions, 1968-1980 Ithaca: Cornell University Press. Elliott, D.S., Ageton, S.S., Huizinga, D., Knowles, B.A., & Canter, R.J. (1983). The prevalence and incidence of delinquent behavior: 1976. Boulder, CO: Behavioral Research Institute. Elliot, D.S., Huizinga, D., & Morse, B. (1986) Self-reported violent offendingA descriptive analysis of juvenile violent offenders and their offending careers. Journal of Interpersonal Violence, 1, 472-514. Elliot, D.S. (1994). Serious violent offenders: Onset, developmental course, and termination The American Society of Criminology 1993 address. Criminology, 32, 1-21. Elliott, D.S., Wilson, W.J., Huizinga, D., Sampson, R.J., Elliott, A. & Rankin, B. (1996). The effects of neighborhood disadvantag e on adolescent development. Journal of Research in Crime and Delinquency, 33 389. Ensminger, M.E., Lamkin, R.P., & Jacobson, N. (1996). School leav ing: A longitudinal perspective including neighborhood effects. Child Development, 67, 2400-2416. Entwisle, D.R., Alexander, K.L., & Olson, L.S. (1994). The gender gap in math: Its possible origins in neighborhood effects. American Sociological Review, 59, 822-838. Farrington, D.P. (1986a). Age and crime. In M. Tonry & N. Morris (eds.), Crime and Justice, Vol. 7 Chicago, IL: University of Chicago Press.

PAGE 164

164 Farrington, D.P. (1986b). Stepping stone s to adult criminal careers. In D. Olweus, J. Block, & M. Radke-Yarrow (eds.), Development of antisocial and prosocial behavior New York: Academic Press. Farrington, D.P. (1995). The development of o ffending and antisocial behaviour from childhood: Key findings from the Cambridge Study in Delinquent Development. Journal of Child Psychology and Psychiatry, 360, 929-964. Farrington, D.P. (1999). A crim inological research agenda for the next millennium. International Journal of Offender Therapy and Comparative Criminology, 43 154-167. Farrington, D.P. (2003). Developmental and lif e-course criminology: Key theoretical and empirical issues-The 2002 Sutherland Address Award. Criminology, 41 221-255. Farrington, D.P., & West, D.J. (1990). The Cambridge Study in deli nquent development: A prospective longitudinal survey of 411 male s, in H. Kerner & G. Kaiser (eds.), Criminality: Personality, behavior, and life history Berlin: Springer-Verlag. Farrington, D.P., Loeber, R., Elliott, D.S., Hawk ins, J.D., Kandel, D.B., Klein, M.W., McCord, J., Rowe, D.C., & Tremblay, R.E. (1990) Advancing knowledge about the onset of delinquency and crime. In B.B. Lahey & A.E. Kazdin (eds.), Advances in clinical child psychology New York: Plenum Press. Farrington, D.P., Sampson, R.J., & Wikstrom, P.H. (1994). Integrating individual and ecological aspects on crime Stockholm, Sweden: Nationa l Council for Crime Prevention. Felson, R.B., & Haynie, D.L. (2002). Pubertal deve lopmental, social factors, and delinquency among adolescent boys. Criminology, 40 967-988. Fergusson, D.M., Horwood, L.J., Caspi, A., Moffitt, T.E., & Silva, P.E. ( 1996). The (artefactual) transmission of reading disability: Psychom etric lessons in the study of stability and change in behavioral development. Developmental Psychology, 32, 132-140. Fergusson, D.M., Horwood, L.J., & Nagin, D.S. ( 2000). Offending trajectories in a New Zealand birth cohort. Criminology, 38, 525-551. Fergusson, D.M., & Horwood, L.J. (2002). Ma le and female offending trajectories. Development and Psychopathology, 14, 159-177. Gephart, M.A. (1997). Neighborhoods and communities as contexts for development. In J. Brooks-Gunn, G. Duncan, & J.L. Aber (eds.), Neighborhood poverty: Context and consequences for children New York: Sage Foundation. Glueck, S., & Glueck, E. (1950). Unraveling juvenile delinquency. Cambridge, MA: Harvard University Press. Glueck, S., & Glueck, E. (1968). Delinquents and nondelinquents in perspective. Cambridge, MA: Harvard University Press.

PAGE 165

165 Goetting, A. (1994). The parenting crime connection. Journal of Primary Prevention, 14, 167184. Goldstein, A.P., Sherman, M., Gershaw, N.J ., Sprafkin, R.P., & Glick, B. (1978). Training aggressive adolescents in prosocial behavior. Journal of Youth and Adolescence, 7 7392. Goldstein, H. (1995). Multilevel statistical models 2nd edition. London, Edward Arnold. Goldsmith, H.H., Bradshaw, D.L., & Reiser-Danne r, L.A. (1984). Temperament as a potential developmental influence on attachment. In J.V. Lerner & R.M. Lerner (eds.), Temperament and social interaction during infancy and childhood San Francisco: Jossey Bass. Gordon, R.A. (1968). Issues in multiple regression. American Journal of Sociology, 73, 592-616. Gottfredson, M., & Hirschi, T. (1990). A general theory of crime Palo Alto: Stanford University Press. Gottfredson, D.C, McNeil, R.J., & Gottfreds on, D.G. (1991). Social area influences on delinquency: A multi-level analysis. Journal of Research in Crime and Delinquency, 28 197-226. Granovetter, M. (1973). The strength of weak ties. American Journal of Sociology, 78, 1360-1380. Grasmick, H., Tittle, C.R., Bursik, R.J., & Arnekl ev, B. (1993). Testing the core implications of Gottfredson and Hirschis general theory of crime. Journal of Research in Crime and Delinquency, 30, 5-29. Greenberg, D.F. (1985). Age, crim e, and social explanation. American Journal of Sociology, 91 1-21. Greene, W.H. (1990). Econometric analysis. New York: Macmillan. Griffiths, E., & Chavez, J.M. (2004). Communiti es, street guns and homicide trajectories in Chicago, 1980-1995: Merging methods for exam ining homicide tren ds across space and time. Criminology, 42, 941-978. Guerra, N.G. (1997). Intervening to prevent chil dhood aggression in the inner city. In J. McCord (ed.), Violence and childhood in the inner city Cambridge, UK: University of Cambridge Press. Hagan, J. (1991). Destiny and drift: Subcultural preferences, status attainments, and the risks and rewards of youth. American Sociological Review, 56, 567-582.

PAGE 166

166 Hagan, J., Gillis, A.R., & Simpson, J. (1985). Th e class structure of ge nder and delinquency: Toward a power-control theory of common delinquent behavior. American Journal of Sociology, 90 1151-1178. Halpern-Felsher, B., Connell, J.P., Spencer, M.B., Aber, J.L., Duncan, G.J., Clifford, E., Crichlow, W., Usinger, P., & Cole, S.S. ( 1997). Neighborhood and family factors predicting educational risk and attainment in African Am erican and White children and adolescents. In J. Brooks-Gunn, G.J. Duncan, & J.L. Aber (eds.), Neighborhood poverty: Vol. 1. Context and consequences for children. New York: Sage Foundation. Hamparian, D.M., Schuster, R., Dinitz, S., & Conrad, J. (1978). The violent few: A study of dangerous juvenile offenders. Lexington, MA: Lexington Books. Hamparian, D.M., Schuster, R., Davis, J., & White, J. (1985). The young criminal careers of the violent few. Report to the U.S. Department of Jus tice, Office of Juvenile Justice and Delinquency Prevention. Harpur, J., & Hare, R.D. (1994). Assessment of psychopathology as a function of age. Journal of Abnormal Psychology, 103, 604-609. Hawkins, J.D., Von Cleve, E., & Catalano, R. F. (1991). Reducing earl y childhood aggression: Results of a primary prevention program. Journal of the American Academy of Child and Adolescent Psychiatry, 30 208-217. Hawkins, D. (ed.). (1995). Ethnicity, race, and crime: Persp ectives across time and space. Albany, NY: State University of New York Press. Heilbron, D. (1994). Zero-altered and other regression models for count data with added zeros. Biometrical Journal, 36 531-547. Heitgerd, J.L., & Bursik, R.J. (1987). Extr a-community dynamics and the ecology of delinquency. American Journal of Sociology, 92 775-787. Hindelang, M.J., Hirschi, T., & Weis, J.G. (1981). Measuring delinquency Beverly Hills, CA: Sage Publications. Hipwell, A.E., Loeber, R., Stouthamer-Loe ber, M., Keenan, K., White, H.R., & Kroneman, L. (2002). Characteristics of girls with early-onset disr uptive and antisocial behaviour. Criminal Behaviour and Mental Health, 12, 99. Hirschi, T., & Hindelang, M. J. (1977). Inte lligence and delinquency: A revisionist review. American Sociological Review, 42, 571-587. Hirschi, T., & Gottfredson, M. (1983). Age and the explanation of crime. American Journal of Sociology, 89 552-584.

PAGE 167

167 Hirschi, T., & Gottfredson, M. (1985). Age a nd crime, logic and scholarship: Comment on Greenberg. The American Journal of Sociology, 91 22-27. Horney, J., Osgood, D.W., & Marshall, I.H. (1995). Criminal careers in the short-term: Intraindividual variability in crime and it s relation to local life circumstances. American Sociological Review, 60, 655-673. Huesmann, L.R., Eron, L.D., Lefkowitz, M.M., & Wa lder, L.O. (1984). Stability of aggression over time and generations. Developmental Psychology, 20, 1120-1134. Huff-Corzine, L., Corzine, J., & Moore, D.C. (1986). Southern expos ure: Deciphering the Souths influence on homicide rates. Social Forces, 64, 906-924. Jencks, C., & Mayer, S.E. (1990). The soci al consequences of growing up in a poor neighborhood. In L. Lynn & M. McGeary (eds.), Inner-city poverty in the United States. Washington: National Academy Press. Jessor, R., Donovan, J.E., & Costa, F.M. (1991). Beyond adolescence: Problem behavior and young adult development. New York: Cambridge University Press. Johnson, D.L., & Walker, T. (1987). The primary prevention of behavior problems in MexicanAmerican children. American Journal of Community Psychology, 15 375-385. Jones, B.L., Nagin, D.S., & Roeder, K. (2001). A SAS procedure based on mixture models for estimating developmental trajectories. Sociological Methods and Research, 29, 374-393. Junger-Tas, J., & Marshall, I.H. (1999). The se lf-report methodology in crime research. In M. Tonry (ed), Crime and Justice (vol. 25) Chicago, IL: University of Chicago Press. Katz, L.F., Kling, J.R., & Liebman, J.B. (2001). M oving to opportunity in Boston: Early results of a randomized mobility experiment. Quarterly Journal of Economics, 116 607. Kazemian, L., & Farrington, D.P. (2006). Exploring residual career length and residual number of offenses for two generations of repeat offenders. Journal of Research in Crime and Delinquency, 43, 89-113. Kennedy, L.W., & Baron, S.W. (1993). Routine activ ities and a subculture of violence:A study of violence on the street. Journal of Research in Crime and Delinquency, 20 88-112. Kim, J., & Mueller, C.W. (1978). Introduction to factor analysis : What it is and how to do it. Thousand Oaks, CA: Sage Publications. King, G. (1989). Event count models for inte rnational relations: Generalizations and applications. International Studies Quarterly 33, 123-147.

PAGE 168

168 Klebanov, P.K., Brooks-Gunn, J., Chase-Lands dale, P., & Gordon, R.A. (1997). Are neighborhood effects on young children mediated by features of the home environment? In J. Brooks-Gunn, G.J. Duncan, & J. Aber (ed.), Neighborhood pove rty: Context and consequences for children. Sage: New York, NY. Klein, M.W. (1971). Street gangs and street workers Englewood Cliffs, NJ: Prentice Hall. Kling, J.R., Liebman, J.B., & Katz, L.F. (2007). E xperimental analysis of neighborhood effects. Econometrica, 75 83. Kornhauser, R.R. (1978). Social sources of delinquency Chicago: University of Chicago Press. Kratzer, L., & Hodgins, S. (1999). A typology of offenders: A test of Moffitts theory among males and females from childhood to age 30. Criminal Behavior and Mental Health, 9, 57-73. Krivo, L.J., & Peterson, R.D. (1996). Extremely disadvantaged neighborhoods and urban crime. Social Forces, 75 619-650. Kroneman, L., Loeber, R., & Hipwell, A.E. (200 4). Is neighborhood context differently related to externalizing problems and delinq uency for girls compared with boys? Clinical Child and Family Psychological Review, 7 109-122. Kupersmidt, J.B., Griesler, P.C., DeRozier, M.E., Patterson, C.J. & Davis, P.W. (1995). Childhood aggression and peer relations in the context of family and neighborhood factors. Child Development, 66, 360-375. Lambert, D. (1992). Zero-inflated Poisson regr ession, with an application to defects in manufacturing. Technometrics 34, 1-14. Land, K. (2000). Influence of neighborhood, peer, and fa mily context: Trajectories of delinquent/criminal offending across the life course, Final Report Washington, DC: National Institute of Justice. Land, K.C., McCall, P.L., & Cohen, L.E. (1990). Stru ctural covariates of homicide rates: Are there any invariances across time and social space? American Journal of Sociology, 95, 922-963. Land, K.C., McCall, P.L., & Nagin, D.S. (1996). A comparison of Poiss on, Negative Binomial, and Semiparametric Mixed Poisson Regression Models with empirica l applications to criminal careers research. Sociological Methods and Research, 24, 387-442. Land, K.C., & Nagin, D.S. (1996). Micro-level mode ls of criminal careers: A synthesis of The criminal careers and life course approach es via Semiparametric Mixed Poisson Regression Models, with empirical models. Journal of Quantitative Criminology, 12, 163-191.

PAGE 169

169 Lander, B. (1954). Toward and understanding of juvenile delinquency New York: Columbia University Press. Lash, T., Sigal, H., & Dudzinski, D. (1980). State of the child: New York City New York: Foundation for Child Development. Laub, J.H., & Sampson, R.J. (1988). Unravelling families and delinquency: A reanalysis of the Gluecks' data. Criminology, 26 355-380. Laub, J.H., & Sampson, R.J. (1993). Turning points in the life course: Why change matters to the study of crime. Criminology, 31, 301-325. Laub, J.H., Nagin, D.S., & Sampson, R.J. (1998). Tr ajectories of change in criminal offending: Good marriages and the desistance process. American Sociological Review, 63, 225-238. Laub, J.H., & Sampson, R.J. (2003). Shared beginnings, divergent li ves: Delinquent boys to age 70. Cambridge, MA: Harvard University Press. Lee, V.E., Brooks-Gunn, J., Schnur, E., & Liaw, F. (1990). Are Head Start effects sustained? A longitudinal follow-up comparison of disa dvantaged children attending Head, no preschool, and other preschool programs. Child Development, 61 495-507. Lempers, J., Clark-Lempers, D., & Simons, R. (1989). Economic hardship, parenting and distress in adolescence. Child Development, 60 25-49. Leventhal, T. & Brooks-Gunn, J. (2000). The neighborhoods they live in: Effects of neighborhood residence on child and adolescent outcomes. Psychological Bulletin, 126, 309-337. Leventhal, T., & Brooks-Gunn, J. (2003). Moving to opportunity: An experimental study of neighborhood effects on mental health. American Journal of Public Health, 93, 15761582. Lindstrom, P. (1996). Family interaction, ne ighborhood context and deviant behaviour. Studies in Crime and Crime Prevention, 5, 113-119. Lochman, J.E., Burch, P.R., Curry, J.F., & Lamp ron, L.B. (1984). Treatment and generalization effects of cognitive behavioral and goal setting interventions with aggressive boys. Journal of Consulting and Clinical Psychology, 52 915-916. Loeber, R. (1982). The stability of antisocial child behavior: A review. Child Development, 53, 1431-1446. Loeber, R., & Wikstrom, P. (1993). Individual pathways to crime in different types of neighborhoods. In Integrating individual and ecological aspects of crime edited by D.P. Farrington, R.J. Sampson, & P. Wikstrom Stockholm: Swedish National Council for Crime Prevention.

PAGE 170

170 Loeber, R., Burke, J.D., Lahey, B.B., Winters, A., & Zera, M. (2000). Oppositional defiant and conduct disorder: A review of the past ten years, part 1. Journal of the American Academy of Child and Adolescent Psychiatry, 39, 1468-1484. Loftin, C., & Hill, R.R. (1974). Regional subculture and homicide: An examination of the GastilHackney thesis. American Sociological Review, 39, 714-724. Loftin, C., & Parker, R.N. (1985). An errors-in-va riable model of the effect of poverty on urban homicide rates. Criminology, 23, 269-287. Loehlin, J.C. (1989). Partitioning environmenta l and genetic contribu tions to behavioral development. American Psychologist, 44, 1285-1292. Long, J.S. (1997). Regression models for categorical and limited dependent variables Thousand Oaks, CA: Sage. Ludwig, J., Duncan, G.J., & Hirschfield, P. (2001). Urban poverty and juvenile crime: Evidence from a randomized housing-mobility experiment. Quarterly Journal of Economics, 116, 655. Lynam, D.R., Moffitt, T.E., & Stouthamer-Loebe r, M. (1993). Explaining the relationship between IQ and delinquency: Clas s, race, test motivation, sch ool failure, or self-control. Journal of Abnormal Psychology, 102, 187-196. Lynam, D.R, Caspi A., Moffitt, T.E., Wikstr om, P.O., Loeber, R., & Novak, S. (2000). The interaction between impulsivity and neighbor hood context on offending: the effects of impulsivity are stronge r in poorer neighborhoods. Journal of Abnormal Psychology, 109, 563-574. Matsueda, R., & Anderson, K. (1998). Th e dynamics of delinquent peers and delinquentbehavior. Criminology, 36 269-308. Maughan, B. (2005). Developmental trajectory modeling: A view from developmental psychopathology. The ANNALS of the American Ac ademy of Political and Social Science, 602, 118-130. Maume, M., Ousey, G., & Beaver, K. (2005). Cut ting the grass: A reexamination of the link between marital attachment, delinquent pe ers and desistance from marijuana use. Journal of Quantitative Criminology, 21, 27-53. Mazzerole, P, Brame, R., Piquero, A.R., & D ean, C. (2001). Onset age, persistence, and offending versatility: Comparisons across gender. Criminology, 38, 1143-1172. McArdle, J.J., & Epstein, D. ( 1987). Latent growth curves within developmental structural equation models. Child Development, 58, 110-113. McCord, W., & McCord, J. (1959). Origins of crime: A new evaluation of the CambridgeSomerville study. New York, NY: Cambridge University Press.

PAGE 171

171 McDermott, S., & Nagin, D.S. (2001). Same or different? Comparing offender groups and covariates over time. Sociological Methods and Research, 29, 282-318. McDermott, S., & Nagin, D. (1998). Same or different over time? In S.O White (ed.), Handbook of Law and Social Science: Youth and Justice New York: Plenum Press. Meredith, W., & Tisak, J. (1990). Latent curve analysis. Psychometrika, 54, 557-585. Messner, S., Krohn, M., & Liska, A. (1989). Theoretical integration in the study of deviance and crime: Problems and prospects Albany, NY: State University of Albany Press. Moffitt, T.E. (1990). Juvenile delinquency a nd attention-deficit disorder: Developmental trajectories from age 3 to 15. Child Development, 61, 893-910. Moffitt, T.E. (1993). Adolescentlimited and life-course-persist ent antisocial behavior: A developmental taxonomy. Psychological Review, 100, 674-701. Moffitt, T.E. (1994). Natural histories of deli nquency. In E.G.M. Weitekamp & Hans-Jurgen Kerner (eds.), Cross-national longitudinal resear ch on human development and criminal behavior. Dordrecht, the Netherlands: Kluwer Academic. Moffitt, T.E. (1997a). Neuropsychology, antisocia l behavior, and neighborhood context. In J. McCord (ed.), Violence and childhood in the inner city Cambridge, UK: University of Cambridge Press. Moffitt, T.E. (1997b). Adolescence-limited and life-course-persistent offending: A complementary pair of developmenta l theories. In T.P Thornberry (ed.), Developmental theories of crime and delinquency: Advances in criminological theory, Volume Seven New Brunswick, NJ: Transaction Publishers. Moffitt, T.E. (2003). Life-course-persistent a nd adolescent-limited antisocial behavior: A tenyear research review and a research agenda. In B.B. Lahey, T.E. Moffitt, & A. Caspi (eds.), Causes of conduct disorder and delinquency. New York: Guilford Press. Moffitt, T.E. (2005). Life-course-persistent versus adolescence-limited antisocial behavior. In D. Cicchetti & D. Cohen (eds.), Developmental psychopathology (2nd ed.). New York: Wiley. Moffitt, T.E., & Henry, B. (1989). Neuropsychological assessment of executive functions in selfreported delinquents. Development and Psychopathology, 1, 105-118. Moffitt, T.E., & Henry, B. (1991). Neuropsychol ogical studies of juvenile delinquency and violence: A review. In J. Milner (ed.), The neuropsychology of aggression Norwell, MA: Kluwer Academic. Moffitt, T.E., Lynam, D.R., & Silva, P.A. (1994) Neuropsychological test s predicting persistent male delinquency. Criminology, 32, 277-300.

PAGE 172

172 Moffitt, T.E., & Caspi, A. (2001). Childhood predicto rs differentiate life-course persistent and adolescent-limited antisocial path ways among males and females. Developmental and Psychopathology, 13, 355-375. Moffitt, T.E., Caspi, A, Rutter, M., & Silva, P.A. (2001). Sex differences in antisocial behavior: Conduct disorder, delinquency, and violen ce in the Dunedin Longitudinal Study. Cambridge, UK: Cambridge University Press. Nagin, D.S. (1999). Analyzing developmental tr ajectories: A semiparametric, group-based approach. Psychological Methods, 4, 139. Nagin, D.S. (2004). Response to Methodological sens itivities to latent class analysis of longterm criminal career trajectories. Journal of Quantitative Criminology, 20, 27-35. Nagin, D.S. (2005). Group-based modeling of development Cambridge, MA: University of Harvard Press. Nagin, D.S., & Paternoster, R. (1991). On the re lationship of past to future delinquency. Criminology, 29, 163-189. Nagin, D.S., & Farrington, D.P. (1992). The onset and persistence of offending. Criminology, 30, 301-523. Nagin, D.S., & Land, K.C. (1993). Age, crimin al Careers, and popul ation heterogeneity: Specification and estimation of a nonpa rametric, mixed Poisson model. Criminology, 31, 327-362. Nagin, D.S., Farrington, D.P., & Moffitt, T.E. (19 95). Life-Course trajector ies of different types of offenders. Criminology, 33, 111-140. Nagin, D.S., & Waldfogel, J. (1998). The effect of conviction on income through the life cycle. International Review of Law and Economics, 18, 25-40. Nagin, D.S., & Tremblay, R.E. (2001). Analyzing developmental trajectori es of distinct but related behaviors: A group-based method. Psychological Methods, 6, 18. Nagin, D.S., & Tremblay, R.E. (2005a). Developm ental trajectory groups: Fact or a useful statistical fiction. Criminology, 43 873-904. Nagin, D.S., & Tremblay, R.E. (2005b). From seduc tion to passion: A response to Sampson and Laub. Criminology, 43, 915-918. Nagin, D.S., & Tremblay, R.E. (2005c). What ha s been learned from group-based trajectory modeling? Examples from physical aggr ession and other problem behaviors. The ANNALS of the American Academy of Political and Social Science, 602, 82-117.

PAGE 173

173 Nagin, D.S., & Tremblay, R.E. (2005d). Furthe r reflections on modeling and analyzing developmental trajectories: A re sponse to Maughan and Raudenbush. The ANNALS of the American Academy of Political and Social Science, 602, 145-154. OBrien, R.M., Stockard, J., & Isaacson, L. (1999) The enduring effects of cohort characteristics on age-specific homicide rates, 1960-1995. The American Journal of Sociology, 104 1061-1095. Olds, D.L., Henderson, C.R., Tatelbaum, R., & Ch amberlin, R. (1986). Improving the delivery of prenatal care and outcomes of pregnancy: A randomized trial of nurse-home visitation. Pediatrics, 77 16-28. Orr, L., Feins, J.D., Jacob, R., Beecroft, E., Sa nbonmatsu, L., Katz, L.F., Liebman, J.B., & Kling, J.R. (2003). Moving to Opportunity: Interim impact evaluations. Washington, DC: U.S. Department of Housing and Urban Developm ent, Office of Policy Development and Research. Pagani, L.S., Tremblay, R.E., Nagin, D.S., Zoccolil lo, M., Vitaro, F., & McDuff, P. (2004). Risk factor models for adolescent verbal a nd physical aggression toward mothers. International Journal of Behavioral Development, 28, 528-537. Park, R.E., Burgess, E.W. (1925). Introduction to the science of sociology Chicago: University of Chicago Press. Parker, R.N., & Smith, M.D. (1980). Deterr ence, poverty, and type of homicide. American Journal of Homicide, 85, 614-624. Parker, K.F., Stults, B.J., & Rice, S.K. (2005). Racial Threat, concentr ated disadvantage and social control: Considering the macro-le vel sources of variation in arrests. Criminology, 43 1111-1134. Paternoster, R., & Brame, R. (1997). Multiple ro utes to delinquency? A test of developmental and general theories of crime. Criminology, 35, 49-84. Paternoster, R., Dean, C.W., Piquero, A.R., Mazerolle, P., & Brame, R. (1997). Generality, continuity, and change in offending. Journal of Quantitative Criminology, 13, 231-266. Petersilia, J., Greenwood, P.W., & Lavin, M. (1977). Criminal careers of habitual felons. Report R-2144-DOJ. Santa Monica, CA: Rand Corporation. Peterson, M.A., & Braiker, H.B. (1980). Doing crime: A survey of California prison inmates. Report R-2200-DOJ. Santa Moni ca, CA: Rand Corporation. Peeples, F., & Loeber, R. (1994). Do individu al factors and neighborhoo d context explain ethnic differences in juvenile delinquency? Journal of Quantitative Criminology, 10, 141-157.

PAGE 174

174 Piquero, A.R. (2007). Taking stock of developmental trajectories of criminal activity over the life course. In A. Liberman (ed.), Longitudinal research on crime and delinquency New York: Springer. Piquero, A.R., & Brezina, T. (2001). Testing Moffitts account of adolescence-limited delinquency. Criminology, 39, 353-370. Piquero, A.R., Blumstein, A., Brame, R., Haapanen, R., Mulvey, E.P., & Nagin, D.S. (2001). Assessing the impact of exposure time and incap acitation on longitudinal trajectories of criminal offending. Journal of Adolescent Research, 16, 54. Piquero, A.R., & Buka, S. (2002). Linking juvenile and adult patterns of cr iminal activity in the Providence Cohort of the National Collaborative Perinatal Project. Journal of Criminal Justice, 30, 259-272. Piquero, A.R., Brame, R., Mazzerole, P., & Haapanen, R. (2002). Crime in emerging adulthood. Criminology, 40, 137-169. Piquero, A.R., Farrington, D.P., & Blumstein, A. (2003). The criminal career paradigm. Crime and Justice, 30, 359-506. Piquero, A.R., & White, N. (2003). On the rela tionship between cognitiv e abilities and lifecourse-persistent offending among a sample of African Americans: A longitudinal test of Moffitt's hypothesis. Journal of Criminal Justice, 31, 399-409. Piquero, A.R., & Moffitt, T.E. (2004). Life-cour se-persistent offending. In J.R. Adler (ed.), Forensic Psychology: Concepts, debates, and practices. Portland, OR: Willan Publishing. Piquero, A.R., Brame, R., & Lynam, D. (2004). Studying career length through early adulthood among serious offenders. Crime & Delinquency, 50, 412-435. Piquero, A.R., Brame, R., & Moffitt, T.E. (2005). Extending the study of continuity and change: Gender differences in the linkage be tween adolescent and adult offending. Journal of Quantitative Criminology, 21, 219-243. Piquero, A.R., Brezina, T., & Turner, M.G. ( 2005). Testing Moffitts ac count of delinquency abstention. Journal of Research in Crime and Delinquency, 42, 27-54. Piquero, A.R., Moffitt, T.E., & Lawton, B. (2005). Race and Crime: The contribution of individual, familial, and neighborhood level risk factors to life-course-persistent offending. In D. Hawkins & K. Kempf-Leonard (eds.), Our children, your children: Confronting race and ethnicity in th e American juvenile justice system Chicago, IL: University of Chicago Press. Plomin, R. (1990). The role of inheritance in behavior. Science, 248, 183-188. Plomin, R., Chipuer, H.M., & Loehlin, J.C. (199 0). Behavioral genetics and personality. In L.A. Pervin (ed.), Handbook of personality theory and research New York: Guilford Press.

PAGE 175

175 Pratt, T.C., Turner, M., & Piquer o, A.R. (2005). Parental sociali zation and community context: A longitudinal analysis of the struct ural sources of low self-control. Journal of Research in Crime and Delinquency, 41, 219-243. Pratt, T.C., & Cullen, F.T. (2005). Assessing macrolevel predictors of theories of crime: A meta-analysis. In M. Tonry (ed.), Crime and Justice (vol. 32) Chicago: University of Chicago Press. Provence, S., & Naylor, A. (1983). Working with disadvantaged parents and children: Scientific issues and practice New Haven, CT: Yale University Press. Raudenbush, S.W. (2005). How do we study "what happens next"? The ANNALS of the American Academy of Political and Social Science, 602, 131-144. Reiss, A.J., & Lewis, A.L. (1961). The distributi on of juvenile delinquency in the social class structure. American Sociological Review, 26, 720-732. Reiss, A.J. (1986). Why are communities important in understanding crime? In A.J. Reiss & M. Tonry (eds.), Communities and Crime Chicago: University of Chicago Press. Reiss, A.J. (1989). Ending criminal careers Prepared for the final report of the DesistancePersistance Working Group of the Progr am on Human Development and Criminal Behavior. MacArthur Foundation and National Institute of Justice. Roncek, D.W. (1981). Dangerous places: Crime and residential environment. Social Forces, 60, 74-96. Rosenbaum, J.E. (1995). Changing the geograp hy of opportunity by expanding residential choice: Lessons from the Gautreaux Program. Housing Policy Debate, 6, 231-269. Rosenbaum, J.E., & Popkin, S. (1991). Employme nt and earnings of low-income Blacks who move to middle-class suburbs. In C. Jencks & P. Peterson (eds.), The urban underclass Washington, DC: Brookings Institution. Rowe, D.C., Vazsonyi, A.T., & Flannery, D.J. ( 1995). Sex differences in crime: Do means and within-sex variation have similar causes? Journal of Research in Crime and Delinquency, 32, 84. Rutter, M. (1983). Developmental neuropsychiatry New York, NY: Guilford. Rutter, M. (1989). Pathways from childhood to adult life. Journal of Child Psychology and Psychiatry, 30, 23-51. Sampson, R.J. (1987). Urban black violence: Th e effect of male joblessness and family disruption. American Journal of Sociology, 93 348-382. Sampson, R.J. (1995). The community. In J.Q. Wilson & J. Petersilia (eds.), Crime San Francisco, CA: ICS, Inc

PAGE 176

176 Sampson, R.J. (1997). The embeddedness of child and adolescent development: A neighborhood-level perspective on urban violence. In J. McCord (ed.), Violence and childhood in the inner city Cambridge, UK: University of Cambridge Press. Sampson, R.J. (2002). Transcending tradition: New directions in community research, Chicago style. Criminology, 40 213-230. Sampson, R.J., & Groves, W.B. (1989). Community structure and crime: Testing socialdisorganization theory. American Journal of Sociology, 94, 774-802. Sampson, R.J., Raudenbush, S.W., & Earls, F. (1997). Neighborhoods and violent crime: A multilevel study of collective efficacy. Science, 277, 918-924. Sampson, R.J., Laub, J.H. (1990). Crime and devian ce over the life c ourse: The salience of adult social bonds. American Sociological Review, 55, 609-627. Sampson, R.J., & Laub, J.H. (1992). Crim e and deviance in the life course. Annual Journal of Sociology, 18, 63-84. Sampson, R.J., & Laub, J.H. (1993). Crime in the making: Pathways and turning points through life. Cambridge, MA: Harvard University Press. Sampson, R.J., & Laub, J.H. (1995). Understa nding variability in lives through time: Contributions of life-course criminology. Studies on Crime and Crime Prevention, 4, 143-158. Sampson, R.J., & Laub, J.H. (1997). A life-course theory of cumulative disadvantage and the stability of delinquency. In T.P. Thornberry (ed.), Developmental theories of crime and delinquency (pp. 131-161). New Brunswi ck, NJ: Transaction. Sampson, R.J., Morenoff, J.D., & Gannon-Rowle y, T. (2002). Assessing neighborhood effects: Social processes and new directions in research. Annual Review of Sociology, 28 443478. Sampson, R.J., & Laub, J.H. (2003). Life-course desisters? Trajectories of crime among delinquent boys followed to age 70. Criminology, 41, 555-592. Sampson, R.J., & Raudenbush, S.W. (2004). Seei ng disorder: Neighborhood stigma and the social construction of broken windows. Social Psychological Quarterly, 67, 319-342. Sampson, R.J., Laub, J.H., & Eggleston, E.P. ( 2004). On the robustness and validity of groups: Response to Daniel Nagin. Journal of Quantitative Criminology, 20, 37-42. Sampson, R.J., & Laub, J.H. (2005a). Seductions of method: Rejoi nder to Nagin and Tremblays Developmental Trajectory Groups: Fact or Fiction. Criminology, 43, 905-914.

PAGE 177

177 Sampson, R.J., & Laub, J.H. (2005b). When pred iction fails: From crime-prone boys to heterogeneity in adulthood. The ANNALS of the American Academy of Political and Social Science, 602, 73-79. Sampson, R.J., & Laub, J.H. (2005c). A life-cour se view of the development of crime. The ANNALS of the American Academy of Political and Social Science, 602, 12-45. Sarason, I.G., & Sarason, B.R. (1981). Teaching cognitive and social skills to high school students. Journal of Clinical and Consulting Psychology, 49 908-918. Schreck, C.J., Fisher, B., & Miller J.M. (2004). The social contex t of delinquent peer networks and violent victimization. Justice Quarterly, 21 23-48. Schuerman, L., & Kobrin, S. (1986). Community caree rs in crime. In A.J. Reiss & M. Tonry (eds.), Crime and justice: A review of research Chicago: University of Chicago Press. Schweinhart, L.J., & Weikart, D.P. (1988). The High/Scope Perry preschool program. In R.H. Price, E.L. Cohen, R.P. Lorion, & J. Ramos-McKay (eds.), 14 ounces of prevention: A casebook for practicioners (pp. 53-65). Washington, DC: American Psychological Association. Seitz, V., Rosenbaum, L.K., & Apfel, N.H. (1985). Effects of family support intervention: A tenyear follow up. Child Development, 56 376-391. Shannon, L.W. (1988). Criminal career continu ity: Its social context New York: Human Sciences Press. Shannon, L.W. (1991). Changing patterns of delinquency and crime : A longitudinal study in Racine Boulder: Westview Press. Shannon, L.W. (1994). Juvenile delinquency and adult crim e, 1948-1977 [Racine, Wisconsin]: City ecological data [Computer file]. 2nd ICPSR versi on. Iowa City, IA: University of Iowa, Iowa Urban Community Research Center [producer], 1980. Ann Arbor, MI: Interuniversity Consortium for Political and Social Research [distributor], 1994. Shavit, Y., & Rattner, A. (1988). Age, crime, and the early life course. The American Journal of Sociology, 93 1457-1470. Shaw, C.D., & McKay, H.D. (1942). Juvenile delinquency and urban areas Chicago: University of Chicago Press. Short, J.F. (1990). Gangs, neighborhoods, and youth crime. Criminal Justice Research Bulletin, 5, 1-11. Silverthorn, P., & Frick, P.J. (1999). Developmen tal pathways to antisocial behavior: The delayed-onset pathway in girls. Development and Psychopathology, 11, 101-126.

PAGE 178

178 Silverthorn, P., Frick, P.J., & Reynolds, R. (2001) Timing of onset and correlates of severe conduct problems in adjudicated girls and boys. Journal of Psychopathology and Behavioral Assessment, 23, 171. Simcha-Fagan, O., & Schwartz, J.E. (1986). Ne ighborhood and delinquency: An assessment of contextual effects. Criminology, 24, 667-703. Singer, J.E., Westphal, M., & Niswander, K.R. (1968). Sex differences in the incidence of neonatal abnormalities and abnormal performance in early childhood. Child Development, 39, 103-112. Skogan, W.G. (1990). Disorder and decline New York: Free Press. Snyder, H. (1988). Court careers of juvenile offenders. Washington, DC: U.S. Department of Justice, Office of Juvenile Ju stice and Delinquency Prevention. Snyder, J., & Patterson, G. (1987). Family inter action and delinquent behavi or. In H. Quay (ed.), Handbook of juvenile delinquency. New York: Wiley. Steffensmeier, D.J., Allan, E.A., Harer, M.D., & Streifel, C. (1989). Age and the distribution of crime. The American Journal of Sociology, 94 803-831. Steffensmeier, D.J., Streifel, C., & Shihadeh, E.S. (1992). Cohort size and arrest rates over the life course: The Easterlin Hypothesis revisited. American Sociological Review, 57 306314. Sutherland, E.H. (1947). Principles of criminology. 4th ed. Philadelphia: Lippincott. Sutherland, E.H. (1939). Principles of criminology 3rd edition. Philade lphia: Lippincott. Taylor, R.B. (1995). The impact of crime on communities. Annals of the American Academy of Political and Social Science, 539, 28-45. Taylor, R.B., & Covington, J. (1988). Nei ghborhood changes in ecology and violence. Criminology, 26, 553-589. Thornberry, T.P., & Krohn, M.D. (2003). The self -report method for measuring delinquency and crime. In D. Duffee (ed.), Criminal justice 2000, volume 4, measurement and analysis of crime Washington, DC: National Institute of Justice. Tibbetts, S.G., & Piquero, A.R. (1999). The in fluence of gender, low birth weight and disadvantaged environment in predicting early-onset of offending: A test of Moffitt's interactional hypothesis. Criminology, 37, 843-878. Tittle, C.R. (1988). Two empirical regularities (ma ybe) in search of explanation: Commentary on the age-crime debate. Criminology, 26, 75-86. Tonry, M., Ohlin, L.E., & Farrington, D.P. (1991). Human development and criminal behavior New York: Springer-Verlag.

PAGE 179

179 Tracy, P.E., Wolfgang, M.E., & Figlio, R.M. (1990). Delinquency in two birth cohorts. New York: Plenum. Uggen, C. (2000). Work as a turning point in the life course of criminal s: A duration model of age, employment, and recidivism. American Sociological Review, 65 529-546. Visher, C.A., & Roth, J.A. (1986). Participation in criminal careers. In A. Blumstein, J. Cohen, J.A. Roth, & C.A. Visher (eds.), Criminal careers and "career criminals, Volume l. Washington, DC: National Academy Press. Warr, M. (1998). Life-course trans itions and desistance from crime. Criminology, 36, 183-216. Warr, M. (1993). Age, peers, and delinquency. Criminology, 31 17-40. Warr, M., & Stafford, M. (1991). The influence of delinquent peers: What they think or what they do? Criminology, 29 851-866. Weissberg, R.P., Gesten, E.L., Rapkin, B.D., Co wen, E.L., Davidson, E., de Apodaca, R.F., & McKim, B.J. (1981). The evaluation of a so cial problem-solving training program for suburban and inner-city third grade children. Journal of Clinical and Consulting Psychology, 49 251-261. Weisburd, D., Bushway, S., Lum, C., &Yang, S. (2004). Trajectories of crime at places: A longitudinal study of street segmen ts in the city of Seattle. Criminology, 42, 283-322. Wellman, B., & Leighton, B. (1979). Networks neighborhoods, and communities: Approaches to the study of the community question. Urban Affairs Quarterly, 14, 363-390. Welsh, W., Greene, J. & Jenkins P. (1999). School disorder: The influence of individual, institutional, and community factors. Criminology, 37 73. Werry, J.S. (1986). Biological factors. In H.C. Quay & J.S. Werry (eds.), Psychopathological disorders of childhood New York, NY: Wiley. West, D.J., & Farrington, D.P. (1973). Who becomes delinquent? London: Heinemann. West, D.J., & Farrington, D.P. (1977). The delinquent way of life. London: Heinemann. White, N., & Piquero, A.R. (2004). A preliminary empirical test of S ilverthorn and Frick's delayed-onset pathway in girls using an urban, African-American, US-based sample. Criminal Behaviour and Mental Health, 14, 291-309. Wikstrom, P.H. (1991a). Urban crime, criminals, and victims New York: Springer-Verlag. Wikstrom. P.H. (1991b). Housing tenure, social class and offending. Criminal Behaviour, 1, 6989.

PAGE 180

180 Wikstrom, P.H., & Loeber, R. (2000). Do di sadvantaged neighborhoods cause well-adjusted children to become adolescent delinquents? A study of male juvenile serious offending, individual risk and protective fa ctors, and neighborhood context. Criminology, 38, 11091142. Wilson, J.Q., & Kelling, G.L. (March, 1982). Broken windows: The police and neighborhood safety. Atlantic Monthly, 29-38. Wilson, J.Q., & Hernstein, R. (1985). Crime and human nature New York: Simon & Schuster. Wilson, W.J. (1987). The truly disadvantaged: The inner city, the underclass, and public policy Chicago: University of Chicago Press. Wirth, L. (1938). Urbanism as a way of life. American Journal of Sociology, 44 1-24. Wolfgang, M.E., Figlio, R.M., & Sellin, T. (1972). Delinquency in a birth cohort Chicago: University of Chicago Press. Wolfgang, M., Thornberry, T.P., & Figlio, R.M. (1987). From boy to man, from delinquency to crime. Chicago: University of Chicago Press. Zorn, C.J. (1998). An analytic and empirical assessment of zero-inflated and hurdle Poisson specifications. Sociological Methods and Research 26, 368-400.

PAGE 181

181 BIOGRAPHICAL SKETCH Wesley Glenn Jennings received his bachelors degree in ps ychology from the University of South Carolina in 2002. After completing hi s undergraduate education he continued on to receive his masters degree in criminal jus tice in 2004 also from the University of South Carolina. Upon finishing his masters degree he moved down to the University of Florida to pursue his doctorate in criminology. His rese arch interests specifically include offense trajectories, sex offenders, and multi-level modeling.