<%BANNER%>

Record for a UF thesis. Title & abstract won't display until thesis is accessible after 2010-08-31.

Permanent Link: http://ufdc.ufl.edu/UFE0021931/00001

Material Information

Title: Record for a UF thesis. Title & abstract won't display until thesis is accessible after 2010-08-31.
Physical Description: Book
Language: english
Creator: Shen, Feng
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2008

Subjects

Subjects / Keywords: Journalism and Communications -- Dissertations, Academic -- UF
Genre: Mass Communication thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Statement of Responsibility: by Feng Shen.
Thesis: Thesis (Ph.D.)--University of Florida, 2008.
Local: Adviser: Morris, Jon D.
Electronic Access: INACCESSIBLE UNTIL 2010-08-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2008
System ID: UFE0021931:00001

Permanent Link: http://ufdc.ufl.edu/UFE0021931/00001

Material Information

Title: Record for a UF thesis. Title & abstract won't display until thesis is accessible after 2010-08-31.
Physical Description: Book
Language: english
Creator: Shen, Feng
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2008

Subjects

Subjects / Keywords: Journalism and Communications -- Dissertations, Academic -- UF
Genre: Mass Communication thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Statement of Responsibility: by Feng Shen.
Thesis: Thesis (Ph.D.)--University of Florida, 2008.
Local: Adviser: Morris, Jon D.
Electronic Access: INACCESSIBLE UNTIL 2010-08-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2008
System ID: UFE0021931:00001


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 E20101221_AAAABY INGEST_TIME 2010-12-21T22:10:32Z PACKAGE UFE0021931_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES
FILE SIZE 3612 DFID F20101221_AABLSB ORIGIN DEPOSITOR PATH shen_f_Page_002.jpg GLOBAL false PRESERVATION BIT MESSAGE_DIGEST ALGORITHM MD5
e259609dc51783e001e6f494c697eb74
SHA-1
334e4cb89a72d48ee12e292a6c71127c07c66b30
25271604 F20101221_AABLRM shen_f_Page_051.tif
35cb8f6766a120c84a257268a0c373d1
3306df42712ef47939a23408e114b328b57d008f
111792 F20101221_AABLQY shen_f_Page_037.jpg
942cb2f04d949fdd4fd28e52888f7ac1
79d153cd8802ef092d22b57c6e977b7f1ac86902
124188 F20101221_AABLSC shen_f_Page_125.jpg
8724d2526e358ec89ced01ed0c80f3fc
285da8172324c42e366abc368858f822fbc2715d
35258 F20101221_AABLRN shen_f_Page_080.QC.jpg
51eb77727a6ae3b54a0c5d7a962eb427
243184cdce944a88a3b55759a7c0a87da5cc4dfd
F20101221_AABLQZ shen_f_Page_105.tif
6cd3682eb0d83553bc74cc42d59d8295
9cd9b30d66240b75505c2fde26b05f3c4fb650ec
F20101221_AABLSD shen_f_Page_114.tif
7b74bf5eaec18e115d4281e98413acf7
51443df0260dbacb8a2d01da7121f0528a2bcfea
7683 F20101221_AABLRO shen_f_Page_065thm.jpg
d3144e027d34fc98c9485da61f7487ef
207b91adf3e249b1c26b6184314dedc0041793a7
7044 F20101221_AABLSE shen_f_Page_094.pro
d2593eabcd255086377f22ff31cb06b2
17b94d5e7b31be865405eed30ba3a7635d67ef4f
1097 F20101221_AABLRP shen_f_Page_098.txt
6a03fc0d4b504518315a01210d51fc36
bf69b0442a478b235ebf0be1c6b7756a0cb03127
117915 F20101221_AABLSF shen_f_Page_026.jpg
bb3ea989222c62c74ba704f3c48c7cf7
2f9a6e1fba5c4845a0c89d110e1161897c03f2c6
F20101221_AABLRQ shen_f_Page_035.tif
a581f238f8568baa118f54fa5a33283a
775107ec4941f1c69148fa133312ba8cda09cf0a
98357 F20101221_AABLSG shen_f_Page_067.jpg
3364358c6192dd54710f118797fd5a8a
faf2f17dda883724b5ffd1eef510f4586d97b0c6
22914 F20101221_AABLRR shen_f_Page_111.QC.jpg
ff896dd8dace3bd19e5d094cfb4132f5
535d7bf3a78d1c6cac7e8a679901b362647c8149
104766 F20101221_AABLSH shen_f_Page_034.jpg
ea93297a5f7b551ac8691d325cddba9f
ff1881e78bfe9521c6ec42c1da778549cdcd53cc
1007 F20101221_AABLRS shen_f_Page_117.txt
f43bbc7b4b2b0700dafd99265918a31b
7c7ae83bb98c9f72f69ff9259d6712a784bf9953
8424 F20101221_AABLSI shen_f_Page_073thm.jpg
ce071726b7740e20dfde80cd68b2a48f
cd92d12e033f55f5aac1bc4748f341b87c3efc4a
F20101221_AABLRT shen_f_Page_014.tif
1141a420ac18997700032c1109782107
2c04588c94856b589d5c7271b0af4dd5d7173b08
F20101221_AABLSJ shen_f_Page_113.tif
0eca8b02f552ac3b4572b5ff996c03c3
e42c7fa576daa1b4f6d5fe0258d847346a6789eb
53447 F20101221_AABLRU shen_f_Page_084.pro
20b2cddec84a7d3046c1e733f0a2a653
9d24035da696fa0fc96b1ca359d15d93b3d34d4b
35992 F20101221_AABLSK shen_f_Page_106.QC.jpg
ca2f3e5925b6e36176311c4fd000fc29
22780b1081a129daf6768a5dcb8225331ce47114
2190 F20101221_AABLRV shen_f_Page_064.txt
12a8bfec69b45ac73d06ef18850ae382
c6e0690c0bb6ac3b404a26d043b0243121edf091
7605 F20101221_AABLSL shen_f_Page_068thm.jpg
6a60ff7bee392afff7b9274628efe2ac
78ad29c2044c8047ce8b74acee1e61556b8734f0
108924 F20101221_AABLRW shen_f_Page_064.jpg
d2c2403f027409d08af4616405120ab1
506ed3601d2fda5fdb413426a8161dc4cb314981
37366 F20101221_AABLRX shen_f_Page_125.QC.jpg
f795db648353ac9fd7305973dce399ab
400fface6c988e9abc126c4592bd40acff8e904c
1051980 F20101221_AABLTA shen_f_Page_008.jp2
5b325c86da2f5394e7b3bafea48f179c
c895e5488ef934f4650978b2fbc9d954404ef3f3
8672 F20101221_AABLSM shen_f_Page_042thm.jpg
289a0aa34d7f384f7b0d3a756b823baa
e7564c60053fe841617795806460ea96d5175dfc
1051966 F20101221_AABLRY shen_f_Page_126.jp2
e63bf62a0001cf61215e3923bd61afea
c46a80acd00172f2563701ca0509b4c3b8e3b0ef
212209 F20101221_AABLTB shen_f_Page_095.jp2
d809e7fe7d2b74542560512d155595f1
b076ec1328033c72c63af88bf1215e122b4d57f5
F20101221_AABLSN shen_f_Page_126.tif
2adf25451132092df4c6592417799976
11f1a5f68a1ee44a035fb824c29f850a0d07e33a
36145 F20101221_AABLRZ shen_f_Page_126.QC.jpg
9d53b4eb515b3c296c0ff97f685c587a
1196c6be24f4d1c78e60b23842302c73359764b6
F20101221_AABLTC shen_f_Page_083.jp2
87060b217e085187b70498081770d644
21e3314dc871b09ee3f73c3f4c1d86801a2a2793
30963 F20101221_AABLSO shen_f_Page_068.QC.jpg
dd8a3e61cbb668f9060a44bb402f1025
26873582681b28da76fc767d7fe2fb361d2c4a19
6828 F20101221_AABLTD shen_f_Page_072thm.jpg
cebfd1ee5306401fbf2bb98db4ab8071
d1e52f6af55006a36dfbdf4c4ddb7f3d074ffe7e
2163 F20101221_AABLSP shen_f_Page_012.txt
76f4ade208c538711d44fa8ac9394e32
ee8e63ef9ba683adb029ee706fde1700a1806531
F20101221_AABLTE shen_f_Page_099.tif
01884ec2158d5e053707eadecd0603a3
b4edc851aea3b73bfe3bafe5e41dcb8553607c7a
F20101221_AABLSQ shen_f_Page_019.tif
82031084db08a135e88966f6252174ce
24c0da84d8a94d48795235fce3bc42269c9fcf51
3191 F20101221_AABLTF shen_f_Page_110thm.jpg
dccf709911ca0393456d203fcd2ebafb
8f472088076cd3a0e9322a6867d27b0618a7d1c7
2175 F20101221_AABLSR shen_f_Page_039.txt
3725ab8a46beb6c468c3a4d0e77df76a
fe64d9fe424a9c3f982bf0e8331b3ae9e8852a58
115622 F20101221_AABLTG shen_f_Page_083.jpg
f89d6f82a410da0b5b2e2d4a25d1fbd5
4ffc3e13442d5c78b38f8ed594f6fa02b74888ce
51405 F20101221_AABLSS shen_f_Page_103.pro
cdd09de2950ad21fa5316dac84ca4766
2c724353148bedba6b37529d86aa4eea7285742d
1051970 F20101221_AABLTH shen_f_Page_022.jp2
1ce797a0bbb48494944d5625070725d5
2d4ab0f06aa16c1b49d25eeed37cf567bbfce3b7
2429 F20101221_AABLST shen_f_Page_095.pro
f892d676eac7c1f837b68ef8a57e7f93
c9070593ffb2c4df8fbb972fbb7ef838f56dd619
36796 F20101221_AABLTI shen_f_Page_043.QC.jpg
c10b893a2b0a446bda0b1de9eb6dca1f
b1ffee29f9a7c6a0070d20c17aa7f5732592dbda
149 F20101221_AABLSU shen_f_Page_095.txt
ad141912634b1eb7c0805104ae6c3a00
4b76437782fa37cb7d782b4d518261b444aa0c90
2055 F20101221_AABLTJ shen_f_Page_070.txt
90e632bbac003dce3a815ae19d206585
04b536e3ed2dd2bf4b84566fdafef32aedee7676
96786 F20101221_AABLSV shen_f_Page_028.jpg
61911a9cf0b4fe2fa8e34b47e3886748
ff1b95ba374607d69bbc0f5ef76601534de57cb4
9133 F20101221_AABLTK shen_f_Page_039thm.jpg
aa7bcc9932cadef6c9cd2d94f262edf0
b76dc64b5153b1233033a9f849b6388e085162c6
62112 F20101221_AABLSW shen_f_Page_123.pro
236b30e427b4cb3d43ec269347d843da
9caa47cd69a19a48bb01b48dda67e6833e349800
F20101221_AABLTL shen_f_Page_034.tif
325cdaf646d8af40cdb7fa3ada5f0164
e7961a00e1408088e43cccc003206e27024c174a
F20101221_AABLSX shen_f_Page_008.tif
f75132112d20f88463456caec9164d43
81453e6c40ffc008ab98d759133a097d9f1b4cc1
9255 F20101221_AABLUA shen_f_Page_125thm.jpg
7fd62a067880d137fe56fef2fb8626bb
9008a53e800653974df3a93c126bf5c441c2955b
F20101221_AABLTM shen_f_Page_017.tif
ce0b1943f5ec1a7694991d4a09966d30
471e082212351669e2455d299766b2937bca13c3
68202 F20101221_AABLSY shen_f_Page_098.jpg
bf95f883bf7f6b8aa7d7cefd3deb5e75
f19ee04c204a198565132e738577fa778cc54781
F20101221_AABLUB shen_f_Page_064.tif
e3a79423c750fb2995ca3916368a2ba7
7dd32f6bba75ec3b5e17ec8208db80c7167a367e
33428 F20101221_AABLSZ shen_f_Page_073.QC.jpg
606a2843a5e24fb73f245ad22eaefa57
e6e100796f9c40e3efdad2e72ca6f7a8c4f46caf
405 F20101221_AABLUC shen_f_Page_074.txt
a629ee53eb9f0477fe7b98c96718b8c5
79f44cb63bc013dffad5260266e09a9eba879b6e
14583 F20101221_AABLTN shen_f_Page_091.QC.jpg
1d95b2d3dcb2219b3110ee8909b7604f
11bf623c04898cf6f4f35658dc53c48eead1bbeb
F20101221_AABLUD shen_f_Page_122.tif
9fb55e756237a438d6fdeedfab44818e
b7105e7ce98d2746500addd30284b1607c1350f1
1051961 F20101221_AABLTO shen_f_Page_129.jp2
25e5effb9caad3fa6f9a4c32cd852107
1d623c0bfecbaa7ff5d57348fea9098430401b38
263509 F20101221_AABLUE shen_f_Page_001.jp2
f05b50de056b0149825114805e181fe7
3f057d5237a12d79929183a84fc37aad9283e12c
1051937 F20101221_AABLTP shen_f_Page_108.jp2
b6f52f7b99742d8dff5fd3c328abe793
ea72efc5b7749a07fd9967ae15e55ecf9831db47
8629 F20101221_AABLUF shen_f_Page_060thm.jpg
cf3330dffebc9a75e00a05191aee2bd6
45ccb7cbd95633afd0c95b7bcd96f60c2c34ba30
35631 F20101221_AABLTQ shen_f_Page_012.QC.jpg
c5e5a258bdf0625801939891f83b97ed
7b23ff4922a46f9b3d50f3c6e2470ad853af4c9d
9033 F20101221_AABLUG shen_f_Page_024thm.jpg
6210e9edc8aa05e58e9e645af39a12d9
9d17bb808f5d763ad17c78d04a72dbb48b4732c6
7945 F20101221_AABLTR shen_f_Page_088.pro
0ed7f1bfe185526f4ad0241c749cc444
5695b8364869651285c686cb69b37c7644124a9c
55679 F20101221_AABMAA shen_f_Page_104.pro
c5cfcb85e7c26b68e8852562a036eeaf
aff2e8c639f9f1949b4a41a3d505c92d73466cad
2244 F20101221_AABLUH shen_f_Page_069.txt
a9aef551f4edc241d0d00413c7672cbd
50cf3d808d30d30b342b5858758f3c065e4ab977
109425 F20101221_AABLTS shen_f_Page_043.jpg
d6a653c1ba5c6c9cb88330311f302610
ad14121cae780a19dcf1ee0438ef246e48239026
F20101221_AABMAB shen_f_Page_096.tif
3a5252bc4f033e9bbbf4a197a67a4f22
281c9d98775effa772f72e865a11147b9281b3a0
F20101221_AABLUI shen_f_Page_081.jp2
fac22724861045f5925d7701ee74958b
391df22a2218d1da885fd1a8418ee60308827e65
23434 F20101221_AABLTT shen_f_Page_008.QC.jpg
f974dc368a77d519e53fade3ecbcde9c
8144756ce5f5bf81ee6ef60e54b45f0496099aa2
9042 F20101221_AABMAC shen_f_Page_095.QC.jpg
4682dd7e2d972c0676a633f1f69fc2b7
68f518fd531b37214d1ba3de12b94d13c6339b89
1051971 F20101221_AABLUJ shen_f_Page_025.jp2
d4ac4c88158016ebea5012d32c6d84ac
2a989d56f5f0b69a6223b43352af8c24db7beca4
2170 F20101221_AABLTU shen_f_Page_074thm.jpg
fff3e25609873fc3221938a6acbebb68
49bcd75ac45a3053d409675137a06db208777533
1994 F20101221_AABMAD shen_f_Page_067.txt
03fd4aec257134cd7bcf8510eb781920
dd4c2edd41ba73ad459897bda3b8d08c79578cca
29645 F20101221_AABLUK shen_f_Page_006.QC.jpg
dcdd21ab949394df61457b9bd69e6d98
ae245647e784275e4aefefa7c6d5ca334780497f
34713 F20101221_AABLTV shen_f_Page_053.QC.jpg
5361eddafcaffd20fb50d0889bf1b636
55c8809e29d271c5a814af62c2078d605baff86a
F20101221_AABMAE shen_f_Page_077.tif
85e40f9549fad341c12c60fa8c4625fd
ef3ae4b3c67f24f058677d5254c5e0216c012219
1051956 F20101221_AABLUL shen_f_Page_053.jp2
74a3e380f6d39e2699480e4265e15c29
47ee3e69e40726861137550d9dfd246fbba00397
2037 F20101221_AABLTW shen_f_Page_103.txt
7d93b6b9615a6b2275f70b026b47b90b
19a90ad620d3f733adf7dc87ecd5dc7544e345dd
8877 F20101221_AABMAF shen_f_Page_069thm.jpg
f5b54c485def8977f806e79778f358d3
458decd310c9a04d8abf2c1427368a7298fca122
10292 F20101221_AABLVA shen_f_Page_099.pro
86731bb2bb6275717976729af30e4fbb
351605446994df7b100e45a461f5b393786c650a
2043 F20101221_AABLUM shen_f_Page_071.txt
c98a5ddd852f010402916175affee3e8
6514ca1701eea403ae26e80274696f113340080c
2159 F20101221_AABLTX shen_f_Page_013.txt
ebfbc8f9305c79ff53375a4033f5466d
90ddc3f02d1490948b2d468ba01e6711bc0d8326
F20101221_AABMAG shen_f_Page_031.tif
4dd7b352f1a71d64125b8b47f3ff9a09
81c08e341ae5cd94bfb2d7d069fdf7b243a6f5e7
57828 F20101221_AABLVB shen_f_Page_082.pro
52bd48946fc469820d13c0a8789d497d
d4fe48cf7d5356f9a0e78128e88a5d00bd4cfb8a
432 F20101221_AABLUN shen_f_Page_099.txt
4c461f3fe979c3a3c383e358c0ce85ad
264283b6a61dcaf03c70975ee3ea4333c786a106
F20101221_AABLTY shen_f_Page_021.tif
702b01129b859e9727a82336e45636ab
5e47752ce3ce5ffacd35df23bd42abfc7e1b1abf
1877 F20101221_AABMAH shen_f_Page_052.txt
53d11a6a631c64cf60e60a584b10916f
f98ac134274b19f968f76b38cac74640a388e1c8
2250 F20101221_AABLVC shen_f_Page_032.txt
830c549b868dc5ce1b8db015bfd09623
88201d7e6987d628c14a8d4e678675a8ede859be
1051943 F20101221_AABLTZ shen_f_Page_064.jp2
c0969327dbba1ae0339cebc9f9f7cdce
f118d6158eb142f27b7ce56c66fcd99a2c6ed29b
34399 F20101221_AABMAI shen_f_Page_129.QC.jpg
a0c486d3ce6e0d9fc05c88d44dab4ffd
99d77f2652263a9376e6f4456ceb655f3384855a
8849 F20101221_AABLVD shen_f_Page_058thm.jpg
67ecf5849f01e792b36fc68c6db772f6
5dff3b215e29b034a7653a2426fbd0aa013cd725
127323 F20101221_AABLUO shen_f_Page_126.jpg
dfeb09acf40a34a7e3dab91cf815510e
202558f035922db6389d6636c68523a7ced9c625
1051949 F20101221_AABMAJ shen_f_Page_042.jp2
f953e538da86a9e84d52348a6bc28252
653c10ee16fb3ad85e530e406eeaf6d0781ead1c
62259 F20101221_AABLVE shen_f_Page_126.pro
e1ffbee1533eb5ac9f9e88eb0e4861f1
b56013d7211aa29c2e276d592d398d8aaea4a4a3
F20101221_AABLUP shen_f_Page_001.tif
7a5dcfba19029796146e62f308e8b24c
973070ad4eda133c9d8573af7f422c638f090e7d
22265 F20101221_AABMAK shen_f_Page_116.QC.jpg
b455886f289952e6a45953cac48fdd51
07c597ca278e0a5e28d38c98a2dd3c805d77b862
F20101221_AABLVF shen_f_Page_020.tif
4962c04f61f995fae6ef8a3a5485954d
26a26ba43cf4f302b8139cf5fe94b34f935ff402
8447 F20101221_AABLUQ shen_f_Page_030thm.jpg
c7151d3caa17374cb2f971d0ac1c0daa
8857774bc4b98e69a4873a4331ddd72d0304973d
2153 F20101221_AABMAL shen_f_Page_020.txt
c61dd62512ced34a9215d069e56a1217
5b148bdb5df6095cbf24000473d5c24882959ac6
F20101221_AABLVG shen_f_Page_013.tif
6989530a0ae026313c3ea437fff6e8d2
5b8b009f8f16dd906330e0559b3ed29845a982cb
8546 F20101221_AABLUR shen_f_Page_034thm.jpg
7510c577fee2500dcae6f5fd29617d11
507f3579593469856cba3f87bfc639beef792ed1
11844 F20101221_AABMBA shen_f_Page_075.jpg
118585518af602855026694f012e3963
187932eeebfa533d5c471e8a0e74e68d1e245ca1
54301 F20101221_AABMAM shen_f_Page_064.pro
94ff0a854244ffae00d4d8a76774d973
5319224a495c739eb8502c798633ce94d6d19470
6927 F20101221_AABLVH shen_f_Page_090.pro
9bb39886de103b610dbf52b22ece1bdc
e2dd3f0d088470e8cead6fc78e941491d51770d1
993 F20101221_AABLUS shen_f_Page_003.QC.jpg
c624f3808ee119f9940df1ee6b26bc70
c9682844dc80fdd5325f29faa64bb5db2467e588
54154 F20101221_AABMBB shen_f_Page_031.pro
8b0f518f4f7b169a1748222e3c042591
85a1fc52d001352509ed8de2ef6f41efc2d46888
50565 F20101221_AABMAN shen_f_Page_062.pro
982d4617e1b4b5a5bc27b9b8cf2a68ee
d96afd25e2697228470eaabf4a207a93577a0a92
1314 F20101221_AABLVI shen_f_Page_002.QC.jpg
898fd12742ea9024b61a5c979aba3656
fbedb085fbec18295d1a3e592d36f3728496e276
101472 F20101221_AABLUT shen_f_Page_047.jpg
dcb114e6d52aab389c9ea719996dc381
d05e322e274c0b673e27f3f21971badf6776ba4a
F20101221_AABMBC shen_f_Page_025.tif
c34d2b558861579e66af6e7209bcbe20
e4e2ff7b5f9e04027528f431bcebf87baba73856
6848 F20101221_AABMAO shen_f_Page_113thm.jpg
499ef99ea1f426a514c0f4b5f3fdc0a4
23611843ca667c135d82060d7378d0b4490e43fa
111543 F20101221_AABLVJ shen_f_Page_104.jpg
65cacc819370fe4a9d13287f9f541522
a26b321233593c3ecaa9e95981cacf3acb02a89b
54744 F20101221_AABLUU shen_f_Page_027.pro
6143a03d7ed42b79b83f9b21d14b77c7
04574c8ce0ba866506c2f5b12df85846f8149df5
1051965 F20101221_AABMBD shen_f_Page_019.jp2
d327319e2398848e64106aca64cb7211
8c7cfa768116ba22535e3a15add6a0ceabb88235
30581 F20101221_AABMAP shen_f_Page_065.QC.jpg
d9770eb60e03232959a9a7802c17dc26
793348a0334936ceaed2b983cfc3573aae70b634
F20101221_AABLVK shen_f_Page_057.txt
d138053f0a777509b66cd76b40324115
c62063e84a1e13587a54415acf1fdc18782845bb
33842 F20101221_AABLUV shen_f_Page_062.QC.jpg
50c921dec7f6611422a9f7cc847a992f
4a4e5c5811f023e58ebcd99f04214e05d5a86346
1051978 F20101221_AABMBE shen_f_Page_046.jp2
0eed551e05abe2179b0bd8ba693708be
0316e4c80d20ef9dd2f3228115a8f87bfa444514
F20101221_AABMAQ shen_f_Page_119.tif
04b6ebe51c542e3ea12770c5e33f047f
36c46e8388096547b4895c7f871889f7f32b5135
113359 F20101221_AABLVL shen_f_Page_082.jpg
8e30e599ada412c6c481384b3833cb18
20d8a4b31a1c119f1731cbdb008510c330908d5e
24307 F20101221_AABLUW shen_f_Page_098.pro
b6e941fcfd6d71796f386cc2bd6d35f1
54aa930664ac872c031184469ce4fb4775ddf959
F20101221_AABMBF shen_f_Page_023.tif
4abc6c2e68f4f935c34813654a573343
520a7bff63efb996541f4026bc229678a743e15c
F20101221_AABMAR shen_f_Page_130.jp2
54e0e1228686054e171f8e03f7aee317
a8d1cbe42c0d437b050f7105c3d01f3dfc452d40
1051913 F20101221_AABLVM shen_f_Page_082.jp2
a56f4bd24c836d07ad2be2b572e6f0ff
5502d870b3d9f57b25d9f92ae26bdcd0a399f1b0
25884 F20101221_AABLUX shen_f_Page_072.QC.jpg
2190e49b8b3ee45641a717b20982ee9b
16ccdcf5a3dc297a70fdfd93b3e79b7591d21313
1051973 F20101221_AABMBG shen_f_Page_036.jp2
13d6ac056a7a520bd3435f4471ed9f51
c381e5dc5d2306dec241fc29853c78827601d3d3
2211 F20101221_AABLWA shen_f_Page_014.txt
fa1806ba49d54098bba05348d9ac49db
840a9de401c1deb41bbc477f07f50c531659f30e
36505 F20101221_AABMAS shen_f_Page_046.QC.jpg
d7ae0348183a940a0b81f22839fa0083
3ff673d61094a9941988999433d1fb29131b3bea
3046 F20101221_AABLVN shen_f_Page_095thm.jpg
e31c7ccd4b9ee1ca798d19aa85f7a885
ed7220d8cb63cf79042ff051a0e0ccf52e1a9f68
2077 F20101221_AABLUY shen_f_Page_079.txt
ae523f94b0b5c414fea8ae06d68646f1
e0e2420a7c3045723aaad31e54ca6857207c9dab
38883 F20101221_AABMBH shen_f_Page_094.jpg
c37e7826f773c931bbbbdce6afb62ddc
689f3205db30facd21df086d5e51fa6674b1335d
435 F20101221_AABLWB shen_f_Page_094.txt
b54c3a7b6bfeceb398a4f9ba26f4b37b
ed8d438bb80de5ac3724d921e4a532642bc2a1c3
F20101221_AABMAT shen_f_Page_127.jp2
586b533a34854c583b708c0d5db08d2b
10b129117d43a414e4feed04688cefcbbe7676ed
91490 F20101221_AABLVO shen_f_Page_068.jpg
5bf5becdb4ed39e4cc1b4ce5b3c8e8c5
007f197f42054d5f7048a805265f7a1f8a6f8f42
8705 F20101221_AABLUZ shen_f_Page_078thm.jpg
a2e80cc6a566e94ec4f878222b497a75
042674e12a2a2bb7c27db79e1e9617b1f5022855
1051931 F20101221_AABMBI shen_f_Page_103.jp2
d3104989efa7568ba780f162fe98ccc4
7b54458928ceeffd547ddc612d8a0c8999d12fb6
451 F20101221_AABLWC shen_f_Page_066.txt
a002e27e092be3c9956500995dcef6e0
c6d2c1f3facaa0ff9344c81c4931fc0af13a4f7c
569831 F20101221_AABMAU shen_f_Page_109.jp2
b489aeb7e79c410701a7a90cf2ae7897
c4b4e497fdc7099aa7c57861329355a86ec9e091
F20101221_AABMBJ shen_f_Page_022.tif
f6363dc33d28713021f700a2b8ece6d5
22b4dcf45a286c52f1b853dd2e51e969a600fc97
F20101221_AABLWD shen_f_Page_055.tif
95ba3f92a37b0c154414624dd1d5ee4a
0543962e80833029cf422e84f21d88b6e64414d2
2128 F20101221_AABMAV shen_f_Page_031.txt
131c9e2197537e31b940097faa84be0a
ac8bd8132946b1c96ab747bee500898e290cb8f6
2072 F20101221_AABLVP shen_f_Page_050.txt
1af896a741bc8492821479198ee09b03
dbae740d4bed09208c19a42c43a3ba4514c279ee
1051974 F20101221_AABMBK shen_f_Page_037.jp2
6c65a69118778ac8e6ab507b1dfc0ae5
b18770c449fabd83e882f6c86ebc51a078327156
F20101221_AABLWE shen_f_Page_067.tif
e49ec90060b7cdcf0e01824284b45f3b
c38a870f4e31776dfbf7db816f864c3e541f7a3a
F20101221_AABMAW shen_f_Page_118.tif
9ffd73101916a682e7eec872a5cf45e8
f428cb569a04840755c7557756fc617253c10220
8401 F20101221_AABLVQ shen_f_Page_077thm.jpg
ea35f3f817d040f1a2d0590826a7287b
e334ca865d764c7b72f58674df9d5c8e59de8b10
9040 F20101221_AABMBL shen_f_Page_026thm.jpg
d83ad9600cbf61e02e6fb3c43c6f584d
476bd2f5f4361c92cbe745006952c883064a0f3c
2069 F20101221_AABLWF shen_f_Page_085.txt
75ad5443c983e505aadb1a54622d7470
17f24860e46a9f48af53ea5d25031c42683ce750
36420 F20101221_AABMAX shen_f_Page_058.QC.jpg
168f463985215c3f99dcc72a134911fb
63885e9e5fe5de177c64b0dc728a23df290230c8
54014 F20101221_AABLVR shen_f_Page_046.pro
324832d0cdad0315aa8c8e77bfe466b2
8d97f0429fee5a7a886df68713b0c9dbe0a941d0
F20101221_AABMCA shen_f_Page_057.tif
3a0f68152e61860479c22116d2545baf
1c0c96bc8d2349d35ae9e02bccb3ae6e055564ba
32863 F20101221_AABMBM shen_f_Page_047.QC.jpg
1f48cb05b0719b2406368f2a738a59d9
c1a612a49b1cc6501863c573936c6a6682f644a0
998 F20101221_AABLWG shen_f_Page_109.txt
5a1cd93167eaf4dde69268f48232d81d
942c591ade5d190a67e22084338389edfbebcc5f
111242 F20101221_AABMAY shen_f_Page_055.jpg
1f4523416118d71771a83feecc239a80
aacbc2c5d2eb4f1abec55efec94e00eebda0d9fd
319 F20101221_AABLVS shen_f_Page_088.txt
483065daf32cce27e343b0ec8eb56810
4d23756318fecece6e2c454a394ae3c9422bf905
460388 F20101221_AABMCB shen_f_Page_007.jp2
3414afa3ccd87b463ef3925b4030473e
e1bbae862929fc3957c7a8d8f55e5095d5dcf3ba
25283 F20101221_AABMBN shen_f_Page_119.QC.jpg
c9e7e8e7b93f272b12741fd56a787e31
e1143c9ab15ddd9862515fa58a02ec66428fc2e4
708248 F20101221_AABLWH shen_f_Page_113.jp2
96f54af417fb91e1e46cc78387f3abe0
f7c109727d5bdc1ca3e7e4a3c871186487c1e36e
33774 F20101221_AABMCC shen_f_Page_048.QC.jpg
226fbab44032fc54e25396a7ddebd57f
861088a99c8903bae5460992a0cdb030fb9c20f2
F20101221_AABMBO shen_f_Page_040.tif
7e705c6e14bafc87ca0d8c1e2887b7c8
653b773debc847793191d51d7be56ae8233b5b0c
34561 F20101221_AABLWI shen_f_Page_084.QC.jpg
7c7710fbcbeaf0ad39c13625a456f91b
26436e5f712d8935bafa177bdd4444294a4a9a8c
1051985 F20101221_AABMAZ shen_f_Page_078.jp2
393b5cad1b47ab622f652e4f1566c754
4e0be0532d58ee30a619a36bf6a99e504b7b13dd
56341 F20101221_AABLVT shen_f_Page_014.pro
819f4731da6da0cedd9beff10803f0ee
c7a358243add6480176a931e79e2ac91574bac9e
9032 F20101221_AABMCD shen_f_Page_037thm.jpg
caa20c579d004d82e472acb3527ca13f
f31e53eb69f95f6647f2df53d593f501da87dc12
235164 F20101221_AABMBP shen_f_Page_074.jp2
1a5d19f311fdf80260fc5457864b0331
69d826a7ac90cca60f85a5e9ff15795f8bab9605
46810 F20101221_AABLWJ shen_f_Page_028.pro
c0b3be6accb726c7ad05a7abdad090aa
f6098d027bd52057f698b0859b42a24760e20554
109568 F20101221_AABLVU shen_f_Page_035.jpg
d5cb53817dc239f0f43181e3fad9132e
6a4fbc84711c57a0cf1a4604bd0259dbe8de48e0
36917 F20101221_AABMCE shen_f_Page_049.QC.jpg
4e36e87da982402bb59bd4d6bdc20333
a99ec4d71b04614d852b8a2864feff0d3ee884f8
32850 F20101221_AABMBQ shen_f_Page_103.QC.jpg
3ce9743b6e18609226fbfa5321efa870
e6f889e94dca507f45f5f59c7b558f1584875655
108878 F20101221_AABLWK shen_f_Page_017.jpg
636c98dc56e57f3fd999b5012191fbcc
8e637ba6d71857ee47cb45167ccae9d0483ac68d
8381 F20101221_AABLVV shen_f_Page_028thm.jpg
2d0b6b7836c376f8b81e675d94d0df29
63ab6175fcee463c8f627be5de263f659b711bb6
55621 F20101221_AABMCF shen_f_Page_081.pro
7485d15f0029379ca53114ab99118ef4
29008bdae98ae10ec662f0b8e7e3f105f0389589
30461 F20101221_AABMBR shen_f_Page_096.pro
91358a787ac7994175f5bee9e4f03205
0b97ce9b6353b52f77756ba4dc62b071813595e2
23916 F20101221_AABLWL shen_f_Page_095.jpg
6243ab7019fde13e5d66def7dfd59ec1
8d2b00382ff416b9c8cfbd69d98af5d6ed7bd33a
2120 F20101221_AABLVW shen_f_Page_043.txt
26c2248271ab769f2e843775550673a4
ba8bc0205d67c84ea4dbd89d32b80b3638287be3
F20101221_AABMCG shen_f_Page_123.tif
2266c6fc4806e67baa93038daebfec22
dae15bc40925698b529925cd55d78493552a510d
F20101221_AABLXA shen_f_Page_033.tif
43c4ccb44cee8264a6060dd176560b92
647e5080512c1f7567ba70d68a043ec94233caad
106250 F20101221_AABMBS shen_f_Page_056.jpg
f0f67239d5994a32af9507ad267a7962
f1592fe348e0b6ac5aaa6450808e6ea850d42db2
1051975 F20101221_AABLWM shen_f_Page_031.jp2
ba6470dc5fb68eef0e3590a8906debe4
9d97f462fe5c61ee38aad78849cd9095020e4c09
17496 F20101221_AABLVX shen_f_Page_109.QC.jpg
e2b9ef4c992f3a52fee38d6d1a80991d
e1c856e041bb4a41a42e0cb270cff88b45e3f4cd
48727 F20101221_AABMCH shen_f_Page_008.pro
caafef4410167560cf617c9abaa1269c
ec104655907469847e1dd645dd3ad13cf42978ee
25021 F20101221_AABLXB shen_f_Page_097.pro
ebd2c9ec9b472f4dc7f36e9173e892f5
850db1fd22dae489ae72ed03129eb45416f480a3
F20101221_AABMBT shen_f_Page_132.jp2
0fc8fc754a533d761a2e0cff5f4237d5
5ca69b9a16aa53a9c6d17a1a7510ecd3ef114ada
753190 F20101221_AABLWN shen_f_Page_004.jp2
28eed029363a3c0cdd97d082c0f754ce
3c24da9762b1fd29014ad0b830d7af483626cd71
111575 F20101221_AABLVY shen_f_Page_108.jpg
084bb6ccd14440595d3be72a15e29071
0a320ea76a1f5c70739f6244c90b3927561adbf7
39283 F20101221_AABMCI shen_f_Page_115.pro
08adfd50aab36b53f0f30ed454b1dba0
d776023de481665c3b70b06a671fd9c2927c0bac
14991 F20101221_AABLXC shen_f_Page_092.QC.jpg
53b2472efd176af350bc9df71db9fb2a
97e8380981eec01f6d29eb604a74532947f4d504
2110 F20101221_AABMBU shen_f_Page_034.txt
3dbb5013e7b3cb040168af029389bfc8
9297226413dd28d34deb2341d5d7fefd1d987357
32572 F20101221_AABLWO shen_f_Page_110.jpg
9eed02ea5e565519f98443ea7485f048
913a725b2a02b3883ea236d3ec6a90fe81aaf09d
5023 F20101221_AABLVZ shen_f_Page_093thm.jpg
0611bf7283bc0f664d54aec18730946c
387c85e4b8aedb4b451768603e9f0a8c3afec8ad
51974 F20101221_AABMCJ shen_f_Page_134.jpg
3bfe18bd8871e8da59fad91e932a7df2
b2629e7f394137ade351f522f276fef1754be33c
8972 F20101221_AABLXD shen_f_Page_124thm.jpg
d367f15612e20bef76d9fdcc8fe9f766
e99a70d60df2a8edf45f3d3dc0a45a71e83b2c99
1051952 F20101221_AABMBV shen_f_Page_049.jp2
cd99c0d27f0c038895165d961b5dc32a
7519e5a5a95298bfc2fa7e14c0c4c40a9f15e409
F20101221_AABLWP shen_f_Page_133.tif
40a45006f75eafdd6dfd1000610f0a0a
47de5bc8b855cbbd49f1c29db9d6590392f7d4be
64091 F20101221_AABMCK shen_f_Page_131.pro
fbf28d9ef1bbbc588a8eeea04262bf39
c0adf4936f86f2e283655de8c19ce951a2080af4
20110 F20101221_AABLXE shen_f_Page_090.QC.jpg
0c02681ca072b87ae2bdfe9316af8441
b6ae01edbc04c41cd80f315dd4e0c071044faf82
54858 F20101221_AABMBW shen_f_Page_044.pro
ec2586f1eb7e8044e9d4ef24709a6da9
cf60cf5288fefd2168b1784aa771845bb1e2ce3b
1051979 F20101221_AABMCL shen_f_Page_047.jp2
e450ab96010da595c46a89aaeb6c449d
59373e54e7ffe6d1b61ee1061eed77365eaf0e1a
F20101221_AABLXF shen_f_Page_120.tif
3fa91e8e5ea1c461931d84578b1cbb72
f8b978bfee5a283e78ac5872ace2384d7feee414
1051977 F20101221_AABMBX shen_f_Page_035.jp2
1537023c82974a76897030375e0dd8cc
0ffe4251abef214587826397ac4ba80f95848f76
F20101221_AABLWQ shen_f_Page_039.tif
a0a121a9b5496a71730d9303f25cbc44
b147dc2d1f13094991e92123babdfdb6db7a8686
8954 F20101221_AABMCM shen_f_Page_120thm.jpg
59abe72e34f254e988b91da8daa69553
347bb7bf46cb4641f3950aafea829baa986ac72b
478 F20101221_AABLXG shen_f_Page_001.txt
5739f1f8e7910cb3654dfeb14ecf88be
67a7089460d2ae066f0e3d49009c8b8eafa1470c
110253 F20101221_AABMBY shen_f_Page_060.jpg
1a149ec5b9a9db2f3e1cda7d44e8f71c
065591bacd0b0ad6d65deb912f9ed1b5ec85ee94
F20101221_AABLWR shen_f_Page_109.tif
f3d59141b14585ebcdf5729f911c7859
48541ed8afa937064e9f115d05a1416eefda743a
62859 F20101221_AABMCN shen_f_Page_116.jpg
bfdea5cc46049d1a67d4c88a98f2e8bb
22e083792b356853bbdeb92276b2fec0ec82b712
F20101221_AABLXH shen_f_Page_041.tif
410d61c06edbd1d9821a28c23a4b6282
bcb16946975c9327812d25c55e84bfd48bf7cc83
F20101221_AABMBZ shen_f_Page_028.jp2
d7709e6f940534897dd9de99c1d1e238
0dd3c02248f0cafcc172900f2e7103a6facf79f9
F20101221_AABLWS shen_f_Page_128.jp2
477df4655bcdfa17fbf4d82222bf1c3f
76662b8c846fe501389657522f1d2cc15c08802a
52584 F20101221_AABMCO shen_f_Page_077.pro
3be1bb05b8bad3f5921246073e21b9b4
39a778fe796a86a622171c047d672dea4794dabf
106663 F20101221_AABLXI shen_f_Page_079.jpg
ba615da0320aed5aeb94fcc67acbe32a
8640e43c94724e4d2986287b32b8d83b6badced1
2533 F20101221_AABLWT shen_f_Page_123.txt
782cc32ae6dae7918ab0289e1974127d
e373bdf90049e35987ddf36db0123e9c4d85931a
27025 F20101221_AABMDC shen_f_Page_001.jpg
6d39f809c65aedb8ef5bdacab3fbf064
19fe237148d60186b8c6e8a9fe651e8857249a2d
F20101221_AABMCP shen_f_Page_015.jp2
ef2eaff0c3e4528873a72d1db076cc66
27241e0f26329288305d130bc8080c128a84dbaf
38055 F20101221_AABLXJ shen_f_Page_019.QC.jpg
f1e7de7269312eaccff967b628f0f731
6004e1b9d5f24d6064dd4b6abf9f7d0bcd078589
100115 F20101221_AABLWU shen_f_Page_102.jpg
f69f3de98101c40dd330604c71390b70
28eaaccfc2a1cf4dec6cabf462a04bed063a36fe
3202 F20101221_AABMDD shen_f_Page_003.jpg
93149fe5d9920af89443f61515a414ef
fa245a8ae05ee055703b7243d312ca0277e28d9c
31411 F20101221_AABMCQ shen_f_Page_086.QC.jpg
0d24135ec245259741a86f7adcc0c540
22e048c20f8a21e4a7e7162a319ab5f78e9c6158
9222 F20101221_AABLXK shen_f_Page_132thm.jpg
34f021f11689fb48fc46de91719b20c3
c805ee1f98a114d598b399d8d1590f9dd89211e5
F20101221_AABLWV shen_f_Page_124.jp2
43b716bc5644f09e3a1ba99b26fd99cb
941ebf99a64d2a48d376c515bf189e7048591de2
130582 F20101221_AABMDE shen_f_Page_005.jpg
c1a62a7157c835778d4b680367fd80cc
ffe856f26b90a66b360a7ca235575c265e4fb47e
54235 F20101221_AABLXL shen_f_Page_013.pro
fc9a04411951e6567a420c65cce1055a
46e673f9a54aa830325b2a39cdb3f4923f7bece6
F20101221_AABLWW shen_f_Page_061.tif
6be17cd5a566341d3565bf1a5bea8518
fbabf27b279e600f08c72686ca01b4355aecade3
138142 F20101221_AABMDF shen_f_Page_006.jpg
361fd637b883e2db82fbad493b8374ea
03d9445ce824e9a1483bddb4390167891268f790
105410 F20101221_AABMCR shen_f_Page_053.jpg
096cd68a3ee6c2a1bc288544e3c5fd07
c69e6e7a94d6ac9b1dbd4c87340a49eee31941cb
55156 F20101221_AABLXM shen_f_Page_039.pro
7a5be54bc95bf07e18057344fc206e38
602c55289dfab95803945864f37faf15081111ec
2039 F20101221_AABLWX shen_f_Page_053.txt
0e0d48be0ed9f6ec45138174f87434ad
3a55b43f6badce5d4a9a058d26c45f4db0a0f919
98901 F20101221_AABMDG shen_f_Page_010.jpg
75dbb205de5ac96eb903aaf9c26b3344
1b6370733fcafb33e909bfa3b4920ca254904933
1051953 F20101221_AABLYA shen_f_Page_059.jp2
8c831428a79451ede43ee2fabae117d7
257366e8dc7824acd94923e4b44fbc6469accbb5
108965 F20101221_AABMCS shen_f_Page_031.jpg
3324d961475e0477e1c02bdba1ce440f
73ef51cdfd2132a660eb0f3f88ba8bfd38f103ba
119929 F20101221_AABLXN shen_f_Page_016.jpg
9edb2d40fdf65d72990905cce775a61f
9028de9571ab687b8fbd42b136a6855159eadad5
1803 F20101221_AABLWY shen_f_Page_133thm.jpg
db8d8f707dcf2268924b020183372a41
92e3edbb55eb6978844085d2e9c45cb1f40280e1
108553 F20101221_AABMDH shen_f_Page_012.jpg
41c8da593784c15535d6efd14b4db696
6fa31e317c2de932e90d5fdb5ced99cd9be235ee
2025 F20101221_AABLYB shen_f_Page_001thm.jpg
12283080c0a4492b39830be21912f267
b2ade3cb9022548e31c1ed77e255bb9567836c5f
467 F20101221_AABMCT shen_f_Page_002thm.jpg
a6e4a39dc3bc191998705318a8e62b4b
39335a18db3b6b49b8500a4aa6a112cef5e9e963
F20101221_AABLXO shen_f_Page_009.jp2
626c7f5db0e69327b3c55d9d1b3e1a42
6d0101b01a2defc25bdc099fbe3197fefa4cd1b0
2093 F20101221_AABLWZ shen_f_Page_023.txt
fd24d6e7be078249abb5fdf9c2269e5c
656f0e3605de9c40cd7ed19f23081341c849caa1
103460 F20101221_AABMDI shen_f_Page_015.jpg
393553e8cdcf5d0f917b4386fc3105cd
727ff04470732f0309b1fe8193d52c82df122768
22943 F20101221_AABLYC shen_f_Page_009.QC.jpg
b92bd5f69b596c8d62d2f6ebedaa46f5
1fc28600898378d942a737038fd53fc79b238b3b
129982 F20101221_AABMCU shen_f_Page_124.jpg
7e7fb6266890de1d21d795006de5e5ff
b880e893fd45bb0c60bf0e67a9099fd4d41303e4
32387 F20101221_AABLXP shen_f_Page_070.QC.jpg
fdd637bdb06ac47439a490667ba70745
b04dbbe6010cddb68833cec50975ec4e7d1e99e9
107956 F20101221_AABMDJ shen_f_Page_018.jpg
1f1b155b1658905c8a2e0df88a739e5c
182174373bbf06b6dc0141d1945d377bff3484ab
F20101221_AABLYD shen_f_Page_028.tif
c9f48d2b2f1eda55c35a7299696e1bbf
2e66a187018f87ea8ff17c662a438591155b2a06
63165 F20101221_AABMCV shen_f_Page_019.pro
61f0da24db2b9bc512cb58b3586ff2d7
0b221f8e4f355129683966ca5e9a51199b808dc1
F20101221_AABLXQ shen_f_Page_007.tif
996011edace444ebd37e9806e97f2fd8
c4bec0e0b8bb3e916e4f54bb6ba97f6b1ccf6acc
125490 F20101221_AABMDK shen_f_Page_019.jpg
141ae5d24f49a588caa413c8a1621a93
1b99735b5f0ccc6d423d1115f277324c059e3fc4
1051983 F20101221_AABLYE shen_f_Page_005.jp2
f0079e02d11da026d28cd54ebe93a0c2
3e4209537637b3abda49e918589735520a5d67b6
34258 F20101221_AABMCW shen_f_Page_034.QC.jpg
e99891f3dc7e90c7b3d2ab7fc34983cc
ef8a2e0f0c69053cb19d6b4a42b31e34e8fd200c
110638 F20101221_AABMDL shen_f_Page_023.jpg
c63551cb0fd054f6cbcb93eff0576b21
3fed3fc0d4085a75e0bf4bcea9156275b70e4e1b
1403 F20101221_AABLYF shen_f_Page_118.txt
e3e772d904c8217b0a475249abca82a4
6dc7b1aaf1febc631224bf7d0b3400b83e1165b0
99688 F20101221_AABMCX shen_f_Page_073.jpg
3d3b8e2d669c7a11fc994e6cae604cf1
98292b577a39441ecdec28a70d62d3078ad44ac2
3295 F20101221_AABLXR shen_f_Page_011thm.jpg
6261967f9b4795124e9821236581a7bc
531b3b6dbc88c763de9e3d4ac2199f195b551ac6
98136 F20101221_AABMEA shen_f_Page_076.jpg
47a5b3f6931ed8884586004a27477f3a
b4e0c22c3dcc8219e7fea50562661ca73fc0f7da
111375 F20101221_AABMDM shen_f_Page_025.jpg
eb7ce9f9d2708d7f0c1dac4020e4ba05
5684f7298fc308e5311d711d3434b632302f6d63
37135 F20101221_AABLYG shen_f_Page_045.QC.jpg
180a831517a21962bd88d7e878ea2f96
a04a2282de5116e5ab3dee28fabad2122af3d4c2
8757 F20101221_AABMCY shen_f_Page_031thm.jpg
b07d2155173b27936f6a0878e504b42c
496362cdc17d87c56fd8ef5556c71a2f0de46ef7
36571 F20101221_AABLXS shen_f_Page_042.QC.jpg
3b0596744944fdde0af018924ead54e4
50313340a7c1b1f355223ba66ae94eaafe31ecb7
107445 F20101221_AABMEB shen_f_Page_077.jpg
e4890ac62061e8f4588c3b8797051a4a
e22cbcfa061799885353cf7541a2f7e4989ed2fe
107030 F20101221_AABMDN shen_f_Page_038.jpg
3b7b47fca9f483ce702db964109c2d7d
2a04bc51a2b4271a63b824e919fb9f8558022d42
2168 F20101221_AABLYH shen_f_Page_008.txt
9deba7f7ecf23224fe33ae193aad2784
e3a5e7452e627025d4f732a67cfa12c28501b848
199150 F20101221_AABMCZ UFE0021931_00001.xml FULL
842c52bd8426f23c2337a76083d4d8bf
ae9106f47163eb61ae244d0c11d4568b91bdfb04
F20101221_AABLXT shen_f_Page_130.tif
3a4f4c58ac3fc12ef7386360f04984f3
b15ab086eb0b1b9f61a099cd839a5781543474a6
107335 F20101221_AABMEC shen_f_Page_080.jpg
b275bfe3bbc784fd11ace597dc6a4811
4d97dff9032b875d02caea039c9dd898570a01fd
114456 F20101221_AABMDO shen_f_Page_039.jpg
2e57460e1596884aae3ab0ae524812c8
485a83d2b65b276b56f4c0813d188a58d1eefbf6
33093 F20101221_AABLYI shen_f_Page_102.QC.jpg
8ad00b9b58394fe469644826b1812698
20f7275037166fb9281f8b7ed11f645670785c77
2044 F20101221_AABLXU shen_f_Page_076.txt
6f0f62d0aff353cfc53728ba7c4497c5
ebad51e0341add6a118b8eb813aa1ab89e90e501
111681 F20101221_AABMED shen_f_Page_081.jpg
a01a71032927be49b368f18be639451a
0b0f820cacdf4e69464ac389bd95ce38ed024e03
109066 F20101221_AABMDP shen_f_Page_040.jpg
fdb7c7ce6c0f2f0c996ef149cd66147c
66e8586b6dd482959faf133399546b20851dcc94
1051954 F20101221_AABLYJ shen_f_Page_033.jp2
510b59285c4f161e267b72037c186910
6cb4b334c77d720721222a8af7824bfd3ee1972a
35186 F20101221_AABLXV shen_f_Page_131.QC.jpg
d269d5fcd4746bdb23c873561b11cf49
c6db69c657cbac0f6d33696da2b26c02363d60a0
90691 F20101221_AABMEE shen_f_Page_086.jpg
03a10116e94c4430f853792f9184d183
a1d817df8e573f32e6558916cf4b8dc18f0cd9cf
109558 F20101221_AABMDQ shen_f_Page_042.jpg
34510a7a24a40e69db6982b8181e1e87
77437c7b8e6eab1e32158b8973550ddf1ae9eb15
6945 F20101221_AABLYK shen_f_Page_005thm.jpg
ec3d8db5b9cbc3c95e2695177e9ee68d
55195f1f647eab9b67174fbe3214becfbcf38897
8264 F20101221_AABLXW shen_f_Page_067thm.jpg
042f32a49bce7e8f89db7474fad8c2d4
9c30672bc0b7a5195c7557ba7a12d5727b6cad7e
50595 F20101221_AABMEF shen_f_Page_090.jpg
ee227ff6d1b0fb13879a29c189958dcc
0d992ee0e96133a7453566d75ce59ae6665f833f
105273 F20101221_AABMDR shen_f_Page_048.jpg
0f73fd439aa57aa9b549b04a76113d54
e9af4ba15eab8097671702feaafd12c2266676b1
977021 F20101221_AABLYL shen_f_Page_086.jp2
2449f8763297ae495f1d55847b1d7261
668b7a5431b486d1ddcd633246d5e26b16720cf0
F20101221_AABLXX shen_f_Page_093.tif
3eda88762965da286f6c44deb4ca4056
8248779f2ebcb625da3c5d171abf66d08d573398
72267 F20101221_AABMEG shen_f_Page_096.jpg
5310ce1c1c223d9b3b8bba3ec1f31b94
54e2bd4aabfc475c1d3945f839095b75b7e76bd2
37594 F20101221_AABLZA shen_f_Page_069.QC.jpg
c523f7e234cec9d13ecadd6080d2fc21
66822a2dceffb6c33bdf85b4d31fa2fbced35fee
106653 F20101221_AABMDS shen_f_Page_050.jpg
27ae4a28e240dc2bfc1b2ff8fc3c20aa
ed0096e834e331297f5eea8edaba63447e251903
2094 F20101221_AABLYM shen_f_Page_061.txt
e005e364e0af49c21a0167102fa23e41
cbb1e6b60d1376b0943ca5c04c5bda4da0c10aca
1051935 F20101221_AABLXY shen_f_Page_055.jp2
c5fc6ea5adad5e77ddd56fc179f173b9
f4fb47cccf885b1a89c591f760153d847b838077
97727 F20101221_AABMEH shen_f_Page_100.jpg
9af31234c9fff2e0d2991012c4795d2a
a052f261b9dc9e55ee0197800880d806640a826f
2164 F20101221_AABLZB shen_f_Page_058.txt
cfce7b30e35c5b1ee6d06e3b0aba1e9f
ce688f375df5293eaf6a33d4b1c41b389279e35e
111544 F20101221_AABMDT shen_f_Page_051.jpg
04b11b1fef4b07d34799a635ddb3c2a0
af4b0e618f2fe894b0eaa56c7864ecd5d07b2bb0
2160 F20101221_AABLYN shen_f_Page_024.txt
9f6979e80f49d295c79cb460b3873cbf
d6ebc28fdad23076c39f388d578484af79f7b684
52648 F20101221_AABLXZ shen_f_Page_038.pro
9198608aa0b4680e1f1b1f3aea58feb4
41f6f3836f7613bcded96fbb92701ab48152f669
104100 F20101221_AABMEI shen_f_Page_105.jpg
b8dc7d3f5d74b3fca7e19da04e389a4f
2a5738799a1c2cd9cfd4bde38820d71ab29d8781
F20101221_AABLZC shen_f_Page_011.tif
fe7a7a60ebdaca12cc9b3f67b1a39a89
538746b7ea37be927e79c8ca763e2787ee00e427
112978 F20101221_AABMDU shen_f_Page_057.jpg
8a6472834f4e5199a2d1cb7498e0c5ab
76d0687598b474f096527ff772fc202e470933ed
32839 F20101221_AABLYO shen_f_Page_076.QC.jpg
48c877a5b42c3998545f0b4977e080af
6a06e65280db5d830631a16737aff27cec4a5614
106440 F20101221_AABMEJ shen_f_Page_106.jpg
bb7e903932e3307ed6dddacc34ca767c
238ebb493b14c39a66e0b6f0596f819b11afbfd2
5570 F20101221_AABLZD shen_f_Page_091.pro
7668c18daac2671efe371fa197431b9a
807e776f73ed8560c71afba9849b79ab8e491404
111200 F20101221_AABMDV shen_f_Page_058.jpg
8856553157afdb29fd853f1aee711e81
fa17ad164856dc61b8dcf947102cc1e5f5efa920
8577 F20101221_AABLYP shen_f_Page_080thm.jpg
3e02ec1e0eb67bb0a730ca20fd7ac0fe
edd77895bb7883a1ab60b57f52e4e8585d5c2ccf
97986 F20101221_AABMEK shen_f_Page_107.jpg
82a652e638abd014b80099fd64d2c72c
3a985d4bf4c4ba7cbf034837bdc92b54c72bf525
F20101221_AABLZE shen_f_Page_083.tif
0a72e54ad712509f02aefa54166c8e34
b6233ac0beebd9cc15ddfc96a4e8d791e7e4bcf1
99219 F20101221_AABMDW shen_f_Page_059.jpg
f16e4af552b4d4e971820adec6daacc9
a900b34d71901938741c6867337b4c2a5488fae2
34933 F20101221_AABLYQ shen_f_Page_056.QC.jpg
1d26d209c786a879b683670145a9ccbd
6ef3d25848898ebba893d674563824b083ae0ceb
53159 F20101221_AABMEL shen_f_Page_109.jpg
d962d5f4e296fad5d6804a8a7da29456
043ff268be246b2455e993319a70b12241135305
31452 F20101221_AABLZF shen_f_Page_122.QC.jpg
9c14befa3d87e91904110ac48c5095b3
d82fa3d56ac02f661018deed144fe04fe717610b
113095 F20101221_AABMDX shen_f_Page_069.jpg
1e6bdb53011be2bf7cd79f8beddf0f49
fc4802f648b43388024e195b0781495b84ad2a84
F20101221_AABLYR shen_f_Page_092.tif
3db092949e84b872eb85d469451a69d1
5f51f9a823571da92a572d3b3dd77d4a4596c0c7
70381 F20101221_AABMEM shen_f_Page_111.jpg
7e5caef669f10cf5d32aa34c09689a5e
ecbcef01af404b950c38b077877228997d80cda3
31715 F20101221_AABLZG shen_f_Page_052.QC.jpg
162685ca99b7df0b12b2c6e8204ca7c2
93d2ef657855a4a228649e1f9eca86a30777d15f
77152 F20101221_AABMDY shen_f_Page_072.jpg
58367960a8ef43505adf6d8c5e403041
3be9d588860ea0f32a458258b28a96e3ffcc413f
F20101221_AABMFA shen_f_Page_021.jp2
c77ee47d4f57fdaeab6fd156a22e5933
9dd1668707ff00df0b010d16ded7b6ebf5783290
88341 F20101221_AABMEN shen_f_Page_112.jpg
5c502489e6410f427a7a8dc9130e6370
0d22627ef1aad92e70492687ee6095a052f01455
66506 F20101221_AABLZH shen_f_Page_097.jpg
5a4286a62a24223e07e050a231793f2a
312335409b823a63c26cbb53aa607f5d5e5ca997
23734 F20101221_AABMDZ shen_f_Page_074.jpg
ad8d15a1bd2d2eadb4bf051d7c59c732
84d4db7c911e368027d47d2a383dc1dffd4147fe
52873 F20101221_AABLYS shen_f_Page_041.pro
cfd4281c2dd1f832457c78ffed26263f
ab81dde3c21453b9fab9cbfeab7e27ec6ef5076c
F20101221_AABMFB shen_f_Page_023.jp2
8e4fdfbf0491c957f237b2dec02509fa
26a3c2763aa3d505d40c7dcd63e308dd8fa8bfeb
69497 F20101221_AABMEO shen_f_Page_113.jpg
054779a381f7b00bbf31c73182fefd6d
8e6bab204e01c2cbeb11912630d7411cef6674b3
F20101221_AABLZI shen_f_Page_005.tif
637dd8ddc04953c29112f0c2220cdbb8
0710fd592cfe48844df5619d372a00ce3a2233f8
F20101221_AABLYT shen_f_Page_077.jp2
27dc60f8c9abd86aca10395b5170ac98
e4e671ec70caa9bd6e430502fbb91cc41a5835ff
F20101221_AABMFC shen_f_Page_024.jp2
55457e154cdde75ea34a08b042354412
4b1040f8bc0596d22a3ffe426c17b9aa1df514e1
79850 F20101221_AABMEP shen_f_Page_114.jpg
a1b6d70bf11b404fc45b9f2cbc4b9aff
1d785e079c60b60183fc8df26f03d9b86186be8b
6836 F20101221_AABLZJ shen_f_Page_088.QC.jpg
3ba1dabf849036277fd599e586bf7759
bf307e33a866d1433a90224cb2671cb5314eb565
F20101221_AABLYU shen_f_Page_102.jp2
84bf1f6223a3208e40ee47b030f85ea3
785f0e7568ae5c91d8fdc86dde2fd2b44f8e8a2d
1051982 F20101221_AABMFD shen_f_Page_026.jp2
312eee3b78772b27dd3ff3892909bb5d
b2ecd1316cad9569355147b20065ab2482c02149
56483 F20101221_AABMEQ shen_f_Page_117.jpg
e8c223c880d0713ac3fae3a1becd23f1
fc3108c97e8112b31c15c3f8310f1e3d0f37da25
112473 F20101221_AABLZK shen_f_Page_033.jpg
21a883cceb009bfcc70bf7f1d90a0b5d
14e0cf1aa16e54bd1a3f5f54f72c207d679333f1
29464 F20101221_AABLYV shen_f_Page_101.QC.jpg
21267828b6f242b356387e8096f4f50f
057b653b4fc7815ca677d9525258585054ef8db6
1051959 F20101221_AABMFE shen_f_Page_029.jp2
e28799141530330c3e00e916e7f6cb50
50dbb60df612d4401cd55fae2f094ca35d81c942
81784 F20101221_AABMER shen_f_Page_118.jpg
77f24bb826a4c2f29380e8403663f1d3
c7669032b6e62b41ae8f01ffc54de7eb657632fe
34638 F20101221_AABLZL shen_f_Page_061.QC.jpg
3c13c547fd88510703cb6b79faa04562
32c47ab78175cdc33f737a8c6e9c9d98f162d329
8436 F20101221_AABLYW shen_f_Page_012thm.jpg
75622c9f8217f7b0ac02e43823cced01
4305769c6c53bc0e3a5cf1c984de491c8561ebfc
1051886 F20101221_AABMFF shen_f_Page_030.jp2
b3f896df4b8838ea76b4fe0b4976af1f
65952b3ef76a0a65726578ad68a82eaf91006087
81251 F20101221_AABMES shen_f_Page_119.jpg
f9c14707561a0d4fafd69033bce79c95
cb42942012a571cabb2d2b108c9225c430d667f4
22137 F20101221_AABLZM shen_f_Page_002.jp2
38f7dd6169de64429305b8e5b3ec2dc7
96ab5001074ab0cd32026d6eb53472588ecd8fa8
F20101221_AABLYX shen_f_Page_037.tif
cff563fd019d5b7407b0bc3380e3a971
2ae02388a655b2f4f30029ac78a9e5fd50b1c576
1051951 F20101221_AABMFG shen_f_Page_044.jp2
905c00bf53b64f192e1fc324dc9ccac4
5cc80c23fe8422f8df4c571f0fe3066d530706ac
134589 F20101221_AABMET shen_f_Page_121.jpg
ab3c6be16de1740c8db8049e14252150
18cb9b93d2f9126a3a0a699d73176456bc08c2cd
36599 F20101221_AABLZN shen_f_Page_124.QC.jpg
923d3bb797d02dabb05947f5baa6e768
4832ea3dbc7b63a8cad13ad7b1ca18740572603d
2066 F20101221_AABLYY shen_f_Page_077.txt
6a25a3f3a800791cd83cbc2ec4c3a4dc
330d61e7530f2836d80f3affe992483addc65a80
1051981 F20101221_AABMFH shen_f_Page_045.jp2
e68847e777595161a73eea205a658d28
f6746f3187f0ba0d8b573ac8cd8f06242709f7bc
133979 F20101221_AABMEU shen_f_Page_127.jpg
6048aa567307488a60f9ec6f6aa96b99
07a1a6c4a8ef555c4ba0e2d2d4409ff1d2efeb9d
45892 F20101221_AABLZO shen_f_Page_101.pro
390ee756fb84568eaec6c6c009cae571
7621fb2dcf11d00dd1c613a82738f12998dee1e4
54458 F20101221_AABLYZ shen_f_Page_022.pro
9deb9e261ea3a1ac8e4d43721d86b389
6f26f9487539435f074765b7ad3d8db580bd0334
F20101221_AABMFI shen_f_Page_051.jp2
79dec5bd2d350afde0bcd2ad6cb1eac5
0a3d4f99fbba3bcb98ce4cd48f43a51abc9e1210
107541 F20101221_AABMEV shen_f_Page_130.jpg
72c676dce5836332813f4ef913ca09fa
f20134c2d993afde5eab1015cbc342095af848ce
F20101221_AABLZP shen_f_Page_111.tif
8db0aa795ad61356551846a6880978b4
883f7ac574560541b5dc7e1a1d1fd38a1dab5ddc
1051906 F20101221_AABMFJ shen_f_Page_054.jp2
c2b24904a69034a353f3c93dfe70b265
ce78b039e4e76c70cc4e292d4a23b6f98b537142
123233 F20101221_AABMEW shen_f_Page_131.jpg
03c22981392b022fd4763468e4fc728f
f0325e406b75df70d09a18a2c6608c0c7653a1fa
8193 F20101221_AABLZQ shen_f_Page_062thm.jpg
dd8615ad4054e0925706607373264b09
cd41af2f1b6276fca5d17ac87dc2a4bde1cc511c
F20101221_AABMFK shen_f_Page_056.jp2
e6ffad91407ac7cb0b597ff8ad6ecd36
574493281f261aaccf5724cd7f473d95cea042c8
131830 F20101221_AABMEX shen_f_Page_132.jpg
8cbe1480619c4aa0a8891e0e59f299a6
0f467c5443f54803000680c1d2c99245b119fc62
F20101221_AABLZR shen_f_Page_127.tif
048ef810d4ef86667d60af895c8a1568
36838643cfc162c87ac86d8081d94f24aa7c2540
1051939 F20101221_AABMFL shen_f_Page_060.jp2
d457890739ab309ffa836af2f52c42d2
710d742b2fa4a4bd4a5ad95b16e51b41d506464d
404413 F20101221_AABMEY shen_f_Page_011.jp2
4f7dd418550dc5ca1f3a186e5eec6fc2
66f87c03d9fb9c7f91f02d75cc6ae676be4b3b7e
54856 F20101221_AABLZS shen_f_Page_108.pro
7e9179db2ac0744ea9dd8992dd87cbee
b1eb1fdd355c6183f5293b0325c9b456970fac28
F20101221_AABMGA shen_f_Page_105.jp2
f988b49733b8d41a14c37d064d71a77d
8c01e1c4c66a33c28fa463d93c8faac8879cb2f6
819140 F20101221_AABMFM shen_f_Page_072.jp2
69f92c5e469e008f1a8cf8ff14efe4ef
b967ad2bc76670d138d12ca37dc4ef2374d7ed17
F20101221_AABMEZ shen_f_Page_017.jp2
10aeeb2458c6d00c3d0fa83a0ef7f71f
8bfa00a7695367c84cc55c4b7cb37148a7a40ce4
F20101221_AABMGB shen_f_Page_106.jp2
bdbecd8e71a7e474cb49e7e4f08da208
e7b497a9d739a1d75885efe7743ee0d668edf67f
95485 F20101221_AABMFN shen_f_Page_075.jp2
5b4291a6369fd25ec81144b8f5946dad
3e3e6cfedf761baa331484e0dd006c69e7da23bb
1234 F20101221_AABLZT shen_f_Page_112.txt
b0f1a7cfe73578313ecec6ec21f7e0d1
ddf167f6595aa21edc12cd854535547edbb09798
1051910 F20101221_AABMGC shen_f_Page_107.jp2
e64272a22d6c0db0d17f416fd3c59364
55dc42c146ce8263d02db5b755cbfd6b035eb891
F20101221_AABMFO shen_f_Page_076.jp2
c6637a39ad3f3c53c303241df1b19eeb
c43d431e3f271f08761f845c8089ca514581f35f
1017807 F20101221_AABLZU shen_f_Page_068.jp2
0a3fd37443ed4bb36692d79c6c14c1de
aca22c51b7d4d77db0a481fea2079c41e04402cf
335095 F20101221_AABMGD shen_f_Page_110.jp2
a2d95a29264d896418f402a1a9986c16
92c566f338e23d764028b1f22326ee87364ab04b
F20101221_AABMFP shen_f_Page_080.jp2
5827daaddfa309d1ab8b7dfbcc2ee12d
e70a52e5f0f58c3a3844aa785e2d38116677e70a
1613 F20101221_AABLZV shen_f_Page_009.txt
7646cecbb1dd43b401e7b65d9a19f4a9
54a91caed83b9ba006c37b6ed8226117e199ddd2
662786 F20101221_AABMGE shen_f_Page_111.jp2
5335de6aa6e82bdd486b8d675f13a981
6185b4777edccb14e1f6eed3b91a586a114288c1
187467 F20101221_AABMFQ shen_f_Page_088.jp2
17f55003f6a0f7554745564b8bbf4621
8a4228463cdf2559667a31763c9e29e3487c613b
49855 F20101221_AABLZW shen_f_Page_036.pro
2fb31a0c1cef2733a27620d24b468141
19be0f3d541905abe8e38717a0eda407a297adbf
847635 F20101221_AABMGF shen_f_Page_112.jp2
5c6de5e51d144d3632b20bedb493f6c4
df598c23ce07f4a32770a9af976ba3102c885723
465255 F20101221_AABMFR shen_f_Page_090.jp2
9b2658df886cf762a52ef9772ce79127
8519d376baffc53d27171ff2262db4702be368f2
108199 F20101221_AABLZX shen_f_Page_041.jpg
3c99d9cb1b0f3e8e66f9173f1eb7cb35
65168dc95b38acf084813b0544e9be1283a70457
618489 F20101221_AABMGG shen_f_Page_116.jp2
f2de32f66064fc0cbd23c19098acde8c
f699e351a0d6765e90f94a9781a4b5c00257a720
354048 F20101221_AABMFS shen_f_Page_092.jp2
503a94a5a42ac80f208ce1818e133335
fd1b6bdf5b262eeb4561b02e8497bc06ab0025aa
F20101221_AABLZY shen_f_Page_081.tif
f83821c9790bbaa89c49722c6352eabe
9afe45712d02647d52e04213cc30c86dcd38fca2
588896 F20101221_AABMGH shen_f_Page_117.jp2
b0c70b4e78e27611e08b3115cfdd1b2a
f30b52ce0671fbb69576b12cdf4b042143b85537
721563 F20101221_AABMFT shen_f_Page_096.jp2
6c1ed3837847bf19f317d63538de3794
1cf11a13e04af1b759e83f1a2ff0bca042baa8a5
1062 F20101221_AABLZZ shen_f_Page_111.txt
09586d59686e22c1e82978e95ae04ee8
a7dfc5bca61b1b9a5eac1402895ffb9e85499ab2
809003 F20101221_AABMGI shen_f_Page_118.jp2
3eb83c9ebc8df24ad612d32d0a5b67e8
7330e26edcee097ed5970d59395c3faaebf7ed0e
638787 F20101221_AABMFU shen_f_Page_097.jp2
6fadd437e367536b2ea47153b29924c4
6fc281c3c255351800ebed6d024e543bf253f471
1051986 F20101221_AABMGJ shen_f_Page_120.jp2
12265481bd1fbfbd5847f47f30942ff6
7b6babe6802b7d6405b3defd9b6681eda252070b
644941 F20101221_AABMFV shen_f_Page_098.jp2
abe7693d91f2a86de2bb78f3a35285ca
786cf46246e96ac436b6044247c2fd246e10614b
F20101221_AABMGK shen_f_Page_123.jp2
3294896a4b403c605e8880fd910876bd
779bb00f9c4d571a871faab3434d37a0f7a62c37
257576 F20101221_AABMFW shen_f_Page_099.jp2
2017daa4d97ee8de3cf1c28bd8971303
c92f740dcbca05316a86df4770767e8dc091fb8c
1051960 F20101221_AABMGL shen_f_Page_131.jp2
25fff6ccee711c345f43d5d9b814d9f7
641505111251cc046d38edb11022ea05f2f89361
F20101221_AABMFX shen_f_Page_100.jp2
0f2e1d64945c5cd8f398a1c4d7cd615e
91fb110358c0c838f95ccbb04682ec6bb2488b05
F20101221_AABMHA shen_f_Page_048.tif
25caa850445d0f5fda0c04e29c7cd70b
ccdcd391d0070b1c0eb538b95c4d1b3728c1eff1
542889 F20101221_AABMGM shen_f_Page_134.jp2
c198e996b73a61398f22dbee6198f516
16745fc5b674f6733c8d10a6a58ccf5cbcd4c485
1024123 F20101221_AABMFY shen_f_Page_101.jp2
0b2748c146747f2c8b591d3daafbc3e1
f7482ed84d9778e7c50aad6e975a3a5b19e19854
F20101221_AABMHB shen_f_Page_049.tif
eec6f40789e36db090b9d7617c7837a8
fed8dcc1adb080b28ac6bf2e44d049d84791be16
F20101221_AABMGN shen_f_Page_002.tif
62deac74ccb4360b1679eb6d1bbb9e67
0c02b95f3f43179a8f4bd2928bf2a790ad92581e
1051940 F20101221_AABMFZ shen_f_Page_104.jp2
e5b99e5c82068a6a2e1df22ca1040793
484d6cf5f5c2c17fa4f1d7cc6761be00eb702569
F20101221_AABMHC shen_f_Page_050.tif
44fffd43dd23ffd765784bbdb41b5001
7d4dea3d7b2c5af30c622da279921706fc3fd3b1
F20101221_AABMGO shen_f_Page_003.tif
f5e1ebf12d551c2e4959e2124ca81d9d
9e621f62e76a156b42e65984ec04bc7414e3610d
F20101221_AABMHD shen_f_Page_052.tif
f7c0a4490c13ecb4b4137227f282ebe6
cd1628ae0ab6f86138c191236359584f8a342795
F20101221_AABMGP shen_f_Page_004.tif
da4e9e48d1db8df77f91344c871f1278
584f84647a0ca5b03d864744658b692986607c9c
F20101221_AABMHE shen_f_Page_060.tif
04f92e9e7bdaf3e3b8f838222de32675
7f5bfe4beb8749cc860968b699e910725de67d5f
F20101221_AABMGQ shen_f_Page_012.tif
2384910748e6175518df7bd4018d612d
5a13ddd9402f9720530a2c42e34902f04b45e25e
F20101221_AABMHF shen_f_Page_063.tif
aa4feca33230e3a7a9c18a49ea58d1f4
abb4f43f7de3b7dfb6e363f55364e34d5deb0bc4
F20101221_AABMGR shen_f_Page_015.tif
95d87a84fe303b1c0bb35988021f43d2
4ffa894ef92a634f8a3506cbb7e12100309fec7a
F20101221_AABMHG shen_f_Page_066.tif
78f57655b70a68eac141a6c57054502d
75143fcb85842b3156ac61350628fe966d706839
F20101221_AABMGS shen_f_Page_016.tif
83a3eb1498a12dcd729198aec4f5a612
424d7d032f09ba4fcdab1a5c41613e5289932fb4
F20101221_AABMHH shen_f_Page_070.tif
707531742a06618b3d35666377c671b6
041596fb34427a5df1c6f3184a1bc2fd17771b68
F20101221_AABMGT shen_f_Page_026.tif
ae7a1d7f1341cfedb6b190e395a246b6
80522aff5b4478b34467203dc181b41189797bfc
F20101221_AABMHI shen_f_Page_073.tif
72d104c082a2d56e4226d89883d70aba
3f0dbf6d2bd8f6ef2bd9561f3a58e020421889bd
F20101221_AABMGU shen_f_Page_029.tif
d56454d6e6f8d775b682e1a2a14675e3
98f2cf0cb0fc6e32a8a5cd6be0c9eda7e61f4e59
F20101221_AABMHJ shen_f_Page_074.tif
fa6de8514c829818cc73bbe38a0d1231
29620dc9e7226bc11c42d897b772a367c1e59d13
F20101221_AABMGV shen_f_Page_032.tif
0a9477e03284b65c809d0ef71f9e1c5b
64c6434407c4e93c8c6e250c0343f041f25c7d41
F20101221_AABMHK shen_f_Page_075.tif
3adcdcd358e0d0cfa1d64da783eb7f6f
bd09794df11c6c29a03db1f5427ae1661c203920
F20101221_AABMGW shen_f_Page_038.tif
e41c9fe8a4aafaa2670e31a2f5c0da29
85fe6bb144edddc64336cbf59afe9ab6469bd73a
F20101221_AABMHL shen_f_Page_076.tif
f71cc3508e8ac0829300996f835aaade
2ee1276ecfbd627fbd091bb3351670246a5c5043
F20101221_AABMGX shen_f_Page_042.tif
0f36f9b674883a76c3647573bd0ca9ed
d286937e9c701d751f62d079f0816874f7c9f23f
F20101221_AABMIA shen_f_Page_110.tif
e41a6b67db358e5efa81cb4197908a53
768763cbae9ba3a932448f5269a31cf4846fd44d
F20101221_AABMHM shen_f_Page_078.tif
456e8926b314918c8f36604f30f68e0c
cf79369f830cd0b964642c1ae255e5d91c554691
F20101221_AABMGY shen_f_Page_045.tif
f1bc7e9b83758e7629e29e4eade6a161
b37850634d91f6c039eec6fbe8569ebc5d07902e
F20101221_AABMHN shen_f_Page_079.tif
9bee96e4836f2e145ecc5f101fe68c77
1b8e584cf2eed34b3f19a1b29f592263f4276164
F20101221_AABMGZ shen_f_Page_046.tif
ff74bc55b827d37060fe1bf52b788446
0648b0db575bf4b8d2890cc4838855f78098aedf
F20101221_AABMIB shen_f_Page_115.tif
8e3e2359c8dd11cf3d354a7bda3a03e8
a644a18d2210a988dbbd330dd0211608acca170d
F20101221_AABMIC shen_f_Page_117.tif
21a0f0851b7541fbfa524ff895b88f93
99c7afda9881ea6889ae8c12ede22d05e618e60e
F20101221_AABMHO shen_f_Page_080.tif
702976427ac6f6f66e73e414c1b71298
a2e8bb91bf00ed84ec2b5141ceeb787bad042d07
F20101221_AABMID shen_f_Page_124.tif
007036be4c6967a2a6df0d077c89b6ea
82bd45ffc47705681aae7721843287c8439c2aad
F20101221_AABMHP shen_f_Page_082.tif
1fff5fb649781ed40790d91adb718ebe
771ac65f5d36b6da2c820fac7f77d220540d9df0
F20101221_AABMIE shen_f_Page_125.tif
6919770036461af43e4fdf047b20511a
0e35d402056ba2673b3dfdac8eae8d187e2d6274
F20101221_AABMHQ shen_f_Page_088.tif
bd7f3212e1274b36a7bfd3e1a932afbc
a78d159d00d10e2ab6b642f5235e902c7e8cdc07
F20101221_AABMIF shen_f_Page_128.tif
56cc9ff3079d46abc6067325692e7f28
5b7e6efc09401cae2a0703b0e28f1f065901f5b4
F20101221_AABMHR shen_f_Page_089.tif
b83f64b5f39231cd6072a2a6e632929a
dc02e8a7e62048870f8f02dda2c3c9f4d75bd2c0
F20101221_AABMIG shen_f_Page_131.tif
2ef783198214a9e5535eb2cf69962309
e6ab4da0fc5755a5616be9f83aeaecbced1ee396
F20101221_AABMHS shen_f_Page_091.tif
6c88f9da8d63b58d842208f66ad0421e
3f869b2b31bcda901633a5d3e464b646df662df6
F20101221_AABMIH shen_f_Page_134.tif
3f6bd4efddba1bdde1f0bc8118579abc
9ca4935a754c8304ce2a35bd19ed8a9d4eb68efc
F20101221_AABMHT shen_f_Page_094.tif
fd67d3babd354037afe20b141f24caf6
a6b6f1c357956bc88d624becb10b6201451ef3f1
8786 F20101221_AABMII shen_f_Page_001.pro
927a8fbeb8ca1596dd0786b480ae74c1
496cb0fabeb387e5600da9b47c528753cfb27d66
F20101221_AABMHU shen_f_Page_097.tif
0f15b59501cde6d5c9db36c4bde40eb6
a2d87a3584dc988351b8437be4270be37faeb91f
594 F20101221_AABMIJ shen_f_Page_003.pro
454d4c99a652a0399810a8b319dd6e6f
989f79a16716cb9ee1054b0d3393b2cd9bf7d6d4
F20101221_AABMHV shen_f_Page_103.tif
77eb3a50b9d2021a128ad9397dab5b95
3a293fc03ee58ab3b179f3761c3d614d2e95bcff
32996 F20101221_AABMIK shen_f_Page_004.pro
58bf9cee8f8aad2eceab4a6d8b271096
69fa421615a78839bfc6884ecf57930af0b95fd2
F20101221_AABMHW shen_f_Page_104.tif
a65fa99ad4dc94fb7625a2c579bbf4bb
e43e37239168d87de9773c0ad738c6dc7a079963
49985 F20101221_AABMJA shen_f_Page_047.pro
70ccfcf776d48c29e02ad70f02c8e013
51286415c8ab796c14f0160165c8624639f52bff
81535 F20101221_AABMIL shen_f_Page_005.pro
b53917cea32dcb0a59889695b27c5994
be4de8dad22e8b5e33ee64e511a9b22bd27e7d2e
F20101221_AABMHX shen_f_Page_106.tif
f2c376596923ee86e675bc7f5e3c564b
8d7814e7deaf8ef1eef76b4992f8909fa977f012
51105 F20101221_AABMJB shen_f_Page_048.pro
4805fbd875fe2d333be7c571d56e3ca2
41c8182ec911992c8f5c8984f71cb5741278887a
39262 F20101221_AABMIM shen_f_Page_009.pro
d6da79d9bc901060d5341bd71169c76c
8e846f632038dbac3eb367d87f96ffd5c16bf679
F20101221_AABMHY shen_f_Page_107.tif
04a7460b29be9e7e34feae82f224583f
832a00748c8b81ff73d804f7bdb910de29451bb2
47501 F20101221_AABMIN shen_f_Page_010.pro
1cee554a5c28ab7a634b211651ee87b7
9b947811325bbae71a71b2921397ee44a63e7d02
F20101221_AABMHZ shen_f_Page_108.tif
3b0cb2a7e57c8840c89c474563f7545d
bad82c3daca0d379684be40fb3b82028ee9164de
2140 F20101221_AABLGA shen_f_Page_007thm.jpg
9c73526e6692f9ba3fdfd852967ecad3
b196a88c4aa4586c29ddc15fc439576ef96a3568
52603 F20101221_AABMJC shen_f_Page_050.pro
1e26f3c7e7668dc7d91b3fc58c513bdc
cdeddf3bdf7cd113630c58364b66e6181d61f9dc
32900 F20101221_AABLFM shen_f_Page_036.QC.jpg
3ea4f5cf10ec4b64e9678232038534ca
4ee7952bf6a26158f0555132b51e8d24050e902d
17651 F20101221_AABMIO shen_f_Page_011.pro
3ede9be2770e8cab98fd3f07756e2f63
df9f08487df20876db5ecb650b28033c87a9fcec
53166 F20101221_AABLGB shen_f_Page_042.pro
f03403efcb8b1c412371bc736869edc7
b316fe103a2f4b1c208de0d8e606161d52f353d2
47367 F20101221_AABMJD shen_f_Page_052.pro
11692e9d8de9346648e3fd82dde07ca7
3ea32874cc4eb646f3aca71c7a226bd7bbce6b89
F20101221_AABLFN shen_f_Page_079.jp2
56f83ec062a072ec168d595c28cd2c5a
b876e270ea7565c124da45f9ad68d7bcca82f2f7
52244 F20101221_AABMIP shen_f_Page_012.pro
ac43d1eccb550f9a1391ca81cb37580a
48740f2100c2bd5cc1c23b70ac2632d7b38ec967
F20101221_AABLGC shen_f_Page_102.tif
2978c3f8b775aa1eada373da447bd46e
e8bc93b821371cb20631b9713ffc9880775c5687
51587 F20101221_AABMJE shen_f_Page_053.pro
2a36b6ce3d4aa7986eacca8a7fa5d205
5bb495f97782227e5c5f4ac553e050f745574c2e
3932 F20101221_AABLFO shen_f_Page_075.QC.jpg
8d32c19939a233c1a63f9f48e8636113
d6314fc6709dc5ff3385d5856987d3738135540d
51307 F20101221_AABMIQ shen_f_Page_015.pro
1b226cdcc7c5dd1bf16196cd9c834004
9ffd8e07010dc1889b5fe2fc444995feabeb2a4c
5032 F20101221_AABLGD shen_f_Page_094thm.jpg
7bd8a5b4614cc548ef8ef53a31864e98
dbd71e21ca21c78720cff759c23a211e44a798ca
55691 F20101221_AABMJF shen_f_Page_055.pro
7c9c7c55ea633ea6a63f9440bf35f4f3
8303c381923fdc8f20827dec4fe7bbf737c50927
F20101221_AABLFP shen_f_Page_125.jp2
c9dce81fa47dc64a70e6065ef3356ec9
63c1950525c2d0af293404d74ccbbc4f20e2e0cc
58575 F20101221_AABMIR shen_f_Page_016.pro
82535e8b4b01cb2fe3f606239070fb4d
7a5b318c6ad19636613b31ccd85f364b483b6684
8504 F20101221_AABLGE shen_f_Page_043thm.jpg
ca22c636e65df5c97dd0b3b8563fe839
3c8454039a7b1cec3ab1451b09fdb0fd6c2c234b
55383 F20101221_AABMJG shen_f_Page_057.pro
7db351b3703e871e38e580635c33afec
f393cb01cff9c8fb3ee6eed749b4fa954c95e8d3
1051984 F20101221_AABLFQ shen_f_Page_032.jp2
50d22977534ff9b59e356f0a977ba6d2
e14e9b95f25121463f53bbee6c000ac56e6fb342
53631 F20101221_AABMIS shen_f_Page_017.pro
92ef6b704d13f9ad3baafa6f3b84aa98
4c4a5c7e684265fb9f6f86b56c558ddded8646bf
35064 F20101221_AABLGF shen_f_Page_077.QC.jpg
f02c3f5265b027e935d92d44739ef27a
21fc45239edca0014a297487f3042dce795ac2ad
55040 F20101221_AABMJH shen_f_Page_058.pro
425f646650db9b7fe83f22c92c92a7ce
282c6cc89454bb414eaf00164967976fe0e30559
17011 F20101221_AABLFR shen_f_Page_134.QC.jpg
d4e9f63899285c43aff2c3a680645897
f4b0009309b0a3151804b5cb60daf2ed6d4fbc62
54215 F20101221_AABMIT shen_f_Page_020.pro
22c25f8875266fe88c25f4509465c3c7
a238db3a0259ebbb6134d809eab4bcd695b1fed7
2107 F20101221_AABLGG shen_f_Page_017.txt
1661979174af395c4363694b2bb26870
bc8c3bd4d811f8e018ab76b03411b14d88b66062
48716 F20101221_AABMJI shen_f_Page_059.pro
f37bdf1e6440c0ba8d3b1554d553ffec
84c927bcc29f49e7eef5d8d70eef4db345fa3305
71213 F20101221_AABLFS shen_f_Page_004.jpg
edf4026fb1712f90a1dcdd6a26ef602f
ab659146669456e0cc85499ffcb23539a7338b79
54202 F20101221_AABMIU shen_f_Page_021.pro
afd0badf50225ca434108045e3b5cbc4
686a11f78eb42302ba042dbac4aaf1389f4340c0
93526 F20101221_AABLGH shen_f_Page_101.jpg
462bc6d24473977f7b9d4b4dbcebef7e
ee8d9fe4dc0d10c200e9003e894e4eda01d5e489
61029 F20101221_AABMJJ shen_f_Page_063.pro
93427e1befad1bc6f5237365417f7730
591baaa48f3d9d0bc774ab9aa6ee25b4fd294192
109444 F20101221_AABLFT shen_f_Page_020.jpg
ee78c06d5180e5d23e8d72a8c422d03f
bbb168e311d000669e4d6ade1888ef11b690f91a
55130 F20101221_AABMIV shen_f_Page_025.pro
18487d33530a3e57301ad9d0156498f3
6cdcc8959cde64e1e95eaaf6edfd7e94da6551ad
2102 F20101221_AABLGI shen_f_Page_078.txt
e9b8883d6a82917b15a1ed4f0406745e
a2c6af3c535b1ed4e24cf62e767f75583f52f532
47969 F20101221_AABMJK shen_f_Page_067.pro
52f04d431bcfffa810e5f05b7b51b707
8be97c85785ab309c5f3b5cf02949807cc9d311b
F20101221_AABLFU shen_f_Page_092thm.jpg
c7b363adcf00ffed4c9ad8567374f41e
24ec08a52040de7a4b2338c5eadd1e17991d28f5
51285 F20101221_AABMIW shen_f_Page_029.pro
dc09e1b23015355c2f4ed70115119a83
e7626de008e481bbe58b3a78551cb64e3cdbeb79
52342 F20101221_AABMKA shen_f_Page_122.pro
069097f89e307327b9fba53949e108a7
d25e7abb12e55425ef1aa6f9a3831e2a92c9adbd
1031223 F20101221_AABLGJ shen_f_Page_087.jp2
9693d5f2b0ced99b1e59d821460db828
9a6eeba6b9911159a426e414f5d24d1aa29b9d05
50227 F20101221_AABMJL shen_f_Page_070.pro
87cee312ebd77fc5e09a2f7f52916b77
90e1e6ed52332d7808e2102bb2222ff62f44049e
F20101221_AABLFV shen_f_Page_016.jp2
dd22b1e7cf47af5809df2e5875e0596d
7437754f15e0c915eb1154da92c0a638486110e6
56761 F20101221_AABMIX shen_f_Page_033.pro
b0183d06acbdea4a2620c99245ddf0d9
dffe06abb804666b65d1dc157b458428dd20f671
64000 F20101221_AABMKB shen_f_Page_124.pro
ff9ac6ffb9a31a4279d3880d93e471e0
e01db8bc4de962e65cd0cd23a35bd7e568a7e00c
35977 F20101221_AABLGK shen_f_Page_054.QC.jpg
9f5e1e0704a521b559e628a1edc5960c
9c154923adf4316369d995e099da49582c9d1bd2
48693 F20101221_AABMJM shen_f_Page_071.pro
4d825478cba6052c788e235fc944f32b
09d581641eb769997fd0494e2cbc151807ef0c2c
53633 F20101221_AABLFW shen_f_Page_034.pro
d6f14af45a22830b151eba16f7f217e9
d3fb77af38a0d240c97f663eacddd2073995984d
54419 F20101221_AABMIY shen_f_Page_040.pro
73c5baf826079d55244b33807b0c2dd4
ada18c2ccf8cdae3aea1fd6762f1a954a9a8525a
67194 F20101221_AABMKC shen_f_Page_128.pro
9305533001575df9d04623c20b84e703
00935a83e6d47ab1a309f2ab4447f73fc492e9f2
110964 F20101221_AABLGL shen_f_Page_045.jpg
97d5434a84f76b069897a7b7e5321086
288ae095e05367076aaebc21c2eb548112fa5014
35363 F20101221_AABMJN shen_f_Page_072.pro
297414f6babbb7b6c2061a075dcad05e
b710552917b937a1cc3905d4f0fb06368b0f476c
F20101221_AABLFX shen_f_Page_068.tif
1d0423bd61d42266cce3b2be2ee802b9
87d570fb37f776a25d3e361377a63653aca1c119
53857 F20101221_AABMIZ shen_f_Page_043.pro
69fe65d01ff80940f17d6172182dea70
bfab012731d07b784ca598718428c1451083835a
8581 F20101221_AABLHA shen_f_Page_064thm.jpg
684c7a13e0c90d96660a8a15e0c72fc1
e0e1c711dc1438a4c71c7647fc9204a2434c7f06
1331 F20101221_AABLGM shen_f_Page_096.txt
bedf381bdb42d38c049838a1210f26e0
60271ab074f04498fa21b996e6e869f11faed94d
47990 F20101221_AABMJO shen_f_Page_076.pro
95567ceba20b6ed30c9ccdce66f40b2d
3e1d529539010dca909c46f23c3e8fa89e6f737b
1051967 F20101221_AABLFY shen_f_Page_052.jp2
8bf3511540b2e0f6c636ba9b29534ecc
c1b2e201c38e115031c2652d1e6ecdd1aa58f922
57959 F20101221_AABMKD shen_f_Page_129.pro
ace7ec93ce486ac43f01dd6748dbe78d
0370552c3843c8bd3f67735a22ffd6b6d8af97ef
110897 F20101221_AABLGN shen_f_Page_046.jpg
aa86eb622256f287d69b5286df0948fb
fae89f8af201e6af01450a85aa8df40722767554
54807 F20101221_AABMJP shen_f_Page_080.pro
b77ab2945251acf9f8152937be6a52db
e89715cfacc14bbc0ac8d6a882123371012e9ed6
55698 F20101221_AABLFZ shen_f_Page_045.pro
c497569e05d9e091a3ed3a9a860d8097
81011e0a00ceb68169c07b136881a7fb7f53e7b9
38144 F20101221_AABLHB shen_f_Page_032.QC.jpg
e91c94b8ddecf9a63955bc3c5f701bc3
400379f8fb8fd277b41a51fc4e8c3ab26a835f4d
63235 F20101221_AABMKE shen_f_Page_132.pro
d809055788f2d7198188a6bca811ad09
0c691de8c1777e00cb70d72fe59ff3697c237938
856319 F20101221_AABLGO shen_f_Page_115.jp2
5c6639e30968be87e221faf0aa5ef55d
39b77fe04539cee067edacc8388a412405b4fd9f
57472 F20101221_AABMJQ shen_f_Page_083.pro
ebef3d1c4a1c4bc390f3dca6cfd1b2cc
02ac7b8d744f579e3d2705a6ca71bee1c865fc4d
14999 F20101221_AABLHC shen_f_Page_094.QC.jpg
503916947876c39889b0f1c08310ee11
6f04ed31cc934fcd6f25deed3deeb3648b9262f9
9115 F20101221_AABMKF shen_f_Page_133.pro
d17aef27358f39acc4a1d010eb1ba8c9
11bda1cfdc4335d24b57c6f19b84e9102fae2f2b
F20101221_AABLGP shen_f_Page_121.jp2
501c45f02fbe997a2b2f6ae85632aefb
49c2e415fdd35338b19268e929a41902fc054990
10114 F20101221_AABMJR shen_f_Page_089.pro
74f046b1f490f59ac4572c99a90482bf
f83acbb4611f7436b474103307cccf9e3b9e0d05
53024 F20101221_AABLHD shen_f_Page_078.pro
8a23abc1c963f618208ed46bd2093fce
f28c5d5e56dc2cba99863f062300de143f1d8702
23442 F20101221_AABMKG shen_f_Page_134.pro
18306feb4310d5a35c452d65777c23ac
1840354c78a547f4f9150e75473b235a22cbe577
6994 F20101221_AABMJS shen_f_Page_092.pro
293fd14ab6b4fb025e9ae10d208384f9
232385d5cd873a8c4d3278294a687c1f748e78ea
2027 F20101221_AABLGQ shen_f_Page_029.txt
a2970775fe884a66cb31fa5f54e06efa
8344f12b5d51ac213a90aa2d595a29094f301a69
52544 F20101221_AABLHE shen_f_Page_085.pro
825d9f46213c45f0a1bd5ae9905f5444
46ab4f6a26661aadbacdb26089720598c79a7ea2
79 F20101221_AABMKH shen_f_Page_003.txt
0e9bf642f69272b5a6a7f86e47849dc5
ecb875f5bdf5e8be5f63ea52b9396ed67f8b7c1a
47213 F20101221_AABMJT shen_f_Page_100.pro
4ae52d3c5d95128fe813a97ebe413a2f
71a676648b82b965c9e7a6ee11b3821fa09ef98a
32409 F20101221_AABLGR shen_f_Page_130.QC.jpg
2881094059238675b0f57dc7d5ff9f05
636c30a9c3b22e1d3e33e070316e15a36fc2db30
43994 F20101221_AABLHF shen_f_Page_089.jpg
1f4f278d38e5c1b2ec01644e46607aad
7e9f4bbc022b26c024be079699ae22a84987b079
1364 F20101221_AABMKI shen_f_Page_004.txt
85f44e7a5a27c9a0b6eb7c2c9a3588b2
e16b834c1e1f38b33c81aa6f0b8282bf629db2f9
49093 F20101221_AABMJU shen_f_Page_102.pro
87a49313f28a2e7d2f9b8ec87a0acb51
0948b01bc29e11fafb3aa8be8e313fca9427bee7
65005 F20101221_AABLGS shen_f_Page_125.pro
9f2894e9bb1681e6589bf376b260a1c9
01098f115a02788fab22ab4cba6ca026c9c16c54
F20101221_AABLHG shen_f_Page_013.jp2
27bf0d83dffd534afc6c262684c65166
2ece793933beb32d4573dc811eba3f18494db16d
3655 F20101221_AABMKJ shen_f_Page_005.txt
06126814b47658ae895995c641d03bc9
f830f8ef987ee25f1d4da1b2e66a2ee2965e684a
52844 F20101221_AABMJV shen_f_Page_106.pro
6976fb65577a1a6c08b9cee93e703d08
10d44256d04286480606a107170b0482a796bfa5
2171 F20101221_AABLGT shen_f_Page_025.txt
e4ff0d9691249d8101ede23a8ad56f49
33c81a61b38858e2282d8568bea0b48a176b2d75
F20101221_AABLHH shen_f_Page_122.jp2
3b144a71354d8fdae8974b364074e9c0
ae0034a4da9d0c77cf145a15064009c2d3bc5305
4035 F20101221_AABMKK shen_f_Page_006.txt
77da7bec2eef225c99b71750567f5be7
27137a22d94f6056ed960a72e77ab0b96e422d5d
23561 F20101221_AABMJW shen_f_Page_111.pro
4c9a56e72e8d5d91af034105e781f150
b6c84540252907ed503fccd9bd63cae8857b7a9a
53254 F20101221_AABLGU shen_f_Page_061.pro
6bfa40da3bb846b24a9772b7e7ebfeef
0ee58d38201aa062efb1f87b48bce9dd6b65f6e4
24444 F20101221_AABLHI shen_f_Page_117.pro
2cbac7d912f658d8435cdae7306d0d76
4cc49beb43f41a3d2bbf8931469ce897b365c3b8
2193 F20101221_AABMLA shen_f_Page_051.txt
812038d52588aaebcffe8a59b6e859da
646b884d7ff07da3e76dfa8ea09da9414f30545b
585 F20101221_AABMKL shen_f_Page_007.txt
068ea6522f8563a834ffda3c3b8b2ceb
2018acec68a5237abcbf1452e12e4af29bbb153a
32252 F20101221_AABMJX shen_f_Page_113.pro
0850f7a766c5470b39b8e21d829a2d2b
c057ffd52ac5353efe45e14c102dc2a2093a3ae4
1925 F20101221_AABLGV shen_f_Page_059.txt
c5f8832895922c610fbb5807ae58c578
4cd27135d4b08bea96c3b14edcca02a57a118f20
9031 F20101221_AABLHJ shen_f_Page_013thm.jpg
0f23c357b61c67ed63e058ba68b2a497
d33f07dac6554f8612f6c557cc1f383095f48c60
2149 F20101221_AABMLB shen_f_Page_054.txt
9883b343b2910ccdcc59dba6ae5a0dbe
9bf94f07d2c05a0448f9bac2fefe55e8144e72b8
2041 F20101221_AABMKM shen_f_Page_010.txt
fa0cf8d1a216349d3c3d339e2a499ac6
2179777e48b400fc5fe815d793aa889d32917fad
35915 F20101221_AABMJY shen_f_Page_119.pro
71c82ba9552d1984120d77c3b36ce6de
5ae0a50b4dab728a1798fad6b45b0c8650fd337e
55522 F20101221_AABLGW shen_f_Page_049.pro
359728c3da74f47d11c8f1d6915dcd90
256cc177b53edfe5d22bbd311cd97fb82c403017
1051968 F20101221_AABLHK shen_f_Page_027.jp2
4b743708b23761f30772f155ca0fe5e5
282316f6e026bc5e81b2e87668612ab183caf6e4
2183 F20101221_AABMLC shen_f_Page_055.txt
735c814f1a656621dff39e8220bd0ef2
ffb443ea3b07b5367b555e012ee1dd03b012cd5f
2114 F20101221_AABMKN shen_f_Page_018.txt
f04053e7ac2a98c21efc94c6457f256a
b501a05edf912c3c9f3038c6732c4dfcf3802ab5
64060 F20101221_AABMJZ shen_f_Page_121.pro
597ac26e0f911e41b4b5bbf7639a36ca
3389e467b0c84ccb5ef95fd3163836c58ec21ba7
27103 F20101221_AABLGX shen_f_Page_099.jpg
92b937d10397714b6029094b24ad7f7d
732ee07eb898cc4a23e0a50491ee0af2fc309595
F20101221_AABLIA shen_f_Page_095.tif
a1b19a2fca18ef1b214314c6958a412e
977a5df60e5841d83ea81ede299a11162028824b
53375 F20101221_AABLHL shen_f_Page_060.pro
d12405c7d7886efc8c72e322aa98dbbd
f085c4675ad49707dd1225951e54ae3dbed88b65
2079 F20101221_AABMLD shen_f_Page_056.txt
453c03e0eae08b33cb6cc83e91051b2d
5b0bd72b2e2d549b9a1fe022f1c29c241146d58e
2515 F20101221_AABMKO shen_f_Page_019.txt
d327da2ba37398d46ef3d0d1cea821d8
647a2b5e087cfc9e488bfeeab74ebbaf4396992b
112912 F20101221_AABLGY shen_f_Page_022.jpg
946156973f2ea08334006727f697edff
414a952a7dd420a68c42275688235fe30f92bb7f
30706 F20101221_AABLIB shen_f_Page_112.pro
058672ff598b806e0db131864e62c36a
64ae418db10b2f5a8eed2dfef8a1f95a7cd08e83
F20101221_AABLHM shen_f_Page_100.tif
027b5ba7037b239effda8e7ed03b99c7
3c138b37249ca96837434cef4bfad8255a6c0b7c
2132 F20101221_AABMKP shen_f_Page_021.txt
30fc7f6bb3d1a2cf4b639e7632572f2a
d9701b554ea040363f0895f9c053101fae46e280
F20101221_AABLGZ shen_f_Page_069.jp2
ebf08cbfb1def62631f0d9631bec26c0
0b7c6113ca8d03a23ca3f3d10d62e1e74f860310
98884 F20101221_AABLHN shen_f_Page_065.jpg
20fef34b4573d01d826523507f9c149f
973f1a7358ae0a3aa5f470b038d21a0225e2ee66
2024 F20101221_AABMLE shen_f_Page_062.txt
102cf6bf3507c878b24edd89e1fdeae1
40f2df9513662d64d6602ed91f1b0c732284a21a
F20101221_AABMKQ shen_f_Page_027.txt
ffbfe02562e1d86b0dc1f842977d3673
2d9d33a4a65ebeb5e922a83ecca364e421f7369d
107937 F20101221_AABLIC shen_f_Page_078.jpg
82bbc1767c2807a5437ffd53d31d827b
3ee49ee82758a64e61bea5796d42f1e5a7e79d86
8717 F20101221_AABLHO shen_f_Page_084thm.jpg
927cca9f3789e57ce53eba252ae86b06
2dde36dcd6c387249612c590f57c89d761d7607b
1927 F20101221_AABMLF shen_f_Page_065.txt
26f0aa4637665a71e8e19dd525f847b3
63f153ef57c87f46c3720513e8850b05202358a7
1860 F20101221_AABMKR shen_f_Page_028.txt
a49522c79bb0eca793ca056f5230c186
884fe8408f6dacb1ea9c7b2c0fe4e27f24f56a8c
353624 F20101221_AABLID shen_f_Page_094.jp2
485bc1d67ecfbd73a5fa9371269b0e65
708bf42652495b955f535c0f21363156afe4b159
F20101221_AABLHP shen_f_Page_036.tif
ccdab6791aed8ad239ef83086f50de9f
5d411b5bbeda3eaf249970f9e2b72e79decfbef3
1826 F20101221_AABMLG shen_f_Page_068.txt
fe7688b712b7f00e15954a9096d684bc
baaa701f42560afba290ed02bebeb78a6f282e21
2229 F20101221_AABMKS shen_f_Page_033.txt
525f1ba5bddac36454cb02f6f90e8421
68d333749b2e294bb3267fa412d14f63d33cb5a3
5832 F20101221_AABLIE shen_f_Page_004thm.jpg
d696325ee5c25384cde4735e58b9a004
8eac7ef265032208529f434f419f86116d85e975
56734 F20101221_AABLHQ shen_f_Page_069.pro
e4d9d9cb67ec37dd7d3ef9928b1a6290
7edb6f057a77bc13d54b1e93f330fb3d86163b6f
2125 F20101221_AABMLH shen_f_Page_073.txt
062997b481edcb4ea0d644fb0d288558
652e425a33d6f686f90979633509037035a98bf1
F20101221_AABMKT shen_f_Page_037.txt
0d1f189186646acd8f5f5cde1f11bbe7
9a9960f9ad7fdcc87dbc5addd1a4f88650f9e085
127479 F20101221_AABLIF shen_f_Page_123.jpg
343425000404e956af5b6b71e72820c4
6a6c972138d86b1de614743b33702762d4bc6810
8435 F20101221_AABLHR shen_f_Page_001.QC.jpg
9a54925e0bf76a6e223ea53b99b6c89a
f0f9adff35067af13835dcd61931695f29efe14a
145 F20101221_AABMLI shen_f_Page_075.txt
481d531dfe8b9ea67bb2ee080e6fe398
7d471ecded5a16d794a434a6eed7065ee16380c5
2073 F20101221_AABMKU shen_f_Page_038.txt
d70d4145d6b0a24bc6e2502402d4744d
bbf49746f8093b8f7a348a0f292977005032113c
F20101221_AABLIG shen_f_Page_040.jp2
4e50e7b4b962c911017eb575af861a2f
6163aca1a723cfdbe9e23ca5ad034cd67fd1e592
53268 F20101221_AABLHS shen_f_Page_023.pro
b9ff94002f77e67792d53fb7841ac7aa
ee730033d6109820a00a76b9eb37c712bbf18bb0
2161 F20101221_AABMLJ shen_f_Page_080.txt
76a51cdcbfeb535a7e3a898ae1807a42
6aa106bfcd3029eeef6fd77601e650a6e5854d60
2167 F20101221_AABMKV shen_f_Page_040.txt
2cb7aaff36f19a361c5dbf9ef14aa804
907b5e6df9a40be07b7207bc8a3b95ff275ebedd
8689 F20101221_AABLIH shen_f_Page_017thm.jpg
2075414f02acfd39c01cf2fd040275d0
57fb4c14d4ff442e38be9e00a279521fbce6a485
109576 F20101221_AABLHT shen_f_Page_054.jpg
02f540e6d39a008c7dcbfb089fe73800
74a9eb12bc3c6805ff1fdd003ca4671b36d3d7c8
2181 F20101221_AABMLK shen_f_Page_081.txt
fd84ff3687c397df7dd7c2c77d4a3bd9
aaae18fcb3daa5baed1ee11dc7502998f9ff22dc
2078 F20101221_AABMKW shen_f_Page_041.txt
b0ae21e9165f4c2f2907cebeac0028dd
dbe6a260ac1057e2a7283a2c226a85a27a22f6bb
5995 F20101221_AABLII shen_f_Page_009thm.jpg
b023c984417bb9872213339f31e3dd7d
2d3e3fa0f4d4d5ec524df17179572a686d9dc105
103105 F20101221_AABLHU shen_f_Page_062.jpg
d6db36a0d6199dada0989958b2513539
60185ca9bd7ab83fec0530ee6751afbcbbe31841
2641 F20101221_AABMMA shen_f_Page_125.txt
3d7b0db6057fbf6b89b529d5c0eb4b04
ad8bedf2a2ad26f4ca3099cc54302d6816afccab
1707 F20101221_AABMLL shen_f_Page_086.txt
6527ec94745ebeaa69c1fa5e0aab8941
2547fb037fe6921d5c4271f0a8b3fbaede21ffcf
2089 F20101221_AABMKX shen_f_Page_042.txt
2e157e4a4fefe17437717b3541658b05
6871e12c10bcad3fdaeb02866a38fbc6a74b4b7b
104572 F20101221_AABLIJ shen_f_Page_029.jpg
2b30b8a67ac29fe7bff5fb14b28353fc
36539cedfe00331198a17993b89966a0f32a32b4
1051963 F20101221_AABLHV shen_f_Page_085.jp2
13cc06ede25d1e14db45dcd12caf4dfa
4c4708012cb72923780bd3d670d71e7adc423030
2539 F20101221_AABMMB shen_f_Page_126.txt
b0942be334eb1eb9b139f0eef2b09e5a
20695be04925b4db745271cb1256049eeabb94c2
1822 F20101221_AABMLM shen_f_Page_087.txt
639b5bdb8313cea13c348ff2c225bbe3
4b523a70719f7abda1704e8f5584fc2c1040a765
2147 F20101221_AABMKY shen_f_Page_044.txt
88c83a86411a2ba185313bdc19ef6786
cb07540b088e904c950277a2ca3ac84449e79dcf
6500 F20101221_AABLIK shen_f_Page_116thm.jpg
4c5adf9abbc00dd839299d99b8b671e0
1ab831ff77c38f1608bf25315b43083061dbaa96
7718 F20101221_AABLHW shen_f_Page_086thm.jpg
3b05678b8904a025c38cd2c452797ee4
0ee1476488d8729ac3e83fed4abd3555191e251d
2783 F20101221_AABMMC shen_f_Page_127.txt
83d766d5842fc718c496a7c56051935d
60d2723d21a3c265b5277adace460c1c2df9afc6
539 F20101221_AABMLN shen_f_Page_089.txt
169df811174afa7765626ad243316bc2
59cc47f5670fbcb65a7832116dbc275de2e663d3
2177 F20101221_AABMKZ shen_f_Page_049.txt
2012084cbe5e9196658ac0e32778a9f7
af208133db69b8108f0f745f3f5a4900dde7f113
410803 F20101221_AABLJA shen_f_Page_089.jp2
d1703432224895412e4cbd1696df1e54
35a169f5e9d0fd55c605bfc46a308d246efe2251
1980 F20101221_AABLIL shen_f_Page_102.txt
3ba1bbf191d7b65f345ea1aad54e528b
c3278e0d6ce0fb45ba821e7b8eda3287552aaf83
37422 F20101221_AABLHX shen_f_Page_055.QC.jpg
66cc0a061bff1023a950e28a5e8d5b3a
be0647423603b5e2a4f25f5b721a16af29c75fac
2266 F20101221_AABMMD shen_f_Page_130.txt
f4f5bf982e5da821d8cdc964a7a7e987
9687e524d6d5e41f09aea4490acb45a73a5d8a53
2009 F20101221_AABMLO shen_f_Page_100.txt
0c02469944ba65cc7f25c2231fdd7fad
2e9d72c05e9a52322d9bf8099488fe9dcd3a1e98
8253 F20101221_AABLIM shen_f_Page_059thm.jpg
607d4d51985cece5592c54bd391855b9
bc3e05df0ab2068d647cd27ac2dbe4d57d88245e
F20101221_AABLHY shen_f_Page_101.tif
8a0de52e6fc2601d3a4ed9a6c3927cf7
d4f481f0861892ca4faed1518b6c1cc9ebdb0749
91265 F20101221_AABLJB shen_f_Page_006.pro
394c93264a2ccab2547146c79925b103
d53df63b9358b2adf4d49008c608024154149c58
2639 F20101221_AABMME shen_f_Page_131.txt
8c8705fb8c69edf0ec5f382ad11c74e6
aebd8c288604648bdb2e0ea191c2e6064a802200
1873 F20101221_AABMLP shen_f_Page_101.txt
8b4775b0632d9786bf4c996b9b7d7ecd
608ad2cf45485b646f3c8e09fb3e8a454afd7129
1051958 F20101221_AABLIN shen_f_Page_073.jp2
3c652f96e782257c5695da3805a40df5
4d4e23e0ff4f318d317b42494149c78966e31aa2
8765 F20101221_AABLHZ shen_f_Page_054thm.jpg
a82ff5858c852fd7f9cdfa86c2ab5d9d
e7b39a86b879f85208d5e30c31e914ed46b37ded
6542 F20101221_AABLJC shen_f_Page_066.QC.jpg
2a4137b739e0e228ca2d5a929b46b22b
66df1d528bb111e48d47ce736eebb799d8de1dff
2218 F20101221_AABMLQ shen_f_Page_104.txt
dcc2aa98b8b8f08f402107f4371dfcf2
1e72a7df5743d66ee0fc701e26a3e3036a5d2059
19242 F20101221_AABLIO shen_f_Page_088.jpg
7fdf546d0e6727aa405c1f813d9395f4
3c4ac9d3c2c9ab5f7bcc04c13952969681906a7c
969 F20101221_AABMMF shen_f_Page_134.txt
c0f06c220135fd3835cca31019c785d7
6f7c5888a9890f2c02db2a9a0646c381adbdeece
1910 F20101221_AABMLR shen_f_Page_107.txt
13876001d7237c5dafa0b54564f7e00e
24ec1b2b53d742e9946430cc8fc97246eacb5063
14747 F20101221_AABLIP shen_f_Page_007.pro
5e453d50c2003432ba2fd6a3f9b6b594
783d4085e008ba4ede07cb39c9c14f1c3c36701f
2749 F20101221_AABLJD shen_f_Page_128.txt
4e6ae657585d02c2d718423183078ba3
2fba297a8ea216b76cde1088808a794a698e4a02
1472333 F20101221_AABMMG shen_f.pdf
72675478e5d4afa08705d00bf5301849
4cc0359da4fa70cd4ec107a408bf9ada78ad9ad1
697 F20101221_AABMLS shen_f_Page_110.txt
6a415698e4ae918b6282a4b3307128bf
9826fb75ff9968848631bb92fcea287a6a5bcdf4
31212 F20101221_AABLIQ shen_f_Page_010.QC.jpg
31868a9e6a47c9e2930d10ebf3655b99
9cff27aabfeef60f48d2bc03797c781ba5294042
3346 F20101221_AABLJE shen_f_Page_075.pro
3e13c4643047dc6a458fe6cded7b687a
edb908717d32e5c291404c602e5244b6e8b2c93c
475 F20101221_AABMMH shen_f_Page_003thm.jpg
133c11d4f46f7b5b0929bc3e9d8e6f89
d14f214a0db6213a899d566c92d816438c17d319
1309 F20101221_AABMLT shen_f_Page_113.txt
3f9a185d9260bedf32b2868c6ab22bf0
c39dfff03a3b5125c4a92ad1b1fd1a7284c4c099
2371 F20101221_AABLIR shen_f_Page_129.txt
74422cd9c3914309075c534467f2f982
7a64162337d19de60c28f214341d7064923d08c3
113869 F20101221_AABLJF shen_f_Page_129.jpg
7be30c59bdd36e3998fa689025bc5a08
6105d998efdc7ba3d8fe4d95a46437cac81a55b5
22057 F20101221_AABMMI shen_f_Page_004.QC.jpg
14f4e1c3caa4c375419d28b9ed21ed42
f4ba33aec938183f9dc059120cbc0f9f5b42d80e
1518 F20101221_AABMLU shen_f_Page_114.txt
313dc5e065300a6d878fb7bc3b19a0ae
852dc1e9043699ce2e9ac1f1f757bc7c875ff92c
371 F20101221_AABLIS shen_f_Page_133.txt
8a43abcc9605c0cf1e45fe6965ef46bd
5948b9b8c09182c9a9c74c79731999ae4a036caf
128308 F20101221_AABLJG shen_f_Page_128.jpg
b58915a2834c12c340d20a521d5ec869
456d4824ef37be12c2f7a20758b905fb79c695bb
28839 F20101221_AABMMJ shen_f_Page_005.QC.jpg
ecf36587a267f1770b4439b05f6880af
284748d3127e4dc3e7c10a8f0eada34f0e3a07b3
1008 F20101221_AABMLV shen_f_Page_116.txt
eff42efffb25674b86db69757fd0263a
b33518c39b0c207c4dd02fc9f39c267c0d29d618
35625 F20101221_AABLIT shen_f_Page_085.QC.jpg
40c6f03daed56bb25306f36f013c842c
9349df325ad8d01587b7c8b8fc7465f52264ceca
30288 F20101221_AABLJH shen_f_Page_007.jpg
bcdcac0bc9ccfe3b9cba5b7fb7e4de1c
4f0dbfcdab9c51abca4196d706bac0ac060c3147
6264 F20101221_AABMMK shen_f_Page_008thm.jpg
d79ac366d5421c99406431bbaa0b440e
8d84e6816a6a01338e0b8a8efda1c1bb5a44591c
2385 F20101221_AABMLW shen_f_Page_120.txt
a630d488b309a4da1c363b2bc80dbdbf
dd4b30bd6400ad87a1253b95169f0a76e54ebd16
55649 F20101221_AABLIU shen_f_Page_037.pro
44d7d915554747fbcdcc71ef98b42512
1cb94996c009a79f7fc566ee0bdf7df982054e8a
10180 F20101221_AABLJI shen_f_Page_074.pro
c7b69cb23cef2afc4a69986f0076ec59
2b76c9e8a854b337df891c94b8d34e25f0fed2f4
9122 F20101221_AABMNA shen_f_Page_023thm.jpg
c723c688c2677f826b52114aff98161c
c051d1c081e8b0645ec88b344d19df120c6ac87a
7557 F20101221_AABMML shen_f_Page_010thm.jpg
eed37239bc1a04fb34ef9a66086a662c
60c95d297ad396b2325833c8e27775336327d32d
2622 F20101221_AABMLX shen_f_Page_121.txt
c08b7a1f4c96096ef2f119d34d4775ee
8a805deb3ff0b6ed4a8a83a57a910b8a76b11b26
F20101221_AABLIV shen_f_Page_061.jp2
38056005611104cf2dcd1d522b853651
e3cd1e406119775c3e26d050c63f138340d46369
111462 F20101221_AABLJJ shen_f_Page_120.jpg
e2992c6333516a877a16466546d1340e
131b702b0b858e40d97213ffefa5d437fa6bbab5
36351 F20101221_AABMNB shen_f_Page_024.QC.jpg
1841efa5478a098fe334051c13adcd2b
cf8f92e028d04868c5674662eb548a148daa1bf5
13023 F20101221_AABMMM shen_f_Page_011.QC.jpg
670863b27fdb3158508cb8168ac5079f
fef42e9f98552819cb3b8c40a5678c108a177b34
2157 F20101221_AABMLY shen_f_Page_122.txt
b1e72220718805a78ef73be41c6ecd2d
9accf91d68b94973ee27ebf88dbfdc455b5dd046
35919 F20101221_AABLIW shen_f_Page_023.QC.jpg
450a2c290b04ae079d49ae83a4996f55
b9fbd6217caf44731ad32d63bec449f13114108e
F20101221_AABLJK shen_f_Page_041.jp2
562d6e1f5d84a89b4af63dfa53b88b9c
a36d2b59c1ba584f8b4a857607b52d2eff27fd5b
9018 F20101221_AABMNC shen_f_Page_025thm.jpg
8f13753ccc40878483f3a477a58f3e4d
8091e4a17fea0ca135c1182c9af317964f76ea5c
36342 F20101221_AABMMN shen_f_Page_013.QC.jpg
2d68410745540981950a9e6368f47c31
d5d85facf0baa49b9b1901fbfe01061bd6cbc0a0
2607 F20101221_AABMLZ shen_f_Page_124.txt
b86cab5e2275f0e766cbeb8ced9c18ee
5126c19411d13b856965a32242f20c11f1f02264
F20101221_AABLIX shen_f_Page_132.tif
95ad1d5f2f70eed0f813a279f091f1ef
19bfa2d97444dfc15b33b0863b003222078b1f66
33836 F20101221_AABLKA shen_f_Page_015.QC.jpg
cdce9c2a72f32631ec1056e540d1620a
1d701669b4563701970bd45d7a3a2b60f47a3f0d
8903 F20101221_AABLJL shen_f_Page_105thm.jpg
6a9884c9919403c68b76c6c994ca23dc
121d6f18cbbbd021dc97bc6c3df043fba867037c
37868 F20101221_AABMND shen_f_Page_026.QC.jpg
08cfe093268da9a10fd8d6df243d53a7
dbb51f1b2179ca1f9ba77e326c97d4f9b4d34f11
36694 F20101221_AABMMO shen_f_Page_014.QC.jpg
eb54323df0bcfe999c4be2142f3b381d
548776ff9df6dd3cdf9bd286cb3171e8a9457296
F20101221_AABLIY shen_f_Page_058.tif
a00e70907c658e07ff92b84ec6099e4c
4ceacd9647b57818cc8b276bce038c0fa7f1ced1
100160 F20101221_AABLKB shen_f_Page_036.jpg
7e2fa298fc52317f5ac35663621c999a
bb79e51a96eded7d186e2419de1883e287642d34
F20101221_AABLJM shen_f_Page_121.tif
3b17ad578aae74a2e17d34defef24ffd
3f1c60848b620bbd756ff147794019b4a84740e8
8906 F20101221_AABMNE shen_f_Page_027thm.jpg
a0021425032e8eec1c836a45eb811269
111aa91d7b5ec5f270241f2a3240a818136ea1bf
8970 F20101221_AABMMP shen_f_Page_014thm.jpg
3b63ae67c90633e165172b05fe854b18
18c11b2d3224ac80362ef1913c26a3867d5a8584
8456 F20101221_AABLIZ shen_f_Page_103thm.jpg
45c4143089a4b515a259dc515debd2f5
ee81e17ca967a822c62143c47a2f77abc5678af4
9274 F20101221_AABLKC shen_f_Page_121thm.jpg
0a52c12b8c03436d1bd3a84e9b0c5a6c
f8651789c3ecf85643ae71715ea300f1dd24d1a5
F20101221_AABLJN shen_f_Page_024.tif
f968e08d8c9df0ce10b78889d37474fb
c31eda1defd5d0631edcc05f35ce2da4e149e3ee
31148 F20101221_AABMNF shen_f_Page_028.QC.jpg
7ef7d0759651b5858c3747716cef1b80
0876d720838b28f2e178bdd473876d5be7596b14
8468 F20101221_AABMMQ shen_f_Page_015thm.jpg
c9cfcf5fef3f8d8d7f305f63c9f5c5ea
403d671e831e0268c9eeec6fb8ab200922e63229
47814 F20101221_AABLKD shen_f_Page_065.pro
098559277d89fb6fca1b080adae59f20
e39231efe7416e3d06177046bd0054ec3e6d37e2
344306 F20101221_AABLJO shen_f_Page_091.jp2
bd7d2c4d7c6ff2d6e08b17d3b341cedd
3b74a1af1e5ca5481dddf1e5a47d8b572a3a003f
36284 F20101221_AABMMR shen_f_Page_016.QC.jpg
0bd6c92b77bfc23651b2ba299b28fadc
2cb1ab7ef8dbc577a9bb129c45208f96cc262484
32226 F20101221_AABLJP shen_f_Page_087.QC.jpg
b60f40adf4b1d27b1cace28e7e76ca14
53e43333c8ae5f1eef214155dd47ceeeba11cc61
33577 F20101221_AABMNG shen_f_Page_029.QC.jpg
e417e4df2bf7947e62c416e58278afda
595bd3d496ce5070151e8fde4dda8e7a7423b109
8632 F20101221_AABMMS shen_f_Page_016thm.jpg
8927bf3ab3935049bdc38fc4426f2302
eac1e6c737e95cabec801787b0f9681c247dcccd
8379 F20101221_AABLKE shen_f_Page_071thm.jpg
45bdf44a4050cad70c78b49973d35fdc
01af4ce8c0ab4ccb1aeca727fe60a8210dd8b1a8
F20101221_AABLJQ shen_f_Page_054.tif
8645f19c3341863127e528879dd452e3
52373a5623ce22ebf8fd7a9acd746028536f728f
9194 F20101221_AABMNH shen_f_Page_032thm.jpg
3ec70df73125bac12c66de4ece0faad4
ae8ebba34c4cd4395c7cad14d1b4dac2a1d7a609
34799 F20101221_AABMMT shen_f_Page_017.QC.jpg
5eebf0083113a3782af93ca078ee729f
815284a9583472044842141983687910a83f80df
6749 F20101221_AABLKF shen_f_Page_133.QC.jpg
4a34b7302a97c952aa9afcbca189b4d3
4834ea6aa6de17ccb8fe7f95f1bcef4144587f02
2427 F20101221_AABLJR shen_f_Page_063.txt
217cac52c0ad675485b6055ddc191403
f86d24495cd996fed23a5e6634e3e4aeece83b82
36611 F20101221_AABMNI shen_f_Page_033.QC.jpg
9c2d00bdb8b5a081a35a98d21af6e9a3
f085acb79855d60e3e82a040d496fc0aa932b246
9238 F20101221_AABMMU shen_f_Page_019thm.jpg
ad6e52524250feba155bcfcaefadac75
4d387db3f9bd27691b8eaa55344687b1b6ee28ec
22271 F20101221_AABLKG shen_f_Page_133.jpg
385a45d509ae9c688638f8c62bbf234b
190c4a5f6559c1c3f2e8deb5197dc0c5a4c79d17
35147 F20101221_AABLJS shen_f_Page_031.QC.jpg
475a32ae2b8e4e608f8b0a85cad1abd2
04c36c1cc28ef2d3cd69f4c3ece326502bd13c69
35401 F20101221_AABMNJ shen_f_Page_035.QC.jpg
64f683873ca496706a747c56cdaf380f
9f2087d70342b7b373a7d2ea8951d0420ae5182e
36074 F20101221_AABMMV shen_f_Page_020.QC.jpg
7580b50ebab6d1bbe0eef66d204deb53
cc8d5b9a92211b74902c77ea888791922079f9d7
24222 F20101221_AABLKH shen_f_Page_113.QC.jpg
f9007928dc2784cb496a9e6c1b9537a9
1c2b116e94d4de8eea82d4e22f09421f5f99f517
F20101221_AABLJT shen_f_Page_129.tif
0aada617ec40fa7458fa244d2a2efa67
e96fe43cf7d362dae0ea935d8e8d796f0747b11d
F20101221_AABMNK shen_f_Page_035thm.jpg
2574c0ffef81ce62298d0aa39bceff3e
3d0aba5732ddae3fd4d416a221361be7d402a53c
8914 F20101221_AABMMW shen_f_Page_020thm.jpg
644b3749cfd9f50d3d124b14c158a2b6
132a925ce31424df1089748f43f2e1db6e510202
7791 F20101221_AABLKI shen_f_Page_107thm.jpg
ed3386543a887110900eae62d23b90e2
241856fec82bf10012002cdee4a24e63d72a099f
38152 F20101221_AABLJU shen_f_Page_114.pro
c2ff0e2e8657dd4c2d2a34b6f86a88e6
798e9d9e2d53f49677feaa41d41414fe4048ce36
8854 F20101221_AABMOA shen_f_Page_055thm.jpg
3d3025edce77acd05a04586adc1912be
adc6d5e33b6514cedb5b5d92c6eb36e03278aaf4
8190 F20101221_AABMNL shen_f_Page_036thm.jpg
206e7d89381cc1b484b87b6899604a51
fc4e3c77591748135268835d451a381282603266
36447 F20101221_AABMMX shen_f_Page_021.QC.jpg
71b6418a52ccf2af6b5ae0586eda604f
6ee49ef35d6a2b5133928b8b8cb0692788e154de
1051972 F20101221_AABLKJ shen_f_Page_065.jp2
c5348b322e71f0769d4feb671cc76a80
a2df1d233078b9665f0423e555672dc1f79e12a3
F20101221_AABLJV shen_f_Page_062.jp2
eead4012aabb75eeb5dba7437daedd32
6b1fd212bd41645ffbd1841b86c7b1d695e27880
36863 F20101221_AABMOB shen_f_Page_057.QC.jpg
896a6f47dc17502df1a2c20871432104
d572292d9149e2429ee454b006c9b9b21ea519da
36378 F20101221_AABMNM shen_f_Page_037.QC.jpg
2771378b887dd6b5198cebbb44f96185
c3dcba4aa7102075326594466313a3a04d38d476
F20101221_AABMMY shen_f_Page_021thm.jpg
cb37e6a7963722beea7c9dbbc8d66404
b918b1a1c46dff0466a19cef0d1269bd3606758c
F20101221_AABLKK shen_f_Page_128thm.jpg
85a098d26b2731fb25f5c36c4c11dc99
babbd36a406b24f82118d3337fd4d8044c187f75
59184 F20101221_AABLJW shen_f_Page_026.pro
9dba86633b29371a08024921dfd59c09
ed613702d50af31c982e2091a148a771b5b4a569
32255 F20101221_AABMOC shen_f_Page_059.QC.jpg
f3b6eb1afe009b3f5d75517bd49c233e
a330b7eca659f5c5bb9b63c819251ac7d30d1161
34492 F20101221_AABMNN shen_f_Page_038.QC.jpg
af0a0d964f05709e539b638c0498381d
3ada1571dc33951be00e19f6847e71e1ce0a742b
37283 F20101221_AABMMZ shen_f_Page_022.QC.jpg
bab8a965ed4141a19ea5efbcafb80634
c140269e9aa212123d39ce4efaff17ccda58b0e1
52793 F20101221_AABLLA shen_f_Page_056.pro
19316a8e0e3b34aca6ce54ff71ce4508
3643465c15efd26334974172125b5cc797cb7af0
1866 F20101221_AABLKL shen_f_Page_088thm.jpg
f97cccd07451a9ea8a989a2846af71e7
f8760c17ffe049d92f00feaba07c173d875c4448
55892 F20101221_AABLJX shen_f_Page_051.pro
cbea3559481dcce724b134a5026a4a53
b1d62d4c720ec148e7faae0a1711f17769884769
35111 F20101221_AABMOD shen_f_Page_060.QC.jpg
98602db2ddfa7ea52591bfee8527b0a0
8c1dd284069f7df6f75c581141f6c13d73a3f912
8885 F20101221_AABMNO shen_f_Page_038thm.jpg
cc3077ae3e6f2262956b8ca7637fd36c
e19bb3cdfdbd4c2d97529d341ed9ce01c6e66671
F20101221_AABLLB shen_f_Page_084.jp2
e017f784d521ea9e2427f32349fa763f
41a9b9ce8c60614a6842c81161c319b9f23e666b
F20101221_AABLKM shen_f_Page_059.tif
ea868c79734100ca7f33631816181320
050a36d636d735f1633f383a41ead3007aeeaffa
2180 F20101221_AABLJY shen_f_Page_045.txt
b6e448224a1bc26393136f80774c0367
fb76f2fcfb36a78e20b1eba0ab7099703660b983
8733 F20101221_AABMOE shen_f_Page_061thm.jpg
96693475d88b358a80c9d29778d53e23
217b5ebd25e89e4f2142a62be36007f6e0bd6fe4
37969 F20101221_AABMNP shen_f_Page_039.QC.jpg
9e84d3884fc1ba588f662cbc2c91b404
9514907ce67d0d377abc50d40401eae719bcefa8
42545 F20101221_AABLLC shen_f_Page_086.pro
ca97c0c73da558bc23fccef636dfb44b
c509f9d0584bbd60d61f6f77b4d48b63cfdd7d30
97145 F20101221_AABLKN shen_f_Page_071.jpg
569c7e619d6ac5bf43b27cabbb4605b6
3044e65526b7437a050cf69f6effed9d5cbac634
14964 F20101221_AABLJZ shen_f_Page_093.QC.jpg
d0694fccb64d26a325ea44cb8c5a3e0c
3be9de86d7af57f53c760d14a4d2f4a81b5b4617
36388 F20101221_AABMOF shen_f_Page_063.QC.jpg
1e4c4fd14a3615c55aa14a1a7bb3fef7
9775fd448f7ca45a13e3ae6a9cec893499b241f3
8679 F20101221_AABMNQ shen_f_Page_040thm.jpg
bffa88d585359d1625c23c786edc3efb
eea22efbaa2fa7d892a66bdb7db3daef94ecca79
819455 F20101221_AABLLD shen_f_Page_114.jp2
694536b9431135a0db95eeb2f34aecd6
3c3c31bb6cd6e19b5a14d422090f7359c8ded017
F20101221_AABLKO shen_f_Page_053.tif
b929f41f2d9a4b7984239b12d1bfad70
2489fe35cae5e64fefba95b7914dee95d7dcc3c9
34054 F20101221_AABMOG shen_f_Page_064.QC.jpg
ce2a39d883fe6449b30773dda2e9bf80
61be9c32c3f008c413d37fb4c5c384cd524a0ce5
35240 F20101221_AABMNR shen_f_Page_041.QC.jpg
278a18bf35060f138e07170d16d7e8ec
0c4220267245c0df31026300beff9d92b981d9ec
95685 F20101221_AABLLE shen_f_Page_087.jpg
f93e302639cf979a40cdb09bc19b2287
585d07db24cbc8c18c71061ce5c818bd69e04bf5
52768 F20101221_AABLKP shen_f_Page_079.pro
6021026d687d5bd71c09b04afd1fc8e4
694cd06514a72f28da3b77c4bb9957134e654552
35918 F20101221_AABMNS shen_f_Page_044.QC.jpg
6482d88d3d8a759c550d9a087a9c0642
2170724b07dd7dac741641440ce3d4e4faaa94fd
19667 F20101221_AABLKQ shen_f_Page_117.QC.jpg
16a3ee8e8378e559fb5d439a5bd59a32
186dba7685c5903390ce3666ef7bd3351f42e4b1
32192 F20101221_AABMOH shen_f_Page_067.QC.jpg
f63bcf879ea60fb6491379a31ac6bd5d
e66fdac4feae86fb3695805ea05de16c6647f181
8565 F20101221_AABMNT shen_f_Page_045thm.jpg
67560628a600110f545b9d47463e3e93
43b24748202e364d7c57f176631c92733517fab2
57413 F20101221_AABLLF shen_f_Page_032.pro
b44f10962348f46fdb020a5320060543
fb27e6bd0f51918a096a0318cdb1d24a8ad4b400
117536 F20101221_AABLKR shen_f_Page_063.jpg
3dc50041f224a0411e05e5dd245ce783
f1d73e48ee37094d1cdc75a054b064cc1f773c87
8270 F20101221_AABMOI shen_f_Page_070thm.jpg
3dc576a792fd1adf114c594ec5de5fe5
fa90f851e8a60ad911b4ee38a0beee072e66fc08
8873 F20101221_AABMNU shen_f_Page_046thm.jpg
030477cf6dbd7dd2f53990698b5ef0e1
36fbc8fcc01620cd0b9a1c320c67760e3685536d
F20101221_AABLLG shen_f_Page_065.tif
932ef22f96c74d3242429c85c5e5d8a7
0fbe9bcb99e37025b260bb95f708f6f57a6ea8b5
F20101221_AABLKS shen_f_Page_063.jp2
27c7d80c565baf120e89e09846873bc7
0ec87a66b01d88ed29d669fed9bc8804d9b9621e
8058 F20101221_AABMOJ shen_f_Page_076thm.jpg
bc0c389672b3b49163fba3abb1403544
f0760d75ad20cd76330415087238811853157de1
8414 F20101221_AABMNV shen_f_Page_048thm.jpg
55ef44cac869b0feab67bd17c23117c4
c8300537255344de3bab17ffc9e12476f143e065
F20101221_AABLLH shen_f_Page_010.jp2
4648e5bf658ac4657bc17ed3194f13a6
703de0b0d099859750a8eaa404bc6153ddc2ec14
F20101221_AABLKT shen_f_Page_014.jp2
c7f251abc4499abf82b2e403bbb1b4a1
68e0e83aa783db0a7bfabb11cd0b55a04349be7f
34313 F20101221_AABMOK shen_f_Page_079.QC.jpg
95a3516bb0c6db45a0cba1e0b0346191
6a5dd91f4370ff45aa08f6a7adb2cb71bb8d2334
F20101221_AABMNW shen_f_Page_049thm.jpg
79a8c07e80ce40a32e788506c41c74ff
e130d4c1ca8d9d7addd94a5cda78f4cdf13bd374
9168 F20101221_AABLLI shen_f_Page_057thm.jpg
09294a2652ee2de361c00030b760a3bb
456cf2ba300c6c0acd318f21f35ef3b17b2a3d36
431 F20101221_AABLKU shen_f_Page_092.txt
3d38d7d08725bc70fa544675ea558f20
10c0a02d67d2bda284c79991cea4cd506f1b4fd0
37167 F20101221_AABMPA shen_f_Page_104.QC.jpg
c22c4c862a4b99b35081c909cd399f2d
490cfe88e9d6b2da3c987e5200e5d9e467559ed4
8437 F20101221_AABMOL shen_f_Page_079thm.jpg
deda225b254ec735ad9e4faaa5dec7c4
0081f2db77a033f80451d65c1dcfc8993e003b15
34338 F20101221_AABMNX shen_f_Page_050.QC.jpg
ecf6134f47e8fe2fd7490876c30ebccc
f423abd7c3a877c8aecc27f0563f87fa4c2a42f1
35446 F20101221_AABLLJ shen_f_Page_018.QC.jpg
826748e981d2811dd7975081649301b5
a703b9edfc17ac88ebe965d506b14958150bc43f
19352 F20101221_AABLKV shen_f_Page_066.jpg
bce345072723b916ea5fe0939d60e3b4
9b61b727a2377ff35c709f27ff2625a7ba0ec4d2
8796 F20101221_AABMPB shen_f_Page_104thm.jpg
e91f5e6847f390410c0f7d757c037585
661c574bbf98dc754de7c37ec6350d703a269653
9012 F20101221_AABMOM shen_f_Page_081thm.jpg
d7c8b10f03708e0912cf253d8be2dc9c
f6d489e911aa062bbfee6b4973b31753ac599b3f
35816 F20101221_AABMNY shen_f_Page_051.QC.jpg
faa511a451542fe8413c94608216ba6a
ff1afca362cd1df096197b181d8722d764d3a682
2182 F20101221_AABLLK shen_f_Page_066thm.jpg
22fa620a664607f2ef5f264c7868f414
a46fde9e14b8ca1e0830f5eb7e2520615756da4a
54453 F20101221_AABLKW shen_f_Page_035.pro
7860ce58c83eb6a61b7e7aa9dd191ec5
595187a124db0ce4e383d73eb1c552c77baa69fd
32347 F20101221_AABMPC shen_f_Page_107.QC.jpg
c2281662291f3622ab3358cb05b7221e
50d7e18e307337bc27470607920b0955be399278
9132 F20101221_AABMON shen_f_Page_082thm.jpg
61805bee04b3b7f8331b794cba61c82f
f08789cd9fd5f0ec97853ed8a72a4c4854ef4de5
8337 F20101221_AABMNZ shen_f_Page_053thm.jpg
70d9be7960f516e0487d3faf0bd37322
0773f44e7d069cdc6fc1d5a629cbc38558236699
7583 F20101221_AABLLL shen_f_Page_074.QC.jpg
26e07b2d3ac4efa9ee7cc57bbdb103ba
43e5f985952f74e09796ac87588c4d83f84726c4
1048667 F20101221_AABLKX shen_f_Page_071.jp2
451884f80cacf46f19c2aa30f333e0e7
42f191f1af83e5ef96b05cfd8a00855546876c31
2588 F20101221_AABLMA shen_f_Page_132.txt
dec476c08b1a517054e6b0cb64e870d5
693a442382a14ae5e8e950adc8857aa156f7f7a0
35712 F20101221_AABMPD shen_f_Page_108.QC.jpg
ab84118f31d3637575126b4e4bd77bc2
199506e6e2c64c8352dcee4dfca2f97f9bbd3fd5
8787 F20101221_AABMOO shen_f_Page_083thm.jpg
63f1c5f7a86f5341c0fde5a8c868a9af
19483ff63f217a50e8ae699b152e18a956e7a03e
698 F20101221_AABLLM shen_f_Page_002.pro
eb37a557166bd8d8fcb2123c6c83ceb3
64aa6cb5fae172e4c82b3d55aa7bd979f15027b0
103912 F20101221_AABLKY shen_f_Page_122.jpg
a9cd677b7ad2080f340c4c66165ef25e
88927467f1862ac3d3e6b01dae56c3094191df44
67774 F20101221_AABLMB shen_f_Page_127.pro
cf55928b496f0693a028b696e189e67f
d8a95ae62084185c32ddef07d41d17fc435ee3c3
6288 F20101221_AABMPE shen_f_Page_111thm.jpg
2b905cf80887b1203028e80fb2e12d42
79c877b36ffea18d1c2d304c989ced770d3f3f4b
15751 F20101221_AABMOP shen_f_Page_089.QC.jpg
63fe0a6419a26f7cb0a443624bed75aa
e397005f752eeddebfeae3372c280a75dd429c9f
111607 F20101221_AABLLN shen_f_Page_049.jpg
88ad430b95188de90e575379696a1951
653d27eb84917a844da480ebdfe1e838667ae345
213487 F20101221_AABLKZ shen_f_Page_133.jp2
8d9a5b98eeea703d83791739d19f061e
7a0bc23f24191eeb844d1c01a9f79ead191cb301
F20101221_AABLMC shen_f_Page_018.tif
d8261b44fc7c8f934fddb0933726f84a
986302ad501ab8afcae802e6338009c9fbb7042f
29326 F20101221_AABMPF shen_f_Page_112.QC.jpg
eb9f49b4b09eca9eee317589b66273ff
dd36a9fc6dc423c7f1e4ce5e172daadf82cf313e
5245 F20101221_AABMOQ shen_f_Page_089thm.jpg
1ee4203b42dbffd0278120c1b819c6fc
fc6f51e2dd7252582b5064382b47ce4d0e2ed260
31139 F20101221_AABLLO shen_f_Page_071.QC.jpg
937020090755ae7f0b30f5f20566db14
cf18bc529aee49e5da943adae7a8b9a29570497f
6406 F20101221_AABLMD shen_f_Page_119thm.jpg
c9676169fc38b3600b03e568b22d7f58
fb923afed960e0f542698d796d57c3f5a4efd70e
7952 F20101221_AABMPG shen_f_Page_114thm.jpg
ef3c47f812119a2219d6deb61dde3d54
0a49b8c762e0dbb00f7c77d2620d1bc5c9a60b1a
6606 F20101221_AABMOR shen_f_Page_090thm.jpg
43de15e8cf198b0d0d398ecc43d08fa0
68a785cb386a9b7c578f3c7e3e54c6fb7825978b
2264 F20101221_AABLLP shen_f_Page_082.txt
6c6d68b8496b145a51af3735d030fa2b
ea49848909b42e8fc5e58a4490e3e510d8935f61
54538 F20101221_AABLME shen_f_Page_054.pro
05f5e334108579e6369fd2d87d79b83a
57302adede95fcfcbd11fa0cbb48eb4ba82e470f
28140 F20101221_AABMPH shen_f_Page_115.QC.jpg
75f6471bb49d58bbfec7108ecd075b96
12d936e5922d3951323bb30f647671654abbd099
23440 F20101221_AABMOS shen_f_Page_096.QC.jpg
66b5ffef496094f11c4a205a0512f8c7
75cd265482f5079b6a9c78480d3cbcc7de7d7aa1
2137 F20101221_AABLLQ shen_f_Page_060.txt
50e359330c015cf16b277381d6b016d7
dd107237119b0048ae38e42821982c5bdfd7bf5f
38762 F20101221_AABLMF shen_f_Page_093.jpg
ca114db77638538476eab157f5624ef1
e3de5547412e3e94917586961ca5170c57112af0
21673 F20101221_AABMOT shen_f_Page_097.QC.jpg
e6f290306c954fe73cb9230f1526e869
0e302ae8f061b8b973bd8f5943544bf745f54ed9
350 F20101221_AABLLR shen_f_Page_091.txt
bf0a15902f7f4aaa4a6d3cbc7df1bd83
1d1da86e7b97cebbb33c34613783eca08959ddcd
7854 F20101221_AABMPI shen_f_Page_115thm.jpg
10e81b1d7b55067a91512fb68fbf22c9
5cb04923c359fe7cb628a7d14ffdfb66a391d71b
6186 F20101221_AABMOU shen_f_Page_097thm.jpg
d1448f20ffbcf6e38612815231ed3959
d95875ae974da263cec30ab6dd881775cf51fbaf
374 F20101221_AABLLS shen_f_Page_093.txt
0b0b7705b7f4d28477ed64df585c9ce1
82bea6005df03822788438618cf6f7f5fbdeb761
8252 F20101221_AABLMG shen_f_Page_087thm.jpg
25b3382dd21776f7ad5f4145c03544b1
0347e978260cda4e7f52a89058be90acf189eb8e
6162 F20101221_AABMPJ shen_f_Page_117thm.jpg
f64c5ead49d476fc25421652cb2286c2
5781e2475f39cfcb8b86e8cb937fac6c3aa3217a
23153 F20101221_AABMOV shen_f_Page_098.QC.jpg
9beb4e2e8801be18541c3983590ee0b7
cc9b5f86d7b63a0d4c8b0c130ed2425f690cdf1b
14228 F20101221_AABLLT shen_f_Page_110.pro
f1a1bafa6a5401e3200bfb21451ccb4e
56efe008274b97ee7b1941629d8acc6c28498c2c
1564 F20101221_AABLMH shen_f_Page_115.txt
9bb4aacc20a0b8ef4991b7576ec9b089
75d82fe8126f6d5de14d69de5825bb305da4bd82
24904 F20101221_AABMPK shen_f_Page_118.QC.jpg
a571e66ecf5cc826942623e29d38a42d
29a7f06e26dcb5f7d1a70386dbc3e4e53af8054a
9185 F20101221_AABMOW shen_f_Page_099.QC.jpg
3146f9aaff66cdbf4dc42ee1c48cec60
31a561eb90906073409cfeb205f932cc5c8cddee
35157 F20101221_AABLLU shen_f_Page_027.QC.jpg
a69e6d32dd05cae662e126df9bd0eec3
e3cf2fb150d45308fb7fb7d77f06cf5daf58b952
7607 F20101221_AABLMI shen_f_Page_101thm.jpg
a4e005dd5a134fafd074e69425517f09
1e9d9a947c5160e7e0351f614d5d9270627896b8
F20101221_AABMPL shen_f_Page_118thm.jpg
774a9f7bc2a7b98ccc67dfe14cc3d01a
22384024c68ea0105b6f6f4f82dbaae8acd116dc
32295 F20101221_AABMOX shen_f_Page_100.QC.jpg
830ece6089246279a40f241514cf81b2
c686a9c745d2a61b08ce557196d44e6470f6df81
107147 F20101221_AABLLV shen_f_Page_061.jpg
293fd16546297a1fc823b339fc269602
3d7b92cf5795b19e9a10215a43dec7c8c712b551
9034 F20101221_AABLMJ shen_f_Page_126thm.jpg
9ea79abe361be2fd2c3ae7e896530ba1
6975a10c764c222e8fb0ed0f2c9e0a17a84f0a8d
8364 F20101221_AABMPM shen_f_Page_122thm.jpg
1d515b4ad3985faef9f8f74eff3d3669
3e5b0f9c736e026fc0051be6f89fe079a0653432
7950 F20101221_AABMOY shen_f_Page_100thm.jpg
5bdc436012cbb1967a346e854d56a7c4
222d7791d6d0eca725a6add277d19a6901aa96e7
8719 F20101221_AABLLW shen_f_Page_130thm.jpg
ba71eb4dadf878a47e7ed43a4946228f
e57dcacc6b2afff5fb34489b106cd334fa2f12dc
35913 F20101221_AABLMK shen_f_Page_081.QC.jpg
c795e06e9051e3d707c81f66f2a03c39
3fb5fcdb98120c3cedb710a9ab3565108eaa4ca9
37002 F20101221_AABMPN shen_f_Page_123.QC.jpg
241ab58ed3d5e502a48aaea3ad0d2a4e
c7e991a38a10f7468fcb7fd1b36b3558185ee6ab
8226 F20101221_AABMOZ shen_f_Page_102thm.jpg
b8e4a70e29812b4b6678cfd0cd66edf2
35a38d6d6fe4116755d769631c0552c34d451ae3
1976 F20101221_AABLLX shen_f_Page_047.txt
da01f4722bc1069119f95d9e09de90ea
ca915caa92ca6fd23c31a20f4ecc3d8e3710c37a
9245 F20101221_AABLNA shen_f_Page_033thm.jpg
b492dfa144d0ae83c977a10324f4f09e
5ca3a37190c166096f584fcc380bf266269e6c9a
8154 F20101221_AABLML shen_f_Page_112thm.jpg
ebd4663b234e4a7f01d0960e65d4f5fb
889424a38fb91205a0fef6b6aa930eb93d66c4d7
9203 F20101221_AABMPO shen_f_Page_123thm.jpg
8ba284dcd11dec19afd7d27757fbd776
918e562635dcb9cdceb8e615a1d2091736a72eb6
F20101221_AABLLY shen_f_Page_043.tif
abf9c2d28cd9d89c95595b892916458b
54dfe5bdb39581f6ba5d16912accb2af1cf37431
F20101221_AABLNB shen_f_Page_069.tif
461e373cf3e915d11996ef969cc13603
340971810c33acb539cc4651469d8fd61bbb49dd
4381 F20101221_AABLMM shen_f_Page_109thm.jpg
dc2ea3b66cd5a2c4bec9fd188289d749
a280ba3cba32f61fcf400f7a3fbe7323d6249703
37115 F20101221_AABMPP shen_f_Page_127.QC.jpg
cf58a938410d3a7a2e55bf51333c1196
0deddd69bd4cd3aae8ddc9683c7dc216fe42c506
8221 F20101221_AABLNC shen_f_Page_047thm.jpg
b76530c6dd288b1f87f2c408a1a18684
b973b053e388016319c729613b57215093b9908e
F20101221_AABLMN shen_f_Page_058.jp2
c7cd2afc34ceb90ce1d5ae365ed4b13f
b51d0777d25972b18f796ff25adab537ba0fa649
46001 F20101221_AABLLZ shen_f_Page_068.pro
1d30a3f21495c790e0e079253d985bf5
88d2b219dff07a93d19b02c19ac6950676c31ccd
F20101221_AABMPQ shen_f_Page_127thm.jpg
ef992ddeabb261a1c5dfdfcb8edb02be
a309d473079e4c4c3320130703ef9c7adf3c4998
2052 F20101221_AABLND shen_f_Page_030.txt
d22e6d08aaeed21876fcc5ab85302b84
62a3c0a1a31baf7dee4f669ce1c3a7dd2e0c7bb9
F20101221_AABLMO shen_f_Page_071.tif
b962fed235a85fdb44a3e8d7fb6d2111
ee9debf759df82936faa6e99240d614235e0892a
37926 F20101221_AABMPR shen_f_Page_128.QC.jpg
9dc400fd59bb5fec11289ce9061b8cd0
a504a38c74d82abf4bef4efb48619032471e50df
F20101221_AABLNE shen_f_Page_034.jp2
649aee11e83d167907f56706f5662214
45b5a6d5511cad8b69b3c8cded064363b118483d
97276 F20101221_AABLMP shen_f_Page_052.jpg
0d305b8d35cc0626237c1b03ba271d7a
636a1619082789eb3905d5d190564e44f1a8cbe9
154006 F20101221_AABMPS UFE0021931_00001.mets
abc2ac776f8f2de19ecafe78370426b0
18c9c2bb06c65f65c7a11a3519746c0ec2aa57f2
7772 F20101221_AABLNF shen_f_Page_052thm.jpg
e646ac442fbaeb4ca22639a5539da9c2
4ed9183da3cece906adaee032b30ce3c2f10198c
37342 F20101221_AABLMQ shen_f_Page_121.QC.jpg
006cfcc81f71804745441e3ef5f8c7d2
b361f77c55414f02de251a965c7c348c15cc8b47
2124 F20101221_AABLNG shen_f_Page_046.txt
6fcf70cc92c1f4944e22ceebe8e39f91
3ce97e05f85529ed4fd1ac6f5be4723bccd131eb
F20101221_AABLMR shen_f_Page_030.tif
e4a8322a661067e091ad124210b4ca5f
63c3136b47f0ad407bf3e40568e3a45e3ba4272d
F20101221_AABLMS shen_f_Page_043.jp2
1bbbd5b9561268380900c7aa9edb35fd
06bdeba4a969060fdf2ac724ba265e174bd6ce6a
104715 F20101221_AABLNH shen_f_Page_030.jpg
5bfb8f92d40af04f6dd5c160ccd58dbf
4b7593e5f4f28f225387ed53476eacd79a5f8156
33622 F20101221_AABLMT shen_f_Page_105.QC.jpg
eb25a7702e9795e3662fa35ae2abc85d
94a71da2c912308179a468fe22975ce0b51e074d
F20101221_AABLNI shen_f_Page_022.txt
839c13b64ad262c8376fa0bf20cc87ab
d1cf9c9da446148c2fc3a0e2d017c8c9337bcc49
F20101221_AABLMU shen_f_Page_006.tif
44850e88e4c882198e877d7f12aa9d4b
1c0c6eb150139c5337e91ca08658c7ecdc0d6969
F20101221_AABLNJ shen_f_Page_086.tif
63ed868ba0d9e711a19b953c948da56c
7b86f690d9f2e36bb3a569fa3c530907fcecfe9c
F20101221_AABLMV shen_f_Page_108.txt
c1ad01c1f3afa471100c359306fecef9
5ef6f1c40b81e563dac1c0e6fe3889b9789d902b
8716 F20101221_AABLNK shen_f_Page_029thm.jpg
b874a1cb4b93ffc1712d2cae4e0293be
90bc8c36f9885c420131316899e264c9d10a4767
86263 F20101221_AABLMW shen_f_Page_008.jpg
50e80b0a2defabaabc82404b9c9b63a3
98ef35295b4f240c8f3734166367f8165d16b7b5
F20101221_AABLNL shen_f_Page_039.jp2
811490b3c371db30b8a0fd30d86b9df2
a50208549d41f4811e4511df6dad99e390c92a66
47991 F20101221_AABLMX shen_f_Page_107.pro
66843fbccef39aa31e209b40af6f37e1
1aa98580577f1fa31f22c1e62b3373534d319581
111038 F20101221_AABLOA shen_f_Page_013.jpg
0858bc665a7c33cfa4d4a912e3490f58
bb706b9668c59d25553d6a66a6fcace13216d13a
24708 F20101221_AABLNM shen_f_Page_116.pro
349385b06e5c0b2f1c9cec92856284f3
7132604d2d8b89f68d2164d9acbc65c3b5a5bd0e
37739 F20101221_AABLMY shen_f_Page_091.jpg
1b397731b1356ddd45fb05810c715fb6
6d7f1642aa6769fb1ad3d0ac90745d791a8a3919
7581 F20101221_AABLOB shen_f_Page_007.QC.jpg
d46915c29d754da2952fdb68f36fcd92
4e21edcbef85d47f680cf6766f59c93c7ba3d0c9
F20101221_AABLNN shen_f_Page_085.tif
cee3f4b78345316be2e645c4b72f0791
e388c1e279385ad8eb7d4e4efe5c37bd7a583b39
1660 F20101221_AABLMZ shen_f_Page_072.txt
90dc8bea9900d0fae0909205cab857a2
540a6f0853c0208f326db067998921c1ae101a2a
F20101221_AABLOC shen_f_Page_027.tif
17cc2e580c23e2e78aa243262329dd31
6ee9896af7f50059f568c886c25d39ec18598183
F20101221_AABLNO shen_f_Page_022thm.jpg
723f5da87660fce183ee61ae25d374e1
88fbb411dc284a698f9d853cc8c723595bdb8e2a
57292 F20101221_AABLOD shen_f_Page_120.pro
0d47518eea450897a21ee66be5b41684
f2ee20b48d40bf6a151e679d3a103f3897acbaf6
F20101221_AABLNP shen_f_Page_018.jp2
bf3df00e50c8da253e7864ecd7e422b2
2e801f63c89b72317de871fb4be25711f7c5f5ff
113295 F20101221_AABLOE shen_f_Page_014.jpg
a0d30541cb12d2c7caa06339c1ad1348
bf5d5b2ab20571ca578f441121113e5dc4d98ed5
8508 F20101221_AABLNQ shen_f_Page_050thm.jpg
66df86393a3729bddef566c7e0de9f7a
6f84688626365bee495280da5610cd4c45ebc234
4204 F20101221_AABLOF shen_f_Page_134thm.jpg
d17f58d31dbc5c2222ed934223644bd5
5ec67cd4025f9516a4e380f92384364691407edc
8714 F20101221_AABLNR shen_f_Page_044thm.jpg
65de55882ee9ec0f2f45e1d10534aba0
2cfaa6fcf3ba78e99718b76ba7c20297809cdd10
9022 F20101221_AABLOG shen_f_Page_051thm.jpg
d73922359b333dd8549ae705e736692b
eb3eb7acdd645b4c44429dfde2d885061cee7feb
F20101221_AABLNS shen_f_Page_009.tif
8fb65a6e04ebd9c7679826aa38832439
e38f92ce50b1f315b52ad9a0945e07f9242cd337
106553 F20101221_AABLOH shen_f_Page_084.jpg
ba8a56a48064aa5ee1805921b4646a7d
fb99a956286d987f4e3b94f7f18dcda27f3f3477
55046 F20101221_AABLNT shen_f_Page_024.pro
1f62c2cf661e26d574db23a8d159a37f
8a92e4ad36181e1dc7eea175b11277d30af20e3c
73 F20101221_AABLNU shen_f_Page_002.txt
69fc161f609b746ef3107f8eb143ebb2
12330cc4774420f750643322834a5ad01f19cd79
406 F20101221_AABLOI shen_f_Page_090.txt
2b4dbf020559c3d742c399ad00bbfa10
ea7ae650c6dc0cd428e1073a7b392d1ef75cf65b
1969 F20101221_AABLNV shen_f_Page_036.txt
919696baec6b5dce6da98f704e4de0a5
567aa571737eaec050f302c310e1c72b3ab78082
F20101221_AABLOJ shen_f_Page_048.jp2
676e7f145e078f4dba1e5dfea3f30b66
3dd9a343f3819f54f0d6c482dd355205aad20018
F20101221_AABLNW shen_f_Page_062.tif
a12c386709fae1425eec445db35fc87b
22330e32195fd8edad733c08166f4c8b9eb7ed79
F20101221_AABLOK shen_f_Page_087.tif
d49fdf66b5782e4793acb0367a142d3b
b2276d1d768bf5d6821a74898f25c53675707495
F20101221_AABLNX shen_f_Page_012.jp2
5455226f44ffc36346c5940e94ef345c
7832b86961aea1d53789c7426124ea9464a57faf
8902 F20101221_AABLPA shen_f_Page_085thm.jpg
caa5039dd20d91f1d218d66ae49bec7c
341e88ee010767eda89dfb25d1e1f41d41af6907
F20101221_AABLOL shen_f_Page_038.jp2
4597f7382de0c5b12325166b65812eac
a0fa4477b9bb9d70d3528bae7b69b4cc9fe4fd3e
159660 F20101221_AABLNY shen_f_Page_066.jp2
c582b2bb740d9b1f1c69d64751c049ec
a2d92e5a8781aaca2e02f637d4742ef3ee6cc5b1
45364 F20101221_AABLPB shen_f_Page_087.pro
dc72013ec64559e0b5420bb9d76a058d
ec8e0407e51828245b3d0e116b02e2927eb16e09
F20101221_AABLOM shen_f_Page_047.tif
34d175bfa5a9672084e738d2b21b0527
fda59c91f9584aee5425a1dd81808759f3ee3790
6130 F20101221_AABLNZ shen_f_Page_096thm.jpg
7eb4ea94de65d0469ee741091ec3679d
286472ac158154f989989ef3a9ae5e440c9c49a6
2416 F20101221_AABLPC shen_f_Page_016.txt
2262314af1931679469247b5adaf5ef3
af8bbf40fdaa46891b21d5630188e6315d8a1f96
F20101221_AABLON shen_f_Page_010.tif
6ed51dbdbef99686389ba34ce43aa952
cddb72678f72cd892b954cd61752a89e74ca95ed
F20101221_AABLPD shen_f_Page_006.jp2
07faf7710d460734a3353258eaba080b
0e3dcd52645a2637b2509eba334b7413dd3cf8e0
111127 F20101221_AABLOO shen_f_Page_024.jpg
1078a43b325dee363a814f368ea46d26
6b3aa39e3fca9ade8525b44e01df7e3600ea9d21
1051932 F20101221_AABLPE shen_f_Page_050.jp2
98800f193fda9bb388d048bb6cc8fdbd
40a90f7cd47483b98ac3f86a61e7c2de4bf16ca7
38040 F20101221_AABLOP shen_f_Page_083.QC.jpg
f36e3a45866ad18c299db04fd6457311
eaf5ef920b2e72e52c3ded285b2e6f006e36545d
6677 F20101221_AABLPF shen_f_Page_098thm.jpg
0df609914334ee370894ae35f87c51cb
c2ace94c3430603880dbaf79e2513d8129f933c2
8700 F20101221_AABLOQ shen_f_Page_063thm.jpg
f34ee7ee08b5965ecd294e5a257b7b80
c8e38c7af86f46ff2c15d976526e98718db314ff
F20101221_AABLPG shen_f_Page_070.jp2
bb713e5ab3ba37431cdb7163786d6831
e84661e7e9857b30c1137179092860b75f23b3aa
F20101221_AABLOR shen_f_Page_098.tif
088545e5da0b6614331320092735152a
1919d1ead31f018ad9ead333c77f8317b0e881f2
36499 F20101221_AABLPH shen_f_Page_025.QC.jpg
9e413d5a5e2727866cda8888ce1a8151
78f7fa460eb037e7eb53bbbe5cad77f60f694120
1433 F20101221_AABLOS shen_f_Page_119.txt
f9bfa56cb27596edd2625da297bf0f23
9b39ceb6b22b94a27d9903909e0bf3bb0102d34e
38922 F20101221_AABLPI shen_f_Page_092.jpg
91690e89f4f783dc3cec02416e3f96ea
036a83e50b574ce3c9e90880d0b33c945609227c
354303 F20101221_AABLOT shen_f_Page_093.jp2
ba5d889e3dd85fa45741c9a1980c3468
1e135de8f99022f6b904f708344d4754d38790f1
4905 F20101221_AABLOU shen_f_Page_091thm.jpg
407e21d195a17baef69b304e84255431
c4bea3d4da0ff558c93699ddaf84097dad867f09
F20101221_AABLPJ shen_f_Page_020.jp2
ce5a34a5a4d893e726ed7de6263163b4
bdb12dbc8a0137ec1556619a9173db54b351fb66
78912 F20101221_AABLOV shen_f_Page_009.jpg
8fe5a5e9c070225d40eb2bbfdbac248b
d22eccf268d34cde37fa8df0c901a90f5146f771
55022 F20101221_AABLPK shen_f_Page_130.pro
45aef210c7f23bcda18112ba6ce4cc99
e015d51a8384c7eb398d3e18351dd084114b6221
17965 F20101221_AABLOW shen_f_Page_003.jp2
00a2c925f3560ff97e916e97596a32ad
7e17071fbd3d95f767a6eea3556f4360fb6a4724
1101 F20101221_AABLQA shen_f_Page_097.txt
9a5b9ac95310549378250d7f8595d651
87e41178bae629d75858b9eef4de5163112e1175
48193 F20101221_AABLPL shen_f_Page_073.pro
5e5b7d723b8279eef487c226792a5f12
a3fd74cf83de1daab47faa22d465e0548c715461
6208 F20101221_AABLOX shen_f_Page_093.pro
ee9d125baba00ddc532646807ae1c6a2
0af867711138efaa76e8772dd7c7a6627256dac2
F20101221_AABLQB shen_f_Page_044.tif
75fbc9999bf7bce43e89033b4fde4d3f
c5e4f30628fda75de9920ec0120f288cfe0a9404
2139 F20101221_AABLPM shen_f_Page_035.txt
4b5374bfce736f1b1b985d8edaac5798
44389bb1ad3ae6de16eddfa64c82e8cb12b6cfa9
33943 F20101221_AABLOY shen_f_Page_030.QC.jpg
d00da0e916b32b087ef525af9ca59833
1dbf4df9ba5444759b743345d6e47c9db831b286
8734 F20101221_AABLQC shen_f_Page_108thm.jpg
adfe1ee1dca3fe531fce3cb293e1d33d
39ac1770115e4f13f167c2078b07d71c94765069
F20101221_AABLPN shen_f_Page_084.txt
89949bbe74fe616b80d455eb85606c4d
d22ea8f5fb9533360b27d6fb2002259902f38757
2119 F20101221_AABLOZ shen_f_Page_106.txt
f175ed4dae5bf5f038ccea08c0072270
23e5dabb6a4d8ef7f5cf2617be72bb7abeac1837
8469 F20101221_AABLQD shen_f_Page_041thm.jpg
09655c20e01234934012ae08b2274589
c48f0dd33f8ae6a49a4c540652975035f9c67af8
8553 F20101221_AABLPO shen_f_Page_056thm.jpg
0d7bcd91f0da04e352227b345ed191e9
7983cd21ab1776dcbd8659dcb846f52955b446f1
704 F20101221_AABLQE shen_f_Page_011.txt
8ea06629fdf4491925fdd1cd0affd229
e1f18445781c8a0748adec904c7f5a7e52cc5cd7
F20101221_AABLPP shen_f_Page_072.tif
da7b0f7ad1d7479868cbf42fd36d148d
62c3e02b763e3416f067ba20e5e42f8aaba23d10
7963 F20101221_AABLQF shen_f_Page_066.pro
7737e96e46888a3772a537b210c962f3
a6f50ea41f57512ff6db08c41ecd961365021fac
51867 F20101221_AABLPQ shen_f_Page_030.pro
39cc9ae9bac60c3de8f4a38d05823442
db65106ecce5199dcb6829471128e516af55b049
25169 F20101221_AABLQG shen_f_Page_109.pro
517fb8f29d10eb9d638f2522fe0f470a
530d40ea45d64c9751c49adacb63d15027b02273
37064 F20101221_AABLPR shen_f_Page_082.QC.jpg
dfdcf1cd51b4be9ae2451ea2a2da005c
2ff43afc36aa3a1d6359a96e4f340bde886f0ab1
10637 F20101221_AABLQH shen_f_Page_110.QC.jpg
efd4f0acac06b1ae9c0848ee4648ad8d
8b62b71d665f098aa4892b9834dfd9de632957ca
2050 F20101221_AABLPS shen_f_Page_048.txt
9a55c339b30142a6fb4387753ba93705
9dcf29730cfd73c0fe44d8ac19b09a1a4664a8e5
84682 F20101221_AABLQI shen_f_Page_115.jpg
11d4156ffb77a5ed09d908232799ac67
467ab19af62253cd4ad76277c04d31c070017bc8
53800 F20101221_AABLPT shen_f_Page_018.pro
fc4a36ae10dc4cfb7d6d76a8eb68c45e
c6e2a8c3cbc90ed9eaecb2339ed43e91180aa7b5
2060 F20101221_AABLQJ shen_f_Page_015.txt
fb8e94144925fb499de47ddfa1a74a2f
cdc8e6f9c9333406468b54a52ab8b98421279abf
2508 F20101221_AABLPU shen_f_Page_099thm.jpg
beb6374384fb1658f80e02b9430efd69
1172b308a1c2f5b5a87e4b8eedb01069ebcdd9d6
2252 F20101221_AABLPV shen_f_Page_083.txt
c4acdec40ccf2515bc2de6510a439621
978e8e06c623f5faf911130b822b918df04c8d5a
9025 F20101221_AABLQK shen_f_Page_131thm.jpg
7414f17098637e16be786b9d557e2795
be78f68e9b4dd7cc86738b208a5f37b863cdb532
F20101221_AABLPW shen_f_Page_084.tif
a22af02ca2aa3914470ce2d1903da4f6
67b96c45339834a607b1d6c00dc5a89f2a58bf8c
34050 F20101221_AABLQL shen_f_Page_120.QC.jpg
71dffa92287fa879909ba4010c95b3cd
a8b3f7b3755aa91d0f2bffb4c4e3f736ed86d1e9
F20101221_AABLPX shen_f_Page_116.tif
be6acf97aeebc71ba6f6874905ce22c9
966656918e0ee5a91b0cbf7a9442b019f657d183
2350 F20101221_AABLRA shen_f_Page_026.txt
9eab9733da59a509acf9d50671ee4a35
2776390d0218693a8c400f3b7ec3ccec0f9986f3
36779 F20101221_AABLQM shen_f_Page_040.QC.jpg
6e554331ea28240b498a151127fbdd92
e69265950dffb3f391e7cb58c350bae4146e1ca4
106442 F20101221_AABLPY shen_f_Page_085.jpg
a8558e126e2d8454154000cb9478d546
5554960aa90cc3d0cc250b5745fcf8ad04f2eec6
27305 F20101221_AABLRB shen_f_Page_114.QC.jpg
591bba6385d17d7ef78656b1154d270d
9003438701ffd1c13a2e1720686076ac794cc566
7400 F20101221_AABLQN shen_f_Page_006thm.jpg
d465eeb93c45cdcf2e1d6110ae401e36
2f425dcc28a7fad572a0b15b2115b04f6f8286fd
35357 F20101221_AABLPZ shen_f_Page_118.pro
81ed99f30e27499cf4c82b3e192cf210
d2cf5fdc3491ade93355c36771cf9fbc4dd522b2
2012 F20101221_AABLRC shen_f_Page_105.txt
05732839344587b3c20e048f8237c934
20161879379bf64587eb172aa6b233172a28cef2
104054 F20101221_AABLQO shen_f_Page_103.jpg
fa07ef88d577e063571351e3f0ff9883
29fc9a8658d8ae5309130f2c5cb276d4feb37eec
50998 F20101221_AABLRD shen_f_Page_105.pro
b59d10751441d9052162458b3d39c3f3
061614f4f26511588831532de0a62069b0fc3479
8806 F20101221_AABLQP shen_f_Page_129thm.jpg
913bde868aeb25a08136e5b92b98ead2
2b0ffe53338da4d6bc7a2169335008b32dfa0531
36895 F20101221_AABLRE shen_f_Page_132.QC.jpg
b81aabfdc0c9c11e32d70606609639f5
101aab3fb159eacaa2468ca481e14182f669b8e9
F20101221_AABLQQ shen_f_Page_056.tif
33ed4f42408347879b165fe303c61b67
5393b62bf9343973f4f865ca2c1b49d3d3c7eb6b
110910 F20101221_AABLRF shen_f_Page_021.jpg
05a387159080fa2720245bea7508776c
b471d41695b0c3beff370a5e885137854c89c4a8
F20101221_AABLQR shen_f_Page_112.tif
11ff1b0d5f44f1b494600ad19c03af4b
54864463e5425afea64ef26ea312c4e2b77f77c5
98910 F20101221_AABLRG shen_f_Page_070.jpg
a668844d4b1d9405d53ee09a5cc4a62b
c63b42fae9c732b363fb40ef01bbd8ce36abc8f8
F20101221_AABLQS shen_f_Page_057.jp2
7a4e10fd4d7648d9918a101230746a1b
2c81de6b808d1c01d0366e06c594dcd4248d7281
114868 F20101221_AABLRH shen_f_Page_032.jpg
1a2b69b9a3ba92dfa3f024e7327518db
1376c14c5098b8fbbbfaf851bfe76a141d1de975
8946 F20101221_AABLQT shen_f_Page_106thm.jpg
79a81ef9255659bbd4cc1218bb6728cc
a4c03f7fbd418f33af54e203523545ceb3d201bf
110753 F20101221_AABLRI shen_f_Page_044.jpg
c42ce95cf130821155f49b281f028683
b273b66ad40e7ccddf37ef19e1cbb5a4a2bac528
38249 F20101221_AABLQU shen_f_Page_011.jpg
4b020d7f5b7ecf62194c62de483a7436
f485d6933dd47bfa791e5b72b7212317ea460d8f
34970 F20101221_AABLRJ shen_f_Page_078.QC.jpg
22851c81616dc203a851ac776020adf8
a3c25ef7e1c3c6108fed264a310058adb3baa253
8981 F20101221_AABLQV shen_f_Page_018thm.jpg
6cb601f16e74d70480a4c182baad8ff2
a70cb815f98f3e728ad2a86cd4e39749b931ff96
1231 F20101221_AABLRK shen_f_Page_075thm.jpg
ad60e284599e18f6c2855f24f7f4f6ac
81ecb6e93fa669652b45b0340d7e2bbe8ecb9797
1051936 F20101221_AABLQW shen_f_Page_067.jp2
606934af1b7ef53efbe5d239a74446aa
3065e4a34709c849ed634fd28a07ee2e01c8d2f9
109074 F20101221_AABLSA shen_f_Page_027.jpg
5b8c395f281270d553ec0ac3772d337c
15b056b138df54165b736ac7cb546fa839a59e7d
798385 F20101221_AABLRL shen_f_Page_119.jp2
c3e6c913f6e2ffd651368e7ec037a81d
a75aaf616e24ad135661b9236224461ab8e1229d
F20101221_AABLQX shen_f_Page_090.tif
a600216e1c468b8111ca000748d5be22
ff6f2ce8acf003957060e6a0bbf08a2742bd35d6







CHAIN REACTION MODEL OF POLITICAL ADVERTISING EFFECTS:
ADVERTISING CONTENT, MEMORY, AND CANDIDATE EVALUATION















By

FENG SHEN


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

2008





























O 2008 Feng Shen


































To my family









ACKNOWLEDGMENTS

I am deeply indebted to many people in my completion of this dissertation and all

the thanks I express herein are meager substitutes for the gratitude and appreciation I

owe. Heartfelt thanks go to Dr. Jon Morris (my supervisory committee chair) for his

invaluable guidance, mentorship, and the unfathomable emotional support he provided.

For her untiring guidance on research design, I also wish to thank Dr. Lynda Kaid.

In addition, this dissertation has benefited immensely from the constructive comments I

received from Dr. Jorge Villegas and Dr. David Miller. Without their efforts this would

not have been nearly as rewarding an accomplishment.

I would also like to take this opportunity to thank my friends and faculty at the

University of Florida, especially Dr. John Sutherland and Dr. Debbie Treise for their

mentorship. I also truly thank Dr. Michael Weigold, Dr. Justin Brown, Seth Over, Annie

Sugar, Sunny Hughes, Kenneth Kim, Tom Mueller, and Michael Clayton for their sincere

support.

The love of my parents enabled me to start and complete the otherwise impossible

dream of attaining this degree. Without the tireless encouragement of my father and

mother, I could not have completed this study. I deeply admire them.











TABLE OF CONTENTS


page

ACKNOWLEDGMENTS .............. ...............4.....


LIST OF TABLES ............ ...... .__. ...............8....

LIST OF FIGURES .............. ...............9.....


AB S TRAC T ........._. ............ ..............._ 10...


CHAPTER


1 INTRODUCTION ................. ...............12.......... ......


Need for the Study ................... ........... ...............12......
Purpose and Importance of Study ................. ...............13........... ...
O utline .............. ...............15....


2 LITERATURE REVIEW ................. ...............16................


History of Modern Political Marketing ................. ......... ...............16.....
Verbal Content and Effects of Political Advertising ................. .............. ......... .....20
Issue and Image in Political Advertising ................. ...............21...............
Positive and Negative Political Advertising ................. ...............22........... ...
Effects of Political Advertising: Memory .............. ...... ...............22.
Effects of Political Advertising: Candidate Electability .............. ....................3
Summary of the Effects of Political Advertising ................. ............... ......... ...39
Informational and Transformational Advertising ................. ...............40........... ...
Antecedent 1: Newness of the Product ................. ...............40........... ..
Antecedent 2: Level of Risk ................. ...............41........... ...
Antecedent 3: Product Involvement ................ ...............41........... ...
Antecedent 4: Product Conspicuousness............... ............4
Antecedent 5: Product versus Service ................... ......__ ....__. ......_.........42
Typology of Informational/Transformational Strategies. ....._.___ .... ... .__ ..............43
Summary .................. .... ........ ... ........... ... ..... ..........4
Survey Scale of Informational/Transformational Advertising ................. ................. .46
Emotion in Advertising............... ..............4
Rational and Emotional Responses .............. ...............48....
M easuring Emotion .............. ...............52....
Physiological measures .............. ...............52....
Self-report measures ................. ...............54......__ ......
Discrete approach to emotion............... ...............54
Dimensional approach to emotion............... ...............55
Measuring emotion with Ad SAM" ....._._ ........_.........._...__ ............5
Summary of Literature Review ..........._..._ ...............60...._.._ .....
Hypotheses and Research Questions .............. ...............62....












Perception of Advertising Content ............_....._ ....._ ....... ....6
Memory ..................... ...............63.
Candidate Electability .............. ...............64....


3 MATERIALS AND METHODS .............. ...............67....


Material s and Measures .............. ...............67....

Independent Variables ................. ...............67.......... .....
Perception of ad content ................... .... ......... ... ...............67....
Prior exposure to ad and familiarity with candidate .............. .....................6
Dependent Variables .............. ...............68....
M message recall .............. ...............68....
Emotional response .............. ...............68....
Candidate electability ............... .... ....... .... .............. .............6
Graded Response Model (GRM) of Item Response Theory (IRT) ................. ............... ...69

4 RE SULT S .............. ...............76....


Pretest ................... ...... .. .... .. .. .. .. .......7
Main Study: Model Estimation and Results ................ ...............76........... ..
Descriptive Statistics of Independent Variables ................. ............. ........ .......77
Prior exposure to ad and familiarity with candidate .............. .....................7
Perception of ad content ................. ............. ...............77. ....
Descriptive Statistics of Dependent Variables ................. ............. ......... .......79
M message recall .............. ...............79....
Candidate electability ................. ...............80.......... ......
Pleasure- arou sal -domi nance ................. ...............83...............
M odel Estimation .............. ...............85....
Positive image condition .............. ...............86....
Positive issue condition ............... ... .............. ...............86......

Negative image condition for the Candidate ................. ............... ......... ...86
Negative image condition for the Opponent .............. ...............87....
Negative issue condition for the Candidate............... ...............8
Negative issue condition for the Opponent ................. .............. ......... .....87

5 DI SCUS SSION ................. ................. 100........ ....


Overview ................... .. .. ....... ...............100.....
Perception of Advertising Content .............. .....................100
M emory ................. ............ ...............101......
Candidate Electability ................... .. .............. ...............102......
Issue, Characteristics, and Voting Intention............... ...............10
Cognition and Emotion .............. ................... ...............102 .....
Memory, Emotion, and Candidate Electability .............. ...............104....
Limitations and Future Research ................. ...............106...............


6 IMPLICATIONS .............. ...............108....












APPENDIX


A QUE STIONNAIRE S ................. ................. 110........ .....

Pretest ................. ..... ..... ..... ..... ......... ..... ..........1
Main Survey Sample (Negative Advertisement) ................. ...............111....._.__....

B CORRELATION MATRICES ............ ............ ...............118.....


LIST OF REFERENCES ............_........... ...............120...


BIOGRAPHICAL SKETCH ............ ............ ...............134...











LIST OF TABLES


Table Dage


3-1 Experimental conditions .............. ...............75....


4-1 Personal qualities and political issues ................. ...............96...............

4-2 Familiarity with candidates ................. ...............96................

4-3 Information and transformation ................ ...............96................


4-4 Reliability coefficients of informational/transformational scale ................... ...............96


4-5 M message recall ................. ...............97........... ....

4-6 Candidate issue competence .............. ...............97....

4-7 Candidate characteristics .............. ...............97....


4-8 Reliability coefficients of candidate characteristics scale .............. ....................9


4-9 Voting intention .............. ...............98....

4-10 Pleasure ................. ...............98.................


4-11 Arousal ................. ...............98.......... ......

4-12 Dominance ................. ...............98.......... .....


4-13 Fit indices in experimental conditions .............. ...............99....

B-1 Positive image condition............... ...............11

B-2 Positive issue condition ................. ...............118...............


B-3 Negative image condition for the candidate ................. ...............118........... ..


B-4 Negative image condition for the opponent ................. ...............119........... ..

B-5 Negative issue condition for the candidate ................ ...............119........... ..


B-6 Negative issue condition for the opponent ................. ...............119........... ..










LIST OF FIGURES


Finr page

2-1 Russell's affect circumplex............... ...............6

4-1 Full path diagram of PIFC, PTFC, MR, ER, and CE ................. .....__ ........_.....89

4-2 Measurement model for CE and ER ..........._.....___ ...............89

4-3 Factor loadings in positive image condition .....__.....___ ..........__ ..........9

4-4 Path coeffcients in positive image condition .....__.....___ .......... .. .........9

4-5 Factor loadings in positive issue condition............... ...............9

4-6 Path coeffcients in positive issue condition............... ...............9

4-7 Factor loadings in negative image condition for the candidate .............. .....................9

4-8 Path coeffcients in negative image condition for the candidate ................... ...............92

4-9 Factor loadings in negative image condition for the opponent............. .._.........___....92

4-10 Path coeffcients in negative image condition for the opponent .............. ............. ..93

4-11 Factor loadings in negative issue condition for the candidate............__ ..........___.....93

4-12 Path coeffcients in negative issue condition for the candidate .............. ....................94

4-13 Factor loadings in negative issue condition for the opponent.............__ ..........___.....94

4-14 Path coeffcients in negative issue condition for the opponent............. ... .........___...95









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

CHAIN REACTION MODEL OF POLITICAL ADVERTISING EFFECTS:
ADVERTISING CONTENT, MEMORY, AND CANDIDATE EVALUATION

By

Feng Shen
August 2008
Chair: Jon Morris
Major: Mass Communication

Despite conceptual similarity between informational/transformational product

advertising and issue/image political advertising, little has been done to connect those

two lines of the research. This research attempts to incorporate informational and

transformational advertising into the effects of political advertising. To further

strengthen the theoretical bond between product advertising and political advertising, a

chain reaction model of advertising effects for products is used to identify constructs in

testing the effects of political advertising. The psychological process of advertising

effects in Positive or Negative and Issue or Image conditions was studied with four latent

constructs (i.e., perception of informational cues, perception of transformational cues,

emotional response, and candidate electability), and one observed construct (i.e., message

recall). With an observation sample size of 1000, and 250 in each experimental

condition, the study tested the proposed chain reaction model to account for the dynamics

from message learning to candidate evaluation.

The results validate the use of the informational/transformational scale to measure

Issue and Image. The results also confirm the conceptual similarity among issue-

information-cognition and image-transformation-emotion. The results also indicate that

emotional response is a more powerful predictor of candidate electability than cognitive










response. In other words, it is what the respondents feel about the candidate rather than

what they remember about the candidate shapes their attitude toward the candidate.

Meanwhile, the results suggest that testing the comparative advantages of Positive vs.

Negative or Issue vs. Image is likely to be biased by the measurement focus of an attitude

scale. Perception of Informational Cues and Transformational Cues impacts message

recall, emotional response, and candidate electability in a predictable fashion for the

positive issue and image ads and the negative image ad. Nevertheless, the effects in the

negative issue ad condition suggest a different model.









CHAPTER 1
INTRODUCTION

Need for the Study

The idea that political candidates can be advertised and marketed in a way similar to

products is prevalent in political campaign literature (Baines, Worcester, Jarrett, & Mortimore,

2005; Ben-Ur & Newman, 2002; Bulter & Collins, 1994; Henneberg, 2004; Kotler, 1975; Lock

& Harris, 1996; Newman, 2001; Niffenegger, 1989; O'Cass, 1996; Omura, 1979; Schiffman,

Sherman, & Kirpalani, 2002; Shama, 1975, 1976; Wring, 1997). All key concepts of consumer

marketing had their counterparts in political marketing (Shama, 1976); for instance, a consumer

marketing transaction through which sellers exchanged products or services with buyers for sales

income was analogous to a political marketing process in which political candidates offered their

service and ideas about policy issues to voters for their support (Shama, 1975). The most visible

similarity between those two types of marketing was probably the extensive use of advertising by

sellers and candidates to inform and persuade target consumers and voters respectively (Shama,

1976). In fact, the ever-increasing importance of political advertising had substantially

contributed to the interest in applying consumer marketing theories and practices to political

campaigns (Dermody & Scullion, 2001; Kotler, 1975).

The significance of political advertising, particularly television political advertising, was

also evident in the magnitude of its expenditure (Devlin, 2005). The last five U. S. presidential

campaigns between 1988 and 2004 witnessed a dramatic increase of advertising spending. A

total of $79 million was spend on television advertising in the 1988 election (Devlin, 1989),

followed by $133 million in 1992 (Devlin, 1993), $173 million in 1996 (Devlin, 1997), $240

million in 2000 (Devlin, 2001), and $620 million in 2004 (Devlin, 2005). The enormous size of

spending was not limited to presidential campaigns (Kaid, 2004). In 2004, U.S. political









advertising spending for races at all level accounted for 67.5% of all political media spending

and reached $1.85 billion, which was equivalent to a 58.9% increase from 2002 and 112.1%

increase from 2000 (Quinn & Kivij ary, 2005). Given the amount of investment in political

advertising, it was highly desirable to better understand the impact of advertising (Kahn & Geer,

1994). One line of research examined the verbal content in political advertising and the effect of

the verbal content on voting behavior (Kaid, 2004). The verbal content often referred to

emphasis, i.e., whether an ad emphasized policy issues or candidate personality, and valence, i.e.,

positive if the ad supported a candidate or negative if it attacked an opponent (Kaid & Johnston,

2001). The effect often referred to recall of advertising message, candidate evaluation, and

intention to vote (Kaid, 2004).

Purpose and Importance of Study

The birth of modern political marketing was indispensable of the use of advertising

through mass media (Shama, 1976). The effects of political marketing on voting behavior could

be best explained with a chain reaction model of advertising effectiveness that incorporated not

only memory of candidate characteristics and issue stands but also candidate evaluation based on

voters' memory (Lavidge & Steiner, 1961; Shama, 1976).

As noted above, a substantial body of research investigated the impact of advertising

verbal content on memory and candidate evaluation when image and issue were used to support

or attack a candidate (Kaid, 2004). Nevertheless, previous research did not generate consistent

findings about the impacts of advertising on memory and candidate evaluation. Furthermore,

memory and candidate evaluation were often examined as two separate effects, although ad

content could influence memory, and memory in turn could influence candidate evaluation in a

chain reaction manner (Lavidge & Steiner, 1961; Shama, 1976). Therefore, there was a missing

link in previous studies on political advertising.









This research attempts to clarify the effects of issue/image and positive/negative ads on

memory and candidate evaluation and explore the relationship between those effects with

enhanced measurement of issue and image content and responses to ads. The enhanced

measurement of issue and image content is achieved through the introduction of Puto and Wells

(1984) informational/transformational survey scale (hereinafter referred to as "the Scale"). The

Scale was constructed on the premise that each ad contained both informational and

transformational cues, and an ad was regarded as informational if it was perceived to contain

more informational cues than transformational cues or as transformational if it was perceived to

contain more transformational cues than informational cues (Puto & Wells, 1984). Likewise, an

issue political ad contained both issue and image content, but it was the issue information that

dominated the content; conversely, an image political ad contained both image and issue content,

but it was the image information that dominated the content (Kaid, 2004). The issue content was

found conceptually similar to the informational cues and the image content close to the

transformational cues (Shen, 2007; Shen & Kim, 2006). Furthermore, issue ads were found to

contain more informational cues than transformational cues and image ads were found to contain

more transformational cues than informational cues (Shen, 2007). In the current study, the Scale

is again used to calibrate the perception of issue and image content in an ad. The perception in

turn is used to assess the impact of content on memory and candidate evaluation.

Emotional response was an integral component of candidate evaluation and was measured

with bipolar adjective pairs such as happy versus unhappy (Thorson, Christ, & Caywood, 1991).

Nevertheless, such an approach was problematic because the use of adj ectives led to a distorted

reporting process that was subj ect to an individual's linguistic habits, vocabulary, memory, and

social desirability (Lazarus, 1966), and language was too insensitive a medium for









communicating private experiences (Mandler, 1962). AdSAM"(Morris, Woo, Geason, & Kim,

2002), a nonverbal measure of emotional response, is used in this study to assess the emotional

component of candidate evaluation with enhanced precision. Moreover, structural equation

modeling (SEM) is used to assess the chain reaction starting from the advertising content to

candidate evaluation. Additionally, item response theory (IRT) is used to estimate latent

responses to the Scale in the current study because IRT was proved an extraordinary tool to

capture the cognitive process of responding to survey scales (Embretson & Reise, 2000;

Hambleton & Swaminathan, 1985; Linden & Hambleton, 1996; Singh, 2004; Singh, Howell, &

Rhoads, 1990).

The enhanced measurement of the content and responses eventually helps answer how

issue/image and positive/negative ads influence memory and candidate evaluation and identify a

manner in which the verbal content can be managed that will lead to optimal mental processing

of political ads and favorable candidate evaluation, which is of vital importance to both

practitioners and scholars.

Outline

This dissertation is structured as follows. Chapter 2 reviews the relevant literature

regarding the development of political marketing and advertising and the effects of political

advertising. The review includes theoretical models of political marketing, the effects of issue

and image political advertising, the effects of support and attack political advertising,

informational and transformational advertising, and emotional response to advertising. The

literature review leads to the formulation of hypotheses and research questions. Chapter 3 details

analytic approaches and methodology. Chapter 4 reviews the results of the study. Chapter 5

includes the discussion of the findings and suggestions for future research. Finally, Chapter 6

addresses the contribution of this research.









CHAPTER 2
LITERATURE REVIEW

History of Modern Political Marketing

Although advertising was just the promotion component of marketing, it was often the key

to the success of a marketing campaign (Hoyer & Maclnnis, 2004). The importance of mass

media and political advertising was also evident in reviews of the development of modern

political marketing (Blumenthal, 1980; Kotler, 1975; Maarek, 1995; Nimmo, 1999; Shama,

1976). It was generally agreed that modern political marketing was "a derivative of the post-

World War II era... and...evolved in stages" (Nimmo, 1999, p. 75). Nevertheless, a major issue

in conceptualizing political marketing, as pointed out by Nimmo (1999), was that researchers

often treated "marketing not as a commercial activity but rather as a form of rhetoric" (p. 76).

However, Shama' s (1976) three-stage framework of political marketing might be an exception

because his work was deeply rooted in King' s three phases of American business (King, 1965)

and was supported by a number of scholars' (Blumenthal, 1980; Maarek, 1995; Nimmo, 1999)

vision of modern political marketing.

These three stages are (1) candidate-orientation, (2) sales-management orientation, and (3)
marketing orientation ... these stages parallel the development in the orientation of
American business as depicted by King ... from (1) production orientation whose main
concern was to produce large volumes of products, to (2) sales-management orientation
whose main concern was to maximize sales through better coordination of production and
mass promotion, and finally to (3) a marketing concept orientation whose main concern
was to offer goods and services that satisfy consumer needs and wants. (Shama, 1976, p.
766)

Specifically, the first stage "covered the period of 1940 1960 ... The candidate was viewed as

a product needing mass exposure, resulting in increased public awareness which was assumed to

be positively connected to voter preference ... number of exposures and length of exposures

were taken as the key for victory in the campaign" (Shama, 1976, p. 766). Similarly, Maarek

(1995) noted that during this time period communication consultants "made massive use of short









advertising spots, strategically broadcast as close as possible to the most popular television

shows. In this way...the candidate could benefit, at no extra expense, from the efforts made by

the television networks to increase their audience" (p. 13). Many researchers (e.g., Blumenthal,

1980; Maarek, 1995; Nimmo, 1999; Shama, 1976) agreed that the landmark of this stage was the

1952 Dwight Eisenhower campaign. It "was the first to hire an advertising firm in a national

contest" (Nimmo, 1999, p. 75).

The candidate/product orientation substantially influenced the development of modern

political advertising; for instance, the preference of frequent but short mass exposures led to

"growing use of short political ads rather than long political speeches" (Shama, 1976, p. 767) and

"the role of the mass media specialists grew from a minor to a maj or place in the campaign ... In

most cases those specialists were advertising executives or advertising agencies" (p. 767).

Essentially, the production/candidate orientation shared the same powerful media effects

assumption with early communication research (Rogers, 1997). "Political marketing of the

1950's ... was nothing but promotion and distribution of candidates through increased exposure

by mass media" (Shama, 1976, p. 767).

In the second stage of political marketing, beginningig in the early 1960's, political

campaigning was viewed and practiced as the selling of candidates" (Shama, 1976, p. 767). "Ad

agencies sold candidates like soap" (Blumenthal, 1980, p. 2). The landmark of this sales-

management orientation stage was the 1960 John Kennedy campaign (Maarek, 1995; Shama,

1976) because this campaign incorporated all the main elements of sales management, namely

"sophisticated voter study ... market segmentation, promotion appeals and media selection

(Shama, 1976, p. 770)." Voter study was also called field analysis or geodemographics (Maarek,

1995) and included "systematic benchmark surveys of the campaign field, which involve an in-










depth study of voters and inhabitants of the geographical area in which the campaign is to be

carried out" (Maarek, 1995, p. 33). Although the surveys could be used for "remodeling existing

candidates to appear more congruent with the voters' desires" (Shama, 1976, p. 768), it was

impossible for a candidate to satisfy the needs and wants of all voters. As a result, another

marketing activity, voter segmentation, was developed to locate "those segments of the general

population which might be subj ect to the influence of political communication" (Maarek, 1995,

p. 34) and the candidate could therefore "determine to which groups he would make his

maximum adaptation and appeal" (Shama, 1976, p. 768). Voter studies also indicated that voter

characteristics such as involvement (Shama, 1976) and selective media exposure (Maarek, 1995)

influenced the effectiveness of political communication. Accordingly, promotional strategies in

this stage focused on "the effectiveness of different promotion appeals and media in reaching the

voters" (Shama, 1976, p. 769).

Many changes in this stage still had substantial influence on today's political advertising

research; for example, survey was still be one of the most important research methods (Graber,

2004; Maarek, 1995) and the search for effective appeals and media was manifest in all research

initiatives nowadays (Nimmo, 1999). The current study is one of those media effects studies that

aim to identify circumstances under which advertising will have the largest impact on voting

behavior.

Shama (1976) also commented on the third stage of political marketing and argued

presentlytl, political marketing is experiencing a transition from sales management orientation

to marketing orientation ... the marketing of political candidates is becoming more ... voter

oriented ... it is the voter who decides which candidate that is, which 'parcel' of personality,

issues, party, etc. to 'buy' with his vote" (p. 770). There was certain ambiguity about the exact









start of the third stage because of the word "presently" in Shama' s (1976) work. Other

researchers held that the third stage began with the 1980 Ronald Reagan presidential campaign

(e.g., Maarek, 1995; Nimmo, 1999) and this campaign delivered to voters a well-defined parcel

about Ronald Reagan: "The candidate was portrayed as having a very clear and simple

conservative profile, with a very sparse political platform: reduce taxes and federal government

intervention, except for Defense, whose budget was increased" (Maarek, 1995, p. 20).

Another significant change in the third stage was the emergence of"[i]ntegrated political

marketing...it involves the simultaneous study, analysis, and segmentation of voters; candidate

positioning, and candidate development; promotion and use of media" (Shama, 1976, p. 772).

"Combining a series of communications through different channels is much more effective than

the single predominant use of the leading audio-visual networks" (Maarek, 1995, p. 62).

Interestingly enough, Shama (1976) used a chain reaction model of advertising effectiveness

(Lavidge & Steiner, 1961) to define the intended effects of integrated political marketing.

The multiple output ... may be viewed in terms of Lavidge' s and Steiner' s (1961)
advertising model calling for increased (1) awareness of the candidate existence in the
political market 4 (2) knowledge of the candidate' s personal and political background, the
office for which he is running, his personality traits and the issues he stands for 4 (3)
liking the candidate on the basis of available information about him and possibly on the
basis of additionally sought information 4 (4) preference of the candidate over other
candidates running for the same office 4 (5) a conviction that the candidate is the best
available one 4 (6) support of the candidate by contributions and/or volunteer work and
voting. (Shama, 1976, p. 772)

The interest in those above-mentioned effects was well recorded in political advertising research.

There were a number of studies on message learning/knowledge/memory (e.g., Geer & Geer,

2003; Martinelli & Chaffee, 1995; Ridout, Shah, Goldstein, & Franz, 2004; West, 1994),

candidate evaluation (e.g., Chang, 2001; Kaid, 1991; Kaid, 2001; Kaid & Tecesco, 1999), and

voting behavior (e.g., Ben-Ur & Newman, 2002; Faber, Tims, & Schmitt, 1993; Kaid & Sanders,










1978; Weaver-Lariscy & Tinkham, 1996). Nevertheless, there were few studies that investigated

the relationships between those effects in a chain-reaction manner. The current study attempts to

address the relationships in question, particularly those between content, memory, and candidate

evaluation with structural equation modeling.

In summary, political marketing was intertwined with political advertising. Heavy reliance

on mass media and advertising practitioners marked the birth of modern political marketing.

Media planning, audience research, advertising effectiveness studies, and integration of all those

efforts emerged in the development of political marketing and are now the backbone of political

campaigns. The literature review above also indicated that 1) content of political ads such as

candidate characteristics and issue stands was the key information for voters' decision-making

and 2) the perception of the content and candidate evaluation based on the perception were the

key predictors of voting behavior. The current study focuses on both the content and its impacts

on voting behavior. The section below provides an overview of previous research on those two

aspects as recorded in the literature of political advertising.

Verbal Content and Effects of Political Advertising

The verbal content of political advertising was often analyzed in terms of issue versus

image and/or positive versus negative (Joslyn, 1980; Kaid & Johnston, 2001). An issue ad

focused on the specific policy issues or positions mentioned in an advertisement whereas an

image ad focused on a candidate's personal qualities, background, or characteristics (Kaid &

Johnston, 2001). A positive ad supported a candidate whereas a negative ad attacked an

opponent (Kaid & Johnston, 2001). In real-life practice, the use of issue and image was also

associated with support or attack; for example, studies on presidential campaign ads (Kaid, 2002,

2005; Kaid & Dimitrova, 2005; Kaid & Johnston, 2001) indicated that both image and issue ads









were more frequently positive than negative; but in recent years there was a clear trend of more

use of negative attack against the opponent, particularly in issue ads.

Issue and Image in Political Advertising

The interest in issue and image in political advertising can be traced back to voting

behavior studies since the 1950s. Researchers at the University of Michigan identified party

loyalty and group identification as the long-term factors and feelings about candidates and issues

as the short-term factors influencing voting behavior in both presidential and congressional

elections (Campbell, Converse, Miller, & Stokes, 1960; Campbell, Gurin, & Miller, 1954).

Nevertheless, for decades political advertising was criticized for emphasizing image and

trivializing issues, and therefore damaging the very foundation of democratic voting, i.e., voting

behavior should result from rational decision-making based on policy issues rather than

candidate personalities or images (Berelson, 1966). However, studies on political advertising

content, especially those on television political advertising content, suggested that such an

accusation was largely unfounded or mythical, since political advertising content was largely

issue-based (Joslyn, 1980; Kaid, 2004; Kaid & Johnston, 2001).

The dominance of issues television political ads (Kaid, 2004). For example, from 1952

through 1996, 66% of the presidential campaign ads were issue-oriented whereas 34% of them

were image-oriented (Kaid & Johnston, 2001). In 2000, 78% of the presidential campaign ads

were issue-oriented and 22% of them were image-oriented (Kaid, 2002). In 2004, 85% of the

presidential campaign ads were issue-oriented and 15% of them were image-oriented (Kaid,

2005; Kaid & Dimitrova, 2005). Although the percentage of issue ads varied over time, it was

below half (47%) only in 1968 (Kaid & Johnston, 2001). The focus on issue was manifest not

only in presidential campaigns but also in gubernatorial and senatorial campaigns (Joslyn, 1980;

Vavreck, 2001).









Positive and Negative Political Advertising

Negativity was a growing aspect of campaign advertising in the U.S. and the past few

election cycles witnessed a real increase in the number of negative spots in presidential

campaigns (Kaid, 2004). Although the percentage of negative ads in presidential campaigns

from 1952 through 1996 was only about 38% (Kaid & Johnston, 2001), in the last four

presidential election cycles in the U. S. (1992, 1996, 2000 and 2004) negative ads made up

substantial percentages of the advertising content of both major party candidates. Both of Bill

Clinton' s campaigns (1992 and 1996) established new highs in the number of negative ads in a

presidential campaign with 69% of his ads being negative in 1992 and 68% being negative in

1996 (Kaid, 1994, 1998; Kaid, DeRosa, & Tedesco, 2002; Kaid & Johnston, 2001). Al Gore's

presidential campaign used only slightly fewer negative ads; 62% of his ads were negative,

compared to 37% of Bush's ads (Kaid, 2002). But Bush resorted to more negative ads (58%) in

2004, as he defended his presidency against challenger John Kerry (Kaid & Dimitrova, 2005).

Negative advertising was also a growing feature of political ads campaign strategy for

candidates in election below the presidential level; nevertheless, there was no strong evidence

that political advertising was predominantly negative in state and local election campaigns (Kaid,

2004). For instance, Kahn and Kenny (2000) analyzed 594 ads from Senate campaign in 1988,

1990 and 1992 and found that less than half (41%) of all ads contained negative attacks.

Vavreck (2001) studied 290 candidate ads in the 1998 elections and found 64% were positive

ads. Furthermore, negative ads were usually focused on issues rather than candidate

personalities (Johnston & Kaid, 2002; Kaid & Johnston, 2001; West, 1993).

Effects of Political Advertising: Memory

There was certain ambiguity about the impact of image and issue ads on memory as

measured by recall. On the one hand, there was some empirical finding that image ads










outperformed issue ads in generating higher recall (Schleuder, 1990). On the other hand, there

was evidence supporting the superiority of issue ads over images ads (Kaid, Chanslor, & Hovind,

1992).

The notion that image ads might facilitate recall originated from some previous research

about message complexity. First, memory of image and issue ads was influenced by complexity

(Natchez & Bupp, 1968; Schleuder, 1990). Second, image ads by nature were less complex than

issue ads (Natchez & Bupp, 1968) because image ads contained fewer visual and verbal cues in a

commercial per time unit than issue ads (Schleuder, 1990). Third, simple commercials were

found to facilitate mental processing and lead to more attention and better memory than complex

commercials (Thorson, Reeves, & Schleuder, 1985, 1987); therefore, it should be easier to

process image ads than issue ads and to recall content from image ads than from issue ads

(Natchez & Bupp, 1968; O'Keefe & Sheinkopf, 1974). Some researchers also argued that an ad,

image or issue, could be made simpler or more complex by including fewer or more cues, and

complex issue ads might work particularly well for an unknown candidate since voters might be

more curious about the new candidate' s issue stands (Schleuder, 1990).

In one empirical study designed to test those ideas, Schlender (1990) showed 22

participants four image ads, two for George Bush, a well-known presumed incumbent in the

1988 presidential campaign, and the other two for Michael Dukakis, a nationally unknown

challenger in the same campaign, and another 22 participants four issue ads, two for each

candidate respectively. The two image or issue ads for each candidate were pre-coded as high or

low complexity. In other words, ad type (issue vs. image) was a between-subjects variable, and

complexity (simple vs. complex) and candidate type (well-known Bush vs. unknown Dukakis)

were two within-subj ects variables. Mental processing was the dependent construct and was









indicated by three variables. The first variable was attention to the commercials and it was

measured with the participants' secondary task reaction time to periodic tones. Longer reaction

time meant more attention to the commercials. The second variable was memory of visual cues

and it was measured with a frame recognition test regarding scene change, dissolve, edit, camera

pan or zoom, person and object movement, and superimposed image (Schleuder, 1990). Larger

numbers of correct recognition meant better visual memory. The third variable was memory of

verbal cues and it was measured with a multiple-choice test about the verbal content including

music, voice-over and natural sound, and macro proposition or idea (Schleuder, 1990). Higher

scores on the test meant better verbal memory.

Schlender (1990) provided some evidence that image ads were associated with better

mental processing. It took the participants watching the image ads longer to respond to the

secondary task than those watching the issue ads, and therefore, the image ads attracted more

attention than the issue ads. Verbal cues were better remembered for image ads than for issue

ads. Nevertheless, memory of visual cues was not influenced by ad type.

The author also explored some interaction effects, but there was lack of consistency in the

effects on the three indicators of mental processing. For instance, complexity had no effect on

attention for both types of the Bush ads; but low complexity outperformed high complexity in

producing more attention to the Dukakis image ads, and high complexity outperformed low

complexity in producing more attention to the Dukakis issue ads. By sharp contrast, complexity

had no effect on visual memory for both types of the Dukakis ads; but low complexity

outperformed high complexity in producing higher recall of the visual cues of the Bush issue ads,

and high complexity outperformed low complexity in producing higher recall of the visual cues

of the Bush image ads. Finally, complexity had no effect on verbal memory for a particular type









of the Bush or Dukakis ads; but low complexity outperformed high complexity in producing

higher recall of the verbal cues of the Bush ads, and high complexity outperformed low

complexity in producing higher recall of the Dukakis ads.

Inconsistencies also existed in other comparisons. Candidate type had no effect on

attention or verbal memory but some effects on visual memory. The Dukakis ads outperformed

the Bush ads in producing higher recall of the visual cues, and the Dukakis issue ads

outperformed the Bush issue ads in producing higher recall of the visual cues.

In the absence of any explanation about the inconsistencies in the mental processing, those

interaction effects could just be experimental artifacts, or effects caused by the ordering of the

ads rather than the characteristics of the ads. In fact, the author was aware of ordering effects

when placing four ads one after another within one condition and tried to make the study more

realistic by embedding the political ads into a situation comedy with consumer ads. Therefore,

Schlender (1990) created two more between-subj ects variables, each with three levels for both

the order of commercial breaks and the position of each political ad in each break. With those

two additional between-subj ects variables, there were 18 between-subj ects conditions and two or

three participants per condition with a total sample size of 44 participants. This small sample

size per condition made any comparison among individual ads problematic. The author did not

report any analyses about the orders although the orders were part of the study.

The author' s research ideas might be theoretically sound, but the way to test the research

ideas was methodologically flawed. Additionally, the manipulation of visual and verbal

complexity was not indeed successful. The inherent complexity of image and issue ads had a

more profound effect on mental processing than manipulated complexity of visual or verbal cues

because visually/verbally complex image ads still attracted more attention than visually/verbally










simple issue ads. The only conclusion from the study was that image ads attracted more

attention and produced higher recall of verbal cues than issue ads. But even that finding was

limited in identifying the exact impact of image and issue ads because it remained unclear

whether the verbal cues focused on image or issue or whether the focus varied by ad type.

Interestingly enough, those above-mentioned methodological issues did not exist in the

study that favored issue ads (Kaid et al., 1992). The main purpose of the study was to test how

the type of TV programs and the type of political ads would influence advertising effectiveness

including recall of advertising content. Three types of ads (issue, image, and negative ads from

Kit Bond's campaign for the U.S. Senate in Missouri) were inserted into three types of TV

programs (situation comedy, drama, and local news) to create nine experimental conditions. A

total of 283 participants were recruited on university campus and from local communities and

were assigned to the nine conditions. Memory was measured with cued recall rather than

recognition tests and the participants were instructed to complete "open-ended questions about

the issues and candidate characteristics stressed in the political advertisement" (Kaid et al., 1992,

p. 311).

The open-ended issue question was worded: "What were the maj or ideas or issues stressed
in the political advertisement you saw?" The image question read: "What personal
characteristics of Kit Bond were stressed in the political advertisement you saw?" Recall
on each item was measured by counting the number of words, phrases, or sentences which
constituted distinguishably different ideas. (p. 312)

The results indicated that there was no main effect of program type or interaction effect of ad

type by program type on recall, but ad type had a significant main effect on recall. The issue ad

outperformed both the image and negative ads in generating higher number of recalled issues and

personal characteristics (Kaid et al., 1992).

In summary, image ads were found more memorable than issue ads when measured with

recognition of verbal cues in both types of ads (Schleuder, 1990), but issue ads were found more









memorable than image ads when measured with recall of issues and personal characteristics in

both types of ads (Kaid et al., 1992). The latter Einding had more important implications for

political advertising research because the chain reaction model (Lavidge & Steiner, 1961;

Shama, 1976) posited an impact of issue and image content rather than verbal or visual content

on candidate evaluation and voting behavior. In the current study, the cued recall questions of

issue and image content (Kaid et al., 1992) are used to measure memory.

The issue and image ads in the 1992 study (Kaid et al., 1992) were positive ads because the

issue ad "dealt with the budget deficit explicitly, with the candidate espousing his views on the

subj ect" (p. 3 10) and the image ad "similarly discussed the budget deficit, but in more human

terms, reflecting the image or persona of the candidate" (p. 310). But it was hard to determine

whether the negative ad in the same study was a negative issue or image ad because it "used the

budget deficit, but this time as an opportunity to paint the candidate's opponent in a negative

light" (p. 310). The positive issue and image ads did not really have their negative counterparts,

and there was only one tentative conclusion that favored positive ads in generating more recall of

issue and image content.

Nevertheless, there was a substantial body of research that specifically addresses the

effects of negative or attack ads and compares negative ad exposure to positive ad exposure

(Kaid, 2004). Much of the conceptual framework about negative advertising came from

evolutionary psychology (A. Lang, 1991). On the one hand, humans, through evolution, were

programmed to assign more attention to negative stimuli for survival (Zajonc, 1984a), and the

increased attention led to better memory. On the other hand, humans, by nature, tended to

approach positive stimuli and avoid negative stimuli (Fiske & Taylor, 1984), and such a

tendency led to increased attention and better memory of positive stimuli. However, there was









no decisive evidence to support either of the two explanations. The actual impact of the valence

of ad, positive or negative, on memory became increasingly perplexing when researchers

operationalized memory in different manners.

Inconsistent results were found in two studies on radio ads (Geer & Geer, 2003; Shapiro &

Rieger, 1992). In one study (Shapiro & Rieger, 1992), two image ads (honesty and leadership)

and two issue ads (rent control and financial aid) were created in both positive and negative

versions and were used for two races. Each participant listened to either two image or two issue

ads for the two races, and one ad was positive and the other ad was negative. Eight experimental

conditions were created and 106 college student participants were recruited for the eight

conditions.

Memory was measured with cued recall of issues and the participants were instructed to

"list all the specific ideas or arguments you can remember from the ad" (Shapiro & Rieger, 1992,

p. 140). The results indicated that negative ads outperformed positive ads in generating higher

number of recalled arguments, but the order of the ads, image/issue content, and interaction of

image/issue content by negative/positive content had no influence on memory (Shapiro &

Rieger, 1992).

In another radio ad study (Geer & Geer, 2003), four versions of party ads instead of

candidate ads were created, namely negative/positive Republican ads and negative/positive

Democratic ads. All the ads were issues ads and contained the same issues. Furthermore, two

different copies of the four versions were created to minimize unusual effects of one particular

ad, which led to eight experimental conditions. A total of 121 college students were assigned to

the eight conditions.









Memory was measured with free recall and the participants were instructed to "write as

completely and accurately as possible the political ad that you heard a few minutes ago" (Geer &

Geer, 2003, p. 86). The recall was further "divided into 'idea units' ... defined as meaningful

divisions in the ads ... A correct memory unit was the presence in the recall protocol of an idea

unit that had been present in the ad" (p. 86). Because party sponsorship was found to have no

influence on memory, the party sponsorship variable was excluded and comparisons were made

between positive and negative ads only. Contrary to finding in the above-mentioned study

(Shapiro & Rieger, 1992), the results indicated that negative ads did not differ from positive ads

in generating the number of memory units but positive ads outperformed negative ads in

generating more correct memory units (Geer & Geer, 2003).

By sharp contrast, studies on TV ads (A. Lang, 1991; Newhagen & Reeves, 1991)

produced consistent results and suggested that negative ads were more memorable than positive

ads. In one study (A. Lang, 1991), four positive and four negative ads were used to create four

orders of ads and 67 college student participants were assigned to one of the four orders.

Memory was measured with three methods, free recall, cued recall, and recognition with

multiple-choice test.

For the free recall, the participants were instructed to "write down everything they could

remember about the commercials they had seen" (A. Lang, 1991, p. 228). For the cued recall,

the participants were "given ... a list of the candidates in the commercials. They were told that if

this list caused them to remember any commercials they had not previously recalled they should

now write everything they could remember about those commercials" (p. 228). For the multiple-

choice test, the participants were given "12 questions about each commercial ... 6 questions










about information available only in the audio track of the commercial and 6 questions about

information available only in the video portion ... of the commercial" (p. 228).

The free and cued recall was further coded into visual and verbal ideas. "Visual ideas were

defined as any single image recalled from the commercial or any global visual description of the

commercial. Verbal ideas were defined as any single thought including names, states, office the

candidate was running for, or global audio description" (p. 229). The results indicated that

negative ads outperformed positive ads in generating higher number of ideas as indicated in the

free recall and more accurate audio-visual memory as indicated in the multiple-choice test.

In another study (Newhagen & Reeves, 1991), a total of 28 ads were shown to 30 adults

recruited from the local community. The 28 ads included not only negative/attack and

positive/support ads but also comparative ads that were "made to lessen the image of the

opponent and promote a positive image of the sponsor as an alternative" (p. 198) within one ad.

Those 28 ads were rotated in three orders.

Memory was measured with a recognition test. After watching the 28 ads, the participants

were shown a second videotape "that contained pictures and sound bites from the

advertisements" (p, 205) and were instructed to "press a 'yes' or a 'no' button to indicate

whether the material was from the messages they had viewed" (p. 205). Out of the 28 ads, 18

ads were retained for the analysis with six ads for the three categories. The results indicated that

negative ads outperformed both comparative and positive ads in generating more accurate

recognition. Nevertheless, there were obviously some methodological issues with that study.

First, three orders were not enough to rotate the 28 ads. Second, with three orders and 30

participants, there was only 10 participants each order, which limited the value of order










manipulation. Third, excluding part of the ads based on some preliminary analysis reduced the

credibility of the research.

In summary, studies on memory differed from each other in the measurement of memory

and the difference in measurements led to mixed results. Two studies (Newhagen & Reeves,

1991; Schleuder, 1990) used recognition tests, and the participants were instructed to judge

whether certain verbal, audio, and visual cues were present in the ads they had watched.

Although those two studies provided evidence that image ads outperformed issue ads in

generating better recognition of verbal cues (Schleuder, 1990) and negative ads outperformed

positive ads in generating better recognition of audio-visual cues (Newhagen & Reeves, 1991),

there were a number of methodological issues with the two studies. Furthermore, recognition

tests were criticized for overlooking "materials in memory not included in the recognition

alternatives" (Geer & Geer, 2003, p. 72) and inflating "the rate of accuracy due to successful

guessing" (p. 72). Recall was recommended to measure memory because it "allows for the full

range of subj ect' s memory to be expressed" (p. 72).

Four studies (Geer & Geer, 2003; Kaid et al., 1992; A. Lang, 1991; Shapiro & Rieger,

1992) used recall, but they differed in the type of ads being recalled, the type of recall being used

and the type of cues being recalled. In one study (Kaid et al., 1992), cued recall was used to

measure memory of issues and candidate personality in TV ads, and the results indicated that

positive issue ads outperformed positive image ads and negative ads in generating higher number

of recalled issues and personal characteristics. In another study (Shapiro & Rieger, 1992), cued

recall was used to measure memory of issues in radio ads, and the results indicated that negative

ads outperformed positive ads in generating higher number of recalled issues. In the third study

(Geer & Geer, 2003), free recall was used to measure idea units in radio ads, and the results









indicated that negative ads did not differ from positive ads in generating the number of idea units

but positive ads outperformed negative ads in generating more correct idea units. The problem

with the third study was the ambiguity of idea units because it remained unclear whether the idea

units were about issues or image. In the last study (A. Lang, 1991), cued recall, free recall, and

recognition test were all used to measure verbal, audio, and visual cues in TV ads, and the results

indicated that negative ads outperformed positive ads in generating higher number of verbal,

audio, and visual cues as indicated in the free recall and more accurate audio-visual memory as

indicated in the recognition multiple-choice test. Again, it remained unclear whether the cues

and memory were about issue or image.

This study uses the chain reaction model of advertising effects (Lavidge & Steiner, 1961;

Shama, 1976) and knowledge of issue and image was an important link in the model. Research

findings about verbal, audio, and visual cues did not specify whether the cues focused on issue

and image; therefore, the findings were not very relevant to the current study. There was strong

evidence that issue ads outperformed image ads in generating higher number of recalled issues

and personal characteristics (Kaid et al., 1992). Meanwhile, there was certain evidence that

positive ads outperformed negative ads in generating higher number of recalled issues and

personal characteristics (Kaid et al., 1992), but it remained unclear whether the negative ad in the

study (Kaid et al., 1992) focused on issue or image. The exact impacts of issue/image and

positive/negative ads are further explored in the current study.

Effects of Political Advertising: Candidate Electability

According to the chain reaction model of advertising effectiveness (Lavidge & Steiner,

1961; Shama, 1976), increased knowledge about a candidate was just the first step of the reaction

chain, and the increased knowledge or memory in turn contributed to the formation of voters'

opinion on the candidate or the candidate's electability. Quite a few studies explored the impact










of political advertising content on candidate electability. Issues were more socially acceptable

for political advertising than image (Roddy & Garramone, 1988; Thorson et al., 1991); therefore,

issue ads led to more positive candidate evaluation than image ads (Kaid & Sanders, 1978).

Positive ads also led to more positive candidate evaluation than negative ads (Ben-Ur &

Newman, 2002; Hill, 1989; Kahn & Geer, 1994).

In one study (Thorson et al., 1991), the researchers selected four TV ads for each of four

candidates, i.e., one positive issue ad, one positive image ad, one negative issue ad, and one

negative image ad for each candidate. The researchers also created 16 orders of 13 ads, with one

ad from each of the four candidates, one filler political ad, and another eight filler consumer TV

ads. With a sample size of 161 participants, each order had about ten to eleven participants.

In that study, memory was also examined and was measured with a recognition test. The

participants were asked to indicate whether certain facts really contained in the ads. Those facts

included "issues (or image qualities) ... office for which the candidate was running, the state,

and other people referred to in the ad" (p. 474). Candidate electability was examined with four

variables. The first variable was ad attitude and was measured with eight 9-point bipolar

adjective pairs. The eight items were believable ... unbelievable; persuasive ... unpersuasive;

like ... dislike; pleasant ... unpleasant; truthful ... deceptive; accurate ... inaccurate; ethical ..

unethical; good ... bad" (p. 474). The second variable was attitude toward the candidate and was

measured with nineteen 9-point bipolar adj ective pairs. Through a factor analysis, seven of the

19 adj ective pairs were used to indicate candidate character, and the remaining 12 items were

used to indicate candidate ability. The seven items about candidate character were "warm ..

cold; emotional ... unemotional; friendly ... unfriendly; compassionate ... uncompassionate;

kind ... unkind; honest ... dishonest; and sincere ... insincere" (p. 476). The 12 items about










candidate ability were "stimulating ... boring; articulate ... inarticulate; weak ... strong; stupid

.. smart; persuasive ... unpersuasive; experienced ... inexperienced; and active ... passive" (p.

476). The third variable was the pleasure dimension of emotional response (Russell, 1978) and

was measured with six 9-point bipolar adjective pairs. The six items were "hopeful ..

despairing; pleased ... annoyed; happy ... unhappy; relaxed ... bored; satisfied ... unsatisfied;

contented ... melancholic" (Thorson et al., 1991, p. 475). Fourth variable was the arousal

dimension of emotional response (Russell, 1978) and was measured with another six 9-point

bipolar adjective pairs. The six items were "stimulated ... relaxed; wide awake ... sleep; excited

.. calm; frenzied ... sluggish; jittery ... dull; aroused ... unaroused" (Thorson et al., 1991, p.

475). Voting intent was measured with a 9-point bipolar adj ective scale ranging from "very

likely" to "very unlikely." Additionally, familiarity was measured with a 7-point bipolar

adj ective scale ranging from "familiar" to "unfamiliar," and it turned out that all the four

candidates were unfamiliar to the participants.

In the analysis, political ads from different candidates and in different orders were

collapsed into either image or issue category. There was no difference in the memory of ad

content between image and issue ads. There was also no difference in arousal between image

and issue ads. Nevertheless, issue ads outperformed image ads in producing more positive ad

attitude, overall candidate attitude, candidate character, candidate ability, voting intention, and

higher pleasure.

The ads were also collapsed into either positive or negative category. Positive ads

outperformed negative ads in generating better memory. Positive ads also outperformed

negative ads in generating more positive ad attitude, overall candidate attitude, candidate

character, higher pleasure, and lower arousal. There was no difference in candidate ability or









voting intention between positive and negative ads. There was also no interaction effect of

image/issue by positive/negative on memory, candidate evaluation, or voting intention.

However, the study itself had a few problems. First, similar to the Schlender study (1990),

order was part of the experimental design, but no tests on ordering effects were performed. In

each order, the participants watched all the four types of ads; but the researchers used ANOVA

as if the ad type had been a between-subj ects variable, and they did not report the degree of

freedom of ANOVA. Second, it remained unclear how a 9-point pleasure and arousal scales

resulted in mean scores ranging from -.83 to .20 (Table 1, p. 478). The researchers did not

provide any explanation about the difference in scoring. Third, although there were significant

differences between issue and image ads and between positive and negative ads in candidate

evaluation measures, the scores for those ads were well above the mid-point of their

corresponding 9-point scales and the actual differences ranged from .23 to .39 (Table 1, p. 478).

The statistical significance could be inflated because the scores were collapsed into either the

image/issue or the positive/negative category and each paired comparison involved the same 161

participants. In other words, that study did not provide convincing evidence about the true

difference in candidate evaluation between issue and image ads or between positive and negative

ads for unknown candidates. Finally, it remained unclear what kind of emotional response was

measured. The researchers did not indicate whether it was response to the ads or to the

candidates.

In another experimental study, Kahn and Geer (1994) tried a different approach. They

used between-subj ects design and compared candidate evaluation toward four types of TV ads,

i.e., issue/image by positive/negative ads, from one unknown candidate's political campaign.

Besides creating the first four experimental conditions in which the participants only watched









one type of ads, the researcher also created another five conditions in which the participants

watched 1) positive image and issue ads supporting the candidate, 2) negative image and issue

ads sponsored by the same candidate, 3) a positive image ad of the candidate and a negative

image ad sponsored by his opponent, 4) a positive issue ad of the candidate and a positive issue

ad of his opponent, or 5) a negative image ad sponsored by the candidate and the negative image

ad sponsored by his opponent. With a sample size of 303 participants, each condition had about

24 to 39 participants.

The authors also used candidate competence and characteristics to assess candidate

electability. Candidate competence was measured with a 6-point adjective scale ranging from

very competent to very incompetent. The same scale was used to assess competence in three

issues including education, health, and economy. Education was the focus of the issue ad.

Health and economy were added because the researchers believed that those two were the central

concerns for most political campaigns. Candidate traits was measured with a 4-point scale

ranging from "a great deal" to "not at all." The same scale was used to assess four traits, i.e.,

trustworthy, strong leadership, hardworking, and knowledgeable. The participants were also

asked to indicate their intention to vote for the candidate on a 4-point scale ranging from "very

likely" to "not at all likely."

There was certain level of inconsistency between evaluation measures. Except for the

negative image condition of the first four conditions, the evaluation scores of issue competence

and personality were all above their corresponding scale midpoints by .01 to 1.01 point (Table 2,

p. 103) and were slightly favorable to the candidate. Nevertheless, the overall evaluation scores

were not significant across the four conditions.










Although interaction effects of issue/image by positive/negative were testable with the

experimental design, there was certain ambiguity about the existence of those effects. On the

one hand, the authors did not mention including an issue/image by positive/negative interaction

term in their ANOVA. On the other hand, their comparison of the first four conditions seemed

to suggest the existence of some interaction effects. Positive ads outperformed negative ads in

producing higher evaluation toward the candidate's issue competence in education and economy

and the candidate's personality as a strong leader. When comparisons on the same issue

competence and personality were made within positive ads, issue and image ads did not differ in

the evaluations. But when the same comparisons were made within negative ads, issue ads

outperformed image ads in generating higher candidate evaluations.

The authors asked the participants to evaluate the candidate on three issues. The results

regarding education seemed reasonable since education was the focus in the issue ads. It

remained unclear why there was significant difference in response to economy but no difference

to health after the participants watched an issue ad about education. The authors also asked the

participants to evaluate the candidate on four personality traits. There was significant difference

only in the evaluation of the candidate' s leadership. It remained unclear why the content had no

effect on the other three traits. Finally, although positive ads worked much better than negative

ads in enhancing issue competence and personality evaluation, the negative issue ad in fact

produced significantly higher voting intention than the other three ads.

The authors showed participants two ads in another five conditions. Comparisons were

made between the impact of one ad and the cumulative impact of two ads. Two positive ads

together produced higher evaluation toward issue competence in education and personality as

knowledgeable, higher and above midpoint overall candidate evaluation, and higher intention to









vote than one positive issue ad or one positive image ad only. Nevertheless, exposure to more

positive ads did not enhance candidate evaluation toward issue competence in health and

economy and personality trait as trustworthy, hardworking, and strong leadership.

Comparisons were also made for negative ads. Two negative ads sponsored by the

candidate produced lower evaluation toward personality as trustworthy than one negative issue

ad or one negative image ad only. Those two negative ads together also resulted in lower

intention to vote for the candidate than one negative issue ad. Surprisingly, those two negative

ads together in fact resulted in higher intention to vote for the candidate than one negative image

ad. No additional differences were found with regard to the other aspects of the candidate issue

and personality evaluation.

The authors also showed some participants both a positive image ad of the candidate and a

negative image ad sponsored by the candidate's opponent. The second ad from the opponent had

no effect on the evaluation toward and intention to vote for the candidate. They also showed

other participants both a positive issue ad of the candidate and a positive issue ad of the

candidate's opponent. Again the second ad from the opponent had no effect on the evaluation

toward and intention to vote for the candidate. Nevertheless, when a negative image ad

sponsored by the candidate's opponent was shown after a negative image ad sponsored by the

candidate, the candidate's evaluation in terms of his personality as trustworthy, strong

leadership, and knowledgeable was significantly reduced than when the negative image ad

sponsored by the candidate was shown only. In other words, showing the participants the

candidate and the candidate's opponent attacking each other on image with two ads was the most

likely to lower the evaluation of the candidate than showing the participants only one negative

image ad sponsored by the candidate.









In summary, some researchers (Thorson et al., 1991) found that issue ads generated more

positive candidate electability than image ads for unknown candidates, and other scholars (Kahn

& Geer, 1994) argued that issue ads did not differ from image ads in influencing candidate

electability for unknown candidates. However, previous research (Kahn & Geer, 1994; Thorson

et al., 1991) consistently indicated that positive ads outperformed negatives ads in generating

more positive candidate electability. Furthermore, there was evidence (Kahn & Geer, 1994) that

negative issue ads outperformed negative image ads in generating more positive candidate

electability, and showing negative image ads sponsored by both a candidate and the candidate's

opponent was more likely to lower the electability of the candidate than showing only one

negative ad sponsored by the candidate.

The notion of electability is conceptually similar to the construct of attitude. A number of

attitude studies employ a tripartite framework, which includes cognition, affect, and conation as

the three components of attitude (Greenwald, 1989; Morris, Woo, Geason, & Kim, 2002). The

cognitive element includes beliefs stored and retrieved from memory (Hoyer & Maclnnis, 2004;

McGuire, 1989). The affective element includes evaluative judgment and personality assessment

(Aaker, 1997; McGuire, 1989). The conative element includes behavioral intention (Fishbein &

Ajzen, 1975; Fishbein & Ajzen, 1980; McGuire, 1989). Those three components have been used

in previous research on political advertising, e.g., beliefs about issue competency (Kahn & Geer,

1994), candidate personality (Kahn & Geer, 1994; Thorson et al., 1991), and voting intention

(Kahn & Geer, 1994). In the current study, those three variables will be used to create a

composite measure of candidate electability.

Summary of the Effects of Political Advertising

A few empirical studies (Geer & Geer, 2003; Kahn & Geer, 1994; Kaid et al., 1992; A.

Lang, 1991; Newhagen & Reeves, 1991; Schleuder, 1990; Shapiro & Rieger, 1992; Thorson et










al., 1991) were reviewed in great detail. Those studies did not provide consistent evidence of the

effects of political advertising on memory and candidate electability. The inconsistency could be

caused by different operationalization of memory and candidate evaluation, different content of

ads, and certain methodological problems in some studies. There was lack of specification of the

exact content being measured, for example, the focus of visual, audio, and verbal cues (A. Lang,

1991; Newhagen & Reeves, 1991; Schleuder, 1990), and the target of emotional response

(Thorson et al., 1991).

Additionally, based on the chain reaction model of advertising effects (Lavidge & Steiner,

1961; Shama, 1976), memory and candidate electability were not two types of separate effects.

Memory of an ad could influence voters' opinion on the electability of the candidate advertised

in the ad. A few measures were developed to measure memory of issues and candidate

characteristics (Kaid et al., 1992) and candidate competence, personal traits, and voting intention

(Kahn & Geer, 1994). But those measures were not incorporated into one study to test the chain

reaction model. This missing link in previous research is investigated in the current study with a

path analysis. More details about the path analysis are in Chapter 3.

Informational and Transformational Advertising

Previous research identified five antecedents that influenced the use of informational or

transformational advertising (Swaminathan, Zinkhan, & Reddy, 1996). The relevance of those

five antecedents to political advertising was assessed in a previous study (Shen & Kim, 2006).

Antecedent 1: Newness of the Product

Consumers often have a need to search for information to identify and evaluate a new

product category (Cohen & Basu, 1987). But when consumers became well informed, their

information need declined and their evaluation of the product category was largely based on

actual experience of the product (Swaminathan et al., 1996). The same idea was also found in









some discussion about the relationship between the use of informational or transformational

advertising and product life cycle (Porter, 1985). In the early stages of a consumer product,

informational advertising was recommended to communicate factual information of the product;

whereas in the late stages, transformational advertising was recommended to communicate use

experience of the product (Porter, 1985). Newness of product was relevant to political

advertising (Shen & Kim, 2006). An unknown challenger, or a candidate in the early stage of his

or her political life, tended to build awareness through informational advertising, whereas a well-

known incumbent, or a candidate in the late stage of the political life, tended to capitalize voters'

positive experience with his or her first term through transformational advertising (Schleuder,

1990).

Antecedent 2: Level of Risk

In the context of consumer behavior, risk referred to the uncertainty encountered by

consumers when they were estimating the social and economic consequences of a purchase

decision (Swaminathan et al., 1996). Risk was associated with product newness and information

search (Hoyer & Maclnnis, 2004). A new product was perceived riskier than a familiar product;

therefore, if consumers had to make a decision between the two, one way to make a less risky

decision was to make an informed decision through adequate information search (Darke,

Freedman, & Chaiken, 1995). Additionally, decision-making literature points out that risk was

often associated with involvement (Hoyer & Maclnnis, 2004).

Antecedent 3: Product Involvement

In the context of consumer behavior, involvement referred to the extent to which a product

had important consequences for consumers to form an informed opinion and to justify their

purchase decisions (Chaiken, 1980). Expensive products resulted in both strong perception of

risk and high involvement with the products (Hoyer & Maclnnis, 2004). Voters highly involved









with a political election might believe that their votes would have a significant impact on their

society and tended to search for adequate information to make a well-informed decision

(Nimmo, 1999). Therefore, for those voters, informational advertising might be the only

appropriate approach. In fact, there was evidence that for voters who were interested in or

involved with political campaigns, their candidate evaluation was primarily determined by issue

advertisements rather than image advertisements (Christ, Thorson, & Caywood, 1994).

Antecedent 4: Product Conspicuousness

In the context of consumer behavior, product conspicuousness referred to the intended

communication role played by certain consumption decisions (Belk, Balm, & Mayer, 1982). In

other words, consumers expressed themselves through purchasing and using products that were

congruent with their self-images (Swaminathan et al., 1996). Therefore, the purchase decision

was based on real or upcoming experience of using the product or the transformational aspect

rather than the hard facts or the informational aspect of the product. In the context of political

advertising, a candidate could use transformational advertising to reinforce support from voters

who had positive experience with the candidate and felt well connected to the candidate (Shen &

Kim, 2006).

Antecedent 5: Product versus Service

In the context of consumer behavior, intangibility was a fundamental difference between

product and service (Swaminathan et al., 1996). Because consumers could not perceive service

in the same manner in which consumers could sense products (Zeithaml & Bitner, 1996),

transformational advertising was used to present service offering with real-person experience of

the service (Swaminathan et al., 1996). At first glance, political office might be closer to the

concept of intangible service than tangible product, which made transformational advertising a

better match to political campaigns (Shen & Kim, 2006).









However, research on intangibility suggested that although service itself might be highly

intangible, the service delivery system was well tangible (Mittal, 1999). Similarly, although the

political service a candidate would provide may be highly intangible, the candidate himself or

herself, as the core service delivery system, was well tangible (Shen & Kim, 2006). Basic facts

about the candidate were crucial for voters to make an informed decision (Nimmo, 1999).

Therefore, informational advertising was still an appropriate approach to political campaigns

although political campaigns essentially aimed to sell candidates' political service (Shen & Kim,

2006).

Typology of Informational/Transformational Strategies

Previous research identified five informational strategies and four transformational

strategies in consumer advertising (Laskey, Day, & Crask, 1989). Again those items were

examined in the context of political advertising for relevance assessment in a previous study

(Shen & Kim, 2006).

The first informational item, comparative, referred to "competition explicitly mentioned"

(Laskey et al., 1989, p. 38). It was relevant to political advertising because in political

advertising, especially in negative political advertising, explicit mention of opponent was

basically a common practice (Shen & Kim, 2006). The second item, preemptive, referred to

"testable claim of superiority based on an attribute or benefit" (Laskey et al., 1989, p. 38). It was

also relevant because in political advertising, there were often statements about a candidate's

previous performance in office and/or issue stands, which should be testable or verifiable and

could help indicate that candidate's superiority to other candidates (Shen & Kim, 2006).

Nevertheless, one risk of using the preemptive strategy was that the same record could also be

framed negatively and be used in an opponent's attack ads (Nimmo, 1999). In fact,

negative/attack ads focused more on issues than on image (Johnston & Kaid, 2002; Kaid &









Johnston, 2001; West, 1993). Certain actions based on inoculation theory of persuasion

(Anderson & McGuire, 1965), such as warning of imminent attack from the opponent followed

by a weak version of such an attack, should be taken if some political records were likely to

cause controversies. The third item, hyperbole, referred to "un-testable claim of superiority

based on an attribute or benefit" (Laskey et al., 1989, p. 38). It was also relevant because in

political advertising, there were also superiority claims about new initiatives to be implemented

but those new initiatives were largely un-testable at the moment they were proposed (Shen &

Kim, 2006). The fourth item, unique selling proposition, referred to "explicit claim of

uniqueness" (Laskey et al., 1989, p. 3 8) and involved heavy use of the word or concept of "only"

(p. 38). It was, however, not really relevant because in the U.S. elections, the influence of

independent candidates was so limited that the competition was primarily between the two

candidates from the Republican Party and the Democratic Party respectively (Devlin, 2001).

Because there were just two candidates, any comparisons between the two automatically

involved the concept of "only" (Shen & Kim, 2006). Including this strategy might lead to

possible confusion with other informational items that emphasized the superiority of a candidate,

for example, pre-emptive and hyperbole (Shen & Kim, 2006). Generic, the last informational

item, referred to "focus on product class" (Laskey et al., 1989, p. 38). It was also not relevant

because there was no such a concept of generic candidate in political advertising (Shen & Kim,

2006).

The first transformational item, user image, focused on "the users of a brand and their

lifestyles" (Laskey et al., 1989, p. 39). It was relevant because some political advertisements did

show how the life of certain type of voters or "users" of a political candidate' s service would be

influenced either positively or negatively by their decision to choose a candidate (Shen & Kim,










2006). The second item, brand image, focused on "the image of the brand itself in an attempt to

convey a brand 'personality'" (Laskey et al., 1989, p. 40). It was also relevant because in

political advertising, and it was again a common practice to emphasize the personality or

characteristics of a candidate (Shen & Kim, 2006). The third item, use occasion, focused on "the

experience of using the brand, or on those situations where use of the brand is most appropriate

.. [or] the 'fit' between the brand and the situation; not the fit to a particular user profile"

(Laskey et al., 1989, p. 40). It was also relevant because some political advertisements did show

scenarios that suggested the appropriateness to choose a candidate for a particular issue rather

than for a particular type of voters (Shen & Kim, 2006). Generic, the last transformational item,

again focused on "the product class rather than on a particular brand" (Laskey et al., 1989, p. 40).

This item was not relevant because there was no such a concept of generic candidate (Shen &

Kim, 2006).

Summary

In summary, the antecedents to the use of information or transformation for consumer

advertising were well applicable to political advertising. Furthermore, three out of the five

informational strategies and three out of the four transformational strategies were relevant to

political advertising. In fact, the three informational strategies and three transformational

strategies were used to code the issue and image content in political advertisements (Shen &

Kim, 2006). Issue political ads were found to contain more informational strategies than

transformational strategies, and image political ads were found to contain more transformational

strategies than informational strategies. Nevertheless, content analyses on issue and image in

political advertising might be of little use if there was a lack of reliable and valid methods to

measure voter' s perception of the content and identify the impact of the perception on candidate

evaluation (Shen, 2007). In the current study, the Puto and Wells (1984)









informational/transformational survey scale (hereinafter referred to as "the Scale") is used to

better understand the perception and impact in question.

Survey Scale of Informational/Transformational Advertising

The Scale had 23 items. Eight of them measured the perception of informational cues

(hereinafter referred to as "PIFC") in an advertisement and were developed from previous

research on the informative function of advertisements (Aaker & Norris, 1982; Resnik & Stern,

1977). The other 15 items measured the perception of transformational cues (hereinafter referred

to as "PTFC") and were developed from viewer response profie measures (Schlinger, 1979) and

Needham, Harper, and Steers Advertising's communication measures. The Scale was by far the

most popular measure of informational/transformational advertising and was used for more than

two decades (Menon, Goodnight, & Wayne, 2006), although survey scales of this type were

criticized for being limited in capturing transformational experience and should be replaced by

open-ended proj ective techniques with pictures or narratives (Aaker & Stayman, 1992).

In the original validation study, Puto and Wells (1984) first coded 400 TV commercials

based on the definitions of informational and transformational advertising. Out of those 400

commercials, they selected Hyve primarily informational commercials and eight primarily

transformational commercials. Next, 130 college students watched the 13 commercials and

responded to the survey scale to report PIFC and PTFC of the commercials. The average

responses of the 13 TV commercials were analyzed to assess the performance of the Scale. A set

of Cronbach' s reliability tests indicated that the eight informational survey items reached an

average reliability of .73 and the 15 transformational survey items reached .88. A set of t-tests

was used to compare the difference between each commercial's PIFC and PTFC scores. The

difference was significant for all the commercials and the higher score was in complete










agreement with the prior coding as whether the commercial was primarily informational or

transformational.

Unfortunately, the validation study itself was not flawless. The participants watched all

the 13 commercials and responded to the same scale 13 times, but the researchers used t-tests

rather than repeated measures for this within-subjects design. Moreover, since Cronbach's

reliability test was known to favor scales of larger numbers of items even with the same degree

of inter-item correlations (Hair, Anderson, Tatham, & Black, 1995), the Scale was likely to be in

favor of the transformational items because those items almost outnumbered the informational

item by 2 to 1. Meanwhile, because the Scale had six points from "Strongly Agree" to "Strongly

Disagree," the participants were forced to make a sided choice and could not indicate a neutral

stand. Finally, the participants were also asked to report prior exposure and evaluation of the

commercials, but the relationship between those measures and the Scale was not analyzed in that

validation study.

In one recent study of using the Scale for political ads (Shen, 2007), a few changes were

made to the Scale; for example, "brand" was replaced with "candidate" and "company" with

"party," and a mid-point category "neither agree nor disagree" was added to the Scale to form a

7-point scale, ranging from "strongly disagree" to "strongly agree." That study indicated that the

PIFC score of an issue ad was significantly higher than its PTFC score, whereas the PTFC score

of an image ad was significantly higher than its PIFC score. The result provided additional

evidence to support the conceptual similarity between issue/image ads and

informational/transformational advertising. The Scale is also used in the current study with the

changes of the wording and scale points made in the previous study (Shen, 2007).









Emotion in Advertising

The informational content of advertising was also called the cognitive dimension of

advertising (Puto & Wells, 1984). Content of this dimension included verbal messages such as

rational and logical arguments as well as non-verbal information such as visual imagery, music,

and language variables that complemented the verbal messages (Wells, 1988).

The transformational content of advertising was also called the experiential dimension or

essentially the emotional dimension of advertising because transformation occurred only when

consumers explicitly associated certain emotional feeling to the use of an advertised brand (Puto

& Wells, 1984). Content of this dimension included emotional appeals to induce a wide variety

of human emotions (Wells, 1988).

Emotional appeals in advertising were often classified into positive and negative appeals

(Hoyer & Maclnnis, 2004). Empirical research has addressed the effects of positive appeals such

as happiness (Lazarus, 1991a), hope (MacInnis & de Mello, 2005), humor (C. Duncan & Nelson,

1985; Gelb & Pickett, 1983), pride (Lazarus, 1991a), and warmth (Aaker, Stayman, & Hagerty,

1986) as well as the effects of negative appeals such as anger (Averill, 1982; Frijda, 1986;

Lazarus, 1991a; Roseman, Antoniou, & Jose, 1996), fear (Boster & Mongeau, 1984; Brooker,

1981; Mewborn & Rogers, 1979; Witte, 1995), guilt (Boster et al., 1999), and irritation (Aaker &

Bruzzone, 1985).

Rational and Emotional Responses

Although the informational content produced rational responses whereas the

transformational content evoked emotional responses, the occurrence of the two types of

responses was in fact highly intertwined (Puto & Wells, 1984). On the one hand, rational

responses to an informational ad, such as message elaboration and belief formation, could be

substantially enhanced when appropriate emotional content was added (Deighton, Romer, &










McQueen, 1989; T. Duncan, 2002; Mano, 1997). Even in product categories where there were

substantial technological differences that provided a basis for a strong informational claim of

product superiority, advertisers still attempted to develop emotional bonds between consumers

and an advertised brand (Marshall, 2006).

On the other hand, the occurrence of emotional responses was not necessarily independent

of cognition (Lazarus, 1984; Zajonc, 1984b). There were two distinct routes to the production of

emotional responses (Andreasen, 1997; Izard, 1993; LeDoux, 1995; Teasdale et al., 1999). One

route "involves subcortical structures and generates emotional responses to simple perceptual

and associative aspects of stimuli. The second route involves cortical structures and generates

emotional responses reflecting the meanings of interpretations given stimuli" (Teasdale et al.,

1999, p. 210). In other words, "primitive, genetically prewired circuit entities of the visceral

brain ... circuits which appear to influence information processing in higher cognitive areas of

the cortex" (Panksepp, 1989, p. 8) was not involved in the generation of subcortical emotional

responses but cortical emotional responses. Subcortical emotional responses were well recorded

in the classical conditioning of fear reactions to audio or visual stimuli (LeDoux, Iwata,

Cicchetti, & Reis, 1988; LeDoux, Romanski, & Xagoraris, 1989) and they either interrupted the

cognitive focus of current attention or influenced the context for ongoing cognitive processes

(Simon, 1982). Cortical emotional responses were well recorded in studies that manipulated

semantic or schematic meanings of audio or visual stimuli (Decety & Chaminade, 2003;

Mitchell, Elliott, Barry, Cruttenden, & Woodruff, 2003) and the meanings directly mediated the

generation of emotion through cognitive processing (Teasdale, 1993; Teasdale & Barnard, 1993;

Teasdale et al., 1999). Because ads in general were highly associated with semantic and

schematic meanings (Hoyer & Maclnnis, 2004), cognitive theories of emotion, such as









attribution theories (Weiner, 1985) and appraisal theories (Lazarus, 1991b), were often used to

account for the role of knowledge structures in the processing of emotional responses.

The close interrelationship between cognition and emotion was more evident when the

hard facts or benefits of a product might be emotional in nature; for example, romantic feeling

was used to promote perfume and the feeling of interest was also an internal attribute of the

product (Hoyer & Maclnnis, 2004). Strictly speaking, an informational ad had a primary content

focus on information/cognition and a secondary content focus on transformation/emotion (Shen,

2007). A transformational ad had a primary content focus on transformation/emotion and a

secondary content focus on information/cognition (Shen, 2007). Cognitive and emotional

responses to an informational ad or a transformational ad might be equally important (Shen,

2007); but under certain circumstances, emotional response was more important than cognitive

response in the formulation of attitude (Morris et al., 2002).

Attitude was defined as a positive or negative evaluation about certain person, obj ect or

issue, and was composed of three elements, cognitive, emotional, and conative (Hoyer &

Maclnnis, 2004). Although an element of attitude, conative attitude was often treated as

intention and predicted with the cognitive and emotional elements of attitude (Cacioppo & Petty,

1989; Fishbein & Ajzen, 1975). The relative impact of cognitive and emotional attitude on

conative attitude or intention was examined in one recent study on responses to 240 TV, radio

and print ads in 13 product categories (Morris et al., 2002). Cognitive attitude was

operationalized as cognitive responses including believability and knowledge. Emotional

attitude was operationalized as emotional responses including pleasure, arousal, and dominance.

Conative attitude was operationalized as brand interest and purchase intention. Path analysis









with structural equation modeling indicated that emotional attitude had a much large direct

impact on conative attitude than did cognitive attitude at an aggregate level.

Fellow-up regression analyses by the 13 categories further indicated that emotional attitude

outperformed cognitive attitude in accounting for more variance in conative attitude except for

computers. Emotional attitude explained 3 to 30 percent of the variance in conative attitude

across the 13 categories whereas cognitive attitude accounted for 2 to 13 percent of the variance.

Regression analyses by the three media also suggested the superiority of emotional attitude over

cognitive attitude in predicting conative attitude except for radio. Emotional attitude explained

15 to 22 percent of the variance in conative attitude across the three media whereas cognitive

attitude accounted for 0 to 26 percent of the variance.

Although there was evidence that emotion could be more important than cognition in

attitude formation (Morris et al., 2002), attitude models were primarily cognitive. The best

example was the elaboration likelihood model (ELM) (Petty & Cacioppo, 1986). According to

the ELM, there were two routes to attitude formation: (1) the route taken by individuals who

thought actively about information in a message, or the central route and (2) the route taken by

individuals who did not think actively about the informational content of a message, or the

peripheral route. The central route emphasized cognitive/effortful thinking and minimal

emotional involvement in processing messages. Individuals who traveled this route were

referred to as cognitive elaborators. The peripheral route was almost the exact opposite of the

central route to persuasion. Individuals who traveled this route were referred to as cognitive

misers and process messages based on simple affective cues such as source attractiveness and

message length. Essentially, because of the cognitive base of the ELM, emotion was treated as a

simple peripheral cue that led to less persistent attitude formation and change whereas cognition









was regarded as the main force contributing to a longer-lasting attitude formation and change

(Petty & Cacioppo, 1986).

A recent study (Morris, Woo, & Singh, 2005) reexamined the superiority of cognition over

emotion in attitude formation as claimed in the ELM. Emotional responses of cognitive

elaborators, i.e. people who formed their attitude through the central route, and cognitive misers,

i.e., people who formed their attitude through the peripheral route, were compared in terms of

pleasure, arousal, and dominance for the same ads. Cognitive elaborators in fact had a higher

level of emotional responses, particularly a higher pleasure score, than cognitive misers.

Furthermore, cognitive elaborators also indicated a higher level of purchase intention than

cognitive misers and there was a particularly significant impact from pleasure to purchase

intention.

Measuring Emotion

A maj or issue in cognition and emotion studies was the difficulty measuring emotional

response (Morris, 1995). It was widely accepted that cognitive responses could be measured

with recall or recognition of message content and thought-listing surveys about beliefs (Hoyer &

Maclnnis, 2004). But measuring emotion was extremely controversial owing to a wide range of

emotion types and a variety of competing emotion theories (Simon, 1982). However, most

measures of emotional responses to advertising could be categorized as one of two types:

physiological and self-report (Mackay, 1980).

Physiological measures

Physiological measures stemmed from early emotion theories that posited that emotion

was manifestation of involuntary physiological or biochemical responses (e.g., changes in skin

conductance, blood pressure, skin temperature and heart rate, facial expressions, respiration and










pupil dilation, and so on) (James, 1884). Two types of physiological responses, arousal and

facial expression, were often used as the indicator of emotional experience (Mackay, 1980).

A few physiological techniques were developed to measure arousal. Those techniques

included electroencephalographs that track electrical activity in the brain, pupillometrics that

investigated the change in the size of an eye's pupil, electrocardiogram that measured heart

action, and galvanic skin response that records skin conductance (Davis, 1997). Those

techniques might introduce experimental artifacts because the existence of those measurement

devices might influence arousal; therefore, they were often used with other types of measurement

tools (Davis, 1997). New techniques such as functional Magnetic Resonance Imaging

techniques (or fMLRI) were increasingly being used to assess physiological responses (Morris et

al., 2008). Nevertheless, technical advancement of arousal measurement could not make up for a

theoretical drawback in arousal research, i.e., arousal alone did not fully explain emotional

experience but just the intensity aspect of it (Clark, 1982).

Facial expression could be identified through electromyographs or subtle changes in the

facial musculature (Davis, 1997). But to construct a link between facial expression and specific

emotion, a coding system, such as the Facial Action Coding System or FACS (Ekman & Friesen,

1978), was needed. Coding systems classified facial expressions based on a select set of

dimensions and associated the pattern of musculature with a specific emotional response (Ekman

& Friesen, 1978). Nevertheless, empirical evidence did not support the notion that each specific

facial expression corresponded to a unique emotional response because individuals could

consciously alter facial expressions owing to social norms and cultural considerations (Izard &

Bartlett, 1972; Scherer & Wallbott, 1994)









Self-report measures

Self-report measures stemmed from cognitive theories of emotion with the assumption that

subjective experience of emotion could be sufficiently described or labeled with words (Mackay,

1980). The marketing and advertising literature was filled with written self-reports on rating

scales. For example, Duncan and Nelson (1985) developed a scale to measure humor in an ad.

Edell and Burke (1987) developed a scale to capture warm, upbeat and negative feelings evoked

by an ad. Within the self-report tradition, there were two distinctive approaches to emotion

(Morris et al., 2008). One was a discrete approach, and the other was a dimensional approach.

Discrete approach to emotion

The discrete approach was in fact a Darwinian approach to emotion, which sought

evolutionarily fundamental or discrete emotions to define all subjective emotional experiences

(Mandler, 1984; Plutchik, 1980). According to this approach, there was a finite set of discrete or

basic emotions such as joy, anger, sadness, and fear that were innate to all human beings

across various cultures and developmental stages (Izard, 1992; Plutchik, 1982). Next, certain

patterns or combinations of those basic emotions resulted in the subj ective experience of emotion

(Izard, 1992; Plutchik, 1982). Therefore, emotional responses were well measurable through

investigating the extent to which each of the basic emotions was experienced (Richins, 1997).

One typical way to measure emotional responses with the discrete approach was to give

participants bipolar adj ective pairs of basic emotions, such as happy versus unhappy (Mackay,

1980). Thorson and her colleagues (1991) employed this bipolar adjective technique when

measuring emotional response to political ads.

The basic emotions approach was criticized as merely labeling without a sound theoretical

foundation that explained the experience of emotion (Roseman, 1984). There could be a set of

eight emotions interest, surprise, joy, anguish, fear, shame, disgust, and rage (Tomkins, 1980),









or another set of eight emotions fear, anger, joy, sadness, acceptance, disgust, anticipation,

and surprise (Plutchik, 1984), or even another set of ten emotions interest, joy, surprise,

distress, anger, disgust, contempt, shame, fear, and guilt (Izard, 1977). The disagreement on

either the number or the kinds of emotions across the three lists seemed to suggest a lack of well-

accepted theoretical ground to define the very notion of "discrete" (Mandler, 1984). The status

of an emotion as basic often depended on a researcher' s own criterion, and some so-called basic

emotions were proposed on grounds, such as frequency-related salience in culture and

experience, levels in an emotion hierarchy, etc., which were irrelevant to whether they were

basic in a psychological sense (Ortony & Turner, 1990).

The reliance of the discrete approach on verbal labels such as adjectives to measure

subj ective emotional experiences was also problematic. The use of adj ectives might lead to a

distorted reporting process that was subj ect to an individual's linguistic habits, vocabulary,

memory, and social desirability (Lazarus, 1966), and language was too insensitive a medium for

communicating private experiences (Mandler, 1962). Additionally, some cultures might not

have adj ectives for a specific type of emotion but individuals in those cultures did experience

that emotion, which made it impossible for a cross-cultural comparison on that emotion with

adjectives (Kitayama & Markus, 1990).

Dimensional approach to emotion

The approach originated from early studies on semantics (Osgood, Suci, & Tannenbaum,

1957). Osgood and his colleagues (1957) conducted a set of factor analyses on adjectives and

identified a three-dimensional semantic space composed of evaluation, potency, and activity. A

few other researchers (Averill, 1975; Bush, 1973; Schlosberg, 1952) explored the semantic

structure of adj ectives about emotion and identified pleasantness, activation as well as potency,

aggression, or attention-rejection as the dimensions of emotion. Those Eindings were further









developed into a three-dimensional pleasure-arousal-dominance model of emotion (hereinafter

referred to as "PAD") (Russell & Mehrabian, 1977). Pleasure-displeasure referred to positive

versus negative affective state in response to a stimulus and corresponded to the pleasantness

component of emotion. Arousal-nonarousal referred to physical activity and/or mental alertness

and corresponded to the activation component. Dominance-submissiveness referred to feelings

of control and influence over others and situations versus feeling controlled and influenced by

external circumstances and corresponded to the potency component of emotion (Mehrabian &

Russell, 1974).

However, PAD went through some changes after the original formulation. Russell later

dropped dominance in his emotion research and used pleasure and arousal to define emotion

(Russell, 1980) either because he preferred the simplicity of a two-factor structure (Russell,

1980) or because he failed to replicate the dominance dimension across studies or cultures

(Russell, 1978, 1983). Russell (1980) went on proposing a circumplex model, in which emotion

descriptors were be systematically arranged around the perimeter of a circle. This affect

circumplex contained four bipolar dimensions that were spaced 45 degrees apart, namely high

versus low activation, high activation and pleasantness versus low activation and unpleasantness,

pleasantness versus unpleasantness, and low activation and pleasantness versus high activation

and unpleasantness (Russell, 1980). The circumplex model is presented in Figure 2-1.

Pleasantness and activation, which are depicted in the darker lines, are the basic dimensions of

Russell's affect circumplex.

Watson and Tellegen (1985) also proposed an affect circumplex. Their circumplex model

closely resembled Russell's (1980) model in terms of the structure but differed from it in terms

of the basic dimensions. To be more specific, Watson and Tellegen's circumplex also contained









four bipolar dimensions that were spaced 45 degrees apart but they were high positive affect

versus low positive affect, strong engagement versus disengagement, high negative affect versus

low negative affect, and pleasantness versus unpleasantness. Positive affect and negative affect,

rather than pleasantness and activation, were the two basic dimensions of their affect circumplex

(Watson & Tellegen, 1985, Figure 1, p. 221).

Because the two circumplex models were both based on factor analysis, the difference in

basic dimensions between them might just be a terminological controversy as to how to name the

basic dimensions (Larsen & Diener, 1992). Larsen and Diener (1992) argued that the labels

"positive affect" and "negative affect" were misleading and should be replaced with

"pleasantness" and "activation". Their argument was undoubtedly supported by Russell and his

associates (Feldman, Barrett, & Russell, 1998) but was challenged by Watson, Tellegan, and

their associates (Watson, Wiese, Vaidya, & Tellegen, 1999). However, a real problem for the

models was neither of them fit real-life data in confirmatory factor analysis (Feldman et al.,

1998). In three analyses conducted by Russell and his associates (Feldman et al., 1998), the root

mean square error of approximation (RMSEA) values were .13 5, .131, and .106, which indicated

a consistently poor fit as RMSEA values of .10 or greater demonstrate a poor fit (Browne &

Cudeck, 1993). Similarly, in two analyses conducted by Watson, Tellegan, and their associates

(Watson et al., 1999), the RMSEA values were .234 for momentary data and .294 for general

data in the first analysis and were .094 for momentary data and .123 for general data in the

second analysis, which also indicated a consistently poor fit of their circumplex model.

The consistent poor fit of the circumplex models suggested that something was missing in

the models (Feldman et al., 1998). Lazarus (1991b), from a relational perspective, pointed out

that circumplex models such as the one proposed by Watson and Tellegen (1985) ignored the










person-environment relationship inherent in emotional responses. Interestingly enough, the

dominance dimension dropped by Russell, to some extent, addressed feelings about the person-

environment relationship (Mehrabian, 1995). Shaver and his associates (Shaver, Schwartz,

Kirson, & O' Connor, 1987) conducted a multidimensional scaling analysis to compare two-

dimensional versus three-dimensional emotion models and concluded "[t]he three-dimensional

solution helps to differentiate between what the cluster analysis suggests are separate basic-

emotion categories, and it is clearly more informative as a representation of emotion knowledge

than the two-dimensional solution" (p. 1071).

There were two ways to translate the three-dimension concept into a workable measure, a

verbal method (Mehrabian & Russell, 1977) and a non-verbal method (P. Lang, 1980). With the

verbal method, participants were given a checklist of up to 16 bipolar adj ectives and asked to

indicate their feelings about a given stimulus. Next, the scores on each adj ective were collapsed

into the three dimensions with factor analysis (Mehrabian & Russell, 1977). With the non-verbal

method, participants were given self-assessment manikin (hereinafter referred to as "SAM") (P.

Lang, 1980) to report their feelings. Unlike its verbal counterpart, SAM was a graphic

representation of each PAD dimension with a manikin along a continuous nine-point scale

(Morris, 1995). In the first row of the SAM, the manikin's facial expression varied from smiling

and happy to frowning and unhappy along the pleasure/displeasure dimension. In the second

row, the manikin's facial expression varied from eye-open and eyebrow-elevated to calm and

eye-closed along the arousal/non-arousal dimension. In the third row, the manikin's size varied

from huge to tiny along the dominance/submissiveness dimension (Morris, 1995).

The non-verbal alternative was found to be a better alternative than the verbal checklist

method because it eliminated cognitive processing, and was easier and quicker to use (Morris,









Waine, & Lang, 1993). Meanwhile, it "generated a similar pattern of scale values for these

situations as was obtained for the semantic differential" (P. Lang, 1980, p. 123). SAM could

even be viewed as a combination of self-report and facial expression because its graphic

characters well resembled facial expressions in different scenarios (P. Lang, 1980). All the

advantages of SAM made it a culture free measure of emotion (Bradley, Greenwald, & Hamm,

1993).

Measuring emotion with AdSAMo

Based on SAM, AdSAM" emerged as an effective tool for measuring emotional response

to marketing communications stimuli, and crossed the boundary between the discrete approach

and the dimensional approach (Morris, 1995). The adj ectives that were used to construct PAD

were added to a list of 'emotion adj ectives' found in the literature and measured with the

manikin. Then the adjectives were examined for mean error or consistency in response. The

result was a list of emotion adj ective with PAD scores that could be used to identify variations in

feelings, such as lucky, angry, provocative, and indignant, while maintaining the integrity of the

three dimensions of PAD (Morris, 1995).

AdSAM" was used to assess responses to television advertising (Morris, Bradley, Waine,

& Lang, 1992; Morris, Roberts, & Baker, 1998), pre-production vs. post-production advertising

(Morris & Waine, 1993), and global advertising across cultures (Morris, 1995; Morris, Bradley,

Sutherland, & Wei, 1993; Morris, Strausbaugh, & Nthangeni, 1996). In the two above-

mentioned studies (Morris et al., 2002; Morris et al., 2005), it was used to refine the process of

attitude formation and was a valuable tool for strategic planning, message testing and brand

tracking. In the current research, AdSAM" is used to measure emotional response to political

ads.









Summary of Literature Review

A chain reaction model of advertising effectiveness (Lavidge & Steiner, 1961; Shama,

1976) was identified through the literature review of political marketing. According to that

model, voters learned of candidate characteristics and issue stands from political ads, and the

knowledge in turn substantially impacted their evaluation of a candidate.

Next, a few empirical studies were reviewed to address the impacts of issue/image and

negative/positive political advertising on message learning/memory (Geer & Geer, 2003; Kaid et

al., 1992; A. Lang, 1991; Newhagen & Reeves, 1991; Schleuder, 1990; Shapiro & Rieger, 1992)

and candidate evaluation (Kahn & Geer, 1994; Thorson et al., 1991). Although those studies

provided contradictory outcomes owing to different operationalization of memory and candidate

evaluation, different types of ads, and certain methodological problems, two sets of Eindings

were identified as particularly relevant to the chain reaction model (Lavidge & Steiner, 1961;

Shama, 1976) adopted by the current study. One was about message learning/memory of

candidate characteristics and issue stands, and the Einding was that issue ads outperformed image

ads in generating higher number of recalled issues and personal characteristics (Kaid et al.,

1992). With that Einding, cued recall of issues and personal characteristics (Kaid et al., 1992) is

used as the measure of memory in this study. The other was about candidate electability, and the

findings were positive ads outperformed negatives ads in generating more positive candidate

electability and negative issue ads outperformed negative image ads in generating more positive

candidate electability (Kahn & Geer, 1994).

Based on the chain reaction model of advertising effects (Lavidge & Steiner, 1961; Shama,

1976), memory and candidate electability were not two separate effects. Memory of an ad might

influence the evaluation of the candidate advertised in the ad, but such an influence was not









tested in previous research. This missing link is explored in the current study with a path

analy si s.

Furthermore, conceptual similarity between image/issue political ads and

informational/transformational advertising was identified in the literature review. An image ad

had a primary focus on transformation and a secondary focus on information, whereas an issue

ad had a primary focus on information and a secondary focus on transformation (Shen, 2007).

The Puto and Wells (1984) informational/transformational survey scale is used to calibrate the

perception of informational and transformational content of issue and image ads and to better

understand the impact of perception on advertising effectiveness.

The informational aspect of adverting was associated with rational/cognitive response to

advertising whereas the transformational aspect was associated with experiential/emotional

response to advertising (Puto & Wells, 1984). There was evidence that emotional response was

more important than cognitive response in shaping consumer attitude and purchase intent (Morris

et al., 2002). There was some attempt to investigate emotional response to political advertising

with a discrete approach to emotion and verbal measures (Thorson et al., 1991); but there were

quite a few theoretical problems with both the discrete approach and verbal measures (Kitayama

& Markus, 1990; Lazarus, 1966; Mandler, 1962, 1984; Ortony & Turner, 1990). Dimensional

approach (Averill, 1975; Bush, 1973; Osgood et al., 1957; Russell & Mehrabian, 1977;

Schlosberg, 1952) and non-verbal measures (P. Lang, 1980; Morris, 1995) were a more

promising alternative to assess emotional response (Morris et al., 2008). Specifically, AdSAM"

(Morris et al., 2002; Morris et al., 2005) is used as the measure of emotional response to

advertising or the emotional attitude toward candidate. AdSAM" was a graphic representation of

three emotional dimensions, namely, pleasure, arousal, and dominance, identified in previous









research (Morris 1995; Russell & Mehrabian, 1977). It was used in a number of studies on

advertising effectiveness (Morris, 1995; Morris et al., 1993; Morris et al., 1992; Morris et al.,

1998; Morris et al., 1996; Morris & Waine, 1993; Morris et al., 2002; Morris et al., 2005).

The literature review leads to the formulation of research questions. Fundamentally, this

study attempts to find out 1) how the perception of issue/image content of political ads, measured

with the informational/transformational advertising scale, will influence memory of candidate

issue stands and personal characteristics, 2) whether the relationship between the perception of

issue/image content and memory of candidate issue stands and personal characteristics will vary

across negative and positive ads, 3) how the perception of issue/image content will influence the

impact of cognitive response, measured by memory, and of emotional response, measured by

AdSAM" on voters' opinion on a candidate's electability, and 4) whether the relationship

between the perception of issue/image content and candidate electability will vary across

negative and positive ads.

Hypotheses and Research Questions

The hypotheses and research questions focus on three aspects. The first aspect is about

perception of advertising content. The second aspect is about memory. The third and last aspect

is about candidate electability.

Perception of Advertising Content

According to the literature review of informational and transformational advertising, the

concept of information and transformation was well applicable to issue and image (Shen & Kim,

2006). The scale of informational/transformational advertising (Puto & Wells, 1984) was used

to calibrate the informational and transformational content of issue and image ads; and an image

ad had a primary focus on transformation and a secondary focus on information, whereas an









issue ad had a primary focus on information and a secondary focus on transformation (Shen,

2007). It is therefore proposed:

* Hypothesis 1: Based on previous research about the relationship between issue/image ads
and informati onal/transformati onal advertising, the informational score of issue ads will be
higher than the transformational score of issues ads at .05 significance level as measured
with the informational/transformational advertising scale. Conversely, the
transformational score of image ads will be higher than the informational score of image
ads at .05 significance level as measured with the informational/transformational
advertising scale.

Memory

The key question about memory is whether the characteristics of political ads, issue/image

and negative/positive, influence the memory of ad content. Memory was previously measured

with free recall and recognition test of verbal, visual, and audio cues in ads (A. Lang, 1991;

Newhagen & Reeves, 1991; Schleuder, 1990) or cued recall of issue stands and candidate

characteristics (Kaid et al., 1992). Cued recall of issue stands and candidate characteristics is the

measure of memory in this study because memory of candidate issue stands and personal

characteristics were a maj or component of the chain reaction model of advertising effectiveness

(Lavidge & Steiner, 1961; Shama, 1976) adopted in the current study. There was strong

evidence that issue ads outperformed image ads in generating higher number of recalled issues

and personal characteristics (Kaid et al., 1992). Therefore, it is proposed:

* Hypothesis 2: Based on previous research about memory of issue and image ads, issue ads
will be better remembered than image ads at .05 significance level as measured with cued
recall of candidate issue stands and personal characteristics.

In previous studies on memory of negative and positive ads, negative ads were found to be

better remembered than positive ads (A. Lang, 1991; Newhagen & Reeves, 1991), but the results

were based on recall and recognition test of verbal, visual, and audio cues in ads rather than cues

about issue stands and personal characteristics. It remained unclear whether the valence of

political ads, negative or positive, would influence the recall of candidate issue stands and










personal characteristics. It also remained unclear whether there would be an interaction effect of

issue/image by negative/positive on the recall of candidate issue stands and personal

characteristics. Therefore, it is proposed:

* Research question 1: How will the valence of political ads, positive or negative, influence
memory of ad content as measured with cued recall of candidate issue stands and personal
characteristics?

* Research question 2: How will the valence of political ads, positive or negative, interact
with and the emphasis of political ads, issue or image, on memory of ad content as
measured with cued recall of candidate issue stands and personal characteristics?

Candidate Electability

Previous research did not provide consistent evidence as to how issue and image ads

differed from each other in influencing voters' opinion on a candidate or the candidate

electability (Kahn & Geer, 1994; Thorson et al., 1991). Therefore, it is proposed:

* Research question 3: How will the emphasis of political ads, issue or image, influence
candidate electability?

Nevertheless, there was evidence that positive ads outperformed negatives ads in

generating more positive candidate electability, and negative issue ads outperformed negative

image ads in generating more positive candidate electability (Kahn & Geer, 1994). In the current

study, it is proposed:

* Hypothesis 3: Based on previous research about candidate electability, positive ads will
produce more positive electability toward candidate than negative ads at .05 significance
level .

* Hypothesis 4: Based on previous research about candidate electability, negative issue ads
will produce more positive cognitive electability toward candidate than negative image ads
at .05 significance level.

In this study, candidate electability is composed of issue competency (Kahn & Geer,

1994), candidate personality (Kahn & Geer, 1994; Thorson et al., 1991), and voting intention

(Kahn & Geer, 1994). Furthermore, based on cognitive theories of persuasion (Cacioppo &









Petty, 1989; Fishbein & Ajzen, 1975), the cognitive antecedent to electability is measured with

message recall. The emotional antecedent to electability is measured with AdSAM"(Morris et

al., 2002; Morris et al., 2005). There was evidence that emotional response had a larger impact

than cognitive response on the attitudinal construct of electability (Morris et al., 2002). It is

therefore proposed:

* Hypothesis 5: Based on previous research about attitude, emotional response, measured
with AdSAM", will have a larger impact on electability than cognitive response, measured
with recall, at .05 significance level.

There was some attempt to investigate emotional response to political advertising with a

discrete approach to emotion and verbal measures (Thorson et al., 1991); but such an attempt

was problematic and should be replaced by a dimensional approach and non-verbal measures

such as AdSAM" (Morris et al., 2008). The relationship between ad content, i.e., issue/image

and negative/positive, and emotional response, measured with AdSAM", has never been

addressed in previous research. The informational/transformational scale is used to measure

voters' perception of issue and image, and the perception provides an alternative way to explore

the impact of ad content on attitude toward candidate. Path analysis is used to investigate the

relationships of interest, and two more research questions are proposed:

* Research question 4: How will the perception of issue and image content, measured with
the informati onal/transformati onal advertising scale, influence cognitive and emotional
responses; and next how cognitive and emotional responses influence candidate
electability?

* Research question 5: How will the relationship measured in Research Question 4 vary
across negative and positive ads?





























Figure 2-1. Russell's affect circumplex (Russell, 1980, Figure 1, p. 1164)


I_


Low
activation


High
activatiosn
ation High activation
ant Reasant


I


High activ;
Unplease




Unpleasantness


I


Reasantness


Losw activation
Unpleasant i


Low, activation
_-IReasant









CHAPTER 3
MATERIALS AND IVETHODS

This study adopted a 2 (positive vs. negative) by 2 (issue vs. image) between-subjects

design. There are four experimental conditions, i.e., positive image, positive issue, negative

image, and negative issue. Table 3-1 presents more details about the conditions.

Materials and Measures

TV ads are chosen for the current study because 1) TV advertisements are particularly

capable of functioning as a transformational medium (Puto & Wells, 1984) and 2) TV

advertisements dominate political broadcasting in the U.S. (Kaid & Johnston, 2001). A content

analysis is used to select the right TV political ads from a pool of congressional and senatorial

campaign advertisements for the 2004 and 2006 elections for the current study. The coding

scheme for the political ads was modified from Laskey's typology (Laskey, Day, & Crask, 1989)

and used in a previous content analysis of political TV ads (Shen & Kim, 2006). All the

variables described below are included in Appendix A.

Independent Variables

Perception of ad content

The Scale (Puto & Wells, 1984, p. 641) was used to measure both the perception of

informational cues (hereinafter referred to as "PIFC") and the perception of transformational

cues (hereinafter referred to as "PTFC") in a political ad. The Scale has 23 items. Eight of them

measure PIFC and the other 15 items measure PTFC. The Scale is modified for political ads.

"Brand" is replaced with "candidate" and "company" with "party." A mid-point category

"neither agree nor disagree" is added to the Scale to form a 7-point scale, ranging from "strongly

disagree" to "strongly agree." Additionally, the negative ads include both the candidate and the

opponent, therefore two informational items and one transformational item are asked twice to









address issue and personality learning about both candidates. As a result, there are ten

informational and 16 transformational items for the negative ads. More details about the items

for negative ads are available in Appendix A. PIFC and PTFC also serve as manipulation checks

of issue and image content. They are also exogenous variables in path analysis.

Prior exposure to ad and familiarity with candidate

Survey participants are asked to report their prior exposure to each ad on a three-point

scale (1: never, 2: a few times, 3: many times), and familiarity with the brand/candidate on a

five-point scale ranging from 1 (very unfamiliar) to 5 (very familiar). Prior exposure is used to

filter out participants who may have watched the ads because previous exposure to ads is likely

to bias processing of message content. Low familiarity is a reality for most candidates (Thorson

et al., 1991), and this phenomenon can be checked with the familiarity measure.

Dependent Variables

Message recall

Cued recall is used to measure memory with a two-item instrument employed in previous

research (Kaid et al., 1992). Participants are asked to 1) describe candidate characteristics in

adjective forms, and 2) state political issues associated with each candidate. One adjective is

counted as one recalled message unit of candidate characteristics. One political issue is counted

as one recalled message unit of political issues.

Emotional response

AdSAM" is used to measure emotional candidate attitude. Participants will be asked to

indicate their feeling of pleasure, arousal, and dominance on the three dimensions of the scale.

The scale has been used in previous studies on attitude (Morris et al., 2002; Morris et al., 2005).









Candidate electability

Three verbal scale of candidate evaluation employed in previous research (Kahn & Geer,

1994) are used to measure candidate electability. The first scale is used to assess candidate

competence in the issues mentioned in the ads and is composed of one seven-point item ranging

from 1 (very incompetent) to 7 (very competent). The second scale is used to assess candidate

characteristics in terms of trustworthy, strong leadership, hardworking, and knowledgeable and is

measured with a 7-point Likert-type scale ranging from 1 (strongly disagree) to 7 (strongly

agree). The third scale is voting intention and participants will be asked to indicate how likely

they would vote for the candidate on a 5-point scale ranging from 1 (very unlikely) to 5 (very

likely). Based on the tripartite model of attitude (McGuire, 1989), those three variables will be

used to create a composite measure of candidate electability.

Graded Response Model (GRM) of Item Response Theory (IRT)

A Likert-type scale, i.e., "strongly disagree strongly agree," is used in the current study

to measure learning of informational and transformational cues from the ads. Traditionally, this

type of scales is assumed to have equal distance between two consecutive scale categories and

generate interval type data (Davis, 1997); therefore, "computation of means, correlations, and

other inferential statistics (e.g., reliability coefficients) is justified" (Singh, 2004, p. 188).

However, the interval data assumption of those scales is rarely tested in practice (Singh, 2004),

and responses to scale categories such as "strongly disagree strongly agree" or to "variations of

this general theme can be assumed to involve categorical, rank-ordered data" (p. 188).

IRT is developed to better account for responses to Likert-type scales. With IRT, the

responses to the scales are assumed to be categorical, rank-ordered observable data, and serve as

the indicator of interval-type latent data; furthermore, a probability model connects the

observable data to the latent data (Ewing, Salzberger, & Sinkovics, 2005; Singh, 2004; Singh et









al., 1 9 90). F or ex ample, e ach item of the i nform ati onal/tran sformati on al adv erti si ng s cale has

seven categories, ranging from "strongly disagree" to "strongly agree." A participant' s response

to Item 1 is assumed to be an observable choice out of the seven available categories for Item 1,

and this observable response is also assumed to be the indicator of the participant' s latent

response to Item 1. The observable response is converted to the latent response through a

probability model that takes into account both participant and scale item characteristics

(Embretson & Reise, 2000).

One of the most commonly used IRT model is a two-parameter logistic model (2PL

model) that has an item discrimination parameter (" a parameter) and an item difficulty

parameter (" P parameter) associated with a latent trait (" 6 ") (Hambleton & Swaminathan,

1985; Oishi, 2006).

For the sth participant responding to the ith item, the model can be expressed as

exp[azl(6s it)
P(X~s=16is, p,, a)=


where X;s is a dichotomous response variable to the ith item with "1" denoting correct

response, 8s is the sth participant' s score on the latent trait scale, at is the ith item's discrimination

parameter, P2 is the ith item's difficulty parameter, and exp is the natural log base (2.718). The

discrimination parameter indicates how well the ith item distinguishes between participants with

low versus high standing on the latent trait scale (Oishi, 2006). The difficulty parameter

indicates the score on the latent trait scale when the probability of responding to the ith item

correctly is 50%; in other words, if an participant with a latent score of -1 (1 standard deviation

below the latent trait scale's mean, which is standardized to be "O") has a 50% probability to

answer correctly to the ith item, then the difficulty parameter of the ith item is -1 (Oishi, 2006).









The obvious problem for the 2PL model to be used for survey responses is there is no right

or wrong answer to a survey item. Additionally, survey items are often multiply ordered rather

than following a dichotomous pattern. To address those concerns, the GRM was developed by

converting the multiply ordered item categories into a set of binary response variables to fit the

IRT framework (Samejima, 1969). When a K category scale is used for the ith survey item, the

categories are scored x, = 0, 1,..., m,, respectively, and m, = K, For example, a 3-

point/category survey item is scored x, = 0, 1, 2. The participants' responses to the ith item are

converted in a binary manner so that those less than the x, category are recorded into "O" and

those equal to or greater than the x, category are recorded into "1." As a result, a response of "1"

on the original 0 to 2 scale is recorded as "1" for both the "O" and "1" categories but "O" for the

"2" category.

I~,(8s) is the probability for the sth participant to have the "1" response to the x, category

of the ith item conditional on the sth participant' s score on the latent trait scale (6 ). Through the

binary coding of responses in the scale categories, the 2PL model can used to specify the GRM

and is expressed as

(s=exp[az,(6s- Bi)l
s 1+exp[az,(6s- Be)l

x = j = 1, ..., mi

Note that "1" response to the x, category for 4 s(6s) refers to responding in or above the x,


category rather than responding in the x, category. For example, when x = 1, s~T(6s) refers to the

probability of responding in or above scale point 1 or in fact any categories above scale point 0.

Similarly, when x = j = 2, 4 s (8s) refers to the probability of responding in or above scale point

2 or in any categories above scale point 1. Therefore, the difference between 4 s Os) and









~ ,(8s), which can be computed by subtracting t4 s(6s) from Ris( Os), is the probability of

responding in scale point 1. Also note that in (2) x starts from 1 rather than 0 because, by

definition, the probability of having the "1" response to the lowest scale point/category x, = 0 or

responding in or above scale point 0 is known as

40s 6s)= 1.0

and the probability of responding in scale point 0 can be computed by subtracting f is(8s) from

1. Furthermore, the probability of having the "1" response above the highest scale

point/category x, = m, or responding in or above scale point m, + 1 is known as

S m,+l)s(6sS)= 0.0

simply because the m, + 1 scale point does not exist. As a result, the probability of responding in

the highest scale point m, is the same as the probability of responding in or above m, or



Onc m, )s(6s)-0.est atciat srsose()ino bveasal on/ctgr

Os ncen the probability of the sth participant' s response (x) inor eaboe scale point/category cnb


expressed as

9os (es)= 1.0 47 (6s)

Exs(s)=Exs6s)- x+)s(s)=exp[a,(6s-pei)] exp[az(6s-pro~+l))]
1+exp[azl(6s-pli)] 1+exp[azl(6s -pir+1))]




x = j = 1, ..., mi

8s is still the sth participant's the latent trait score; at is the ith item's discrimination

parameter and it is constrained to be equal across all the response categories of the ith item.









However, pl; has a very different meaning. For the K category scale scored from 0 to m, (m, = K,

-1), there are m, intervals between each two neighboring categories. In the IRT literature, the

intervals are called thresholds. P1; (f= 1... m,) is called "threshold" parameter and refers to the

trait score necessary to respond above the jth threshold of the ith item with .50 probability.

With the above-mentioned 3-point item on the 0 to 2 scale, the probability of responding in

each of the three scale points can be expressed as


0So~(6s)=1. 0-I,(6s)=1.0-

1+ exppaz(6s pri)]

f, is(6)= 47,(6s,)- P21s is)= exa(6-p) _xpa6s-2
1+exp~az(6s-p,,)] 1+ exp~z(6s -p,,2

exp[az,(6s p072
~,(2s i = 2s ies)-0o= -0


and there are two threshold parameters, Pt and P22, OStimated for threshold 1 between scale

points 0 and 1 and threshold 2 between scales points 1 and 2, respectively. If P21 is -1 and P22 1S

1, then respondents whose latent trait score is below -1 are more likely to respond in scale point

0, those whose score is -1 are equally likely to respond in scale points 0 and 1 or threshold 1,

those whose score is between -1 and 1 are more likely to respond in scale point 1, those whose

score is 1 are equally likely to respond in scale points 1 and 2 or threshold 2, and those whose

score is above 1 are more likely to respond in scale point 2.

Essentially, the observed responses are viewed as probabilities of agreeing with an item

(Hambleton & Swaminathan, 1985; Linden & Hambleton, 1996; Oishi, 2006). Such a

probabilistic approach is in line with the cognitive process that "an individual's response [to

questionnaire items] is not based solely on his/her 'true' standing on the latent variable but is

also affected by other cognitive processes ... including] distracting information and events,









mood swings and processes, social desirability and other biases, and random memory processes"

(Singh, 2004, p. 196). Because of this measurement advantage, IRT is used in this study to

estimate latent responses to the Scale. Next, the latent responses are used to test the hypotheses

and research questions about learning of informational and transformational cues in this study.









Table 3-1. Experimental conditions
Condition Ad
1 Positive Image
2 Positive Issue
3 Negative Image
4 Negative Issue









CHAPTER 4
RESULTS

Pretest

A pretest was conducted to select one typical positive image advertisement, one typical

positive issue advertisement, one typical negative image advertisement, and one typical negative

issue advertisement. The items in the pretest are included in Appendix A. Forty-four

undergraduate students in an advertising strategy class were asked to indicate how much political

issue and personal qualities information they learned from five pre-selected advertisements on a

seven-point scale ranging from "Very Little" to "A Great Deal." Table 4-1 provided the scores

of political issue and personal qualities information.

Based on the results, the typical positive image advertisement was an advertisement of

U.S. Congressman Mike Pence of Indiana focusing on the candidate's personal charisma. For

the main study, this positive image advertisement was edited to create a Chris Chocola

(hereinafter referred to as "the Candidate") positive image advertisement. Specifically, the

visuals in the Mike Pence advertisement were replaced with those of the Candidate, but most of

the audio about candidate personality was retained. The typical positive issue advertisement was

an advertisement of the Candidate focusing on his views on small business. The typical negative

image advertisement was an advertisement of the Candidate attacking his opponent Joe Donnelly

(hereinafter referred to as "the Opponent") on his poor tax paying record and character. The

typical negative issue advertisement was an advertisement of the Candidate attacking the

Opponent on tax policies.

Main Study: Model Estimation and Results

A structural equation modeling (SEM) procedure, Lisrel (Joeskog & Sobom, 2005), was

used to test the proposed model. Several assumptions of SEM, i.e., independence, normality,









linearity (Hair, Anderson, Tatham, & Black, 1995), as well as equality of variance-covariance

matrices across the four experimental conditions, were checked. Except for those variables

noted below, there was no significant violation of the assumptions. Specifically, bivariate

scatterplots of the dependent variables by their studentized residuals were used to check the

independence assumption, and there were no systematic patterns that would indicate the

violation of that assumption. The dependent variables' normal probability plots and their

skewness and kurtosis scores were used to check the normality assumption, and there was no

significant deviation from normality. Barlett' s test of sphericity was used to check the linearity

assumption, and there was no significant deviation from linearity. Both univariate tests, i.e.,

Cochran's C, Bartett-Box, and Levene test, and a multivariate test, i.e., Box's M, were used to

check the equal variance-covariance assumption, and there was no significant deviation from the

assumption.

Descriptive Statistics of Independent Variables

Prior exposure to ad and familiarity with candidate

A total of 1087 participants were recruited. Eight-seven participants had watched at least

one of the four advertisements before, and their responses were excluded from the study. The

1000 participants retained in the study were very unfamiliar with both candidates. Table 4-2

provided means and standard deviations of familiarity with the Candidate and the Opponent.

Perception of ad content

The original informati onal/transformati onal scale (Puto & Wells, 1984) has eight

informational items and 15 transformational items. The same items were used to measure

positive issue and image ads in a previous study (Shen, 2007). However, because the negative

ads included both the Candidate and the Opponent, two informational items and one

transformational item were asked twice to address issue and personality learning about both









candidates. As a result, there were ten informational and 16 transformational items for the

negative ads. More details about the items for negative ads are available in Appendix A.

Originally there were nine categories on the scale, but there were missing responses to the

low and high ends ("1" and "9") on the scale. Reducing the number of categories is often used to

solve this problem (Ewing, Salzberger, & Sinkovics, 2005; Singh, 2004). In this study, the nine-

point scale was recorded into a seven-point scale.

For the two positive advertisements, the perception of informational cues (PIFC) was the

average of the eight informational items, and the perception of transformational cues (PTFC) was

the average of the 15 transformational items. For the two negative advertisements, PIFC was the

average of the ten informational items, and the PTFC was the average of the 16 transformational

items. Table 4-3 reported the scores of the four advertisements. The PIFC scores of the two

image advertisements were significantly lower than the PTFC scores. Conversely, the PIFC

scores of the two issue advertisements were significantly higher than the PTFC scores.

The results not only confirmed the validity of the coding scheme to identify informational

and transformational advertisements and proved the experimental manipulation successful, but

also supported Hypothesis 1. Although the statistical significance might be inflated because of

the large sample size, the PIFC scores were above the scale mid-point (4 on a 7-point scale) for

the issue advertisements and below the mid-point for the image advertisements. The PTFC

scores were above the mid-point for the image advertisements and below the mid-point for the

issue advertisements.

One caveat of the results is noteworthy. Although PIFC and PTFC can be measured

separately, they are not mutually exclusive (Puto & Wells, 1984). Voters can learn

transformational cues from an issue political advertisement (Johnston & Kaid, 2002), for










example, a down-to-earth personality of the Candidate owing to his views on small businesses.

Voters can also learn informational cues from an image political advertisement (Christ et al.,

1994; Johnston & Kaid, 2002), for example, the Candidate representing Indiana in the U. S.

Congress. An informational advertisement in fact has a primary focus on information and a

secondary focus on transformation, and a transformational advertisement has a primary focus on

transformation and a secondary focus on information (Puto & Wells, 1984).

Table 4-4 reported the reliability coefficients in the four advertisement conditions. The

informational coeffcients were a little bit higher than those reported in the validation study (Puto

& Wells, 1984) whereas the transformational coeffcients were quite close to those reported in

that study. The informational coeffcients were smaller than the transformational coefficients.

But the smaller coefficients do not necessarily mean that the informational items are less reliable

than the transformational items but indicate that Cronbach's reliability test favors larger numbers

of scale items (Hair et al., 1995).

Descriptive Statistics of Dependent Variables

Message recall

Participants were asked to 1) describe candidate characteristics in adj ective forms, and 2)

state political issues associated with each candidate. Each characteristic adjective or issue was

counted as one unit of image or issue recall. Table 4-5 provided means and standard deviations

of image recall, issue recall, and total recall of both image and issue. Because the negative

advertisements had two candidates whereas the positive advertisements had one candidate only,

the image and issue recall in the two negative advertisement conditions were in fact the mean of

the recall of the two candidates whereas the two types of recall in the two positive advertisement

condition were the exact number of recall of the featured candidate. Also as an effort to make










total recall more comparable across the conditions, the total recall was mean of both image and

issue recall.

MANOVA was used to investigate message recall between advertisement valence (positive

vs. negative) and advertisement emphasis (image vs. issue). Both advertisement valence and

emphasis had a significant direct effect on image recall Frazence(1, 996) = 173.35, p < .05, partial

eta squared = .15; FEmphasts(1, 996) = 12.48, p < .05, partial eta squared = .01; issue recall

Frazence(1, 996) = 135.21, p < .05, partial eta squared = .12; FEmphasts(1, 996) = 253.72, p < .05,

partial eta squared = .20; and total recall Frazence(1, 996) = 254.65, p < .05, partial eta squared =

.20; FEmphasts(1, 996) = 29.76, p < .05, partial eta squared = .03. There was also a significant

valence by type interaction effect on image recall F(1, 996) = 8.77, p < .05, partial eta squared =

.01; issue recall F(1, 996) = 288.67, p < .05, partial eta squared = .23; and total recall F(1, 996) =

119.62, p < .05, partial eta squared = .11.

Based on the results, the issue advertisements generated more issue recall and total recall

than the image advertisement; therefore, Hypothesis 2 was supported. Additionally, the negative

advertisements generated more image, issue, and total recall than the positive advertisement;

therefore, the answer to Research Question 1 was that negative advertisements were better

remembered than positive advertisement. Meanwhile, the valence interacted with the emphasis

on recall; therefore, the answer to Research Question 2 was that the interaction effect did exist.

Candidate electability

Participants were asked to indicate their opinion on three verbal scale of candidate

electability employed in previous research (Kahn & Geer, 1994). The first scale was used to

assess candidate competence in tax, the issue mentioned in the ads, on a seven-point item

ranging from 1 (very incompetent) to 7 (very competent). Table 4-6 provided comparisons of

the two candidates in the two negative advertisement conditions. The Candidate was










significantly more competent than the Opponent in the negative image condition, whereas there

was no difference in the negative issue condition. ANOVA was used to investigate the

Candidate's tax competency between advertisement valence (positive vs. negative) and

advertisement emphasis (image vs. issue). The results indicated that valence did not have a

significant direct impact on issue competency Frazence(1, 996) = .06, p > .05, partial eta squared =

.00; but emphasis had a significant direct impact on issue competency FEmphasts(1, 996) = 33.82, p

< .05, partial eta squared = .03. Additionally, valence and emphasis had a significant interaction

effect on issue competency F(1, 996) = 111.85, p < .05, partial eta squared = .10. So the

Candidate was perceived as more competent in issue advertisements than in image

advertisements, and he was perceived as the most competent in the positive issue condition but

the least competent in the positive image condition. Furthermore, t-tests about issue competency

of the same candidate between the two negative ad conditions indicated that the Candidate was

perceived as more competent in the negative image condition than the negative issue condition

t(498) = 3.09, p < .05, whereas the Opponent was perceived as more competent in the negative

issue condition than the negative image condition t(498) = 8. 15, p < .05.

The second scale was used to assess candidate characteristics in terms of trustworthy,

strong leadership, hardworking, and knowledgeable and is measured with a 7-point Likert-type

scale ranging from 1 (strongly disagree) to 7 (strongly agree). Table 4-7 provided comparisons

of the two candidates in the two negative advertisement conditions. The Candidate was

preferred over the Opponent in the negative image condition, but the Opponent was preferred

over the Candidate in the negative issue condition. ANOVA was used to investigate the

Candidate's characteristics between advertisement valence (positive vs. negative) and

advertisement emphasis (image vs. issue). The results indicated that emphasis did not have a










significant direct impact on candidate characteristics FEmphasts(1, 996) = 2.48, p > .05, partial eta

squared = .00; but valence had a significant direct impact on candidate characteristics Frazence(1,

996) = 226.96, p < .05, partial eta squared = .19. Additionally, valence and emphasis had a

significant interaction effect on issue competency F(1, 996) = 82.43, p < .05, partial eta squared

=.08. So the Candidate was more likable in the positive advertisement conditions than the

negative advertisement conditions, and he was the most likable in the positive issue condition but

the least likable in the negative issue condition. Furthermore, t-tests about candidate

characteristics of the same candidate between the two negative ad conditions indicated that the

Candidate was more likable in the negative image condition than the negative issue condition

t(498) = 4.99, p < .05, whereas the Opponent was more likable in the negative issue condition

than the negative image condition t(498) = 4. 10, p < .05. Table 4-8 reported the reliability

coefficients of the candidate characteristics scale in the four advertisement conditions. Table 4-8

reported the reliability coefficients in the four advertisement conditions.

The third scale was used to assess voting intention on a one-item five-point scale ranging

from 1 (very unlikely) to 5 (very likely). Table 4-9 provided descriptive statistics and

comparisons of voting intention. The voting intention in all conditions was below the scale mid-

point, which indicated a low intention to vote for either candidate. The two candidates did not

differ from each other in voting intention in the two negative conditions. ANOVA was used to

investigate voting intention for the Candidate between advertisement valence (positive vs.

negative) and advertisement emphasis (image vs. issue). The results indicated that emphasis did

not have a significant direct impact on voting intention FEmphasts(1, 996) = 1.17, p > .05, partial

eta squared = .00; but valence had a significant direct impact on voting intention Frazence(1, 996)

= 15.10, p < .05, partial eta squared = .02. Additionally, valence and emphasis had a significant









interaction effect on voting intention F(1, 996) = 78.39, p < .05, partial eta squared = .07.

Furthermore, t-tests about voting intention between the two negative ad conditions indicated that

voting intention for the Candidate was higher in the negative image condition than in the

negative issue condition t(498) = 5.08, p < .05, whereas that for the Opponent indicated no

difference between the negative issue and negative image conditions t(498) = .48, p > .05.

The answer to Research Question 3 varied across the three scales. When electability was

measured in terms of issue competence, issue ads generated more favorable candidate attitude

than image ads. When electability was measured in terms of candidate characteristics and voting

intention, there was no difference between issue and image ads. By contrast, Hypothesis 3 was

not supported when electability was measured in terms of issue competence, but was supported

when the same variable was measured in terms of candidate characteristics and voting intention.

Finally, the t-tests on the Candidate did not support Hypothesis 4, but those on the Opponent

supported Hypothesis 5 in terms of issue competence and candidate characteristics.

Pleasure-arousal-dominance

Participants were asked to indicate their emotional response to the candidates) on the

nine-point AdSAM" scale in term of pleasure, arousal and dominance. Table 4-10 provided

descriptive statistics and comparisons of the pleasure dimension. The pleasure scores were

above the scale mid-point in the two positive conditions, and below the scale mid-point in the

two negative conditions. The Candidate received higher pleasure score than the Opponent in the

two negative conditions. ANOVA was used to investigate pleasure response to the Candidate

between advertisement valence (positive vs. negative) and advertisement emphasis (image vs.

issue). Both advertisement valence and emphasis had a significant direct effect on pleasure

Frazence(1, 996) = 172.84, p < .05, partial eta squared = .15; FEmphasts(1, 996) = 6.30, p < .05,

partial eta squared = .01. There was also a significant valence by emphasis interaction effect on










pleasure F(1, 996) = 11.75, p < .05, partial eta squared = .01. The positive issue ad generated the

highest pleasure response, whereas the negative issue ad generated the lowest pleasure response.

Furthermore, there was no difference between the pleasure response to either the Candidate or

the Opponent in the two negative conditions with t-tests, t(498)Candidate = .61, p > .05;

t(498)Opponent = -.84, p > .05.

Table 4-11 provided descriptive statistics and comparisons of the arousal dimension. All

the arousal scores were below the scale mid-point. The Candidate received lower arousal score

than the Opponent in the negative image condition, but there was no difference between the two

candidates in the negative issue condition. ANOVA was used to investigate arousal response to

the Candidate between advertisement valence (positive vs. negative) and advertisement emphasis

(image vs. issue). Advertisement valence had no significant effect on arousal Frazence(1, 996) =

.04, p > .05, partial eta squared = .00; but emphasis had a significant effect on arousal FEmphasis l,

996) = 17.54, p < .05, partial eta squared = .02. There was also a significant valence by

emphasis interaction effect on arousal F(1, 996) = 45.04, p < .05, partial eta squared = .04. The

positive issue ad generated the highest arousal response, whereas the positive image ad generated

the lowest arousal response. Furthermore, t-tests about arousal response to the same candidate

between the two negative ad conditions indicated that there was no difference for the Candidate

in the two conditions t(498) = 1.70, p > .05, but there was stronger arousal response to the

Opponent in the negative image condition than the negative issue condition t(498) = 3.31, p <

.05.

Table 4-12 provided descriptive statistics and comparisons of the dominance dimension.

The dominance scores were above or close to the scale mid-point. There was no difference

between the two candidates in the two negative conditions. ANOVA was used to investigate










dominance response to the Candidate between advertisement valence (positive vs. negative) and

advertisement emphasis (image vs. issue). Advertisement valence had a significant effect on

dominance Frazence(1, 996) = 8.44, p < .05, partial eta squared = .01; but emphasis had no

significant effect on dominance FEmphasts(1, 996) = .65, p > .05, partial eta squared = .00. There

was also a significant valence by type interaction effect on dominance F(1, 996) = 14.09, p < .05,

partial eta squared = .01. The positive image ad generated the highest dominance response,

whereas the negative image ad generated the lowest dominance response. Furthermore, t-tests

about arousal response to the same candidate between the two negative ad conditions indicated

that there was stronger dominance response to the Candidate in the negative issue condition than

the negative image condition t(498) = 1.98, p < .05, but there was no difference for the Opponent

between the two negative conditions t(498) = 1.17, p > .05.

Model Estimation

The proposed model for this study was specified in Figure 4-1. The error variance of

PIFC and PTFC was correlated because the two constructs overlap conceptually (Puto & Wells,

1984). Item response theory was used to estimate the latent scores from the observed scores of

PIFC and PTFC, and the latent scores were placed in SEM. Correlation matrices were included

in Appendix B.

Figure 4-2 specified the measurement model for candidate electability (hereinafter referred

to as "CE") and emotional response (hereinafter referred to as "ER".) CE was indicated by issue

competence, personal characteristics, and voting intention. Issue competence was measured with

a one-item scale, and its error variance was set to be equal to zero. Personal characteristics was

measured with a four-item scale, and its error variance was set to be equal to subtracting the

scale reliability from 1. Voting intention was measured with a one-item scale, and its error









variance was set to be equal to zero. Emotional response was indicated by pleasure, arousal, and

dominance measured with AdSAM" (Morris, 1995).

Positive image condition

Figure 4-3 reported factor loadings on CE and ER. Table 4-13 reported goodness of fit

statistics. GFI (Goodness of Fit Index), NFI (Normed Fit Index), CFI (Comparative Fit Index),

RMSEA (Root Mean Square Error of Approximation), and SRMR (Standardized Root Mean

Square Residual) were chosen due to their popularity (Hair et al., 1995). The model had a good

fit in this condition

Figure 4-4 reported standardized coefficients of the model. Based on the results, both

PIFC and PTFC had a significant effect on CE. PTFC also had significant effects on MR and

ER. Additionally, ER had a significant effect on CE, but MR had an insignificant effect on CE.

Therefore, Hypothesis 5 was supported in the positive image condition.

Positive issue condition

Figure 4-5 reported factor loadings on CE and ER. Table 4-13 reported goodness of fit

statistics. The model had a good fit.

Figures 4-6 reported standardized coefficients of the model. Based on the results, both

PIFC and PTFC had significant effects on CE. PTFC had a significant effect on ER, but PIFC

had a significant effect on MR. Additionally, both MR and ER had significant effects on CE, but

ER appeared to have a stronger impact than MR. Therefore, Hypothesis 6 was supported in the

positive issue condition.

Negative image condition for the Candidate

Figure 4-7 reported factor loadings on CE and ER. Table 4-13 reported goodness of fit

statistics. The model had a good fit.









Figure 4-8 reported standardized coefficients of the path model. Both PIFC and PTFC had

significant effects on CE. PIFC also had a significant effect on both MR and ER. PTFC had a

significant effect on ER. ER had a significant effect on CE, but MR had an insignificant effect

on CE. Therefore, Hypothesis 5 was supported in the negative image condition for the

Candidate.

Negative image condition for the Opponent

Figure 4-9 reported factor loadings on CE and ER. Table 4-13 reported goodness of fit

statistics. The model had a good fit.

Figures 4-10 reported standardized coefficients of the path model. PIFC had a significant

and negative effect on ER. Both ER and MR had a significant impact on CE, but ER appeared to

have a stronger effect on CE. Therefore, Hypothesis 5 was supported in the negative image

condition for the Opponent.

Negative issue condition for the Candidate

Figure 4-11 reported factor loadings on CE and ER. Table 4-13 reported goodness of fit

statistics. The model had a good fit.

Figures 4-12 reported standardized coefficients of the path model. Both PIFC and PTFC

had a significant effect on CE, and PTFC had significant effects on both MR and ER. PIFC had

a significant effect on MR. Neither MR nor ER had a significant impact on CE. Therefore,

Hypothesis 5 was not supported in the negative issue condition for the Candidate.

Negative issue condition for the Opponent

Figure 4-13 reported factor loadings on CE and ER. Table 4-13 reported goodness of fit

statistics. The model had a good fit.

Figures 4-14 reported standardized coefficients of the path model. Modification indices

suggested a causal arrow from CE to ER, which was opposite to the arrow in the other









conditions. In this condition, PTFC had a significant and negative effect on ER but CE had a

significant and positive effect on ER. The data suggested a different effect hierarchy (Vakratsas

& Ambler, 1999) and Hypothesis 5 was not supported in the negative issue condition for the

Opponent.


























Figure 4-1. Full path diagram of PIFC, PTFC, MR (message really ER (emotional response),
and CE (candidate electability)



CE ER






IC PC VI P A I D



Figure 4-2. Measurement model for CE indicated with IC (issue competence), PC (personal
characteristics), and VI (voting intention) and ER indicated with P (pleasure), A
(arousal), and D (dominance)






















Figure 4-3. Factor loadings in positive image condition
*p<.05


.34*


Figure 4-4. Path coefficients in positive image condition
*p<.05



CE


.59*/ .76* \.64*



IC PC VI



Figure 4-5. Factor loadings in positive issue condition
*p<.05



























Figure 4-6. Path coefficients in positive issue condition
*p<.05



CE


.47*/ .88* \.64* .69*/ .34*



IC PC VI P



Figure 4-7. Factor loadings in negative image condition for the candidate
*p<.05


























.25*


Figure 4-8. Path coeffieients in negative image condition for the candidate
*p<.05



CE El


.59*/ .65* \.48* 51* .07



IC PC VI P A



Figure 4-9. Factor loadings in negative image condition for the opponent
*p<.05




























Figure 4-10. Path coeffieients in negative image condition for the opponent
*p<.05



CE EF


.58*/ .94* \.64* .78*/ .29*



IC PC VI P A



Figure 4-11. Factor loadings in negative issue condition for the candidate
*p<.05


























.29*


Figure 4-12. Path coeffieients in negative issue condition for the candidate
*p<.05



CE El


.40*/ .93* \.52* .50*/ .36*



IC PC ET P A



Figure 4-13. Factor loadings in negative issue condition for the opponent
*p<.05




























Figure 4-14. Path coefficients in negative issue condition for the opponent
*p<.05





Table 4-2. Familiarity with candidates
Candidate Opponent
Mean SD Mean SD
Positive image 1.10 .37
Positive issue 1.05 .22
Negative image 1.17 .44 1.22 .48
Negative issue 1.12 .33 1.19 .40

Table 4-3. Information and transformation
Information Transformation
Mean SD Mean SD
Positive image 3.41 1.09 4.20 1.14
Positive issue 4.46 1.20 3.12 1.10
Negative image 3.06 .99 4.05 1.09
Negative issue 4.33 1.11 3.10 1.07
* p<.05


Table 4-4. Reliability coefficients of informational/transformational scale
Information Transformation
Positive image .81 .93
Positive issue .81 .91
Negative image .78 .87
Negative issue .79 .89


Table 4-1. Personal qualities and political issues
Personal qualities
Mean SD
Positive image 5.61 .97
Positive issue 2.09 1.14
Negative image candidate 4.30 1.58
Negative image opponent 5.20 1.37
Negative issue candidate 1.57 .97
Negative issue opponent 2.41 1.53
*p<.05


Political issues
Mean SD
1.39 .69
5.73 .87
1.41 .79
1.84 1.29
2.27 1.48
5.59 1.23


Paired samples test
t (df = 43)
24.20*
-17.24*
11.74*
13.69*
-3.63*
-11.67*


Paired Samples Test
t (df = 249)
-10.70*
21.19*
-18.26*
16.98*





Image recall Issue recall Total recall
Mean SD Mean SD Mean SD
Positive image 1.83 1.03 .05 .21 1.88 1.11
Positive issue 1.22 1.21 1.48 1.05 2.70 1.48
Negative image 3.08 2.13 .52 .81 3.60 2.03
Negative issue 2.28 2.39 2.23 2.81 4.52 4.09

Table 4-6. Candidate issue competence
Candidate Opponent Paired samples test
Mean SD Mean SD t (df = 249)
Positive image 3.98 .79
Positive issue 4.85 1.14
Negative image 4.71 1.15 3.08 1.17 17.07*
Negative issue 4.02 1.21 4.15 1.49 -1.11
*p<.05

Table 4-7. Candidate characteristics
Candidate Opponent Paired samples test
Mean SD Mean SD t (df = 249)
Positive image 4.33 .76
Positive issue 4.70 1.20
Negative image 3.63 1.03 3.31 .79 4.79*
Negative issue 3.22 1.09 3.77 1.06 -6.98*
*p<.05


Table 4-8. Reliability coefficients of candidate characteristics scale
Candidate Opponent
Positive image .76
Positive issue .93
Negative image .85 .83
Negative issue .87 .93


Table 4-5. Message recall










Table 4-9. Voting intention
Candidate Opponent Paired samples test
Mean SD Mean SD t (df = 249)
Positive image 2.49 .72
Positive issue 2.98 .84
Negative image 2.44 .90 2.36 .80 1.27
Negative issue 2.04 .87 2.40 .98 -4.75*
* p<.05

Table 4-10. Pleasure
Candidate Opponent Paired samples test
Mean SD Mean SD t (df = 249)
Positive image 5.54 1.31
Positive issue 5.98 .78
Negative image 4.82 1.34 3.75 1.48 8.94*
Negative issue 4.75 1.17 3.85 1.37 8.40*
*p<.05

Table 4-11. Arousal
Candidate Opponent Paired samples test
Mean SD Mean SD t (df = 249)
Positive image 3.71 1.91
Positive issue 4.85 1.12
Negative image 4.39 1.75 4.68 1.87 -2.76*
Negative issue 4.13 1.72 4.15 1.69 -.18
*p<.05

Table 4-12. Dominance
Candidate Opponent Paired samples test
Mean SD Mean SD t (df = 249)
Positive image 5.65 1.80
Positive issue 5.20 1.07
Negative image 4.99 1.70 4.93 1.72 .58
Negative issue 5.28 1.59 5.12 1.72 1.71









Table 4-13. Fit indices in experimental conditions
X2 df GFI NFI CFI RMSEA SRMR
Positive image 66.69 22 .94 .94 .95 .10 .05
Positive issue 79.85 22 .94 .93 .94 .10 .07
Negative image candidate 32.33 22 .97 .95 .98 .04 .05
Negative image opponent 64.20 22 .95 .82 .87 .09 .07
Negative issue candidate 52.62 22 .96 .93 .96 .07 .05
Negative issue opponent 51.76 22 .95 .84 .89 .08 .07









CHAPTER 5
DISCUSSION

Overview

This study was designed to investigate the effects of issue/image and positive/negative

political advertisements on message processing and voting behavior with a new measurement

tool of message learning, i.e. the informational/transformational scale (Puto & Wells, 1984). The

constructs in this study were selected based on a chain reaction model of advertising effects

(Lavidge & Steiner, 1961) and the relationships between them were analyzed with t-Tests,

ANOVA, MANOVA, and structural equation modeling. This study generated very interesting

results regarding the mechanism through which different types of political advertisements were

processed. In the following sections, findings related to the hypotheses and research questions

are presented.

Perception of Advertising Content

Hypothesis 1 proposed that issue ads are perceived as more informational than

transformational whereas image ads are perceived as more transformational than informational.

This was supported, as the PIFC and PTFC scores was significantly different for each ad and the

difference was in line with the primary content focus of the ad. Both the positive and negative

image ads were perceived as more transformational than informational, and the transformational

scores were above the scale mid-point whereas the informational scores were below the scale

mid-point. Both the positive and negative issue ads were perceived as more informational than

transformational, and the informational scores were above the scale mid-point whereas the

transformational scores were below the scale mid-point. The findings confirmed the conceptual

similarity between information/transformation and issue/image advocated in previous studies

(Shen, 2007; Shen & Kim, 2006).









Memory

Hypothesis 2 proposed that issue ads are bettered remembered than image ads. This was

supported, as the F-test statistic for emphasis (image vs. issue) was significant and the total recall

of candidate characteristics and candidate issue stands in the two issue ad conditions was higher

than that in the two image ad conditions. The result confirmed previous findings about the

superiority of issue ads in generating higher recall than image ads (Kaid et al., 1992).

Research Question 1 asked how valence (positive vs. negative) influences recall. The F-

test statistic for valence was also significant and the total recall of candidate characteristics and

candidate issue stands in the two negative ad conditions was higher than that in the two positive

ad conditions. Therefore, the answer was that negative ads were better remembered than positive

ads.

Research Question 2 asked how valence interacts with emphasis on recall. The F-test

statistic for the interaction term was not significant. Therefore, the answer was that valence did

not interact with emphasis on recall.

The results about recall should be interpreted with caution. One possible methodological

flaw in this study was that it remained unchecked whether the four ads had the same amount of

message units for recall. Issue ads might be better recalled than image ads because they had

more message units than image ads. Negative ads might be better recalled than positive ads

because they had more candidates (the Candidate and the Opponent) than positive ads (the

Candidate only). It is highly worthwhile to calibrate the message units and keep the amount of

the units constant across experimental conditions in future research on recall of political ads.









Candidate Electability

Issue, Characteristics, and Voting Intention

Candidate electability was measured with issue competence, candidate characteristics, and

voting intention. The results did not provide consistent answer or support to Research Question

3 and Hypotheses 3 and 4. In fact, the results largely depended on the focus of each

measurement and the status of the candidates. For example, in terms of issue competence, issue

ads generated more favorable candidate attitude than image ads, but positive ads did not differ

from negative ads. In terms of candidate characteristics and voting intention, positive ads

generated more favorable candidate attitude than negative ads, but there was no difference

between issue and image ads.

Additionally, the effects of two negative ads varied between the two candidates. For the

Candidate, the negative issue ad led to less favorable candidate electability than the negative

image ad. By sharp contrast, for the Opponent, the negative issue ad led to more favorable

candidate electabililty than the negative image ad. The negative issue ad backfired on the

Candidate in terms of candidate characteristics because the Candidate was less likable than the

Opponent in that condition, which was inconsistent with previous findings about the

effectiveness of negative issue ads to overcome the backlash syndrome (Johnson-Cartee &

Copeland, 1989; Kahn & Geer, 1994; Sonner, 1998). In future research, it is highly worthwhile

to reexamine the relationship between ad content and the backlash effect.

Cognition and Emotion

Hypothesis 5 proposed that emotional response has a larger impact on candidate

electability than cognitive response measured with message recall. This was analyzed through

path analysis, and the hypothesis was supported in most conditions except the negative issue

condition.









In the positive image condition, emotional response had a significant effect on candidate

electability whereas cognitive response had an insignificant effect on candidate electability. In

the positive issue condition, both emotional response and cognitive response had a significant

positive impact on candidate electability, but the emotional response appeared to have a stronger

effect on candidate electability.

In the negative image for the Candidate condition, emotional response had a significant

effect on candidate electability whereas cognitive response had an insignificant effect on

candidate electability. In the negative image for the Opponent condition, both emotional

response and cognitive response had a significant positive impact on candidate electability, but

the emotional response appeared to have a stronger effect on candidate electability.

In the negative issue for the Candidate condition, both emotional and cognitive responses

had an insignificant effect on candidate electability. In the negative issue for the Opponent

condition, again both emotional and cognitive responses had an insignificant effect on candidate

electability. In fact, in that condition, there was a significant effect from candidate electability to

emotional response.

The results in the positive ad condition further confirmed the conceptual similarity among

image, transformation and emotion. Image is analogous to transformation (Shen, 2007) and

transformation focuses on the emotional aspect of advertising (Puto & Wells, 1984); therefore, it

is not surprising that emotional response was much more powerful than cognitive response in

determining candidate electability in the positive image condition. Similarly, issue is analogous

to information (Shen, 2007) and information focuses on the cognitive aspect of advertising (Puto

& Wells, 1984); therefore, cognitive response gained significance in the positive issue condition.









The results in the negative ad condition are also very interesting. In the negative image for

the Candidate condition, the effects of emotional and cognitive response were identical to those

in the positive issue condition. In the negative image for the Opponent condition, the effects

were identical to those in the positive image condition. A close look into the message recall for

this condition indicated that the participants inferred unmentioned issues, such as campaign

finance and tax, from the negative image ad. It was likely that those implicitly inferred issues

influenced the effects of emotional and cognitive responses in a way similar to the explicitly

mentioned tax issue in the positive issue condition.

Furthermore, the insignificant effects of both emotional and cognitive response in the

negative issue for the Candidate condition may be due to limited message about the candidate.

The only message about the Candidate was a still image of the Candidate and a voice-over "I'm

Chris Choloca. I approve this message." The opposite direct effect from candidate electability

to emotional response in the negative issue for the Opponent condition might mean that the

participants had some existing attitude about candidates who advocated higher tax, and this

existing attitude influenced emotional response to the Opponent who also supported higher tax.

Memory, Emotion, and Candidate Electability

Research Questions 4 and 5 investigated the complete chain reaction process from message

learning, message recall, emotion, and electability in all the experimental conditions. This was

also analyzed through path analysis.

In the positive image condition, PTFC had a significant impact on message recall whereas

PIFC had an insignificant impact on message recall. This pattern made good sense because

image and transformational cues were the primary content of the positive image ad. PTFC also

had a significant impact on emotional response whereas PIFC still had an insignificant impact on

emotional response. The relationship between PTFC and emotional response was well









understandable because transformational cues were the emotional message of an ad (Puto &

Wells, 1984). Additionally, both PIFC and PTFC had a significant and direct impact on

candidate electability.

In the positive issue condition, PTFC had an insignificant impact on message recall

whereas PIFC had a significant impact on message recall. This pattern made good sense because

issue and informational cues were the primary content of the positive issue ad. Both PTFC and

PIFC had a significant impact on emotional response. In other words, besides the conceptually

determined relationship between PTFC and emotional response (Puto & Wells, 1984), PIFC, the

primary content in the positive issue ad, served as another important source of emotional

response. Additionally, both PIFC and PTFC had a significant and direct impact on candidate

electability.

In the negative image for the Candidate condition, PTFC had an insignificant impact on

message recall whereas PIFC had a significant impact on message recall. Both PTFC and PIFC

had a significant impact on emotional response. The patterns in this condition were opposite to

those in the positive image condition but were identical to that in the positive issue condition for

the same candidate. There might be inferred issue learning, such as campaign finance and tax,

from the negative image condition, which led to issue-induced effects as recorded in the positive

issue condition. Again, both PIFC and PTFC had a significant and direct impact on candidate

electability in this condition.

In the negative image for the Opponent condition, neither PTFC nor PIFC had a significant

impact on message recall. PTFC had a significant impact on emotional response, which could be

explained with the conceptual similarity between transformation and emotion (Puto & Wells,

1984). PIFC also had a significant impact on emotional response, which was also found in the










positive issue condition. It was likely that the inferred issue learning became a significant source

of emotional response. The impact from PIFC to emotional response was negative, which was in

line with the valence of the ad. Finally, neither PIFC nor PTFC had a significant and direct

impact on candidate electability in this condition.

In the negative issue for the Candidate condition, both PIFC and PTFC had a significant

impact on message recall, and the impact from PTFC was negative. PTFC also had a significant

impact on emotional response, whereas PIFC had an insignificant impact on emotional response.

The patterns in this condition seemed to suggest that transformational cues or the secondary

content in this condition had a stronger impact on message recall and emotional response than

informational cues or the primary content. Finally, both PIFC and PTFC had a significant and

direct impact on candidate electability in this condition.

In the negative issue for the Opponent condition, neither PTFC nor PIFC had a significant

impact on message recall. PTFC had a significant impact on emotional response, whereas PIFC

had an insignificant impact on emotional response. Neither PIFC nor PTFC had a significant and

direct impact on candidate electability in this condition. Finally, unlike the other conditions, the

causal arrow between emotional response and candidate attitude was from candidate electability

to emotional response.

Limitations and Future Research

There are a few methodological and theoretical limitations in this study. They should be

well addressed in future research.

First, message recall was a problematic construct in the study. There was no check on the

amount of message units in each ad, and it is impossible to rule out that some ads were better

recalled because there were more message units in them. It is also questionable to add message

recall into the original chain reaction model. College students are known for low involvement









with political campaigns (Yoon et al., 2005) and low involvement makes recall an inadequate

predictor of memory-based effects (Petty, Cacioppo, & Goldman, 1981). In future research,

more efforts should be spent on equivalent amount of message units for recall and enough

involvement to generate adequate impact from recall to other behavioral variables.

Second, the Scale might not fit negative advertising. Unlike most product advertising and

positive political advertising which features one product or one candidate, there are two

candidates in a negative ad and the perception of informational and transformational cues could

be very different. It is possible that the Scale might not adequately meet this increased

complexity of message content. The participants might experience difficulty in choosing the

right answers to the Scale items. The increased measurement error, as indicated by lower

reliability coefficients in the negative ad condition, led to a bad fit of the model to the data. This

problem was alleviated with the latent PIFC and PTFC scores from IRT because they were

assumed to have zero measurement error. In future research, it is very important to develop new

PIFC and PTFC items for negative advertising.

The use of college students was another limitation. College students generally have low

interests in politics (Yoon et al., 2005). This characteristic is a possible impediment to the

generalizability of the study results in regard to the broader voter population. In future research,

it is very important to include non-student population to test the proposed model.

A further limitation may be inherent in the limited number of ads for this study. In future

research, it is very worthwhile to including a broader spectrum of issues and image at different

levels of political races.









CHAPTER 6
IMPLICATIONS

This study provides various theoretical and managerial implications to political

campaigning practitioners and researchers. In order to investigate the underlying mechanism of

candidate attitude formulation in response to political advertising, this study was among the first

to empirically test the chain reaction theory of advertising effects (Lavidge & Steiner, 1961). In

addition, this study used the informational/transformational scale (Puto & Wells, 1984) to

measure message learning, which has never been used for political advertisements, in spite of its

theoretical applicability to issue and image political advertisements (Shen, 2007). Furthermore,

this study balanced both emotional and cognitive antecedent to candidate electability with the

use of a non-verbal emotional measure known as AdSAM" (Morris, 1995) and message recall

(Petty et al., 1981). This study also balanced the emotional and cognitive perspectives of

candidate electability with a tripartite structure composed of issue competence, candidate

characteristics, and voting intention (Kahn & Geer, 1994). The attempt to link

information/transformation, issue/image, and cognition/emotion proved a legitimate approach to

examine the dynamics of issue/image and positive/negative advertisements in voting behavior.

Finally, besides structural equation modeling, item response theory was used to estimate latent

response to issue and image content in political advertisement and improve model fitting because

of reduced measurement error. This approach is well applicable to scale-based research.

The managerial implications are particularly evident from a copy-testing standpoint for

political campaigning. With some modifications, the informational/transformational scale can be

a valid measure of message learning for issue and image in political advertisements, especially

for a new and unfamiliar candidate. The enhanced calibration of advertising content will

eventually help identify a manner in which issue and image can be managed that will lead to










optimal mental processing and favorable candidate evaluation. The real-life complexity of using

positive or negative issue and image to support or attack a candidate was also investigated in this

study, and the four message conditions in this research can be used as the base category for a

broader range of message strategies. Furthermore, emotional response had a stronger impact

than cognitive response on candidate electability, or it was what the respondents felt about the

candidates rather than what they remembered about the candidates from the political

advertisements that shaped their judgment on the candidate. This finding will provide directions

to create effective advertisements that serve to create the right emotion for the right candidate.

Finally, given the enormous amount of investment in political advertising, this line of research

will become increasing critical in light of the value of democracy and informed voting decision.









APPENDIX A
QUESTIONNAIRES

Pretest

1. How much did you learn about the government policies/political issues related to the
candidate from the ad you just watched?

Very little 1 2 3 4 5 6 7 A great deal

2. How much did you learn about the personal qualities related to the candidate from the ad you
just watched?

Very little 1 2 3 4 5 6 7 A great deal

3. What political issues were mainly associated the candidate in the ad?



4. What characteristics of the candidate stand out in your mind?
Describe the candidate in adj ective forms









Main Survey Sample (Negative Advertisement)


Instructions (continued)
Q1. Have you ever watched this advertisement before?
<1> Never <2> A few times <3> Many times

Q2. Please use the scale below to indicate your familiarity with Chris Chocola, the candidate.
<1> Very unfamiliar
<2> Unfamiliar
<3> Neither unfamiliar nor familiar
<4> Familiar
<5> Very familiar

Q3. Please use the scale below to indicate your familiarity with Joe Donnelly, the candidate's
opponent.
<1> Very unfamiliar
<2> Unfamiliar
<3> Neither unfamiliar nor familiar
<4> Familiar
<5> Very familiar

Please use the scale to indicate how you feel about Chris Chocola after watching. the ad.

HAPPY UNHAPPY
or or
ELATED SAD
000000000

STIMULATEDr dc d CALM

INVOLVED 00 00 00 0 00 BORED


NOT IN de6~IN-CONTROL
CONTROL 00 00 00 0 00 or
EMPOWERED










Please use the scale to indicate how you feel about Joe Donnelly after watching the ad.


HAPPY
or
ELATED


UNHAPPY
or
SAD


STIMULATED
or
INVOLVED


CALM
or
BORED


0 0 0


NOT IN
CONTROL


IN-CONTROL
or
EMPOWERED


Instructions (Continued)
Please circle the number that best indicates the degree of your agreement or disagreement with
each of the statements listed below. There are no correct or incorrect answers to those
statements.

Questions
Q1A. I learned something from this advertisement that I didn't know before about Chris
Chocola, the candidate.
-4 -3 -2 -1 0 1 2 3 4


very strong
disagreement


very strong
agreement


Q1B. I learned something from this advertisement that I didn't know before about Joe
Donnelly, the candidate's opponent.
-4 -3 -2 -1 0 1 2 3 4


very strong
disagreement


very strong
agreement


Q2. I would like to have an experience like the one shown in the advertisement.
-4 -3 -2 -1 0 1 2 3
very strong
disagreement

Q3. The advertisement did not seem to be speaking directly to me.
-4 -3 -2 -1 0 1 2 3
very strong
disagreement


very strong
agreement


very strong
agreement


0 00 0










Q4. There is nothing special about Chris Chocola, the candidate, that makes him different
from Joe Donnelly, the candidate's opponent.
-4 -3 -2 -1 0 1 2 3 4


very strong
disagreement


very strong
agreement


Q5. While I watched this advertisement, I thought how Chris Chocola, the candidate,
might be useful to me.
-4 -3 -2 -1 0 1 2 3 4


very strong
disagreement


very strong
agreement


Chris Chocola, the



very strong
agreement


Q6. The advertisement did not teach me what to look for when choosing
candidate.
-4 -3 -2 -1 0 1 2 3
very strong
disagreement

Q7. This advertisement was meaningful to me.
-4 -3 -2 -1 0 1 2 3
very strong
disagreement

Q8. This advertisement was very uninformative.
-4 -3 -2 -1 0 1 2 3
very strong
disagreement

Q9. Chris Chocola, the candidate, fits my lifestyle very well.
-4 -3 -2 -1 0 1 2 3
very strong
disagreement


very strong
agreement


very strong
agreement


very strong
agreement


Q10. I could really relate to this advertisement.
-4 -3 -2 -1 0
very strong
disagreement


very strong
agreement


Q11. Choosing Chris Chocola, the candidate, makes me feel good about myself.
-4 -3 -2 -1 0 1 2 3
very strong
disagreement


very strong
agreement










Q12. If they had to, Chris Chocola's party could provide evidence to support the claims
made in this advertisement.
-4 -3 -2 -1 0 1 2 3 4


very strong
disagreement


very strong
agreement


Q13. It's hard to give a specific reason, but somehow Chris Chocola, the candidate, is not
really for me.
-4 -3 -2 -1 0 1 2 3 4


very strong
disagreement


very strong
agreement


Q14. This advertisement did not really hold my attention.
-4 -3 -2 -1 0 1
very strong
disagreement


very strong
agreement


Q15A. This advertisement reminded me of some important facts about Chris Chocola, the
candidate, which I already knew.
-4 -3 -2 -1 0 1 2 3 4


very strong
disagreement


very strong
agreement


Q15B. This advertisement reminded me of some important facts about Joe Donnelly, the
candidate's opponent, which I already knew.
-4 -3 -2 -1 0 1 2 3 4


very strong
disagreement


very strong
agreement


Q16. If I could change my lifestyle, I would make it less like the people who choose Chris
Chocola, the candidate.
-4 -3 -2 -1 0 1 2 3 4


very strong
disagreement


very strong
agreement


Q17A. When I think of Chris Chocola, the candidate, I think of this advertisement.
-4 -3 -2 -1 0 1 2 3 4


very strong
disagreement


very strong
agreement


Q17B. When I think of Joe Donnelly, the candidate's opponent, I think of this
advertisement.
-4 -3 -2 -1 0 1 2 3
very strong
disagreement


very strong
agreement










Q18. I felt as though I were right there in the advertisement, experiencing the same thing.
-4 -3 -2 -1 0 1 2 3 4


very strong
disagreement


very strong
agreement


Q19. I can now accurately compare Chris Chocola, the candidate, with Joe Donnelly, the
candidate's opponent, on matters that are important to me.
-4 -3 -2 -1 0 1 2 3 4


very strong
disagreement


very strong
agreement


Q20. This advertisement did not remind me of any experience or feelings I've had in my
own life.
-4 -3 -2 -1 0 1 2 3 4


very strong
disagreement


very strong
agreement


Q21. I would have more confidence in my ability to judge the merits of choosing Chris
Chocola, the candidate, now than before I saw this advertisement.
-4 -3 -2 -1 0 1 2 3 4


very strong
disagreement


very strong
agreement


Q22. It is the kind of advertisement that keeps running through your head after you've
seen it.
-4 -3 -2 -1 0 1 2 3 4


very strong
disagreement


very strong
agreement


Q23. It's hard to put into words, but this advertisement leaves me with a good feeling about
Chris Chocola, the candidate.
-4 -3 -2 -1 0 1 2 3 4


very strong
disagreement


very strong
agreement


Q24. What personal qualities/characteristics of Chris Chocola, the candidate, stand out in
your mind?
Describe the candidate in adj ective forms




Q25. What personal qualities/characteristics of Joe Donnelly, the candidate's opponent,
stand out in your mind?
Describe the candidate in adj ective forms










Q26. What government policies/political issues were mainly associated with Chris Chocola,
the candidate, in the ad?




Q27. What government policies/political issues were mainly associated with Joe Donnelly,
the candidate's opponent, in the ad?




Q28. What is your best guess about Chris Chocola's competence in dealing with issues
related to tax?


very
competent


very
incompetent


Q29. What is your best guess about Joe Donnelly's competence in dealing with issues
related to tax?


very
competent


very
incompetent


Q30. How much would you agree that "trustworthy" fits your impression of Chris
Chocola?


strongly
disagree


strongly
agree


Q31. How much would you agree that "provides strong leadership" fits your impression of
Chris Chocola?


strongly
disagree


strongly
agree


Q32. How much would you agree that "hardworking" fits your impression of Chris
Chocola?


strongly
disagree


strongly
agree









Q33. How much would you agree that "knowledgeable" fits your impression of Chris
Chocola?


strongly
disagree


strongly
agree


Q34. How much would you agree that "trustworthy" fits your impression of Joe Donnelly?


strongly
disagree


strongly
agree


Q35. How much would you agree that "provides strong leadership" fits your impression of
Joe Donnelly?


strongly
disagree


strongly
agree


Q36. How much would you agree that "hardworking" fits your impression of Joe
Donnelly?


strongly
disagree


strongly
agree


Q37. How much would you agree that "knowledgeable" fits your impression of Joe
Donnelly?


strongly
disagree


strongly
agree


Q38. How likely would you vote for Chris Chocola?


4
likely


4
likely


5
very likely


5
very likely


very unlikely


unlikely


have no opinion


Q39. How likely would you vote for Joe Donnelly?


very unlikely


unlikely


have no opinion














Table B-1. Positive image condition
Variable 1 2 3 4 5 6 7 8 9
1. PIFC 1.00
2. PTFC 0.56 1.00
3. Message Recall 0.06 0.14 1.00
4. Issue Competency 0.34 0.27 0.02 1.00
5. Personal Characteristics 0.50 0.57 0.17 0.43 1.00
6. Vote Intention 0.44 0.58 0.03 0.37 0.57 1.00
7. Pleasure 0.36 0.58 0.04 0.29 0.44 0.45 1.00
8. Arousal 0.29 0.55 0.09 0.16 0.34 0.32 0.56 1.00
9. Dominance 0.26 0.29 0.03 0.29 0.25 0.22 0.55 0.34 1.00

Table B-2. Positive issue condition
Variable 1 2 3 4 5 6 7 8 9
1. PIFC 1.00
2. PTFC 0.67 1.00
3. Message Recall 0.16 0.08 1.00
4. Issue Competency 0.28 0.29 0.23 1.00
5. Personal Characteristics 0.49 0.52 0.21 0.47 1.00
6. Vote Intention 0.42 0.52 0.20 0.35 0.47 1.00
7. Pleasure 0.42 0.55 0.05 0.32 0.44 0.44 1.00
8. Arousal 0.34 0.43 0.14 0.28 0.39 0.30 0.58 1.00
9. Dominance -0.04 -0.01 -0.14 0.15 -0.02 0.05 0.20 0.15 1.00

Table B-3. Negative image condition for the candidate
Variable 1 2 3 4 5 6 7 8 9
1. PIFC 1.00
2. PTFC 0.60 1.00
3. Message Recall 0.17 0.10 1.00
4. Issue Competency 0.32 0.26 0.15 1.00
5. Personal Characteristics 0.48 0.50 0.09 0.42 1.00
6. Vote Intention 0.37 0.43 0.08 0.25 0.57 1.00
7. Pleasure 0.24 0.28 -0.08 0.25 0.37 0.29 1.00
8. Arousal 0.22 0.22 0.06 0.11 0.26 0.22 0.19 1.00
9. Dominance 0.12 0.08 -0.01 0.13 0.11 0.15 0.30 0.13 1.00


APPENDIX B
CORRELATION MATRICES










Table B-4. Negative image condition for the opponent
Variable 1 2 3 4 5 6 7 8 9
1.00
2. PTFC 0.60 1.00
3. Message Recall 0.19 0.16 1.00
4. Issue Competency -0.18 -0.20 -0.17 1.00
5. Personal Characteristics -0.17 -0.04 -0.21 0.38 1.00
6. Vote Intention -0.07 -0.02 -0.02 0.23 0.36 1.00
7. Pleasure -0.30 -0.10 -0.18 0.31 0.15 0.16 1.00
8. Arousal 0.18 0.20 0.09 0.08 0.01 0.07 0.03 1.00
9. Dominance -0.07 -0.01 -0.01 0.17 0.20 0.13 0.22 -0.01 1.00

Table B-5. Negative issue condition for the candidate
Variable 1 2 3 4 5 6 7 8 9
1. PIFC 1.00
2. PTFC 0.62 1.00
3. Message Recall 0.19 -0.01 1.00
4. Issue Competency 0.29 0.38 -0.02 1.00
5. Personal Characteristics 0.46 0.48 -0.02 0.54 1.00
6. Vote Intention 0.38 0.49 -0.12 0.42 0.59 1.00
7. Pleasure 0.31 0.38 -0.05 0.17 0.29 0.24 1.00
8. Arousal 0.16 0.26 0.11 0.03 0.13 0.08 0.22 1.00
9. Dominance 0.04 0.05 0.02 -0.04 0.12 0.17 0.26 0.11 1.00

Table B-6. Negative issue condition for the opponent
Variable 1 2 3 4 5 6 7 8 9
1. PIFC 1.00
2. PTFC 0.64 1.00
3. Message Recall 0.01 0.06 1.00
4. Issue Competency -0.09 -0.04 -0.06 1.00
5. Personal Characteristics 0.11 0.07 -0.05 0.36 1.00
6. Vote Intention -0.06 -0.12 -0.03 0.35 0.47 1.00
7. Pleasure -0.09 -0.21 -0.12 0.12 0.12 0.22 1.00
8. Arousal 0.08 0.09 0.04 -0.01 0.13 0.02 0.14 1.00
9. Dominance 0.04 -0.04 -0.03 -0.04 -0.03 0.01 0.29 0.22 1.00










LIST OF REFERENCES

Aaker, D. A., & Bruzzone, D. E. (1985). Causes of irritation in advertising. Journal of
Marketing, 49, 47-57.

Aaker, D. A., & Norris, D. (1982). Characteristics of TV commercials percieved as informative.
Journal ofAdvertising Research, 22, 61-70.

Aaker, D. A., & Stayman, D. M. (1992). Implementing the concept of transformational
advertising. Psychology & Marketing, 9, 237-253.

Aaker, D. A., Stayman, D. M., & Hagerty, M. R. (1986). Warmth in advertising: Measurement,
impact, and sequence effects. Journal of Consumer Research, 12, 365-381.

Aaker, J. L. (1997). Dimensions of brand personality. Journal of2\a~rketing Research, 34, 347-
356.

Anderson, L. R., & McGuire, W. J. (1965). Prior reassurance of group consensus as a factor in
producing resistance to persuasion. Socionsetry, 28, 44-56.

Andreasen, N. C. (1997). Linking mind and brain in the study of mental illnesses: A proj ect for a
scientific psychopathology. Science, 275, 1586-1593.

Averill, J. R. (1975). A semantic atlas of emotional concepts. JSAS Catalogue of Selected
Documents in Psychology, 5, 330.

Averill, J. R. (1982). Anger and aggress.~ionr: An essay on emotion. New York: Springer.

Baines, P. R., Worcester, R. M., Jarrett, D., & Mortimore, R. (2005). Product attribute-based
voter segmentation and resource advantage theory. Journal of2arketing Management,
21, 1079-1115.

Belk, R. W., Balm, K. D., & Mayer, R. N. (1982). Development recognition of consumption
symbolism. Journal of Consumer Research, 9, 4-16.

Ben-Ur, J., & Newman, B. I. (2002). Motives, perceptions and voting intention of voters in the
2000 U.S. presidential election. Psychology & Marketing, 19, 1047-1065.

Berelson, B. (1966). Democratic theory and public opinion. In B. Berelson & M. Janowitz
(Eds.), Reader in public opinion and conanunication (pp. 489-540). New York: Free
Press.

Blumenthal, S. (1980). The permanent campaign: hIside the world of elite political operatives.
Boston: Beacon.

Boster, F. J., & Mongeau, P. (1984). Fear-arousing persuasive messages. In R. N. Bostrom & B.
H. Westley (Eds.), Conanunication Yearbook (Vol. 8, pp. 330-375). Newbury Park, Ca:
Sage.










Bradley, M. M., Greenwald, M. K., & Hamm, A. O. (1993). Affective picture processing. In
N.Birbaumer & A. Ohman (Eds.), The structure of emotion: Psychophysiological,
cognitive and clinical aspects. Toronto: Hogrefe & Huber.

Brooker, G. (1981). A comparison of the persuasive effects of mild humor and mild fear appeals.
Journal ofAdvertising, 10, 29-40.

Browne, M. W., & Cudeck, R. (1993). Alternative ways of assessing model fit. In K. A. Bollen
& J. S. Long (Eds.), Testing structural equation models (pp. 136-162). Newbury Park,
CA: Sage.

Bulter, P., & Collins, N. (1994). Political marketing: Structure and process. European Journal of
Marketing, 28, 19-34.

Bush, L. E., II. (1973). Individual differences in multidimensional scaling of adjectives denoting
feelings. Journal ofPersonality and Social Psychology, 25, 50-57.

Cacioppo, J. T., & Petty, R. E. (1989). The Elaboration Likelihood Model: The role of affect and
affect-laden information processing in persuasion. In P. Cafferata & A. Tybout (Eds.),
Cognitive and A~ffective Responses to Advertising (pp. 69-89). Lexington, MA: Lexington
Books .

Campbell, A., Converse, P., Miller, W. E., & Stokes, D. E. (1960). The American voter. New
York: John Wiley and Sons.

Campbell, A., Gurin, G., & Miller, W. E. (1954). The voter decides. Evanston, 1L: Row, Peterson
and Co.

Chaiken, S. (1980). Heuristic versus systematic information processing and the use of source
versus message cues in persuasion. Journal ofPersonality and Social Psychology, 39,
752-766.

Chang, C. (2001). The impacts of emotion elicited by print political advertising on candidate
evaluation. Media Psychology, 3, 9 1-1 19.

Christ, W. G., Thorson, E., & Caywood, C. (1994). Do attitudes toward political advertising
affect information processing of televised political commercials? Journal of
Broadcasting & Electronic M~edia, 38, 251-270.

Clark, M. S. (1982). A role for arousal in the link between feeling states, judgments, and
behavior. In M. S. Clark & S. T. Fiske (Eds.). Hillsdale, NJ: Erlbaum.

Cohen, J., & Basu, K. (1987). Alternative models of categorization: Toward a contingent
processing framework. Journal of Consumer Research, 13, 455-472.

Darke, P. R., Freedman, J. L., & Chaiken, S. (1995). Percentage discounts, initial price, and
bargain hunting: A heuristic-systematic approach to price search behavior. Journal of
Applied Psychology, 80, 580-586.










Davis, J., J. (1997). Advertising Research: Theory and Practice. Upper Saddle River, NJ.:
Prentice Hall.

Decety, J., & Chaminade, T. (2003). Neural correlates of feeling sympathy. Neuropsychologia,
41, 127-138.

Deighton, J., Romer, D., & McQueen, J. (1989). Using drama to persuade. Journal of Consumer
Research, 16, 335-343.

Dermody, J., & Scullion, R. (2001). An exploration of the advertising ambitions and strategies of
the 2001 British General Election. Journal of2\a~rketing2\anan~gement, 1 7, 969-987.

Devlin, L. P. (1989). Contrasts in presidential campaign commercials of 1988. American
Behavioral Scientist, 32, 389-414.

Devlin, L. P. (1993). Contrasts in presidential campaign commercials of 1992. American
Behavioral Scientist, 37, 272-290.

Devlin, L. P. (1997). Contrast in presidential campaign commercials of 1996. American
Behavioral Scientist, 40, 1058-1084.

Devlin, L. P. (2001). Contrast in presidential campaign commercials of 2000. American
Behavioral Scientist, 44, 2338-2369.

Devlin, L. P. (2005). Contrasts in presidential campaign commercials of 2004. American
Behavioral Scientist, 49, 279-3 13.

Duncan, C. P., & Nelson, J. E. (1985). Effects of humor in a radio advertising experiment.
Journal ofAdvertising, 14, 33-40.

Duncan, T. (2002). IM~C: Using advertising and promotion to build brands. New York:
McGraw-Hill.

Edell, J. A., & Burke, M. C. (1987). The power of feelings in understanding advertising effects.
Journal of Consumer Research, 14, 42 1-433 .

Ekman, P., & Friesen, W. V. (1978). Facial action coding system (FACS): A technique for the
measurement of facial action. Palo Alto, CA: Consulting Psychologists Press.

Embretson, S. E., & Reise, S. P. (2000). Item response theory for psychologists. Mahwah, NJ:
Lawrence Erlbaum Associates.

Ewing, M. T., Salzberger, T., & Sinkovics, R. R. (2005). An alternate approach to assessing
cross-cultural measurement equivalence in advertising research. Journal ofAdvertising,
34, 17-36.










Faber, R. J., Tims, A. R., & Schmitt, K. G. (1993). Negative political advertising and voting
intent: The role of involvement and alternative information sources. Journal of
Advertising, 22, 67-76.

Feldman, L. A., Barrett, L., & Russell, J. A. (1998). Independence and bipolarity in the structure
of affect. Journal ofPersonality and Social Psychology, 74, 967-984.

Fishbein, M., & Ajzen, I. (1975). BeliefJ attitude, intention and behavior: An introduction to
theory and research. Reading, MA: Addison-Wesley.

Fishbein, M., & Ajzen, I. (1980). Predicting and understanding consumer behavior: Attitude-
behavior correspondence. In I. Ajzen & M. Fishbein (Eds.), Understanding attitudes and
predicting social behavior (pp. 148-172). Eaglewood Cliffs, NJ: Prentice-Hall.

Fiske, S. T., & Taylor, S. E. (1984). Social cognition. New York: Random House.

Frijda, N. H. (1986). The emotions. New York: Cambridge University Press.

Geer, J. G., & Geer, J. H. (2003). Remembering attack ads: An experimental investigation of
radio. Political Behavior, 25, 69-95.

Gelb, B. D., & Pickett, C. M. (1983). Attitude-toward-the-ad: Links to humor and to advertising
effectiveness. Journal ofAdvertising, 12, 34-42.

Graber, D. A. (2004). Methodological developments in political communication research. In L.
L. Kaid (Ed.), H~andbook of political communication research (pp. 45-67). Mahwah, New
Jersey: Lawrence Erlbaum Associates.

Greenwald, A. G. (1989). Why are attitudes important? In A. G. Greenwald (Ed.), Attitude
structure and function (pp. 1-10). Hillsdale, New Jersey: Lawrence Erlbaum Associates.

Hair, J. F. J., Anderson, R. E., Tatham, R. L., & Black, W. C. (1995). M~ultivariate data analysis.
Upper Saddle River, New Jersey: Prentice-Hall.

Hambleton, R. K., & Swaminathan, H. (1985). Item response theory: Principles and
applications. Boston: Kluwer-Nijhoff.

Henneberg, S. C. (2004). The views of an advocates dei: Political marketing and its critics.
Journal of Public Affairs, 4, 225-243.

Hill, R. P. (1989). An exploration of voter responses to political advertisements. Journal of
Advertising, 18, 14-22.

Hoyer, W. D., & Maclnnis, D. J. (2004). Consumer behavior (3rd ed.). Boston, MA: Houghton
Mifflin.

Izard, C. E. (1977). The emotions in life and science. In C. E. Izard (Ed.), H~uman emotions (pp.
1-18). New York: Plenum Press.










Izard, C. E. (1992). Basic emotions, relations among emotions, and emotion-cognition relations.
Psychological Review, 99, 561-565.

Izard, C. E. (1993). Four systems for emotion activation: Cognitive and non-cognitive processes.
Psychology Review, 100, 68-90.

Izard, C. E., & Bartlett, E. S. (1972). Patterns ofemotions: A new analysis ofanxiety and
depression. Oxford, England: Academic Press, 1972.

James, W. (1884). What is an emotion? M~ind, 9, 188-205.

Johnson-Cartee, K. S., & Copeland, G. A. (1989). Southern voters' reaction to negative political
ads in 1986 election. Journalism and Ma~ss Conanunication Quarterly, 66, 888-893, 986.

Johnston, A., & Kaid, L. L. (2002). Image ads and issue ads in U.S. presidential advertising:
Using videostyle to explore stylistic differences in televised political ads from 1952 to
2000. Journal of Conanunication, 52, 28 1-3 00.

Joreskog, K. G., & Sorbom, D. (2005). Lisrel 8.72. Lincolnwood, IL: Scientific Software
International, Inc.

Joslyn, R. A. (1980). The content of political spot ads. Journalism Quarterly, 57, 92-98.

Kahn, K. F., & Geer, J. G. (1994). Creating impressions: An experimental investigation of
political advertising on television. Political Behavior, 16, 93-116.

Kahn, K. F., & Kenney, P. J. (2000). How negative campaigning enhances knowledge of Senate
elections. In J. A. Thurber, C. J. Nelson & D. A. Dulio (Eds.), Crowdedairwaves:
Campaign advertising in elections (pp. 65-95). Washtington, DC: Brooking Institution.

Kaid, L. L. (2001). Technodistortions and effects of the 2000 political advertising. American
Behavioral Scientist, 44, 2370-2378.

Kaid, L. L. (2002). Videostyle and political advertising effects in the 2000 presidential
campaign. In R. E. J. Denton (Ed.), The 2000 presidential campaign: A conanunication
perspective (pp. 183-197). Westport, CT: Praeger.

Kaid, L. L. (2004). Political advertising. In L. L. Kaid (Ed.), Handbook ofpolitical
conanunication research (pp. 155-202). Mahwah, NJ: Lawrence Erlbaum Associates.

Kaid, L. L. (2005). Videostyle in the 2004 political advertising. In R. E. J. Denton (Ed.), The
2004 presidential campaign: A conanunication perspective (pp. 283 -299). Lanham, MD:
Rowman & Littlefield.

Kaid, L. L., Chanslor, M., & Hovind, M. (1992). The influence of program and commercial type
on political advertising effectiveness. Journal of Broadcasting & Electronic M~edia, 36,
303-320.










Kaid, L. L., & Johnston, A. (2001). Videostyle in presidential campaigns: Style and content of
televised political advertising. Westport, CT: Praeger.

Kaid, L. L., & Postelnicu, M. (2005). Political advertising in the 2004 Election: Comparison of
traditional television and Internet message. American Behavioral Scientist, 49, 265-278.

Kaid, L. L., & Sanders, K. R. (1978). Political television commercials: An experimental study of
type and length. Conanunication Research, 5, 57-70.

Kaid, L. L., & Tecesco, J. C. (1999). Tracking voter reactions to television advertising. In L. L.
Kaid & D. G. Bystrom (Eds.), The electronic election: Perspectives on the 1996
campaign conanunication (pp. 233-246). Mahwah, NJ: Lawrence Earlbaum.

King, L. R. (1965). The marketing concept. In G. Schwartz (Ed.), Science in marketing. New
York: John Wiley and Sons.

Kitayama, S., Markus, H. R., Tummala, P., Kurokawa, M., & Kato, K. (1990). Culture and self-
cognition.Unpublished manuscript.

Kotler, P. (1975). Overview of political candidate marketing. Advances in Consumer Research,
2, 761-769.

Lang, A. (1991). Emotion, formal features, and memory for televised political advertisements. In
F. Biocca (Ed.), Television and political advertising: Psychological processes (Vol. 1,
pp. 221-243). Hillsdale, New Jersey: Lawrence Erlbaum Associates.

Lang, P. J. (1980). Behavioral treatment and bio-behavioral assessment: Computer applications
b. In J. B. Sidowski, J. H. Johnson & T. A. Williams (Eds.), Technology in nzental health
care delivery systems (pp. 119-137). Norwood, NJ: Ablex.

Larsen, R. J., & Diener, E. (1992). Promises and problems with the circumplex model of
emotion. In M. S. Clark (Ed.), Review of personality and social psychology: Emotion
(Vol. 13, pp. 25-59).

Laskey, H. A., Day, E., & Crask, M. R. (1989). Typology of main message strategies for
television commercials. Journal ofAdvertising, 18, 36-41.

Lavidge, R. J., & Steiner, G. A. (1961). A model for predictive measurements of advertising
effectiveness. Journal of2arketing, 25, 59-62.

Lazarus, R. S. (1966). Psychological stress and the coping process. New York: McGraw-Hill.

Lazarus, R. S. (1984). On the primacy of cognition. American Psychologist, 39, 124-129.

Lazarus, R. S. (1991a). Emotion and adaptation. New York: Oxford University Press.

Lazarus, R. S. (1991b). Progress on a cognitive-motivational-relational theory of emotion.
American Psychologist, 46, 8 19-834.










LeDoux, J. E. (1995). Emotion: Clues from the brain. Annual Review ofPsychology, 46, 209-
235.

LeDoux, J. E., Iwata, J., Cicchetti, P., & Reis, D. J. (1988). Different projections of the central
amygdaloid nucleus mediate autonomic and behavioral correlates of conditioned fear.
Journal ofNeuroscience, 8, 2517-2529.

LeDoux, J. E., Romanski, L., & Xagoraris, A. (1989). Indelibility of subcortical emotional
memories. Journal of Cognitive Neuroscience, 1, 23 8-243.

Linden, W. J. v. d., & Hambleton, R. K. (1996). Handbook of modern item response theory.
New York: Springer-Verlag.

Lock, A., & Harris, P. (1996). Political marketing vive la difference! European Journal of
Marketing, 30, 21-31.

Maarek, P. J. (1995). Political marketing and communication. London: John Libby.

MacInnis, D. J., & de Mello, G. E. (2005). The concept of hope and its relevance to product
evaluation and choice. Journal of2arketing, 69, 1-14.

Mackay, C. J. (1980). The measurement of mood and psychophysiological activity using self-
report techniques. In I. Mating & P. Venebles (Eds.), Techniques in psychophysiology
(pp. 501-562). New York: Wiley.

Mandler, G. (1962). Emotion. In R. Brown, E. Galanter, E. Hess & G. Mandler (Eds.), New
directions in psychology (pp. 267-339). New York: Holt, Rinehart & Winston.

Mandler, G. (1984). M~ind and body. New York: Norton.

Mano, H. (1997). Affect and persuasion: The influence of pleasantness and arousal on attitude
formation and message elaboration. Psychology and2arketing, 14, 315-335.

Marshall, S. W. (2006). Advertising message strategies and executional devices in television
commercials fr~om award-winning "effective campaigns fr~om 1999 to 2004.Unpublished
manuscript, University of Florida, Gainesville, Florida.

Martinelli, K. A., & Chaffee, S. H. (1995). Measuring new-voter learning via three channels of
political information. Journalism and Ma~ss Communication Quarterly, 72, 18-32.

McGuire, W. J. (1989). Why are attitudes important? In A. R. Pratkanis, S. J. Breckler & A. G.
Greenwald (Eds.), Attitude structure and function (pp. 37-69). Hillsdale, New Jersey:
Lawrence Erlbaum Associates.

Mehrabian, A. (1977). A questionnaire measure of individual differences in stimulus screening
and associated differences in arousability. Environmental Psychology andNonverbal
Behavior, 1, 89-103.










Mehrabian, A. (1995). Framework for a comprehensive description and measurement of
emotional states. Genetic, Social, and General Psychology M~onographs, 121, 339-361.

Mehrabian, A., & Russell, J. A. (1974). An approach to environmentalpsychology b.
Cambridge, MA: Oelgesschlager, Gunn. and Hain.

Menon, M. K., Goodnight, J. M., & Wayne, R. J. (2006). Assessing advertising content in a
hospital advertising campaign: An application of Puto and Wells (1984) measure of
informational and transformational advertising content. Journal ofHospital\~'arketing &
Public Relations, 17, 21-44.

Merritt, S. (1984). Negative political advertising: Some empirical findings. Journal of
Advertising, 13, 27-38.

Mewborn, R. C., & Rogers, R. W. (1979). Effects of threatening and reassuring components of
fear appeals and the physiological and verbal measures of emotions and attitudes. Journal
ofExperimental and Social Psychology, 15, 242-253.

Mitchell, R. L. C., Elliott, R., Barry, M., Cruttenden, A., & Woodruff, P. W. R. (2003). The
neural response to emotional prosody, as revealed by functional magnetic resonance
imaging. Neuropsychologia, 41, 1410-1421.

Mittal, B. (1999). The advertising of service: Meeting the challenge of intangibility. Journal of
Service Research, 2, 98-116.

Morris, J. D. (1995). Observations: SAM: The Self-Assessment Manikin: An efficient cross-
cultural measurement of emotional response. Journal ofAdvertising Research, 35, 63-68.

Morris, J. D., Bradley, M. M., Sutherland, J., & Wei, L. (1993). Assessing cross cultural
ttrttrtransfet~t~t~t~rablty~ of standard~~ddizedd~~~~ddddd global advertising: An emotional response approach.
Paper presented at the Annual Conference of the Association of Education in Journalism
and Mass Communications, Kansas City.

Morris, J. D., Bradley, M. M., Waine, C. A., & Lang, J. B. (1992). Assessing affective reactions
to advertisements nI ithr the self-assessment manikin (SAM). Paper presented at the
Southern Marketing Association Conference.

Morris, J. D., Klahr, N. J., Shen, F., Villegas, J., Wright, P., He, A. G., et al. (2008). fMLRI for
emotional responses to TV commercials: A dimensional approach for locating emotion
processors in human brain. Human Brain Mapping.

Morris, J. D., Roberts, M. S., & Baker, G. F. (1998). The effects of music on emotional response,
brand attitude, and purchase intent in an emotional advertising condition. In J. W. Alba &
J. W. Hutchinson (Eds.), Advances in consumer research (Vol. 25, pp. 518-526). Provo,
UT: Association for Consumer Research.










Morris, J. D., Strausbaugh, K. L., & Nthangeni, M. (1996). Emotional response to
advertisements (or commercials) across culture. In G. B. Wilcox (Ed.), Proceedings of
the 1996 Conference of the American Academy of Advertising (pp. 47-54). Austin, TX:
American Academy of Advertising.

Morris, J. D., & Waine, C. A. (1993). Managing the creative effort: Pre- production and post-
production measures of emotional responses. In E. Thorson (Ed.), Proceedings of the
1993 Conference of the American Academy ofAdvertising (pp. 1 58-176). Columbia, MO.

Morris, J. D., Waine, C. A., & Lang, J. B. (1993). Assessing emotional responses to
advertisements with (SAM) the self-assessment manikin. In Proceedings of the Allied
Senrlhli n Business Association 1993 Annual Meeting: Allied Southern Business
Association.

Morris, J. D., Woo, C., Geason, J. A., & Kim, J. (2002). The power of affect: Predicting
intention. Journal ofAdvertising Resea~rch, 42, 7-17.

Morris, J. D., Woo, C., & Singh, A. J. (2005). Elaboration likelihood model: A missing intrinsic
emotional implication. Journal of Targeting, M~easurement & Analysis for Marketing, 14,
79-98.

Natchez, P. B., & Bupp, I. C. (1968). Candidates, issues, and voters. Public Policy, 1 7, 409-437.

Newhagen, J. E., & Reeves, B. (1991). Emotion and memory responses for negative political
advertising: A study of television commercials used in the 1988 presidential election. In
F. Biocca (Ed.), Television and political advertising: Psychological processes (Vol. 1,
pp. 197-220). Hillsdale, New Jersey: Lawrence Erlbaum Associates.

Newman, B. I. (2001). An assessment of the 2000 US presidential election: A set of political
marketing guidelines. Journal of Public Affairs, 1, 210-216.

Niffenegger, P. B. (1989). Strategies for success from the political marketers. Journal of
Consumer Marketing, 6, 45-51.

Nimmo, D. (1999). The permanent campaign: Marketing as a governing tool. In B. I. Newman
(Ed.), Handbook ofpolitical marketing (pp. 73-86). Thousand Oaks, CA: Sage.

O'Cass, A. (1996). Political marketing and the marketing concept. European Journal of
Marketing, 30, 45-61.

Oishi, S. (2006). The concept of life satisfaction across cultures: An IRT analysis. Journal of
Research in Personality, 40, 347-462.

O'Keefe, M. T., & Sheinkopf, K. G. (1974). The voter decides: Candidate image or campaign
issue. Journal ofBi Iackatl(fing. 18, 403-412.

Omura, G. S. (1979). Role of attribute generality in cognitions of political candidates. Advances
in Consumer Research, 6, 635-640.










Ortony, A., & Turner, T. J. (1990). What's basic about basic emotions? Psychological Review,
97, 315-331.

Osgood, C. E., Suci, G. J., & Tannenbaum, P. H. (1957). The nzeasurentent of meaning. Urbana,
IL: University of Illinois Press.

Panksepp, J. (1989). The neurobiology of emotions. In T. Manstead & H. Wagner (Eds.),
Handbook of social psychophysiology (pp. 5-26). New York: Wiley.

Petty, R. E., & Cacioppo, J. T. (1986). The elaboration likelihood model of persuasion. In L.
Berkowitz (Ed.), Advances in experimental social psychology (Vol. 19, pp. 123-205).
New York: Academic Press.

Petty, R. E., Cacioppo, J. T., & Goldman, R. (1981). Personal involvement as a determinant of
argument-based persuasion. Journal ofPersonality and Social Psychology, 41, 847-855.

Plutchik, R. (1980). A general psychoevolutionary theory of emotion. In R. Plutchik & H.
Kellerman (Eds.), Enrotion:Theory, research, and experience (Vol. 1, pp. 3-33). New
York: Academic.

Plutchik, R. (1982). A psycho evolutionary theory of emotions. Social Science Information, 21,
529-553.

Plutchik, R. (1984). Emotions: A general psychoevolutionary theory b. In K. R. Scherer & P.
Ekman (Eds.), Approaches to emotion (pp. 197-219). Hillsdale, NJ: Lawrence Erlbaum
Associates.

Porter, M. E. (1985). Competitive advantage. New York: The Free Press.

Puto, C. P., & Wells, W. D. (1984). Informational and transformational advertising: The
differential effects of time. Advances in Consumer Research, 11, 63 8-643.

Quinn, P., & Kivij ary, L. (2005). US political media buying 2004. International Journal of
Advertising, 24, 131-140.

Resnik, A., & Stern, B. L. (1977). An analysis of information content in television advertising.
Journal of2arketing, 41, 50-53.

Richins, M. L. (1997). Measuring emotions in consumption experience. Journal of Consumer
Research, 24, 127-146.

Ridout, T. N., Shah, D. V., Goldstein, K. M., & Franz, M. M. (2004). Evaluating measures of
campaign advertising exposure on political learning. Political Behavior, 26, 201-225.

Roddy, B. L., & Garramone, G. M. (1988). Appeals and strategies of negative political
advertising. Journal ofBroadcasting & Electronic M~edia, 32, 415-427.










Rogers, E. M. (1997). A history of conanunication study: A biographical approach. New York:
Free Press.

Roseman, I. J. (1984). Cognitive determinants of emotion : A structural theory. Review of
Personality and Social Psychology, 5, 1 1-36.

Roseman, I. J., Antoniou, A. A., & Jose, P. E. (1996). Appraisal determinants of emotions:
Constructing a more accurate and comprehensive theory. Cognition & Emotion, 10, 241-
277.

Russell, J. A. (1978). Evidence of convergent validity on the dimension of affect. Journal of
Personality and Social Psychology, 36, 1152-1168.

Russell, J. A. (1980). A circumplex model of affect. Journal of Personality and Social
Psychology, 42, 1161-1178.

Russell, J. A., & Mehrabian, A. (1977). Evidence for a three-factor theory of emotions. Journal
ofResearch in Personality, 11, 273-294.

Samejima, F. (1969). Estimation of latent ability using a response pattern of graded scores.
Psychonsetrika M~onoguraph, 17, 1-100.

Scherer, K. R., & Wallbott, H. G. (1994). Evidence for universality and cultural variation of
differential emotion response patterning. Journal ofPersonality and Social Psychology,
66, 310-328.

Schiffman, L. G., Sherman, E., & Kirpalani, N. (2002). Trusing souls: A segmentation of the
voting public. Psychology & Marketing, 19, 993-1007.

Schleuder, J. (1990). Effects of commercial complexity, the candidate, and issue vs. image
strategies in political ads. Advances in Consumer Research, 17, 159-168.

Schlinger, M. J. (1979). A profile of responses to commercials. Journal of Advertising Research,
19, 37-46.

Schlosberg, H. (1952). The description of facial expression in terms of two dimensions. Journal
ofExperintental Psychology, 44, 229-237.

Shama, A. (1975). Applications of marketing concepts to candidate marketing. Advances in
Consumer Research, 2, 793-801.

Shama, A. (1976). The marketing of political candidates. Journal of the Academy of2arketing
Science, 4, 764-777.

Shapiro, M. A., & Rieger, R. H. (1992). Comparing positive and negative political advertising on
radio. Journalism Quarterly, 69, 135-145.










Shaver, P., Schwartz, J., Kirson, D., & O' Connor, C. (1987). Emotion knowledge: Further
exploration of a prototype approach. Journal ofPersonality and Social Psychology, 52,
1061-1086.

Shen, F. (2007). Measuring issue and image in political advertising: An
informational/transformational approach. Paper presented at the Annual Convention of
Association for Education in Journalism and Mass Communication, Washintgton, D. C.

Shen, F., & Kim, K. (2006). Videostyle in Presidential Camnpaigns 2000 & 2004: An
informational versus transformational approach. Paper presented at the Association for
Education in Journalism and Mass Communication, San Francisco, California.

Simon, H. A. (1982). Comments. In M. S. Clarke & S. T. Fiske (Eds.), Affect and cognition: The
seventeenth annual Calrnegie symposium on cognition (pp. 333-342). Hillsdale, NJ:
Erlbaum.

Singh, J. (2004). Tackling measurement problems with Item Response Theory: Principles,
characteristics, and assessment, with an illustrative example. Journal of Business
Research, 57, 184-208.

Singh, J., Howell, R. D., & Rhoads, G. K. (1990). Adaptive designs for Likert-type data: An
approach for implementing marketing surveys. Journal of2arketing Research, 27, 304-
321.

Sonner, B. S. (1998). The effectiveness of negative political advertising: A case study. Journal of
Advertising Resea~rch, 38, 37-42.

Swaminathan, V., Zinkhan, G. M., & Reddy, S. K. (1996). The evolution and antecedents of
transformational advertising: A conceptual model. Advances in Consumer Research, 23,
49-55.

Teasdale, J. D. (1993). Emotion and two kinds of meaning: Cognitive therapy and applied
cognitive science. Behavior Research Theory, 31, 339-354.

Teasdale, J. D., & Barnard, P. J. (1993). Affect, cognition, and change: Remodelling depressive
thought. Hove, England: Lawrence Erlbaum Associates.

Teasdale, J. D., Howard, R. J., Cox, S. G., Ha, Y., Brammer, M. J., Williams, S. C. R., et al.
(1999). Functional MRI study of the cognitive generation of affect. American Journal of
Psychology, 156, 209-215.

Thorson, E., Christ, W. G., & Caywood, C. (1991). Effects of issue-image strategies, attack and
support appeals, music, and visual content in political commercials. Journal of
Broadcasting & Electronic M~edia, 35, 465-486.

Thorson, E., Reeves, B., & Schleuder, J. (1985). Message complexity and attention to television.
Communication Research, 12, 427-454.










Thorson, E., Reeves, B., & Schleuder, J. (1987). Attention to local and global complexity in
television messages. In M. McLaughlin (Ed.), Conanunication Yearbook 10. Beverly
Hills, CA: Sage.

Tomkins, S. S. (1980). Affect as amplification: Some modifications in theory. In R. Plutchik &
H. Kellerman (Eds.), Emotion: Theory, research and experience (pp. 141-164). New
York: Academic Press.

Vakratsas, D., & Ambler, T. (1999). How advertising works: What do we really know? Journal
of2arketing, 63, 26-43.

Vavreck, L. (2001). The reasoning voter meets the strategic candidate: Signals and specificity in
campaign advertising, 1998. American Politics Research, 29, 507-529.

Watson, D., & Tellegen, A. (1985). Toward a consensual structure of mood. Psychological
Bulletin, 98, 219-235.

Watson, D., Wiese, D., Vaidya, J., & Tellegen, A. (1999). The two general activation systems of
affect: Structural findings, evolutinory considerations, and psychological evidence.
Journal of Personality and Social Psychology, 76, 820-838.

Weaver-Lariscy, R. A., & Tinkham, S. F. (1996). Advertising message strategies in U.S.
congressional campaigns: 1982, 1990. Journal of Current Issues and Research in
Advertising, 18, 53-66.

Weiner, B. (1985). An attributional theory of achievement motivation and emotion.
Psychological Review, 92, 548-573.

Wells, W. D. (1988). Lectures and dramas. In P. Cafferata & A. Tybout (Eds.), Cognitive and
affective responses to advertising (pp. 13-20). Lexington, MA: D. C. Heath.

West, D. M. (1993). Air wars: Television advertising in election campaigns, 1952-1992.
Washington, DC: Congressional Quarterly.

West, D. M. (1994). Political advertising and news coverage in the 1992 California U.S. Senate
campaigns. The Journal ofPolitics, 56, 1053-1075.

Witte, K. (1995). Generating effective risk messages: How scary should your risk
communication be? In B. R. Burleson (Ed.), Conanunication Yearbook (Vol. 18, pp. 229-
254). Thousand Oaks, CA: Sage.

Wring, D. (1997). Reconciling marketing with political science: Theories of political marketing.
Journal of2arketing Management, 13, 65 1-663 .

Yoon, K., Pinkleton, B., & Ko, W. (2005). Effects of negative political advertising on voting
intention: An exploration of the roles of involvement and source credibility in the
development of voter cynicism. Journal of2arketing Conanunications, 11, 95-1 12.










Zajonc, R. B. (1984a). The interaction of affect and cognition. In K. R. Scherer & P. Ekman
(Eds.), Approaches to emotion (pp. 239-246). Hillsdale, NJ: Lawrence Erlbaum.

Zajonc, R. B. (1984b). On the primacy of affect. American Psychologist, 39, 1 17-123.

Zeithaml, V. A., & Bitner, M. J. (1996). Service marketing. New York: McGraw-Hill.









BIOGRAPHICAL SKETCH

Feng Shen was born in China. He received his Bachelor of Arts degree in english at

Shanghai International Studies University (Shanghai, China) in 1995. After five years of

industry experience with public health and marketing campaigns in China, he came to the United

States and completed a Master of Science degree in communication at Cornell University, Ithaca,

New York in 2002. After spending one year in China managing international investment, he

joined the Ph.D. program at the University of Florida, Gainesville, Florida in 2005. He has been

collaborating primarily with Dr. Jon Morris on several research projects. In 2006 and 2007, he

taught Advertising Research at the University of Florida. He has now accepted a position as an

Assistant Professor of Marketing at Erivan K. Haub School of Business, Saint Joseph' s

University, Philadelphia, Pennsylvania.





PAGE 1

CHAIN REACTION MODEL OF POLITICAL ADVERTISING EFFECTS: ADVERTISING CONTENT, MEMORY, AND CANDIDATE EVALUATION By FENG SHEN 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 2008 1

PAGE 2

2008 Feng Shen 2

PAGE 3

To my family 3

PAGE 4

ACKNOWLEDGMENTS I am deeply indebted to many people in my completion of this dissertation and all the thanks I express herein are meager substitutes for the gratitude and appreciation I owe. Heartfelt thanks go to Dr. Jon Morris (my supervisory committee chair) for his invaluable guidance, mentorship, and the unfathomable emotional support he provided. For her untiring guidance on research design, I also wish to thank Dr. Lynda Kaid. In addition, this dissertation has benefited immensely from the constructive comments I received from Dr. Jorge Villegas and Dr. David Miller. Without their efforts this would not have been nearly as rewarding an accomplishment. I would also like to take this opportunity to thank my friends and faculty at the University of Florida, especially Dr. John Sutherland and Dr. Debbie Treise for their mentorship. I also truly thank Dr. Michael Weigold, Dr. Justin Brown, Seth Oyer, Annie Sugar, Sunny Hughes, Kenneth Kim, Tom Mueller, and Michael Clayton for their sincere support. The love of my parents enabled me to start and complete the otherwise impossible dream of attaining this degree. Without the tireless encouragement of my father and mother, I could not have completed this study. I deeply admire them. 4

PAGE 5

TABLE OF CONTENTS page ACKNOWLEDGMENTS...............................................................................................................4 LIST OF TABLES...........................................................................................................................8 LIST OF FIGURES.........................................................................................................................9 ABSTRACT...................................................................................................................................10 CHAPTER 1 INTRODUCTION..................................................................................................................12 Need for the Study..................................................................................................................12 Purpose and Importance of Study...........................................................................................13 Outline....................................................................................................................................15 2 LITERATURE REVIEW.......................................................................................................16 History of Modern Political Marketing..................................................................................16 Verbal Content and Effects of Political Advertising..............................................................20 Issue and Image in Political Advertising.........................................................................21 Positive and Negative Political Advertising....................................................................22 Effects of Political Advertising: Memory.......................................................................22 Effects of Political Advertising: Candidate Electability.................................................32 Summary of the Effects of Political Advertising............................................................39 Informational and Transformational Advertising...................................................................40 Antecedent 1: Newness of the Product............................................................................40 Antecedent 2: Level of Risk............................................................................................41 Antecedent 3: Product Involvement................................................................................41 Antecedent 4: Product Conspicuousness.........................................................................42 Antecedent 5: Product versus Service.............................................................................42 Typology of Informational/Transformational Strategies.................................................43 Summary..........................................................................................................................45 Survey Scale of Informational/Transformational Advertising........................................46 Emotion in Advertising...........................................................................................................48 Rational and Emotional Responses.................................................................................48 Measuring Emotion.........................................................................................................52 Physiological measures............................................................................................52 Self-report measures.................................................................................................54 Discrete approach to emotion...................................................................................54 Dimensional approach to emotion............................................................................55 Measuring emotion with AdSAM..........................................................................59 Summary of Literature Review..............................................................................................60 Hypotheses and Research Questions......................................................................................62 5

PAGE 6

Perception of Advertising Content..................................................................................62 Memory...........................................................................................................................63 Candidate Electability.....................................................................................................64 3 MATERIALS AND METHODS...........................................................................................67 Materials and Measures..........................................................................................................67 Independent Variables.....................................................................................................67 Perception of ad content...........................................................................................67 Prior exposure to ad and familiarity with candidate................................................68 Dependent Variables.......................................................................................................68 Message recall..........................................................................................................68 Emotional response..................................................................................................68 Candidate electability...............................................................................................69 Graded Response Model (GRM) of Item Response Theory (IRT)........................................69 4 RESULTS...............................................................................................................................76 Pretest.....................................................................................................................................76 Main Study: Model Estimation and Results...........................................................................76 Descriptive Statistics of Independent Variables..............................................................77 Prior exposure to ad and familiarity with candidate................................................77 Perception of ad content...........................................................................................77 Descriptive Statistics of Dependent Variables................................................................79 Message recall..........................................................................................................79 Candidate electability...............................................................................................80 Pleasure-arousal-dominance.....................................................................................83 Model Estimation............................................................................................................85 Positive image condition..........................................................................................86 Positive issue condition............................................................................................86 Negative image condition for the Candidate............................................................86 Negative image condition for the Opponent............................................................87 Negative issue condition for the Candidate..............................................................87 Negative issue condition for the Opponent..............................................................87 5 DISCUSSION.......................................................................................................................100 Overview...............................................................................................................................100 Perception of Advertising Content.......................................................................................100 Memory.................................................................................................................................101 Candidate Electability...........................................................................................................102 Issue, Characteristics, and Voting Intention..................................................................102 Cognition and Emotion..................................................................................................102 Memory, Emotion, and Candidate Electability....................................................................104 Limitations and Future Research..........................................................................................106 6 IMPLICATIONS..................................................................................................................108 6

PAGE 7

APPENDIX A QUESTIONNAIRES............................................................................................................110 Pretest...................................................................................................................................110 Main Survey Sample (Negative Advertisement)..................................................................111 B CORRELATION MATRICES.............................................................................................118 LIST OF REFERENCES.............................................................................................................120 BIOGRAPHICAL SKETCH.......................................................................................................134 7

PAGE 8

LIST OF TABLES Table page 3-1 Experimental conditions....................................................................................................75 4-1 Personal qualities and political issues................................................................................96 4-2 Familiarity with candidates................................................................................................96 4-3 Information and transformation.........................................................................................96 4-4 Reliability coefficients of informational/transformational scale.......................................96 4-5 Message recall....................................................................................................................97 4-6 Candidate issue competence..............................................................................................97 4-7 Candidate characteristics...................................................................................................97 4-8 Reliability coefficients of candidate characteristics scale.................................................97 4-9 Voting intention.................................................................................................................98 4-10 Pleasure..............................................................................................................................98 4-11 Arousal...............................................................................................................................98 4-12 Dominance.........................................................................................................................98 4-13 Fit indices in experimental conditions...............................................................................99 B-1 Positive image condition..................................................................................................118 B-2 Positive issue condition....................................................................................................118 B-3 Negative image condition for the candidate....................................................................118 B-4 Negative image condition for the opponent.....................................................................119 B-5 Negative issue condition for the candidate......................................................................119 B-6 Negative issue condition for the opponent.......................................................................119 8

PAGE 9

LIST OF FIGURES Figure page 2-1 Russells affect circumplex................................................................................................66 4-1 Full path diagram of PIFC, PTFC, MR, ER, and CE........................................................89 4-2 Measurement model for CE and ER..................................................................................89 4-3 Factor loadings in positive image condition......................................................................90 4-4 Path coefficients in positive image condition....................................................................90 4-5 Factor loadings in positive issue condition........................................................................90 4-6 Path coefficients in positive issue condition......................................................................91 4-7 Factor loadings in negative image condition for the candidate.........................................91 4-8 Path coefficients in negative image condition for the candidate.......................................92 4-9 Factor loadings in negative image condition for the opponent..........................................92 4-10 Path coefficients in negative image condition for the opponent........................................93 4-11 Factor loadings in negative issue condition for the candidate...........................................93 4-12 Path coefficients in negative issue condition for the candidate.........................................94 4-13 Factor loadings in negative issue condition for the opponent............................................94 4-14 Path coefficients in negative issue condition for the opponent..........................................95 9

PAGE 10

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 CHAIN REACTION MODEL OF POLITICAL ADVERTISING EFFECTS: ADVERTISING CONTENT, MEMORY, AND CANDIDATE EVALUATION By Feng Shen August 2008 Chair: Jon Morris Major: Mass Communication Despite conceptual similarity between informational/transformational product advertising and issue/image political advertising, little has been done to connect those two lines of the research. This research attempts to incorporate informational and transformational advertising into the effects of political advertising. To further strengthen the theoretical bond between product advertising and political advertising, a chain reaction model of advertising effects for products is used to identify constructs in testing the effects of political advertising. The psychological process of advertising effects in Positive or Negative and Issue or Image conditions was studied with four latent constructs (i.e., perception of informational cues, perception of transformational cues, emotional response, and candidate electability), and one observed construct (i.e., message recall). With an observation sample size of 1000, and 250 in each experimental condition, the study tested the proposed chain reaction model to account for the dynamics from message learning to candidate evaluation. The results validate the use of the informational/transformational scale to measure Issue and Image. The results also confirm the conceptual similarity among issue-information-cognition and image-transformation-emotion. The results also indicate that emotional response is a more powerful predictor of candidate electability than cognitive 10

PAGE 11

response. In other words, it is what the respondents feel about the candidate rather than what they remember about the candidate shapes their attitude toward the candidate. Meanwhile, the results suggest that testing the comparative advantages of Positive vs. Negative or Issue vs. Image is likely to be biased by the measurement focus of an attitude scale. Perception of Informational Cues and Transformational Cues impacts message recall, emotional response, and candidate electability in a predictable fashion for the positive issue and image ads and the negative image ad. Nevertheless, the effects in the negative issue ad condition suggest a different model. 11

PAGE 12

CHAPTER 1 INTRODUCTION Need for the Study The idea that political candidates can be advertised and marketed in a way similar to products is prevalent in political campaign literature (Baines, Worcester, Jarrett, & Mortimore, 2005; Ben-Ur & Newman, 2002; Bulter & Collins, 1994; Henneberg, 2004; Kotler, 1975; Lock & Harris, 1996; Newman, 2001; Niffenegger, 1989; O'Cass, 1996; Omura, 1979; Schiffman, Sherman, & Kirpalani, 2002; Shama, 1975, 1976; Wring, 1997). All key concepts of consumer marketing had their counterparts in political marketing (Shama, 1976); for instance, a consumer marketing transaction through which sellers exchanged products or services with buyers for sales income was analogous to a political marketing process in which political candidates offered their service and ideas about policy issues to voters for their support (Shama, 1975). The most visible similarity between those two types of marketing was probably the extensive use of advertising by sellers and candidates to inform and persuade target consumers and voters respectively (Shama, 1976). In fact, the ever-increasing importance of political advertising had substantially contributed to the interest in applying consumer marketing theories and practices to political campaigns (Dermody & Scullion, 2001; Kotler, 1975). The significance of political advertising, particularly television political advertising, was also evident in the magnitude of its expenditure (Devlin, 2005). The last five U.S. presidential campaigns between 1988 and 2004 witnessed a dramatic increase of advertising spending. A total of $79 million was spend on television advertising in the 1988 election (Devlin, 1989), followed by $133 million in 1992 (Devlin, 1993), $173 million in 1996 (Devlin, 1997), $240 million in 2000 (Devlin, 2001), and $620 million in 2004 (Devlin, 2005). The enormous size of spending was not limited to presidential campaigns (Kaid, 2004). In 2004, U.S. political 12

PAGE 13

advertising spending for races at all level accounted for 67.5% of all political media spending and reached $1.85 billion, which was equivalent to a 58.9% increase from 2002 and 112.1% increase from 2000 (Quinn & Kivijarv, 2005). Given the amount of investment in political advertising, it was highly desirable to better understand the impact of advertising (Kahn & Geer, 1994). One line of research examined the verbal content in political advertising and the effect of the verbal content on voting behavior (Kaid, 2004). The verbal content often referred to emphasis, i.e., whether an ad emphasized policy issues or candidate personality, and valence, i.e., positive if the ad supported a candidate or negative if it attacked an opponent (Kaid & Johnston, 2001). The effect often referred to recall of advertising message, candidate evaluation, and intention to vote (Kaid, 2004). Purpose and Importance of Study The birth of modern political marketing was indispensable of the use of advertising through mass media (Shama, 1976). The effects of political marketing on voting behavior could be best explained with a chain reaction model of advertising effectiveness that incorporated not only memory of candidate characteristics and issue stands but also candidate evaluation based on voters memory (Lavidge & Steiner, 1961; Shama, 1976). As noted above, a substantial body of research investigated the impact of advertising verbal content on memory and candidate evaluation when image and issue were used to support or attack a candidate (Kaid, 2004). Nevertheless, previous research did not generate consistent findings about the impacts of advertising on memory and candidate evaluation. Furthermore, memory and candidate evaluation were often examined as two separate effects, although ad content could influence memory, and memory in turn could influence candidate evaluation in a chain reaction manner (Lavidge & Steiner, 1961; Shama, 1976). Therefore, there was a missing link in previous studies on political advertising. 13

PAGE 14

This research attempts to clarify the effects of issue/image and positive/negative ads on memory and candidate evaluation and explore the relationship between those effects with enhanced measurement of issue and image content and responses to ads. The enhanced measurement of issue and image content is achieved through the introduction of Puto and Wells (1984) informational/transformational survey scale (hereinafter referred to as the Scale). The Scale was constructed on the premise that each ad contained both informational and transformational cues, and an ad was regarded as informational if it was perceived to contain more informational cues than transformational cues or as transformational if it was perceived to contain more transformational cues than informational cues (Puto & Wells, 1984). Likewise, an issue political ad contained both issue and image content, but it was the issue information that dominated the content; conversely, an image political ad contained both image and issue content, but it was the image information that dominated the content (Kaid, 2004). The issue content was found conceptually similar to the informational cues and the image content close to the transformational cues (Shen, 2007; Shen & Kim, 2006). Furthermore, issue ads were found to contain more informational cues than transformational cues and image ads were found to contain more transformational cues than informational cues (Shen, 2007). In the current study, the Scale is again used to calibrate the perception of issue and image content in an ad. The perception in turn is used to assess the impact of content on memory and candidate evaluation. Emotional response was an integral component of candidate evaluation and was measured with bipolar adjective pairs such as happy versus unhappy (Thorson, Christ, & Caywood, 1991). Nevertheless, such an approach was problematic because the use of adjectives led to a distorted reporting process that was subject to an individuals linguistic habits, vocabulary, memory, and social desirability (Lazarus, 1966), and language was too insensitive a medium for 14

PAGE 15

communicating private experiences (Mandler, 1962). AdSAM(Morris, Woo, Geason, & Kim, 2002), a nonverbal measure of emotional response, is used in this study to assess the emotional component of candidate evaluation with enhanced precision. Moreover, structural equation modeling (SEM) is used to assess the chain reaction starting from the advertising content to candidate evaluation. Additionally, item response theory (IRT) is used to estimate latent responses to the Scale in the current study because IRT was proved an extraordinary tool to capture the cognitive process of responding to survey scales (Embretson & Reise, 2000; Hambleton & Swaminathan, 1985; Linden & Hambleton, 1996; Singh, 2004; Singh, Howell, & Rhoads, 1990). The enhanced measurement of the content and responses eventually helps answer how issue/image and positive/negative ads influence memory and candidate evaluation and identify a manner in which the verbal content can be managed that will lead to optimal mental processing of political ads and favorable candidate evaluation, which is of vital importance to both practitioners and scholars. Outline This dissertation is structured as follows. Chapter 2 reviews the relevant literature regarding the development of political marketing and advertising and the effects of political advertising. The review includes theoretical models of political marketing, the effects of issue and image political advertising, the effects of support and attack political advertising, informational and transformational advertising, and emotional response to advertising. The literature review leads to the formulation of hypotheses and research questions. Chapter 3 details analytic approaches and methodology. Chapter 4 reviews the results of the study. Chapter 5 includes the discussion of the findings and suggestions for future research. Finally, Chapter 6 addresses the contribution of this research. 15

PAGE 16

CHAPTER 2 LITERATURE REVIEW History of Modern Political Marketing Although advertising was just the promotion component of marketing, it was often the key to the success of a marketing campaign (Hoyer & Maclnnis, 2004). The importance of mass media and political advertising was also evident in reviews of the development of modern political marketing (Blumenthal, 1980; Kotler, 1975; Maarek, 1995; Nimmo, 1999; Shama, 1976). It was generally agreed that modern political marketing was a derivative of the post-World War II era andevolved in stages (Nimmo, 1999, p. 75). Nevertheless, a major issue in conceptualizing political marketing, as pointed out by Nimmo (1999), was that researchers often treated marketing not as a commercial activity but rather as a form of rhetoric (p. 76). However, Shamas (1976) three-stage framework of political marketing might be an exception because his work was deeply rooted in Kings three phases of American business (King, 1965) and was supported by a number of scholars (Blumenthal, 1980; Maarek, 1995; Nimmo, 1999) vision of modern political marketing. These three stages are (1) candidate-orientation, (2) sales-management orientation, and (3) marketing orientation these stages parallel the development in the orientation of American business as depicted by King from (1) production orientation whose main concern was to produce large volumes of products, to (2) sales-management orientation whose main concern was to maximize sales through better coordination of production and mass promotion, and finally to (3) a marketing concept orientation whose main concern was to offer goods and services that satisfy consumer needs and wants. (Shama, 1976, p. 766) Specifically, the first stage covered the period of 1940 1960 The candidate was viewed as a product needing mass exposure, resulting in increased public awareness which was assumed to be positively connected to voter preference number of exposures and length of exposures were taken as the key for victory in the campaign (Shama, 1976, p. 766). Similarly, Maarek (1995) noted that during this time period communication consultants made massive use of short 16

PAGE 17

advertising spots, strategically broadcast as close as possible to the most popular television shows. In this waythe candidate could benefit, at no extra expense, from the efforts made by the television networks to increase their audience (p. 13). Many researchers (e.g., Blumenthal, 1980; Maarek, 1995; Nimmo, 1999; Shama, 1976) agreed that the landmark of this stage was the 1952 Dwight Eisenhower campaign. It was the first to hire an advertising firm in a national contest (Nimmo, 1999, p. 75). The candidate/product orientation substantially influenced the development of modern political advertising; for instance, the preference of frequent but short mass exposures led to growing use of short political ads rather than long political speeches (Shama, 1976, p. 767) and the role of the mass media specialists grew from a minor to a major place in the campaign In most cases those specialists were advertising executives or advertising agencies (p. 767). Essentially, the production/candidate orientation shared the same powerful media effects assumption with early communication research (Rogers, 1997). Political marketing of the 1950s was nothing but promotion and distribution of candidates through increased exposure by mass media (Shama, 1976, p. 767). In the second stage of political marketing, [b]eginning in the early 1960s, political campaigning was viewed and practiced as the selling of candidates (Shama, 1976, p. 767). Ad agencies sold candidates like soap (Blumenthal, 1980, p. 2). The landmark of this sales-management orientation stage was the 1960 John Kennedy campaign (Maarek, 1995; Shama, 1976) because this campaign incorporated all the main elements of sales management, namely sophisticated voter study market segmentation, promotion appeals and media selection (Shama, 1976, p. 770). Voter study was also called field analysis or geodemographics (Maarek, 1995) and included systematic benchmark surveys of the campaign field, which involve an in17

PAGE 18

depth study of voters and inhabitants of the geographical area in which the campaign is to be carried out (Maarek, 1995, p. 33). Although the surveys could be used for remodeling existing candidates to appear more congruent with the voters desires (Shama, 1976, p. 768), it was impossible for a candidate to satisfy the needs and wants of all voters. As a result, another marketing activity, voter segmentation, was developed to locate those segments of the general population which might be subject to the influence of political communication (Maarek, 1995, p. 34) and the candidate could therefore determine to which groups he would make his maximum adaptation and appeal (Shama, 1976, p. 768). Voter studies also indicated that voter characteristics such as involvement (Shama, 1976) and selective media exposure (Maarek, 1995) influenced the effectiveness of political communication. Accordingly, promotional strategies in this stage focused on the effectiveness of different promotion appeals and media in reaching the voters (Shama, 1976, p. 769). Many changes in this stage still had substantial influence on todays political advertising research; for example, survey was still be one of the most important research methods (Graber, 2004; Maarek, 1995) and the search for effective appeals and media was manifest in all research initiatives nowadays (Nimmo, 1999). The current study is one of those media effects studies that aim to identify circumstances under which advertising will have the largest impact on voting behavior. Shama (1976) also commented on the third stage of political marketing and argued [p]resently, political marketing is experiencing a transition from sales management orientation to marketing orientation the marketing of political candidates is becoming more voter oriented it is the voter who decides which candidate that is, which parcel of personality, issues, party, etc. to buy with his vote (p. 770). There was certain ambiguity about the exact 18

PAGE 19

start of the third stage because of the word presently in Shamas (1976) work. Other researchers held that the third stage began with the 1980 Ronald Reagan presidential campaign (e.g., Maarek, 1995; Nimmo, 1999) and this campaign delivered to voters a well-defined parcel about Ronald Reagan: The candidate was portrayed as having a very clear and simple conservative profile, with a very sparse political platform: reduce taxes and federal government intervention, except for Defense, whose budget was increased (Maarek, 1995, p. 20). Another significant change in the third stage was the emergence of [i]ntegrated political marketingit involves the simultaneous study, analysis, and segmentation of voters; candidate positioning, and candidate development; promotion and use of media (Shama, 1976, p. 772). Combining a series of communications through different channels is much more effective than the single predominant use of the leading audio-visual networks (Maarek, 1995, p. 62). Interestingly enough, Shama (1976) used a chain reaction model of advertising effectiveness (Lavidge & Steiner, 1961) to define the intended effects of integrated political marketing. The multiple output may be viewed in terms of Lavidges and Steiners (1961) advertising model calling for increased (1) awareness of the candidate existence in the political market (2) knowledge of the candidates personal and political background, the office for which he is running, his personality traits and the issues he stands for (3) liking the candidate on the basis of available information about him and possibly on the basis of additionally sought information (4) preference of the candidate over other candidates running for the same office (5) a conviction that the candidate is the best available one (6) support of the candidate by contributions and/or volunteer work and voting. (Shama, 1976, p. 772) The interest in those above-mentioned effects was well recorded in political advertising research. There were a number of studies on message learning/knowledge/memory (e.g., Geer & Geer, 2003; Martinelli & Chaffee, 1995; Ridout, Shah, Goldstein, & Franz, 2004; West, 1994), candidate evaluation (e.g., Chang, 2001; Kaid, 1991; Kaid, 2001; Kaid & Tecesco, 1999), and voting behavior (e.g., Ben-Ur & Newman, 2002; Faber, Tims, & Schmitt, 1993; Kaid & Sanders, 19

PAGE 20

1978; Weaver-Lariscy & Tinkham, 1996). Nevertheless, there were few studies that investigated the relationships between those effects in a chain-reaction manner. The current study attempts to address the relationships in question, particularly those between content, memory, and candidate evaluation with structural equation modeling. In summary, political marketing was intertwined with political advertising. Heavy reliance on mass media and advertising practitioners marked the birth of modern political marketing. Media planning, audience research, advertising effectiveness studies, and integration of all those efforts emerged in the development of political marketing and are now the backbone of political campaigns. The literature review above also indicated that 1) content of political ads such as candidate characteristics and issue stands was the key information for voters decision-making and 2) the perception of the content and candidate evaluation based on the perception were the key predictors of voting behavior. The current study focuses on both the content and its impacts on voting behavior. The section below provides an overview of previous research on those two aspects as recorded in the literature of political advertising. Verbal Content and Effects of Political Advertising The verbal content of political advertising was often analyzed in terms of issue versus image and/or positive versus negative (Joslyn, 1980; Kaid & Johnston, 2001). An issue ad focused on the specific policy issues or positions mentioned in an advertisement whereas an image ad focused on a candidates personal qualities, background, or characteristics (Kaid & Johnston, 2001). A positive ad supported a candidate whereas a negative ad attacked an opponent (Kaid & Johnston, 2001). In real-life practice, the use of issue and image was also associated with support or attack; for example, studies on presidential campaign ads (Kaid, 2002, 2005; Kaid & Dimitrova, 2005; Kaid & Johnston, 2001) indicated that both image and issue ads 20

PAGE 21

were more frequently positive than negative; but in recent years there was a clear trend of more use of negative attack against the opponent, particularly in issue ads. Issue and Image in Political Advertising The interest in issue and image in political advertising can be traced back to voting behavior studies since the 1950s. Researchers at the University of Michigan identified party loyalty and group identification as the long-term factors and feelings about candidates and issues as the short-term factors influencing voting behavior in both presidential and congressional elections (Campbell, Converse, Miller, & Stokes, 1960; Campbell, Gurin, & Miller, 1954). Nevertheless, for decades political advertising was criticized for emphasizing image and trivializing issues, and therefore damaging the very foundation of democratic voting, i.e., voting behavior should result from rational decision-making based on policy issues rather than candidate personalities or images (Berelson, 1966). However, studies on political advertising content, especially those on television political advertising content, suggested that such an accusation was largely unfounded or mythical, since political advertising content was largely issue-based (Joslyn, 1980; Kaid, 2004; Kaid & Johnston, 2001). The dominance of issues television political ads (Kaid, 2004). For example, from 1952 through 1996, 66% of the presidential campaign ads were issue-oriented whereas 34% of them were image-oriented (Kaid & Johnston, 2001). In 2000, 78% of the presidential campaign ads were issue-oriented and 22% of them were image-oriented (Kaid, 2002). In 2004, 85% of the presidential campaign ads were issue-oriented and 15% of them were image-oriented (Kaid, 2005; Kaid & Dimitrova, 2005). Although the percentage of issue ads varied over time, it was below half (47%) only in 1968 (Kaid & Johnston, 2001). The focus on issue was manifest not only in presidential campaigns but also in gubernatorial and senatorial campaigns (Joslyn, 1980; Vavreck, 2001). 21

PAGE 22

Positive and Negative Political Advertising Negativity was a growing aspect of campaign advertising in the U.S. and the past few election cycles witnessed a real increase in the number of negative spots in presidential campaigns (Kaid, 2004). Although the percentage of negative ads in presidential campaigns from 1952 through 1996 was only about 38% (Kaid & Johnston, 2001), in the last four presidential election cycles in the U.S. (1992, 1996, 2000 and 2004) negative ads made up substantial percentages of the advertising content of both major party candidates. Both of Bill Clintons campaigns (1992 and 1996) established new highs in the number of negative ads in a presidential campaign with 69% of his ads being negative in 1992 and 68% being negative in 1996 (Kaid, 1994, 1998; Kaid, DeRosa, & Tedesco, 2002; Kaid & Johnston, 2001). Al Gores presidential campaign used only slightly fewer negative ads; 62% of his ads were negative, compared to 37% of Bushs ads (Kaid, 2002). But Bush resorted to more negative ads (58%) in 2004, as he defended his presidency against challenger John Kerry (Kaid & Dimitrova, 2005). Negative advertising was also a growing feature of political ads campaign strategy for candidates in election below the presidential level; nevertheless, there was no strong evidence that political advertising was predominantly negative in state and local election campaigns (Kaid, 2004). For instance, Kahn and Kenny (2000) analyzed 594 ads from Senate campaign in 1988, 1990 and 1992 and found that less than half (41%) of all ads contained negative attacks. Vavreck (2001) studied 290 candidate ads in the 1998 elections and found 64% were positive ads. Furthermore, negative ads were usually focused on issues rather than candidate personalities (Johnston & Kaid, 2002; Kaid & Johnston, 2001; West, 1993). Effects of Political Advertising: Memory There was certain ambiguity about the impact of image and issue ads on memory as measured by recall. On the one hand, there was some empirical finding that image ads 22

PAGE 23

outperformed issue ads in generating higher recall (Schleuder, 1990). On the other hand, there was evidence supporting the superiority of issue ads over images ads (Kaid, Chanslor, & Hovind, 1992). The notion that image ads might facilitate recall originated from some previous research about message complexity. First, memory of image and issue ads was influenced by complexity (Natchez & Bupp, 1968; Schleuder, 1990). Second, image ads by nature were less complex than issue ads (Natchez & Bupp, 1968) because image ads contained fewer visual and verbal cues in a commercial per time unit than issue ads (Schleuder, 1990). Third, simple commercials were found to facilitate mental processing and lead to more attention and better memory than complex commercials (Thorson, Reeves, & Schleuder, 1985, 1987); therefore, it should be easier to process image ads than issue ads and to recall content from image ads than from issue ads (Natchez & Bupp, 1968; O'Keefe & Sheinkopf, 1974). Some researchers also argued that an ad, image or issue, could be made simpler or more complex by including fewer or more cues, and complex issue ads might work particularly well for an unknown candidate since voters might be more curious about the new candidates issue stands (Schleuder, 1990). In one empirical study designed to test those ideas, Schlender (1990) showed 22 participants four image ads, two for George Bush, a well-known presumed incumbent in the 1988 presidential campaign, and the other two for Michael Dukakis, a nationally unknown challenger in the same campaign, and another 22 participants four issue ads, two for each candidate respectively. The two image or issue ads for each candidate were pre-coded as high or low complexity. In other words, ad type (issue vs. image) was a between-subjects variable, and complexity (simple vs. complex) and candidate type (well-known Bush vs. unknown Dukakis) were two within-subjects variables. Mental processing was the dependent construct and was 23

PAGE 24

indicated by three variables. The first variable was attention to the commercials and it was measured with the participants secondary task reaction time to periodic tones. Longer reaction time meant more attention to the commercials. The second variable was memory of visual cues and it was measured with a frame recognition test regarding scene change, dissolve, edit, camera pan or zoom, person and object movement, and superimposed image (Schleuder, 1990). Larger numbers of correct recognitions meant better visual memory. The third variable was memory of verbal cues and it was measured with a multiple-choice test about the verbal content including music, voice-over and natural sound, and macro proposition or idea (Schleuder, 1990). Higher scores on the test meant better verbal memory. Schlender (1990) provided some evidence that image ads were associated with better mental processing. It took the participants watching the image ads longer to respond to the secondary task than those watching the issue ads, and therefore, the image ads attracted more attention than the issue ads. Verbal cues were better remembered for image ads than for issue ads. Nevertheless, memory of visual cues was not influenced by ad type. The author also explored some interaction effects, but there was lack of consistency in the effects on the three indicators of mental processing. For instance, complexity had no effect on attention for both types of the Bush ads; but low complexity outperformed high complexity in producing more attention to the Dukakis image ads, and high complexity outperformed low complexity in producing more attention to the Dukakis issue ads. By sharp contrast, complexity had no effect on visual memory for both types of the Dukakis ads; but low complexity outperformed high complexity in producing higher recall of the visual cues of the Bush issue ads, and high complexity outperformed low complexity in producing higher recall of the visual cues of the Bush image ads. Finally, complexity had no effect on verbal memory for a particular type 24

PAGE 25

of the Bush or Dukakis ads; but low complexity outperformed high complexity in producing higher recall of the verbal cues of the Bush ads, and high complexity outperformed low complexity in producing higher recall of the Dukakis ads. Inconsistencies also existed in other comparisons. Candidate type had no effect on attention or verbal memory but some effects on visual memory. The Dukakis ads outperformed the Bush ads in producing higher recall of the visual cues, and the Dukakis issue ads outperformed the Bush issue ads in producing higher recall of the visual cues. In the absence of any explanation about the inconsistencies in the mental processing, those interaction effects could just be experimental artifacts, or effects caused by the ordering of the ads rather than the characteristics of the ads. In fact, the author was aware of ordering effects when placing four ads one after another within one condition and tried to make the study more realistic by embedding the political ads into a situation comedy with consumer ads. Therefore, Schlender (1990) created two more between-subjects variables, each with three levels for both the order of commercial breaks and the position of each political ad in each break. With those two additional between-subjects variables, there were 18 between-subjects conditions and two or three participants per condition with a total sample size of 44 participants. This small sample size per condition made any comparison among individual ads problematic. The author did not report any analyses about the orders although the orders were part of the study. The authors research ideas might be theoretically sound, but the way to test the research ideas was methodologically flawed. Additionally, the manipulation of visual and verbal complexity was not indeed successful. The inherent complexity of image and issue ads had a more profound effect on mental processing than manipulated complexity of visual or verbal cues because visually/verbally complex image ads still attracted more attention than visually/verbally 25

PAGE 26

simple issue ads. The only conclusion from the study was that image ads attracted more attention and produced higher recall of verbal cues than issue ads. But even that finding was limited in identifying the exact impact of image and issue ads because it remained unclear whether the verbal cues focused on image or issue or whether the focus varied by ad type. Interestingly enough, those above-mentioned methodological issues did not exist in the study that favored issue ads (Kaid et al., 1992). The main purpose of the study was to test how the type of TV programs and the type of political ads would influence advertising effectiveness including recall of advertising content. Three types of ads (issue, image, and negative ads from Kit Bonds campaign for the U.S. Senate in Missouri) were inserted into three types of TV programs (situation comedy, drama, and local news) to create nine experimental conditions. A total of 283 participants were recruited on university campus and from local communities and were assigned to the nine conditions. Memory was measured with cued recall rather than recognition tests and the participants were instructed to complete open-ended questions about the issues and candidate characteristics stressed in the political advertisement (Kaid et al., 1992, p. 311). The open-ended issue question was worded: What were the major ideas or issues stressed in the political advertisement you saw? The image question read: What personal characteristics of Kit Bond were stressed in the political advertisement you saw? Recall on each item was measured by counting the number of words, phrases, or sentences which constituted distinguishably different ideas. (p. 312) The results indicated that there was no main effect of program type or interaction effect of ad type by program type on recall, but ad type had a significant main effect on recall. The issue ad outperformed both the image and negative ads in generating higher number of recalled issues and personal characteristics (Kaid et al., 1992). In summary, image ads were found more memorable than issue ads when measured with recognition of verbal cues in both types of ads (Schleuder, 1990), but issue ads were found more 26

PAGE 27

memorable than image ads when measured with recall of issues and personal characteristics in both types of ads (Kaid et al., 1992). The latter finding had more important implications for political advertising research because the chain reaction model (Lavidge & Steiner, 1961; Shama, 1976) posited an impact of issue and image content rather than verbal or visual content on candidate evaluation and voting behavior. In the current study, the cued recall questions of issue and image content (Kaid et al., 1992) are used to measure memory. The issue and image ads in the 1992 study (Kaid et al., 1992) were positive ads because the issue ad dealt with the budget deficit explicitly, with the candidate espousing his views on the subject (p. 310) and the image ad similarly discussed the budget deficit, but in more human terms, reflecting the image or persona of the candidate (p. 310). But it was hard to determine whether the negative ad in the same study was a negative issue or image ad because it used the budget deficit, but this time as an opportunity to paint the candidates opponent in a negative light (p. 310). The positive issue and image ads did not really have their negative counterparts, and there was only one tentative conclusion that favored positive ads in generating more recall of issue and image content. Nevertheless, there was a substantial body of research that specifically addresses the effects of negative or attack ads and compares negative ad exposure to positive ad exposure (Kaid, 2004). Much of the conceptual framework about negative advertising came from evolutionary psychology (A. Lang, 1991). On the one hand, humans, through evolution, were programmed to assign more attention to negative stimuli for survival (Zajonc, 1984a), and the increased attention led to better memory. On the other hand, humans, by nature, tended to approach positive stimuli and avoid negative stimuli (Fiske & Taylor, 1984), and such a tendency led to increased attention and better memory of positive stimuli. However, there was 27

PAGE 28

no decisive evidence to support either of the two explanations. The actual impact of the valence of ad, positive or negative, on memory became increasingly perplexing when researchers operationalized memory in different manners. Inconsistent results were found in two studies on radio ads (Geer & Geer, 2003; Shapiro & Rieger, 1992). In one study (Shapiro & Rieger, 1992), two image ads (honesty and leadership) and two issue ads (rent control and financial aid) were created in both positive and negative versions and were used for two races. Each participant listened to either two image or two issue ads for the two races, and one ad was positive and the other ad was negative. Eight experimental conditions were created and 106 college student participants were recruited for the eight conditions. Memory was measured with cued recall of issues and the participants were instructed to list all the specific ideas or arguments you can remember from the ad (Shapiro & Rieger, 1992, p. 140). The results indicated that negative ads outperformed positive ads in generating higher number of recalled arguments, but the order of the ads, image/issue content, and interaction of image/issue content by negative/positive content had no influence on memory (Shapiro & Rieger, 1992). In another radio ad study (Geer & Geer, 2003), four versions of party ads instead of candidate ads were created, namely negative/positive Republican ads and negative/positive Democratic ads. All the ads were issues ads and contained the same issues. Furthermore, two different copies of the four versions were created to minimize unusual effects of one particular ad, which led to eight experimental conditions. A total of 121 college students were assigned to the eight conditions. 28

PAGE 29

Memory was measured with free recall and the participants were instructed to write as completely and accurately as possible the political ad that you heard a few minutes ago (Geer & Geer, 2003, p. 86). The recall was further divided into idea units defined as meaningful divisions in the ads A correct memory unit was the presence in the recall protocol of an idea unit that had been present in the ad (p. 86). Because party sponsorship was found to have no influence on memory, the party sponsorship variable was excluded and comparisons were made between positive and negative ads only. Contrary to finding in the above-mentioned study (Shapiro & Rieger, 1992), the results indicated that negative ads did not differ from positive ads in generating the number of memory units but positive ads outperformed negative ads in generating more correct memory units (Geer & Geer, 2003). By sharp contrast, studies on TV ads (A. Lang, 1991; Newhagen & Reeves, 1991) produced consistent results and suggested that negative ads were more memorable than positive ads. In one study (A. Lang, 1991), four positive and four negative ads were used to create four orders of ads and 67 college student participants were assigned to one of the four orders. Memory was measured with three methods, free recall, cued recall, and recognition with multiple-choice test. For the free recall, the participants were instructed to write down everything they could remember about the commercials they had seen (A. Lang, 1991, p. 228). For the cued recall, the participants were given a list of the candidates in the commercials. They were told that if this list caused them to remember any commercials they had not previously recalled they should now write everything they could remember about those commercials (p. 228). For the multiple-choice test, the participants were given questions about each commercial 6 questions 29

PAGE 30

about information available only in the audio track of the commercial and 6 questions about information available only in the video portion of the commercial (p. 228). The free and cued recall was further coded into visual and verbal ideas. Visual ideas were defined as any single image recalled from the commercial or any global visual description of the commercial. Verbal ideas were defined as any single thought including names, states, office the candidate was running for, or global audio description (p. 229). The results indicated that negative ads outperformed positive ads in generating higher number of ideas as indicated in the free recall and more accurate audio-visual memory as indicated in the multiple-choice test. In another study (Newhagen & Reeves, 1991), a total of 28 ads were shown to 30 adults recruited from the local community. The 28 ads included not only negative/attack and positive/support ads but also comparative ads that were made to lessen the image of the opponent and promote a positive image of the sponsor as an alternative (p. 198) within one ad. Those 28 ads were rotated in three orders. Memory was measured with a recognition test. After watching the 28 ads, the participants were shown a second videotape that contained pictures and sound bites from the advertisements (p, 205) and were instructed to press a yes or a no button to indicate whether the material was from the messages they had viewed (p. 205). Out of the 28 ads, 18 ads were retained for the analysis with six ads for the three categories. The results indicated that negative ads outperformed both comparative and positive ads in generating more accurate recognition. Nevertheless, there were obviously some methodological issues with that study. First, three orders were not enough to rotate the 28 ads. Second, with three orders and 30 participants, there was only 10 participants each order, which limited the value of order 30

PAGE 31

manipulation. Third, excluding part of the ads based on some preliminary analysis reduced the credibility of the research. In summary, studies on memory differed from each other in the measurement of memory and the difference in measurements led to mixed results. Two studies (Newhagen & Reeves, 1991; Schleuder, 1990) used recognition tests, and the participants were instructed to judge whether certain verbal, audio, and visual cues were present in the ads they had watched. Although those two studies provided evidence that image ads outperformed issue ads in generating better recognition of verbal cues (Schleuder, 1990) and negative ads outperformed positive ads in generating better recognition of audio-visual cues (Newhagen & Reeves, 1991), there were a number of methodological issues with the two studies. Furthermore, recognition tests were criticized for overlooking materials in memory not included in the recognition alternatives (Geer & Geer, 2003, p. 72) and inflating the rate of accuracy due to successful guessing (p. 72). Recall was recommended to measure memory because it allows for the full range of subjects memory to be expressed (p. 72). Four studies (Geer & Geer, 2003; Kaid et al., 1992; A. Lang, 1991; Shapiro & Rieger, 1992) used recall, but they differed in the type of ads being recalled, the type of recall being used and the type of cues being recalled. In one study (Kaid et al., 1992), cued recall was used to measure memory of issues and candidate personality in TV ads, and the results indicated that positive issue ads outperformed positive image ads and negative ads in generating higher number of recalled issues and personal characteristics. In another study (Shapiro & Rieger, 1992), cued recall was used to measure memory of issues in radio ads, and the results indicated that negative ads outperformed positive ads in generating higher number of recalled issues. In the third study (Geer & Geer, 2003), free recall was used to measure idea units in radio ads, and the results 31

PAGE 32

indicated that negative ads did not differ from positive ads in generating the number of idea units but positive ads outperformed negative ads in generating more correct idea units. The problem with the third study was the ambiguity of idea units because it remained unclear whether the idea units were about issues or image. In the last study (A. Lang, 1991), cued recall, free recall, and recognition test were all used to measure verbal, audio, and visual cues in TV ads, and the results indicated that negative ads outperformed positive ads in generating higher number of verbal, audio, and visual cues as indicated in the free recall and more accurate audio-visual memory as indicated in the recognition multiple-choice test. Again, it remained unclear whether the cues and memory were about issue or image. This study uses the chain reaction model of advertising effects (Lavidge & Steiner, 1961; Shama, 1976) and knowledge of issue and image was an important link in the model. Research findings about verbal, audio, and visual cues did not specify whether the cues focused on issue and image; therefore, the findings were not very relevant to the current study. There was strong evidence that issue ads outperformed image ads in generating higher number of recalled issues and personal characteristics (Kaid et al., 1992). Meanwhile, there was certain evidence that positive ads outperformed negative ads in generating higher number of recalled issues and personal characteristics (Kaid et al., 1992), but it remained unclear whether the negative ad in the study (Kaid et al., 1992) focused on issue or image. The exact impacts of issue/image and positive/negative ads are further explored in the current study. Effects of Political Advertising: Candidate Electability According to the chain reaction model of advertising effectiveness (Lavidge & Steiner, 1961; Shama, 1976), increased knowledge about a candidate was just the first step of the reaction chain, and the increased knowledge or memory in turn contributed to the formation of voters opinion on the candidate or the candidates electability. Quite a few studies explored the impact 32

PAGE 33

of political advertising content on candidate electability. Issues were more socially acceptable for political advertising than image (Roddy & Garramone, 1988; Thorson et al., 1991); therefore, issue ads led to more positive candidate evaluation than image ads (Kaid & Sanders, 1978). Positive ads also led to more positive candidate evaluation than negative ads (Ben-Ur & Newman, 2002; Hill, 1989; Kahn & Geer, 1994). In one study (Thorson et al., 1991), the researchers selected four TV ads for each of four candidates, i.e., one positive issue ad, one positive image ad, one negative issue ad, and one negative image ad for each candidate. The researchers also created 16 orders of 13 ads, with one ad from each of the four candidates, one filler political ad, and another eight filler consumer TV ads. With a sample size of 161 participants, each order had about ten to eleven participants. In that study, memory was also examined and was measured with a recognition test. The participants were asked to indicate whether certain facts really contained in the ads. Those facts included issues (or image qualities) office for which the candidate was running, the state, and other people referred to in the ad (p. 474). Candidate electability was examined with four variables. The first variable was ad attitude and was measured with eight 9-point bipolar adjective pairs. The eight items were believable unbelievable; persuasive unpersuasive; like dislike; pleasant unpleasant; truthful deceptive; accurate inaccurate; ethical unethical; good bad (p. 474). The second variable was attitude toward the candidate and was measured with nineteen 9-point bipolar adjective pairs. Through a factor analysis, seven of the 19 adjective pairs were used to indicate candidate character, and the remaining 12 items were used to indicate candidate ability. The seven items about candidate character were warm cold; emotional unemotional; friendly unfriendly; compassionate uncompassionate; kind unkind; honest dishonest; and sincere insincere (p. 476). The 12 items about 33

PAGE 34

candidate ability were stimulating boring; articulate inarticulate; weak strong; stupid smart; persuasive unpersuasive; experienced inexperienced; and active passive (p. 476). The third variable was the pleasure dimension of emotional response (Russell, 1978) and was measured with six 9-point bipolar adjective pairs. The six items were hopeful despairing; pleased annoyed; happy unhappy; relaxed bored; satisfied unsatisfied; contented melancholic (Thorson et al., 1991, p. 475). Fourth variable was the arousal dimension of emotional response (Russell, 1978) and was measured with another six 9-point bipolar adjective pairs. The six items were stimulated relaxed; wide awake sleep; excited calm; frenzied sluggish; jittery dull; aroused unaroused (Thorson et al., 1991, p. 475). Voting intent was measured with a 9-point bipolar adjective scale ranging from very likely to very unlikely. Additionally, familiarity was measured with a 7-point bipolar adjective scale ranging from familiar to unfamiliar, and it turned out that all the four candidates were unfamiliar to the participants. In the analysis, political ads from different candidates and in different orders were collapsed into either image or issue category. There was no difference in the memory of ad content between image and issue ads. There was also no difference in arousal between image and issue ads. Nevertheless, issue ads outperformed image ads in producing more positive ad attitude, overall candidate attitude, candidate character, candidate ability, voting intention, and higher pleasure. The ads were also collapsed into either positive or negative category. Positive ads outperformed negative ads in generating better memory. Positive ads also outperformed negative ads in generating more positive ad attitude, overall candidate attitude, candidate character, higher pleasure, and lower arousal. There was no difference in candidate ability or 34

PAGE 35

voting intention between positive and negative ads. There was also no interaction effect of image/issue by positive/negative on memory, candidate evaluation, or voting intention. However, the study itself had a few problems. First, similar to the Schlender study (1990), order was part of the experimental design, but no tests on ordering effects were performed. In each order, the participants watched all the four types of ads; but the researchers used ANOVA as if the ad type had been a between-subjects variable, and they did not report the degree of freedom of ANOVA. Second, it remained unclear how a 9-point pleasure and arousal scales resulted in mean scores ranging from -.83 to .20 (Table 1, p. 478). The researchers did not provide any explanation about the difference in scoring. Third, although there were significant differences between issue and image ads and between positive and negative ads in candidate evaluation measures, the scores for those ads were well above the mid-point of their corresponding 9-point scales and the actual differences ranged from .23 to .39 (Table 1, p. 478). The statistical significance could be inflated because the scores were collapsed into either the image/issue or the positive/negative category and each paired comparison involved the same 161 participants. In other words, that study did not provide convincing evidence about the true difference in candidate evaluation between issue and image ads or between positive and negative ads for unknown candidates. Finally, it remained unclear what kind of emotional response was measured. The researchers did not indicate whether it was response to the ads or to the candidates. In another experimental study, Kahn and Geer (1994) tried a different approach. They used between-subjects design and compared candidate evaluation toward four types of TV ads, i.e., issue/image by positive/negative ads, from one unknown candidates political campaign. Besides creating the first four experimental conditions in which the participants only watched 35

PAGE 36

one type of ads, the researcher also created another five conditions in which the participants watched 1) positive image and issue ads supporting the candidate, 2) negative image and issue ads sponsored by the same candidate, 3) a positive image ad of the candidate and a negative image ad sponsored by his opponent, 4) a positive issue ad of the candidate and a positive issue ad of his opponent, or 5) a negative image ad sponsored by the candidate and the negative image ad sponsored by his opponent. With a sample size of 303 participants, each condition had about 24 to 39 participants. The authors also used candidate competence and characteristics to assess candidate electability. Candidate competence was measured with a 6-point adjective scale ranging from very competent to very incompetent. The same scale was used to assess competence in three issues including education, health, and economy. Education was the focus of the issue ad. Health and economy were added because the researchers believed that those two were the central concerns for most political campaigns. Candidate traits was measured with a 4-point scale ranging from a great deal to not at all. The same scale was used to assess four traits, i.e., trustworthy, strong leadership, hardworking, and knowledgeable. The participants were also asked to indicate their intention to vote for the candidate on a 4-point scale ranging from very likely to not at all likely. There was certain level of inconsistency between evaluation measures. Except for the negative image condition of the first four conditions, the evaluation scores of issue competence and personality were all above their corresponding scale midpoints by .01 to 1.01 point (Table 2, p. 103) and were slightly favorable to the candidate. Nevertheless, the overall evaluation scores were not significant across the four conditions. 36

PAGE 37

Although interaction effects of issue/image by positive/negative were testable with the experimental design, there was certain ambiguity about the existence of those effects. On the one hand, the authors did not mention including an issue/image by positive/negative interaction term in their ANOVA. On the other hand, their comparison of the first four conditions seemed to suggest the existence of some interaction effects. Positive ads outperformed negative ads in producing higher evaluation toward the candidates issue competence in education and economy and the candidates personality as a strong leader. When comparisons on the same issue competence and personality were made within positive ads, issue and image ads did not differ in the evaluations. But when the same comparisons were made within negative ads, issue ads outperformed image ads in generating higher candidate evaluations. The authors asked the participants to evaluate the candidate on three issues. The results regarding education seemed reasonable since education was the focus in the issue ads. It remained unclear why there was significant difference in response to economy but no difference to health after the participants watched an issue ad about education. The authors also asked the participants to evaluate the candidate on four personality traits. There was significant difference only in the evaluation of the candidates leadership. It remained unclear why the content had no effect on the other three traits. Finally, although positive ads worked much better than negative ads in enhancing issue competence and personality evaluation, the negative issue ad in fact produced significantly higher voting intention than the other three ads. The authors showed participants two ads in another five conditions. Comparisons were made between the impact of one ad and the cumulative impact of two ads. Two positive ads together produced higher evaluation toward issue competence in education and personality as knowledgeable, higher and above midpoint overall candidate evaluation, and higher intention to 37

PAGE 38

vote than one positive issue ad or one positive image ad only. Nevertheless, exposure to more positive ads did not enhance candidate evaluation toward issue competence in health and economy and personality trait as trustworthy, hardworking, and strong leadership. Comparisons were also made for negative ads. Two negative ads sponsored by the candidate produced lower evaluation toward personality as trustworthy than one negative issue ad or one negative image ad only. Those two negative ads together also resulted in lower intention to vote for the candidate than one negative issue ad. Surprisingly, those two negative ads together in fact resulted in higher intention to vote for the candidate than one negative image ad. No additional differences were found with regard to the other aspects of the candidate issue and personality evaluation. The authors also showed some participants both a positive image ad of the candidate and a negative image ad sponsored by the candidates opponent. The second ad from the opponent had no effect on the evaluation toward and intention to vote for the candidate. They also showed other participants both a positive issue ad of the candidate and a positive issue ad of the candidates opponent. Again the second ad from the opponent had no effect on the evaluation toward and intention to vote for the candidate. Nevertheless, when a negative image ad sponsored by the candidates opponent was shown after a negative image ad sponsored by the candidate, the candidates evaluation in terms of his personality as trustworthy, strong leadership, and knowledgeable was significantly reduced than when the negative image ad sponsored by the candidate was shown only. In other words, showing the participants the candidate and the candidates opponent attacking each other on image with two ads was the most likely to lower the evaluation of the candidate than showing the participants only one negative image ad sponsored by the candidate. 38

PAGE 39

In summary, some researchers (Thorson et al., 1991) found that issue ads generated more positive candidate electability than image ads for unknown candidates, and other scholars (Kahn & Geer, 1994) argued that issue ads did not differ from image ads in influencing candidate electability for unknown candidates. However, previous research (Kahn & Geer, 1994; Thorson et al., 1991) consistently indicated that positive ads outperformed negatives ads in generating more positive candidate electability. Furthermore, there was evidence (Kahn & Geer, 1994) that negative issue ads outperformed negative image ads in generating more positive candidate electability, and showing negative image ads sponsored by both a candidate and the candidates opponent was more likely to lower the electability of the candidate than showing only one negative ad sponsored by the candidate. The notion of electability is conceptually similar to the construct of attitude. A number of attitude studies employ a tripartite framework, which includes cognition, affect, and conation as the three components of attitude (Greenwald, 1989; Morris, Woo, Geason, & Kim, 2002). The cognitive element includes beliefs stored and retrieved from memory (Hoyer & Maclnnis, 2004; McGuire, 1989). The affective element includes evaluative judgment and personality assessment (Aaker, 1997; McGuire, 1989). The conative element includes behavioral intention (Fishbein & Ajzen, 1975; Fishbein & Ajzen, 1980; McGuire, 1989). Those three components have been used in previous research on political advertising, e.g., beliefs about issue competency (Kahn & Geer, 1994), candidate personality (Kahn & Geer, 1994; Thorson et al., 1991), and voting intention (Kahn & Geer, 1994). In the current study, those three variables will be used to create a composite measure of candidate electability. Summary of the Effects of Political Advertising A few empirical studies (Geer & Geer, 2003; Kahn & Geer, 1994; Kaid et al., 1992; A. Lang, 1991; Newhagen & Reeves, 1991; Schleuder, 1990; Shapiro & Rieger, 1992; Thorson et 39

PAGE 40

al., 1991) were reviewed in great detail. Those studies did not provide consistent evidence of the effects of political advertising on memory and candidate electability. The inconsistency could be caused by different operationalization of memory and candidate evaluation, different content of ads, and certain methodological problems in some studies. There was lack of specification of the exact content being measured, for example, the focus of visual, audio, and verbal cues (A. Lang, 1991; Newhagen & Reeves, 1991; Schleuder, 1990), and the target of emotional response (Thorson et al., 1991). Additionally, based on the chain reaction model of advertising effects (Lavidge & Steiner, 1961; Shama, 1976), memory and candidate electability were not two types of separate effects. Memory of an ad could influence voters opinion on the electability of the candidate advertised in the ad. A few measures were developed to measure memory of issues and candidate characteristics (Kaid et al., 1992) and candidate competence, personal traits, and voting intention (Kahn & Geer, 1994). But those measures were not incorporated into one study to test the chain reaction model. This missing link in previous research is investigated in the current study with a path analysis. More details about the path analysis are in Chapter 3. Informational and Transformational Advertising Previous research identified five antecedents that influenced the use of informational or transformational advertising (Swaminathan, Zinkhan, & Reddy, 1996). The relevance of those five antecedents to political advertising was assessed in a prevous study (Shen & Kim, 2006). Antecedent 1: Newness of the Product Consumers often have a need to search for information to identify and evaluate a new product category (Cohen & Basu, 1987). But when consumers became well informed, their information need declined and their evaluation of the product category was largely based on actual experience of the product (Swaminathan et al., 1996). The same idea was also found in 40

PAGE 41

some discussion about the relationship between the use of informational or transformational advertising and product life cycle (Porter, 1985). In the early stages of a consumer product, informational advertising was recommended to communicate factual information of the product; whereas in the late stages, transformational advertising was recommended to communicate use experience of the product (Porter, 1985). Newness of product was relevant to political advertising (Shen & Kim, 2006). An unknown challenger, or a candidate in the early stage of his or her political life, tended to build awareness through informational advertising, whereas a well-known incumbent, or a candidate in the late stage of the political life, tended to capitalize voters positive experience with his or her first term through transformational advertising (Schleuder, 1990). Antecedent 2: Level of Risk In the context of consumer behavior, risk referred to the uncertainty encountered by consumers when they were estimating the social and economic consequences of a purchase decision (Swaminathan et al., 1996). Risk was associated with product newness and information search (Hoyer & Maclnnis, 2004). A new product was perceived riskier than a familiar product; therefore, if consumers had to make a decision between the two, one way to make a less risky decision was to make an informed decision through adequate information search (Darke, Freedman, & Chaiken, 1995). Additionally, decision-making literature points out that risk was often associated with involvement (Hoyer & Maclnnis, 2004). Antecedent 3: Product Involvement In the context of consumer behavior, involvement referred to the extent to which a product had important consequences for consumers to form an informed opinion and to justify their purchase decisions (Chaiken, 1980). Expensive products resulted in both strong perception of risk and high involvement with the products (Hoyer & Maclnnis, 2004). Voters highly involved 41

PAGE 42

with a political election might believe that their votes would have a significant impact on their society and tended to search for adequate information to make a well-informed decision (Nimmo, 1999). Therefore, for those voters, informational advertising might be the only appropriate approach. In fact, there was evidence that for voters who were interested in or involved with political campaigns, their candidate evaluation was primarily determined by issue advertisements rather than image advertisements (Christ, Thorson, & Caywood, 1994). Antecedent 4: Product Conspicuousness In the context of consumer behavior, product conspicuousness referred to the intended communication role played by certain consumption decisions (Belk, Balm, & Mayer, 1982). In other words, consumers expressed themselves through purchasing and using products that were congruent with their self-images (Swaminathan et al., 1996). Therefore, the purchase decision was based on real or upcoming experience of using the product or the transformational aspect rather than the hard facts or the informational aspect of the product. In the context of political advertising, a candidate could use transformational advertising to reinforce support from voters who had positive experience with the candidate and felt well connected to the candidate (Shen & Kim, 2006). Antecedent 5: Product versus Service In the context of consumer behavior, intangibility was a fundamental difference between product and service (Swaminathan et al., 1996). Because consumers could not perceive service in the same manner in which consumers could sense products (Zeithaml & Bitner, 1996), transformational advertising was used to present service offering with real-person experience of the service (Swaminathan et al., 1996). At first glance, political office might be closer to the concept of intangible service than tangible product, which made transformational advertising a better match to political campaigns (Shen & Kim, 2006). 42

PAGE 43

However, research on intangibility suggested that although service itself might be highly intangible, the service delivery system was well tangible (Mittal, 1999). Similarly, although the political service a candidate would provide may be highly intangible, the candidate himself or herself, as the core service delivery system, was well tangible (Shen & Kim, 2006). Basic facts about the candidate were crucial for voters to make an informed decision (Nimmo, 1999). Therefore, informational advertising was still an appropriate approach to political campaigns although political campaigns essentially aimed to sell candidates political service (Shen & Kim, 2006). Typology of Informational/Transformational Strategies Previous research identified five informational strategies and four transformational strategies in consumer advertising (Laskey, Day, & Crask, 1989). Again those items were examined in the context of political advertising for relevance assessment in a previous study (Shen & Kim, 2006). The first informational item, comparative, referred to competition explicitly mentioned (Laskey et al., 1989, p. 38). It was relevant to political advertising because in political advertising, especially in negative political advertising, explicit mention of opponent was basically a common practice (Shen & Kim, 2006). The second item, preemptive, referred to testable claim of superiority based on an attribute or benefit (Laskey et al., 1989, p. 38). It was also relevant because in political advertising, there were often statements about a candidates previous performance in office and/or issue stands, which should be testable or verifiable and could help indicate that candidates superiority to other candidates (Shen & Kim, 2006). Nevertheless, one risk of using the preemptive strategy was that the same record could also be framed negatively and be used in an opponents attack ads (Nimmo, 1999). In fact, negative/attack ads focused more on issues than on image (Johnston & Kaid, 2002; Kaid & 43

PAGE 44

Johnston, 2001; West, 1993). Certain actions based on inoculation theory of persuasion (Anderson & McGuire, 1965), such as warning of imminent attack from the opponent followed by a weak version of such an attack, should be taken if some political records were likely to cause controversies. The third item, hyperbole, referred to un-testable claim of superiority based on an attribute or benefit (Laskey et al., 1989, p. 38). It was also relevant because in political advertising, there were also superiority claims about new initiatives to be implemented but those new initiatives were largely un-testable at the moment they were proposed (Shen & Kim, 2006). The fourth item, unique selling proposition, referred to explicit claim of uniqueness (Laskey et al., 1989, p. 38) and involved heavy use of the word or concept of only (p. 38). It was, however, not really relevant because in the U.S. elections, the influence of independent candidates was so limited that the competition was primarily between the two candidates from the Republican Party and the Democratic Party respectively (Devlin, 2001). Because there were just two candidates, any comparisons between the two automatically involved the concept of only (Shen & Kim, 2006). Including this strategy might lead to possible confusion with other informational items that emphasized the superiority of a candidate, for example, pre-emptive and hyperbole (Shen & Kim, 2006). Generic, the last informational item, referred to focus on product class (Laskey et al., 1989, p. 38). It was also not relevant because there was no such a concept of generic candidate in political advertising (Shen & Kim, 2006). The first transformational item, user image, focused on the users of a brand and their lifestyles (Laskey et al., 1989, p. 39). It was relevant because some political advertisements did show how the life of certain type of voters or users of a political candidates service would be influenced either positively or negatively by their decision to choose a candidate (Shen & Kim, 44

PAGE 45

2006). The second item, brand image, focused on the image of the brand itself in an attempt to convey a brand personality (Laskey et al., 1989, p. 40). It was also relevant because in political advertising, and it was again a common practice to emphasize the personality or characteristics of a candidate (Shen & Kim, 2006). The third item, use occasion, focused on the experience of using the brand, or on those situations where use of the brand is most appropriate [or] the fit between the brand and the situation; not the fit to a particular user profile (Laskey et al., 1989, p. 40). It was also relevant because some political advertisements did show scenarios that suggested the appropriateness to choose a candidate for a particular issue rather than for a particular type of voters (Shen & Kim, 2006). Generic, the last transformational item, again focused on the product class rather than on a particular brand (Laskey et al., 1989, p. 40). This item was not relevant because there was no such a concept of generic candidate (Shen & Kim, 2006). Summary In summary, the antecedents to the use of information or transformation for consumer advertising were well applicable to political advertising. Furthermore, three out of the five informational strategies and three out of the four transformational strategies were relevant to political advertising. In fact, the three informational strategies and three transformational strategies were used to code the issue and image content in political advertisements (Shen & Kim, 2006). Issue political ads were found to contain more informational strategies than transformational strategies, and image political ads were found to contain more transformational strategies than informational strategies. Nevertheless, content analyses on issue and image in political advertising might be of little use if there was a lack of reliable and valid methods to measure voters perception of the content and identify the impact of the perception on candidate evaluation (Shen, 2007). In the current study, the Puto and Wells (1984) 45

PAGE 46

informational/transformational survey scale (hereinafter referred to as the Scale) is used to better understand the perception and impact in question. Survey Scale of Informational/Transformational Advertising The Scale had 23 items. Eight of them measured the perception of informational cues (hereinafter referred to as PIFC) in an advertisement and were developed from previous research on the informative function of advertisements (Aaker & Norris, 1982; Resnik & Stern, 1977). The other 15 items measured the perception of transformational cues (hereinafter referred to as PTFC) and were developed from viewer response profile measures (Schlinger, 1979) and Needham, Harper, and Steers Advertisings communication measures. The Scale was by far the most popular measure of informational/transformational advertising and was used for more than two decades (Menon, Goodnight, & Wayne, 2006), although survey scales of this type were criticized for being limited in capturing transformational experience and should be replaced by open-ended projective techniques with pictures or narratives (Aaker & Stayman, 1992). In the original validation study, Puto and Wells (1984) first coded 400 TV commercials based on the definitions of informational and transformational advertising. Out of those 400 commercials, they selected five primarily informational commercials and eight primarily transformational commercials. Next, 130 college students watched the 13 commercials and responded to the survey scale to report PIFC and PTFC of the commercials. The average responses of the 13 TV commercials were analyzed to assess the performance of the Scale. A set of Cronbachs reliability tests indicated that the eight informational survey items reached an average reliability of .73 and the 15 transformational survey items reached .88. A set of t-tests was used to compare the difference between each commercials PIFC and PTFC scores. The difference was significant for all the commercials and the higher score was in complete 46

PAGE 47

agreement with the prior coding as whether the commercial was primarily informational or transformational. Unfortunately, the validation study itself was not flawless. The participants watched all the 13 commercials and responded to the same scale 13 times, but the researchers used t-tests rather than repeated measures for this within-subjects design. Moreover, since Cronbachs reliability test was known to favor scales of larger numbers of items even with the same degree of inter-item correlations (Hair, Anderson, Tatham, & Black, 1995), the Scale was likely to be in favor of the transformational items because those items almost outnumbered the informational item by 2 to 1. Meanwhile, because the Scale had six points from Strongly Agree to Strongly Disagree, the participants were forced to make a sided choice and could not indicate a neutral stand. Finally, the participants were also asked to report prior exposure and evaluation of the commercials, but the relationship between those measures and the Scale was not analyzed in that validation study. In one recent study of using the Scale for political ads (Shen, 2007), a few changes were made to the Scale; for example, brand was replaced with candidate and company with party, and a mid-point category neither agree nor disagree was added to the Scale to form a 7-point scale, ranging from strongly disagree to strongly agree. That study indicated that the PIFC score of an issue ad was significantly higher than its PTFC score, whereas the PTFC score of an image ad was significantly higher than its PIFC score. The result provided additional evidence to support the conceptual similarity between issue/image ads and informational/transformational advertising. The Scale is also used in the current study with the changes of the wording and scale points made in the previous study (Shen, 2007). 47

PAGE 48

Emotion in Advertising The informational content of advertising was also called the cognitive dimension of advertising (Puto & Wells, 1984). Content of this dimension included verbal messages such as rational and logical arguments as well as non-verbal information such as visual imagery, music, and language variables that complemented the verbal messages (Wells, 1988). The transformational content of advertising was also called the experiential dimension or essentially the emotional dimension of advertising because transformation occurred only when consumers explicitly associated certain emotional feeling to the use of an advertised brand (Puto & Wells, 1984). Content of this dimension included emotional appeals to induce a wide variety of human emotions (Wells, 1988). Emotional appeals in advertising were often classified into positive and negative appeals (Hoyer & Maclnnis, 2004). Empirical research has addressed the effects of positive appeals such as happiness (Lazarus, 1991a), hope (MacInnis & de Mello, 2005), humor (C. Duncan & Nelson, 1985; Gelb & Pickett, 1983), pride (Lazarus, 1991a), and warmth (Aaker, Stayman, & Hagerty, 1986) as well as the effects of negative appeals such as anger (Averill, 1982; Frijda, 1986; Lazarus, 1991a; Roseman, Antoniou, & Jose, 1996), fear (Boster & Mongeau, 1984; Brooker, 1981; Mewborn & Rogers, 1979; Witte, 1995), guilt (Boster et al., 1999), and irritation (Aaker & Bruzzone, 1985). Rational and Emotional Responses Although the informational content produced rational responses whereas the transformational content evoked emotional responses, the occurrence of the two types of responses was in fact highly intertwined (Puto & Wells, 1984). On the one hand, rational responses to an informational ad, such as message elaboration and belief formation, could be substantially enhanced when appropriate emotional content was added (Deighton, Romer, & 48

PAGE 49

McQueen, 1989; T. Duncan, 2002; Mano, 1997). Even in product categories where there were substantial technological differences that provided a basis for a strong informational claim of product superiority, advertisers still attempted to develop emotional bonds between consumers and an advertised brand (Marshall, 2006). On the other hand, the occurrence of emotional responses was not necessarily independent of cognition (Lazarus, 1984; Zajonc, 1984b). There were two distinct routes to the production of emotional responses (Andreasen, 1997; Izard, 1993; LeDoux, 1995; Teasdale et al., 1999). One route involves subcortical structures and generates emotional responses to simple perceptual and associative aspects of stimuli. The second route involves cortical structures and generates emotional responses reflecting the meanings of interpretations given stimuli (Teasdale et al., 1999, p. 210). In other words, primitive, genetically prewired circuit entities of the visceral brain circuits which appear to influence information processing in higher cognitive areas of the cortex (Panksepp, 1989, p. 8) was not involved in the generation of subcortical emotional responses but cortical emotional responses. Subcortical emotional responses were well recorded in the classical conditioning of fear reactions to audio or visual stimuli (LeDoux, Iwata, Cicchetti, & Reis, 1988; LeDoux, Romanski, & Xagoraris, 1989) and they either interrupted the cognitive focus of current attention or influenced the context for ongoing cognitive processes (Simon, 1982). Cortical emotional responses were well recorded in studies that manipulated semantic or schematic meanings of audio or visual stimuli (Decety & Chaminade, 2003; Mitchell, Elliott, Barry, Cruttenden, & Woodruff, 2003) and the meanings directly mediated the generation of emotion through cognitive processing (Teasdale, 1993; Teasdale & Barnard, 1993; Teasdale et al., 1999). Because ads in general were highly associated with semantic and schematic meanings (Hoyer & Maclnnis, 2004), cognitive theories of emotion, such as 49

PAGE 50

attribution theories (Weiner, 1985) and appraisal theories (Lazarus, 1991b), were often used to account for the role of knowledge structures in the processing of emotional responses. The close interrelationship between cognition and emotion was more evident when the hard facts or benefits of a product might be emotional in nature; for example, romantic feeling was used to promote perfume and the feeling of interest was also an internal attribute of the product (Hoyer & Maclnnis, 2004). Strictly speaking, an informational ad had a primary content focus on information/cognition and a secondary content focus on transformation/emotion (Shen, 2007). A transformational ad had a primary content focus on transformation/emotion and a secondary content focus on information/cognition (Shen, 2007). Cognitive and emotional responses to an informational ad or a transformational ad might be equally important (Shen, 2007); but under certain circumstances, emotional response was more important than cognitive response in the formulation of attitude (Morris et al., 2002). Attitude was defined as a positive or negative evaluation about certain person, object or issue, and was composed of three elements, cognitive, emotional, and conative (Hoyer & Maclnnis, 2004). Although an element of attitude, conative attitude was often treated as intention and predicted with the cognitive and emotional elements of attitude (Cacioppo & Petty, 1989; Fishbein & Ajzen, 1975). The relative impact of cognitive and emotional attitude on conative attitude or intention was examined in one recent study on responses to 240 TV, radio and print ads in 13 product categories (Morris et al., 2002). Cognitive attitude was operationalized as cognitive responses including believability and knowledge. Emotional attitude was operationalized as emotional responses including pleasure, arousal, and dominance. Conative attitude was operationalized as brand interest and purchase intention. Path analysis 50

PAGE 51

with structural equation modeling indicated that emotional attitude had a much large direct impact on conative attitude than did cognitive attitude at an aggregate level. Fellow-up regression analyses by the 13 categories further indicated that emotional attitude outperformed cognitive attitude in accounting for more variance in conative attitude except for computers. Emotional attitude explained 3 to 30 percent of the variance in conative attitude across the 13 categories whereas cognitive attitude accounted for 2 to 13 percent of the variance. Regression analyses by the three media also suggested the superiority of emotional attitude over cognitive attitude in predicting conative attitude except for radio. Emotional attitude explained 15 to 22 percent of the variance in conative attitude across the three media whereas cognitive attitude accounted for 0 to 26 percent of the variance. Although there was evidence that emotion could be more important than cognition in attitude formation (Morris et al., 2002), attitude models were primarily cognitive. The best example was the elaboration likelihood model (ELM) (Petty & Cacioppo, 1986). According to the ELM, there were two routes to attitude formation: (1) the route taken by individuals who thought actively about information in a message, or the central route and (2) the route taken by individuals who did not think actively about the informational content of a message, or the peripheral route. The central route emphasized cognitive/effortful thinking and minimal emotional involvement in processing messages. Individuals who traveled this route were referred to as cognitive elaborators. The peripheral route was almost the exact opposite of the central route to persuasion. Individuals who traveled this route were referred to as cognitive misers and process messages based on simple affective cues such as source attractiveness and message length. Essentially, because of the cognitive base of the ELM, emotion was treated as a simple peripheral cue that led to less persistent attitude formation and change whereas cognition 51

PAGE 52

was regarded as the main force contributing to a longer-lasting attitude formation and change (Petty & Cacioppo, 1986). A recent study (Morris, Woo, & Singh, 2005) reexamined the superiority of cognition over emotion in attitude formation as claimed in the ELM. Emotional responses of cognitive elaborators, i.e. people who formed their attitude through the central route, and cognitive misers, i.e., people who formed their attitude through the peripheral route, were compared in terms of pleasure, arousal, and dominance for the same ads. Cognitive elaborators in fact had a higher level of emotional responses, particularly a higher pleasure score, than cognitive misers. Furthermore, cognitive elaborators also indicated a higher level of purchase intention than cognitive misers and there was a particularly significant impact from pleasure to purchase intention. Measuring Emotion A major issue in cognition and emotion studies was the difficulty measuring emotional response (Morris, 1995). It was widely accepted that cognitive responses could be measured with recall or recognition of message content and thought-listing surveys about beliefs (Hoyer & Maclnnis, 2004). But measuring emotion was extremely controversial owing to a wide range of emotion types and a variety of competing emotion theories (Simon, 1982). However, most measures of emotional responses to advertising could be categorized as one of two types: physiological and self-report (Mackay, 1980). Physiological measures Physiological measures stemmed from early emotion theories that posited that emotion was manifestation of involuntary physiological or biochemical responses (e.g., changes in skin conductance, blood pressure, skin temperature and heart rate, facial expressions, respiration and 52

PAGE 53

pupil dilation, and so on) (James, 1884). Two types of physiological responses, arousal and facial expression, were often used as the indicator of emotional experience (Mackay, 1980). A few physiological techniques were developed to measure arousal. Those techniques included electroencephalographs that track electrical activity in the brain, pupillometrics that investigated the change in the size of an eyes pupil, electrocardiogram that measured heart action, and galvanic skin response that records skin conductance (Davis, 1997). Those techniques might introduce experimental artifacts because the existence of those measurement devices might influence arousal; therefore, they were often used with other types of measurement tools (Davis, 1997). New techniques such as functional Magnetic Resonance Imaging techniques (or fMRI) were increasingly being used to assess physiological responses (Morris et al., 2008). Nevertheless, technical advancement of arousal measurement could not make up for a theoretical drawback in arousal research, i.e., arousal alone did not fully explain emotional experience but just the intensity aspect of it (Clark, 1982). Facial expression could be identified through electromyographs or subtle changes in the facial musculature (Davis, 1997). But to construct a link between facial expression and specific emotion, a coding system, such as the Facial Action Coding System or FACS (Ekman & Friesen, 1978), was needed. Coding systems classified facial expressions based on a select set of dimensions and associated the pattern of musculature with a specific emotional response (Ekman & Friesen, 1978). Nevertheless, empirical evidence did not support the notion that each specific facial expression corresponded to a unique emotional response because individuals could consciously alter facial expressions owing to social norms and cultural considerations (Izard & Bartlett, 1972; Scherer & Wallbott, 1994) 53

PAGE 54

Self-report measures Self-report measures stemmed from cognitive theories of emotion with the assumption that subjective experience of emotion could be sufficiently described or labeled with words (Mackay, 1980). The marketing and advertising literature was filled with written self-reports on rating scales. For example, Duncan and Nelson (1985) developed a scale to measure humor in an ad. Edell and Burke (1987) developed a scale to capture warm, upbeat and negative feelings evoked by an ad. Within the self-report tradition, there were two distinctive approaches to emotion (Morris et al., 2008). One was a discrete approach, and the other was a dimensional approach. Discrete approach to emotion The discrete approach was in fact a Darwinian approach to emotion, which sought evolutionarily fundamental or discrete emotions to define all subjective emotional experiences (Mandler, 1984; Plutchik, 1980). According to this approach, there was a finite set of discrete or basic emotions such as joy, anger, sadness, and fear that were innate to all human beings across various cultures and developmental stages (Izard, 1992; Plutchik, 1982). Next, certain patterns or combinations of those basic emotions resulted in the subjective experience of emotion (Izard, 1992; Plutchik, 1982). Therefore, emotional responses were well measurable through investigating the extent to which each of the basic emotions was experienced (Richins, 1997). One typical way to measure emotional responses with the discrete approach was to give participants bipolar adjective pairs of basic emotions, such as happy versus unhappy (Mackay, 1980). Thorson and her colleagues (1991) employed this bipolar adjective technique when measuring emotional response to political ads. The basic emotions approach was criticized as merely labeling without a sound theoretical foundation that explained the experience of emotion (Roseman, 1984). There could be a set of eight emotions interest, surprise, joy, anguish, fear, shame, disgust, and rage (Tomkins, 1980), 54

PAGE 55

or another set of eight emotions fear, anger, joy, sadness, acceptance, disgust, anticipation, and surprise (Plutchik, 1984), or even another set of ten emotions interest, joy, surprise, distress, anger, disgust, contempt, shame, fear, and guilt (Izard, 1977). The disagreement on either the number or the kinds of emotions across the three lists seemed to suggest a lack of well-accepted theoretical ground to define the very notion of discrete (Mandler, 1984). The status of an emotion as basic often depended on a researchers own criterion, and some so-called basic emotions were proposed on grounds, such as frequency-related salience in culture and experience, levels in an emotion hierarchy, etc., which were irrelevant to whether they were basic in a psychological sense (Ortony & Turner, 1990). The reliance of the discrete approach on verbal labels such as adjectives to measure subjective emotional experiences was also problematic. The use of adjectives might lead to a distorted reporting process that was subject to an individuals linguistic habits, vocabulary, memory, and social desirability (Lazarus, 1966), and language was too insensitive a medium for communicating private experiences (Mandler, 1962). Additionally, some cultures might not have adjectives for a specific type of emotion but individuals in those cultures did experience that emotion, which made it impossible for a cross-cultural comparison on that emotion with adjectives (Kitayama & Markus, 1990). Dimensional approach to emotion The approach originated from early studies on semantics (Osgood, Suci, & Tannenbaum, 1957). Osgood and his colleagues (1957) conducted a set of factor analyses on adjectives and identified a three-dimensional semantic space composed of evaluation, potency, and activity. A few other researchers (Averill, 1975; Bush, 1973; Schlosberg, 1952) explored the semantic structure of adjectives about emotion and identified pleasantness, activation as well as potency, aggression, or attention-rejection as the dimensions of emotion. Those findings were further 55

PAGE 56

developed into a three-dimensional pleasure-arousal-dominance model of emotion (hereinafter referred to as PAD) (Russell & Mehrabian, 1977). Pleasure-displeasure referred to positive versus negative affective state in response to a stimulus and corresponded to the pleasantness component of emotion. Arousal-nonarousal referred to physical activity and/or mental alertness and corresponded to the activation component. Dominance-submissiveness referred to feelings of control and influence over others and situations versus feeling controlled and influenced by external circumstances and corresponded to the potency component of emotion (Mehrabian & Russell, 1974). However, PAD went through some changes after the original formulation. Russell later dropped dominance in his emotion research and used pleasure and arousal to define emotion (Russell, 1980) either because he preferred the simplicity of a two-factor structure (Russell, 1980) or because he failed to replicate the dominance dimension across studies or cultures (Russell, 1978, 1983). Russell (1980) went on proposing a circumplex model, in which emotion descriptors were be systematically arranged around the perimeter of a circle. This affect circumplex contained four bipolar dimensions that were spaced 45 degrees apart, namely high versus low activation, high activation and pleasantness versus low activation and unpleasantness, pleasantness versus unpleasantness, and low activation and pleasantness versus high activation and unpleasantness (Russell, 1980). The circumplex model is presented in Figure 2-1. Pleasantness and activation, which are depicted in the darker lines, are the basic dimensions of Russells affect circumplex. Watson and Tellegen (1985) also proposed an affect circumplex. Their circumplex model closely resembled Russells (1980) model in terms of the structure but differed from it in terms of the basic dimensions. To be more specific, Watson and Tellegens circumplex also contained 56

PAGE 57

four bipolar dimensions that were spaced 45 degrees apart but they were high positive affect versus low positive affect, strong engagement versus disengagement, high negative affect versus low negative affect, and pleasantness versus unpleasantness. Positive affect and negative affect, rather than pleasantness and activation, were the two basic dimensions of their affect circumplex (Watson & Tellegen, 1985, Figure 1, p. 221). Because the two circumplex models were both based on factor analysis, the difference in basic dimensions between them might just be a terminological controversy as to how to name the basic dimensions (Larsen & Diener, 1992). Larsen and Diener (1992) argued that the labels positive affect and negative affect were misleading and should be replaced with pleasantness and activation. Their argument was undoubtedly supported by Russell and his associates (Feldman, Barrett, & Russell, 1998) but was challenged by Watson, Tellegan, and their associates (Watson, Wiese, Vaidya, & Tellegen, 1999). However, a real problem for the models was neither of them fit real-life data in confirmatory factor analysis (Feldman et al., 1998). In three analyses conducted by Russell and his associates (Feldman et al., 1998), the root mean square error of approximation (RMSEA) values were .135, .131, and .106, which indicated a consistently poor fit as RMSEA values of .10 or greater demonstrate a poor fit (Browne & Cudeck, 1993). Similarly, in two analyses conducted by Watson, Tellegan, and their associates (Watson et al., 1999), the RMSEA values were .234 for momentary data and .294 for general data in the first analysis and were .094 for momentary data and .123 for general data in the second analysis, which also indicated a consistently poor fit of their circumplex model. The consistent poor fit of the circumplex models suggested that something was missing in the models (Feldman et al., 1998). Lazarus (1991b), from a relational perspective, pointed out that circumplex models such as the one proposed by Watson and Tellegen (1985) ignored the 57

PAGE 58

person-environment relationship inherent in emotional responses. Interestingly enough, the dominance dimension dropped by Russell, to some extent, addressed feelings about the person-environment relationship (Mehrabian, 1995). Shaver and his associates (Shaver, Schwartz, Kirson, & O' Connor, 1987) conducted a multidimensional scaling analysis to compare two-dimensional versus three-dimensional emotion models and concluded [t]he three-dimensional solution helps to differentiate between what the cluster analysis suggests are separate basic-emotion categories, and it is clearly more informative as a representation of emotion knowledge than the two-dimensional solution (p. 1071). There were two ways to translate the three-dimension concept into a workable measure, a verbal method (Mehrabian & Russell, 1977) and a non-verbal method (P. Lang, 1980). With the verbal method, participants were given a checklist of up to 16 bipolar adjectives and asked to indicate their feelings about a given stimulus. Next, the scores on each adjective were collapsed into the three dimensions with factor analysis (Mehrabian & Russell, 1977). With the non-verbal method, participants were given self-assessment manikin (hereinafter referred to as SAM) (P. Lang, 1980) to report their feelings. Unlike its verbal counterpart, SAM was a graphic representation of each PAD dimension with a manikin along a continuous nine-point scale (Morris, 1995). In the first row of the SAM, the manikins facial expression varied from smiling and happy to frowning and unhappy along the pleasure/displeasure dimension. In the second row, the manikins facial expression varied from eye-open and eyebrow-elevated to calm and eye-closed along the arousal/non-arousal dimension. In the third row, the manikins size varied from huge to tiny along the dominance/submissiveness dimension (Morris, 1995). The non-verbal alternative was found to be a better alternative than the verbal checklist method because it eliminated cognitive processing, and was easier and quicker to use (Morris, 58

PAGE 59

Waine, & Lang, 1993). Meanwhile, it generated a similar pattern of scale values for these situations as was obtained for the semantic differential (P. Lang, 1980, p. 123). SAM could even be viewed as a combination of self-report and facial expression because its graphic characters well resembled facial expressions in different scenarios (P. Lang, 1980). All the advantages of SAM made it a culture free measure of emotion (Bradley, Greenwald, & Hamm, 1993). Measuring emotion with AdSAM Based on SAM, AdSAM emerged as an effective tool for measuring emotional response to marketing communications stimuli, and crossed the boundary between the discrete approach and the dimensional approach (Morris, 1995). The adjectives that were used to construct PAD were added to a list of emotion adjectives found in the literature and measured with the manikin. Then the adjectives were examined for mean error or consistency in response. The result was a list of emotion adjective with PAD scores that could be used to identify variations in feelings, such as lucky, angry, provocative, and indignant, while maintaining the integrity of the three dimensions of PAD (Morris, 1995). AdSAM was used to assess responses to television advertising (Morris, Bradley, Waine, & Lang, 1992; Morris, Roberts, & Baker, 1998), pre-production vs. post-production advertising (Morris & Waine, 1993), and global advertising across cultures (Morris, 1995; Morris, Bradley, Sutherland, & Wei, 1993; Morris, Strausbaugh, & Nthangeni, 1996). In the two above-mentioned studies (Morris et al., 2002; Morris et al., 2005), it was used to refine the process of attitude formation and was a valuable tool for strategic planning, message testing and brand tracking. In the current research, AdSAM is used to measure emotional response to political ads. 59

PAGE 60

Summary of Literature Review A chain reaction model of advertising effectiveness (Lavidge & Steiner, 1961; Shama, 1976) was identified through the literature review of political marketing. According to that model, voters learned of candidate characteristics and issue stands from political ads, and the knowledge in turn substantially impacted their evaluation of a candidate. Next, a few empirical studies were reviewed to address the impacts of issue/image and negative/positive political advertising on message learning/memory (Geer & Geer, 2003; Kaid et al., 1992; A. Lang, 1991; Newhagen & Reeves, 1991; Schleuder, 1990; Shapiro & Rieger, 1992) and candidate evaluation (Kahn & Geer, 1994; Thorson et al., 1991). Although those studies provided contradictory outcomes owing to different operationalization of memory and candidate evaluation, different types of ads, and certain methodological problems, two sets of findings were identified as particularly relevant to the chain reaction model (Lavidge & Steiner, 1961; Shama, 1976) adopted by the current study. One was about message learning/memory of candidate characteristics and issue stands, and the finding was that issue ads outperformed image ads in generating higher number of recalled issues and personal characteristics (Kaid et al., 1992). With that finding, cued recall of issues and personal characteristics (Kaid et al., 1992) is used as the measure of memory in this study. The other was about candidate electability, and the findings were positive ads outperformed negatives ads in generating more positive candidate electability and negative issue ads outperformed negative image ads in generating more positive candidate electability (Kahn & Geer, 1994). Based on the chain reaction model of advertising effects (Lavidge & Steiner, 1961; Shama, 1976), memory and candidate electability were not two separate effects. Memory of an ad might influence the evaluation of the candidate advertised in the ad, but such an influence was not 60

PAGE 61

tested in previous research. This missing link is explored in the current study with a path analysis. Furthermore, conceptual similarity between image/issue political ads and informational/transformational advertising was identified in the literature review. An image ad had a primary focus on transformation and a secondary focus on information, whereas an issue ad had a primary focus on information and a secondary focus on transformation (Shen, 2007). The Puto and Wells (1984) informational/transformational survey scale is used to calibrate the perception of informational and transformational content of issue and image ads and to better understand the impact of perception on advertising effectiveness. The informational aspect of adverting was associated with rational/cognitive response to advertising whereas the transformational aspect was associated with experiential/emotional response to advertising (Puto & Wells, 1984). There was evidence that emotional response was more important than cognitive response in shaping consumer attitude and purchase intent (Morris et al., 2002). There was some attempt to investigate emotional response to political advertising with a discrete approach to emotion and verbal measures (Thorson et al., 1991); but there were quite a few theoretical problems with both the discrete approach and verbal measures (Kitayama & Markus, 1990; Lazarus, 1966; Mandler, 1962, 1984; Ortony & Turner, 1990). Dimensional approach (Averill, 1975; Bush, 1973; Osgood et al., 1957; Russell & Mehrabian, 1977; Schlosberg, 1952) and non-verbal measures (P. Lang, 1980; Morris, 1995) were a more promising alternative to assess emotional response (Morris et al., 2008). Specifically, AdSAM (Morris et al., 2002; Morris et al., 2005) is used as the measure of emotional response to advertising or the emotional attitude toward candidate. AdSAM was a graphic representation of three emotional dimensions, namely, pleasure, arousal, and dominance, identified in previous 61

PAGE 62

research (Morris 1995; Russell & Mehrabian, 1977). It was used in a number of studies on advertising effectiveness (Morris, 1995; Morris et al., 1993; Morris et al., 1992; Morris et al., 1998; Morris et al., 1996; Morris & Waine, 1993; Morris et al., 2002; Morris et al., 2005). The literature review leads to the formulation of research questions. Fundamentally, this study attempts to find out 1) how the perception of issue/image content of political ads, measured with the informational/transformational advertising scale, will influence memory of candidate issue stands and personal characteristics, 2) whether the relationship between the perception of issue/image content and memory of candidate issue stands and personal characteristics will vary across negative and positive ads, 3) how the perception of issue/image content will influence the impact of cognitive response, measured by memory, and of emotional response, measured by AdSAM on voters opinion on a candidates electability, and 4) whether the relationship between the perception of issue/image content and candidate electability will vary across negative and positive ads. Hypotheses and Research Questions The hypotheses and research questions focus on three aspects. The first aspect is about perception of advertising content. The second aspect is about memory. The third and last aspect is about candidate electability. Perception of Advertising Content According to the literature review of informational and transformational advertising, the concept of information and transformation was well applicable to issue and image (Shen & Kim, 2006). The scale of informational/transformational advertising (Puto & Wells, 1984) was used to calibrate the informational and transformational content of issue and image ads; and an image ad had a primary focus on transformation and a secondary focus on information, whereas an 62

PAGE 63

issue ad had a primary focus on information and a secondary focus on transformation (Shen, 2007). It is therefore proposed: Hypothesis 1: Based on previous research about the relationship between issue/image ads and informational/transformational advertising, the informational score of issue ads will be higher than the transformational score of issues ads at .05 significance level as measured with the informational/transformational advertising scale. Conversely, the transformational score of image ads will be higher than the informational score of image ads at .05 significance level as measured with the informational/transformational advertising scale. Memory The key question about memory is whether the characteristics of political ads, issue/image and negative/positive, influence the memory of ad content. Memory was previously measured with free recall and recognition test of verbal, visual, and audio cues in ads (A. Lang, 1991; Newhagen & Reeves, 1991; Schleuder, 1990) or cued recall of issue stands and candidate characteristics (Kaid et al., 1992). Cued recall of issue stands and candidate characteristics is the measure of memory in this study because memory of candidate issue stands and personal characteristics were a major component of the chain reaction model of advertising effectiveness (Lavidge & Steiner, 1961; Shama, 1976) adopted in the current study. There was strong evidence that issue ads outperformed image ads in generating higher number of recalled issues and personal characteristics (Kaid et al., 1992). Therefore, it is proposed: Hypothesis 2: Based on previous research about memory of issue and image ads, issue ads will be better remembered than image ads at .05 significance level as measured with cued recall of candidate issue stands and personal characteristics. In previous studies on memory of negative and positive ads, negative ads were found to be better remembered than positive ads (A. Lang, 1991; Newhagen & Reeves, 1991), but the results were based on recall and recognition test of verbal, visual, and audio cues in ads rather than cues about issue stands and personal characteristics. It remained unclear whether the valence of political ads, negative or positive, would influence the recall of candidate issue stands and 63

PAGE 64

personal characteristics. It also remained unclear whether there would be an interaction effect of issue/image by negative/positive on the recall of candidate issue stands and personal characteristics. Therefore, it is proposed: Research question 1: How will the valence of political ads, positive or negative, influence memory of ad content as measured with cued recall of candidate issue stands and personal characteristics? Research question 2: How will the valence of political ads, positive or negative, interact with and the emphasis of political ads, issue or image, on memory of ad content as measured with cued recall of candidate issue stands and personal characteristics? Candidate Electability Previous research did not provide consistent evidence as to how issue and image ads differed from each other in influencing voters opinion on a candidate or the candidate electability (Kahn & Geer, 1994; Thorson et al., 1991). Therefore, it is proposed: Research question 3: How will the emphasis of political ads, issue or image, influence candidate electability? Nevertheless, there was evidence that positive ads outperformed negatives ads in generating more positive candidate electability, and negative issue ads outperformed negative image ads in generating more positive candidate electability (Kahn & Geer, 1994). In the current study, it is proposed: Hypothesis 3: Based on previous research about candidate electability, positive ads will produce more positive electability toward candidate than negative ads at .05 significance level. Hypothesis 4: Based on previous research about candidate electability, negative issue ads will produce more positive cognitive electability toward candidate than negative image ads at .05 significance level. In this study, candidate electability is composed of issue competency (Kahn & Geer, 1994), candidate personality (Kahn & Geer, 1994; Thorson et al., 1991), and voting intention (Kahn & Geer, 1994). Furthermore, based on cognitive theories of persuasion (Cacioppo & 64

PAGE 65

Petty, 1989; Fishbein & Ajzen, 1975), the cognitive antecedent to electability is measured with message recall. The emotional antecedent to electability is measured with AdSAM(Morris et al., 2002; Morris et al., 2005). There was evidence that emotional response had a larger impact than cognitive response on the attitudinal construct of electability (Morris et al., 2002). It is therefore proposed: Hypothesis 5: Based on previous research about attitude, emotional response, measured with AdSAM, will have a larger impact on electability than cognitive response, measured with recall, at .05 significance level. There was some attempt to investigate emotional response to political advertising with a discrete approach to emotion and verbal measures (Thorson et al., 1991); but such an attempt was problematic and should be replaced by a dimensional approach and non-verbal measures such as AdSAM (Morris et al., 2008). The relationship between ad content, i.e., issue/image and negative/positive, and emotional response, measured with AdSAM, has never been addressed in previous research. The informational/transformational scale is used to measure voters perception of issue and image, and the perception provides an alternative way to explore the impact of ad content on attitude toward candidate. Path analysis is used to investigate the relationships of interest, and two more research questions are proposed: Research question 4: How will the perception of issue and image content, measured with the informational/transformational advertising scale, influence cognitive and emotional responses; and next how cognitive and emotional responses influence candidate electability? Research question 5: How will the relationship measured in Research Question 4 vary across negative and positive ads? 65

PAGE 66

Figure 2-1. Russells affect circumplex (Russell, 1980, Figure 1, p. 1164) 66

PAGE 67

CHAPTER 3 MATERIALS AND METHODS This study adopted a 2 (positive vs. negative) by 2 (issue vs. image) between-subjects design. There are four experimental conditions, i.e., positive image, positive issue, negative image, and negative issue. Table 3-1 presents more details about the conditions. Materials and Measures TV ads are chosen for the current study because 1) TV advertisements are particularly capable of functioning as a transformational medium (Puto & Wells, 1984) and 2) TV advertisements dominate political broadcasting in the U.S. (Kaid & Johnston, 2001). A content analysis is used to select the right TV political ads from a pool of congressional and senatorial campaign advertisements for the 2004 and 2006 elections for the current study. The coding scheme for the political ads was modified from Laskeys typology (Laskey, Day, & Crask, 1989) and used in a previous content analysis of political TV ads (Shen & Kim, 2006). All the variables described below are included in Appendix A. Independent Variables Perception of ad content The Scale (Puto & Wells, 1984, p. 641) was used to measure both the perception of informational cues (hereinafter referred to as PIFC) and the perception of transformational cues (hereinafter referred to as PTFC) in a political ad. The Scale has 23 items. Eight of them measure PIFC and the other 15 items measure PTFC. The Scale is modified for political ads. Brand is replaced with candidate and company with party. A mid-point category neither agree nor disagree is added to the Scale to form a 7-point scale, ranging from strongly disagree to strongly agree. Additionally, the negative ads include both the candidate and the opponent, therefore two informational items and one transformational item are asked twice to 67

PAGE 68

address issue and personality learning about both candidates. As a result, there are ten informational and 16 transformational items for the negative ads. More details about the items for negative ads are available in Appendix A. PIFC and PTFC also serve as manipulation checks of issue and image content. They are also exogenous variables in path analysis. Prior exposure to ad and familiarity with candidate Survey participants are asked to report their prior exposure to each ad on a three-point scale (1: never, 2: a few times, 3: many times), and familiarity with the brand/candidate on a five-point scale ranging from 1 (very unfamiliar) to 5 (very familiar). Prior exposure is used to filter out participants who may have watched the ads because previous exposure to ads is likely to bias processing of message content. Low familiarity is a reality for most candidates (Thorson et al., 1991), and this phenomenon can be checked with the familiarity measure. Dependent Variables Message recall Cued recall is used to measure memory with a two-item instrument employed in previous research (Kaid et al., 1992). Participants are asked to 1) describe candidate characteristics in adjective forms, and 2) state political issues associated with each candidate. One adjective is counted as one recalled message unit of candidate characteristics. One political issue is counted as one recalled message unit of political issues. Emotional response AdSAM is used to measure emotional candidate attitude. Participants will be asked to indicate their feeling of pleasure, arousal, and dominance on the three dimensions of the scale. The scale has been used in previous studies on attitude (Morris et al., 2002; Morris et al., 2005). 68

PAGE 69

Candidate electability Three verbal scale of candidate evaluation employed in previous research (Kahn & Geer, 1994) are used to measure candidate electability. The first scale is used to assess candidate competence in the issues mentioned in the ads and is composed of one seven-point item ranging from 1 (very incompetent) to 7 (very competent). The second scale is used to assess candidate characteristics in terms of trustworthy, strong leadership, hardworking, and knowledgeable and is measured with a 7-point Likert-type scale ranging from 1 (strongly disagree) to 7 (strongly agree). The third scale is voting intention and participants will be asked to indicate how likely they would vote for the candidate on a 5-point scale ranging from 1 (very unlikely) to 5 (very likely). Based on the tripartite model of attitude (McGuire, 1989), those three variables will be used to create a composite measure of candidate electability. Graded Response Model (GRM) of Item Response Theory (IRT) A Likert-type scale, i.e., strongly disagree strongly agree, is used in the current study to measure learning of informational and transformational cues from the ads. Traditionally, this type of scales is assumed to have equal distance between two consecutive scale categories and generate interval type data (Davis, 1997); therefore, computation of means, correlations, and other inferential statistics (e.g., reliability coefficients) is justified (Singh, 2004, p. 188). However, the interval data assumption of those scales is rarely tested in practice (Singh, 2004), and responses to scale categories such as strongly disagree strongly agree or to variations of this general theme can be assumed to involve categorical, rank-ordered data (p. 188). IRT is developed to better account for responses to Likert-type scales. With IRT, the responses to the scales are assumed to be categorical, rank-ordered observable data, and serve as the indicator of interval-type latent data; furthermore, a probability model connects the observable data to the latent data (Ewing, Salzberger, & Sinkovics, 2005; Singh, 2004; Singh et 69

PAGE 70

al., 1990). For example, each item of the informational/transformational advertising scale has seven categories, ranging from strongly disagree to strongly agree. A participants response to Item 1 is assumed to be an observable choice out of the seven available categories for Item 1, and this observable response is also assumed to be the indicator of the participants latent response to Item 1. The observable response is converted to the latent response through a probability model that takes into account both participant and scale item characteristics (Embretson & Reise, 2000). One of the most commonly used IRT model is a two-parameter logistic model (2PL model) that has an item discrimination parameter ( parameter) and an item difficulty parameter ( parameter) associated with a latent trait () (Hambleton & Swaminathan, 1985; Oishi, 2006). For the sth participant responding to the ith item, the model can be expressed as isiisiiisisXPexp1exp),,1( where Xis is a dichotomous response variable to the ith item with denoting correct response,is the sth participants score on the latent trait scale,is the ith items discrimination parameter,is the ith items difficulty parameter, and exp is the natural log base (2.718). The discrimination parameter indicates how well the ith item distinguishes between participants with low versus high standing on the latent trait scale (Oishi, 2006). The difficulty parameter indicates the score on the latent trait scale when the probability of responding to the ith item correctly is 50%; in other words, if an participant with a latent score of (1 standard deviation below the latent trait scales mean, which is standardized to be ) has a 50% probability to answer correctly to the ith item, then the difficulty parameter of the ith item is (Oishi, 2006). si i 70

PAGE 71

The obvious problem for the 2PL model to be used for survey responses is there is no right or wrong answer to a survey item. Additionally, survey items are often multiply ordered rather than following a dichotomous pattern. To address those concerns, the GRM was developed by converting the multiply ordered item categories into a set of binary response variables to fit the IRT framework (Samejima, 1969). When a K category scale is used for the ith survey item, the categories are scored xi = 0, 1,, mi, respectively, and mi = Ki 1. For example, a 3-point/category survey item is scored xi = 0, 1, 2. The participants responses to the ith item are converted in a binary manner so that those less than the xi category are recoded into and those equal to or greater than the xi category are recoded into As a result, a response of on the original 0 to 2 scale is recoded as for both the and categories but for the category. sixsP* is the probability for the sth participant to have the response to the xi category of the ith item conditional on the sth participants score on the latent trait scale (). Through the binary coding of responses in the scale categories, the 2PL model can used to specify the GRM and is expressed as ijsiijsisixsPexp1exp* x = j = 1, mi Note that response to the xi category for refers to responding in or above the xi category rather than responding in the xi category. For example, when x = 1, refers to the probability of responding in or above scale point 1 or in fact any categories above scale point 0. Similarly, when x = j = 2, refers to the probability of responding in or above scale point 2 or in any categories above scale point 1. Therefore, the difference between and sixsP* ssiP*1ssiP*1 ssiP*2 71

PAGE 72

ssiP*2*)(ssmiiP which can be computed by subtracting from is the probability of responding in scale point 1. Also note that in (2) x starts from 1 rather than 0 because, by definition, the probability of having the response to the lowest scale point/category xi = 0 or responding in or above scale point 0 is known as ssiP*2sssi)10.1 ssiP*1 1exp0 expexpisi*simsiP s = 1.0 siP*0 and the probability of responding in scale point 0 can be computed by subtracting from 1. Furthermore, the probability of having the response above the highest scale point/category xi = mi or responding in or above scale point mi + 1 is known as ssiP*1 = 0.0 miP*( simply because the mi + 1 scale point does not exist. As a result, the probability of responding in the highest scale point mi is the same as the probability of responding in or above mi or .0 Once the probability of the sth participants response (x) in or above a scale point/category is known, the probability of the sth participants response (x) to each scale point/category can be expressed as ssissiPP*10 )1()1(*)1(*exp1jisijisiijsijssxisixssixsPPP 0ssimiP x = j = 1, mi is still the sth participants the latent trait score; is the ith items discrimination parameter and it is constrained to be equal across all the response categories of the ith item. i 72

PAGE 73

However, has a very different meaning. For the K category scale scored from 0 to mi (mi = Ki 1), there are mi intervals between each two neighboring categories. In the IRT literature, the intervals are called thresholds. (j = 1 mi) is called threshold parameter and refers to the trait score necessary to respond above the jth threshold of the ith item with .50 probability. ij ij s s 2s With the above-mentioned 3-point item on the 0 to 2 scale, the probability of responding in each of the three scale points can be expressed as 11*10exp1exp0.10.1isiisissisiPP 2211*21*11exp1expexp1expisiisiisiisisssissiPPP 0exp1exp022*2isiisissisiPP and there are two threshold parameters, and estimated for threshold 1 between scale points 0 and 1 and threshold 2 between scales points 1 and 2, respectively. If is and is 1, then respondents whose latent trait score is below are more likely to respond in scale point 0, those whose score is are equally likely to respond in scale points 0 and 1 or threshold 1, those whose score is between and 1 are more likely to respond in scale point 1, those whose score is 1 are equally likely to respond in scale points 1 and 2 or threshold 2, and those whose score is above 1 are more likely to respond in scale point 2. 1i 2i 1i 2i Essentially, the observed responses are viewed as probabilities of agreeing with an item (Hambleton & Swaminathan, 1985; Linden & Hambleton, 1996; Oishi, 2006). Such a probabilistic approach is in line with the cognitive process that an individuals response [to questionnaire items] is not based solely on his/her true standing on the latent variable but is also affected by other cognitive processes includ[ing] distracting information and events, 73

PAGE 74

mood swings and processes, social desirability and other biases, and random memory processes (Singh, 2004, p. 196). Because of this measurement advantage, IRT is used in this study to estimate latent responses to the Scale. Next, the latent responses are used to test the hypotheses and research questions about learning of informational and transformational cues in this study. 74

PAGE 75

Table 3-1. Experimental conditions Condition Ad 1 Positive Image 2 Positive Issue 3 Negative Image 4 Negative Issue 75

PAGE 76

CHAPTER 4 RESULTS Pretest A pretest was conducted to select one typical positive image advertisement, one typical positive issue advertisement, one typical negative image advertisement, and one typical negative issue advertisement. The items in the pretest are included in Appendix A. Forty-four undergraduate students in an advertising strategy class were asked to indicate how much political issue and personal qualities information they learned from five pre-selected advertisements on a seven-point scale ranging from Very Little to A Great Deal. Table 4-1 provided the scores of political issue and personal qualities information. Based on the results, the typical positive image advertisement was an advertisement of U.S. Congressman Mike Pence of Indiana focusing on the candidates personal charisma. For the main study, this positive image advertisement was edited to create a Chris Chocola (hereinafter referred to as the Candidate) positive image advertisement. Specifically, the visuals in the Mike Pence advertisement were replaced with those of the Candidate, but most of the audio about candidate personality was retained. The typical positive issue advertisement was an advertisement of the Candidate focusing on his views on small business. The typical negative image advertisement was an advertisement of the Candidate attacking his opponent Joe Donnelly (hereinafter referred to as the Opponent) on his poor tax paying record and character. The typical negative issue advertisement was an advertisement of the Candidate attacking the Opponent on tax policies. Main Study: Model Estimation and Results A structural equation modeling (SEM) procedure, Lisrel (Joeskog & Sobom, 2005), was used to test the proposed model. Several assumptions of SEM, i.e., independence, normality, 76

PAGE 77

linearity (Hair, Anderson, Tatham, & Black, 1995), as well as equality of variance-covariance matrices across the four experimental conditions, were checked. Except for those variables noted below, there was no significant violation of the assumptions. Specifically, bivariate scatterplots of the dependent variables by their studentized residuals were used to check the independence assumption, and there were no systematic patterns that would indicate the violation of that assumption. The dependent variables normal probability plots and their skewness and kurtosis scores were used to check the normality assumption, and there was no significant deviation from normality. Barletts test of sphericity was used to check the linearity assumption, and there was no significant deviation from linearity. Both univariate tests, i.e., Cochrans C, Bartett-Box, and Levene test, and a multivariate test, i.e., Boxs M, were used to check the equal variance-covariance assumption, and there was no significant deviation from the assumption. Descriptive Statistics of Independent Variables Prior exposure to ad and familiarity with candidate A total of 1087 participants were recruited. Eight-seven participants had watched at least one of the four advertisements before, and their responses were excluded from the study. The 1000 participants retained in the study were very unfamiliar with both candidates. Table 4-2 provided means and standard deviations of familiarity with the Candidate and the Opponent. Perception of ad content The original informational/transformational scale (Puto & Wells, 1984) has eight informational items and 15 transformational items. The same items were used to measure positive issue and image ads in a previous study (Shen, 2007). However, because the negative ads included both the Candidate and the Opponent, two informational items and one transformational item were asked twice to address issue and personality learning about both 77

PAGE 78

candidates. As a result, there were ten informational and 16 transformational items for the negative ads. More details about the items for negative ads are available in Appendix A. Originally there were nine categories on the scale, but there were missing responses to the low and high ends ( and ) on the scale. Reducing the number of categories is often used to solve this problem (Ewing, Salzberger, & Sinkovics, 2005; Singh, 2004). In this study, the nine-point scale was recoded into a seven-point scale. For the two positive advertisements, the perception of informational cues (PIFC) was the average of the eight informational items, and the perception of transformational cues (PTFC) was the average of the 15 transformational items. For the two negative advertisements, PIFC was the average of the ten informational items, and the PTFC was the average of the 16 transformational items. Table 4-3 reported the scores of the four advertisements. The PIFC scores of the two image advertisements were significantly lower than the PTFC scores. Conversely, the PIFC scores of the two issue advertisements were significantly higher than the PTFC scores. The results not only confirmed the validity of the coding scheme to identify informational and transformational advertisements and proved the experimental manipulation successful, but also supported Hypothesis 1. Although the statistical significance might be inflated because of the large sample size, the PIFC scores were above the scale mid-point (4 on a 7-point scale) for the issue advertisements and below the mid-point for the image advertisements. The PTFC scores were above the mid-point for the image advertisements and below the mid-point for the issue advertisements. One caveat of the results is noteworthy. Although PIFC and PTFC can be measured separately, they are not mutually exclusive (Puto & Wells, 1984). Voters can learn transformational cues from an issue political advertisement (Johnston & Kaid, 2002), for 78

PAGE 79

example, a down-to-earth personality of the Candidate owing to his views on small businesses. Voters can also learn informational cues from an image political advertisement (Christ et al., 1994; Johnston & Kaid, 2002), for example, the Candidate representing Indiana in the U.S. Congress. An informational advertisement in fact has a primary focus on information and a secondary focus on transformation, and a transformational advertisement has a primary focus on transformation and a secondary focus on information (Puto & Wells, 1984). Table 4-4 reported the reliability coefficients in the four advertisement conditions. The informational coefficients were a little bit higher than those reported in the validation study (Puto & Wells, 1984) whereas the transformational coefficients were quite close to those reported in that study. The informational coefficients were smaller than the transformational coefficients. But the smaller coefficients do not necessarily mean that the informational items are less reliable than the transformational items but indicate that Cronbachs reliability test favors larger numbers of scale items (Hair et al., 1995). Descriptive Statistics of Dependent Variables Message recall Participants were asked to 1) describe candidate characteristics in adjective forms, and 2) state political issues associated with each candidate. Each characteristic adjective or issue was counted as one unit of image or issue recall. Table 4-5 provided means and standard deviations of image recall, issue recall, and total recall of both image and issue. Because the negative advertisements had two candidates whereas the positive advertisements had one candidate only, the image and issue recall in the two negative advertisement conditions were in fact the mean of the recall of the two candidates whereas the two types of recall in the two positive advertisement condition were the exact number of recall of the featured candidate. Also as an effort to make 79

PAGE 80

total recall more comparable across the conditions, the total recall was mean of both image and issue recall. MANOVA was used to investigate message recall between advertisement valence (positive vs. negative) and advertisement emphasis (image vs. issue). Both advertisement valence and emphasis had a significant direct effect on image recall FValence(1, 996) = 173.35, p < .05, partial eta squared = .15; FEmphasis(1, 996) = 12.48, p < .05, partial eta squared = .01; issue recall FValence(1, 996) = 135.21, p < .05, partial eta squared = .12; FEmphasis(1, 996) = 253.72, p < .05, partial eta squared = .20; and total recall FValence(1, 996) = 254.65, p < .05, partial eta squared = .20; FEmphasis(1, 996) = 29.76, p < .05, partial eta squared = .03. There was also a significant valence by type interaction effect on image recall F(1, 996) = 8.77, p < .05, partial eta squared = .01; issue recall F(1, 996) = 288.67, p < .05, partial eta squared = .23; and total recall F(1, 996) = 119.62, p < .05, partial eta squared = .11. Based on the results, the issue advertisements generated more issue recall and total recall than the image advertisement; therefore, Hypothesis 2 was supported. Additionally, the negative advertisements generated more image, issue, and total recall than the positive advertisement; therefore, the answer to Research Question 1 was that negative advertisements were better remembered than positive advertisement. Meanwhile, the valence interacted with the emphasis on recall; therefore, the answer to Research Question 2 was that the interaction effect did exist. Candidate electability Participants were asked to indicate their opinion on three verbal scale of candidate electability employed in previous research (Kahn & Geer, 1994). The first scale was used to assess candidate competence in tax, the issue mentioned in the ads, on a seven-point item ranging from 1 (very incompetent) to 7 (very competent). Table 4-6 provided comparisons of the two candidates in the two negative advertisement conditions. The Candidate was 80

PAGE 81

significantly more competent than the Opponent in the negative image condition, whereas there was no difference in the negative issue condition. ANOVA was used to investigate the Candidates tax competency between advertisement valence (positive vs. negative) and advertisement emphasis (image vs. issue). The results indicated that valence did not have a significant direct impact on issue competency FValence(1, 996) = .06, p > .05, partial eta squared = .00; but emphasis had a significant direct impact on issue competency FEmphasis(1, 996) = 33.82, p < .05, partial eta squared = .03. Additionally, valence and emphasis had a significant interaction effect on issue competency F(1, 996) = 111.85, p < .05, partial eta squared = .10. So the Candidate was perceived as more competent in issue advertisements than in image advertisements, and he was perceived as the most competent in the positive issue condition but the least competent in the positive image condition. Furthermore, t-tests about issue competency of the same candidate between the two negative ad conditions indicated that the Candidate was perceived as more competent in the negative image condition than the negative issue condition t(498) = 3.09, p < .05, whereas the Opponent was perceived as more competent in the negative issue condition than the negative image condition t(498) = 8.15, p < .05. The second scale was used to assess candidate characteristics in terms of trustworthy, strong leadership, hardworking, and knowledgeable and is measured with a 7-point Likert-type scale ranging from 1 (strongly disagree) to 7 (strongly agree). Table 4-7 provided comparisons of the two candidates in the two negative advertisement conditions. The Candidate was preferred over the Opponent in the negative image condition, but the Opponent was preferred over the Candidate in the negative issue condition. ANOVA was used to investigate the Candidates characteristics between advertisement valence (positive vs. negative) and advertisement emphasis (image vs. issue). The results indicated that emphasis did not have a 81

PAGE 82

significant direct impact on candidate characteristics FEmphasis(1, 996) = 2.48, p > .05, partial eta squared = .00; but valence had a significant direct impact on candidate characteristics FValence(1, 996) = 226.96, p < .05, partial eta squared = .19. Additionally, valence and emphasis had a significant interaction effect on issue competency F(1, 996) = 82.43, p < .05, partial eta squared = .08. So the Candidate was more likable in the positive advertisement conditions than the negative advertisement conditions, and he was the most likable in the positive issue condition but the least likable in the negative issue condition. Furthermore, t-tests about candidate characteristics of the same candidate between the two negative ad conditions indicated that the Candidate was more likable in the negative image condition than the negative issue condition t(498) = 4.99, p < .05, whereas the Opponent was more likable in the negative issue condition than the negative image condition t(498) = 4.10, p < .05. Table 4-8 reported the reliability coefficients of the candidate characteristics scale in the four advertisement conditions. Table 4-8 reported the reliability coefficients in the four advertisement conditions. The third scale was used to assess voting intention on a one-item five-point scale ranging from 1 (very unlikely) to 5 (very likely). Table 4-9 provided descriptive statistics and comparisons of voting intention. The voting intention in all conditions was below the scale mid-point, which indicated a low intention to vote for either candidate. The two candidates did not differ from each other in voting intention in the two negative conditions. ANOVA was used to investigate voting intention for the Candidate between advertisement valence (positive vs. negative) and advertisement emphasis (image vs. issue). The results indicated that emphasis did not have a significant direct impact on voting intention FEmphasis(1, 996) = 1.17, p > .05, partial eta squared = .00; but valence had a significant direct impact on voting intention FValence(1, 996) = 15.10, p < .05, partial eta squared = .02. Additionally, valence and emphasis had a significant 82

PAGE 83

interaction effect on voting intention F(1, 996) = 78.39, p < .05, partial eta squared = .07. Furthermore, t-tests about voting intention between the two negative ad conditions indicated that voting intention for the Candidate was higher in the negative image condition than in the negative issue condition t(498) = 5.08, p < .05, whereas that for the Opponent indicated no difference between the negative issue and negative image conditions t(498) = .48, p > .05. The answer to Research Question 3 varied across the three scales. When electability was measured in terms of issue competence, issue ads generated more favorable candidate attitude than image ads. When electability was measured in terms of candidate characteristics and voting intention, there was no difference between issue and image ads. By contrast, Hypothesis 3 was not supported when electability was measured in terms of issue competence, but was supported when the same variable was measured in terms of candidate characteristics and voting intention. Finally, the t-tests on the Candidate did not support Hypothesis 4, but those on the Opponent supported Hypothesis 5 in terms of issue competence and candidate characteristics. Pleasure-arousal-dominance Participants were asked to indicate their emotional response to the candidate(s) on the nine-point AdSAM scale in term of pleasure, arousal and dominance. Table 4-10 provided descriptive statistics and comparisons of the pleasure dimension. The pleasure scores were above the scale mid-point in the two positive conditions, and below the scale mid-point in the two negative conditions. The Candidate received higher pleasure score than the Opponent in the two negative conditions. ANOVA was used to investigate pleasure response to the Candidate between advertisement valence (positive vs. negative) and advertisement emphasis (image vs. issue). Both advertisement valence and emphasis had a significant direct effect on pleasure FValence(1, 996) = 172.84, p < .05, partial eta squared = .15; FEmphasis(1, 996) = 6.30, p < .05, partial eta squared = .01. There was also a significant valence by emphasis interaction effect on 83

PAGE 84

pleasure F(1, 996) = 11.75, p < .05, partial eta squared = .01. The positive issue ad generated the highest pleasure response, whereas the negative issue ad generated the lowest pleasure response. Furthermore, there was no difference between the pleasure response to either the Candidate or the Opponent in the two negative conditions with t-tests, t(498)Candidate = .61, p > .05; t(498)Opponent = -.84, p > .05. Table 4-11 provided descriptive statistics and comparisons of the arousal dimension. All the arousal scores were below the scale mid-point. The Candidate received lower arousal score than the Opponent in the negative image condition, but there was no difference between the two candidates in the negative issue condition. ANOVA was used to investigate arousal response to the Candidate between advertisement valence (positive vs. negative) and advertisement emphasis (image vs. issue). Advertisement valence had no significant effect on arousal FValence(1, 996) = .04, p > .05, partial eta squared = .00; but emphasis had a significant effect on arousal FEmphasis(1, 996) = 17.54, p < .05, partial eta squared = .02. There was also a significant valence by emphasis interaction effect on arousal F(1, 996) = 45.04, p < .05, partial eta squared = .04. The positive issue ad generated the highest arousal response, whereas the positive image ad generated the lowest arousal response. Furthermore, t-tests about arousal response to the same candidate between the two negative ad conditions indicated that there was no difference for the Candidate in the two conditions t(498) = 1.70, p > .05, but there was stronger arousal response to the Opponent in the negative image condition than the negative issue condition t(498) = 3.31, p < .05. Table 4-12 provided descriptive statistics and comparisons of the dominance dimension. The dominance scores were above or close to the scale mid-point. There was no difference between the two candidates in the two negative conditions. ANOVA was used to investigate 84

PAGE 85

dominance response to the Candidate between advertisement valence (positive vs. negative) and advertisement emphasis (image vs. issue). Advertisement valence had a significant effect on dominance FValence(1, 996) = 8.44, p < .05, partial eta squared = .01; but emphasis had no significant effect on dominance FEmphasis(1, 996) = .65, p > .05, partial eta squared = .00. There was also a significant valence by type interaction effect on dominance F(1, 996) = 14.09, p < .05, partial eta squared = .01. The positive image ad generated the highest dominance response, whereas the negative image ad generated the lowest dominance response. Furthermore, t-tests about arousal response to the same candidate between the two negative ad conditions indicated that there was stronger dominance response to the Candidate in the negative issue condition than the negative image condition t(498) = 1.98, p < .05, but there was no difference for the Opponent between the two negative conditions t(498) = 1.17, p > .05. Model Estimation The proposed model for this study was specified in Figure 4-1. The error variance of PIFC and PTFC was correlated because the two constructs overlap conceptually (Puto & Wells, 1984). Item response theory was used to estimate the latent scores from the observed scores of PIFC and PTFC, and the latent scores were placed in SEM. Correlation matrices were included in Appendix B. Figure 4-2 specified the measurement model for candidate electability (hereinafter referred to as CE) and emotional response (hereinafter referred to as ER.) CE was indicated by issue competence, personal characteristics, and voting intention. Issue competence was measured with a one-item scale, and its error variance was set to be equal to zero. Personal characteristics was measured with a four-item scale, and its error variance was set to be equal to subtracting the scale reliability from 1. Voting intention was measured with a one-item scale, and its error 85

PAGE 86

variance was set to be equal to zero. Emotional response was indicated by pleasure, arousal, and dominance measured with AdSAM (Morris, 1995). Positive image condition Figure 4-3 reported factor loadings on CE and ER. Table 4-13 reported goodness of fit statistics. GFI (Goodness of Fit Index), NFI (Normed Fit Index), CFI (Comparative Fit Index), RMSEA (Root Mean Square Error of Approximation), and SRMR (Standardized Root Mean Square Residual) were chosen due to their popularity (Hair et al., 1995). The model had a good fit in this condition Figure 4-4 reported standardized coefficients of the model. Based on the results, both PIFC and PTFC had a significant effect on CE. PTFC also had significant effects on MR and ER. Additionally, ER had a significant effect on CE, but MR had an insignificant effect on CE. Therefore, Hypothesis 5 was supported in the positive image condition. Positive issue condition Figure 4-5 reported factor loadings on CE and ER. Table 4-13 reported goodness of fit statistics. The model had a good fit. Figures 4-6 reported standardized coefficients of the model. Based on the results, both PIFC and PTFC had significant effects on CE. PTFC had a significant effect on ER, but PIFC had a significant effect on MR. Additionally, both MR and ER had significant effects on CE, but ER appeared to have a stronger impact than MR. Therefore, Hypothesis 6 was supported in the positive issue condition. Negative image condition for the Candidate Figure 4-7 reported factor loadings on CE and ER. Table 4-13 reported goodness of fit statistics. The model had a good fit. 86

PAGE 87

Figure 4-8 reported standardized coefficients of the path model. Both PIFC and PTFC had significant effects on CE. PIFC also had a significant effect on both MR and ER. PTFC had a significant effect on ER. ER had a significant effect on CE, but MR had an insignificant effect on CE. Therefore, Hypothesis 5 was supported in the negative image condition for the Candidate. Negative image condition for the Opponent Figure 4-9 reported factor loadings on CE and ER. Table 4-13 reported goodness of fit statistics. The model had a good fit. Figures 4-10 reported standardized coefficients of the path model. PIFC had a significant and negative effect on ER. Both ER and MR had a significant impact on CE, but ER appeared to have a stronger effect on CE. Therefore, Hypothesis 5 was supported in the negative image condition for the Opponent. Negative issue condition for the Candidate Figure 4-11 reported factor loadings on CE and ER. Table 4-13 reported goodness of fit statistics. The model had a good fit. Figures 4-12 reported standardized coefficients of the path model. Both PIFC and PTFC had a significant effect on CE, and PTFC had significant effects on both MR and ER. PIFC had a significant effect on MR. Neither MR nor ER had a significant impact on CE. Therefore, Hypothesis 5 was not supported in the negative issue condition for the Candidate. Negative issue condition for the Opponent Figure 4-13 reported factor loadings on CE and ER. Table 4-13 reported goodness of fit statistics. The model had a good fit. Figures 4-14 reported standardized coefficients of the path model. Modification indices suggested a causal arrow from CE to ER, which was opposite to the arrow in the other 87

PAGE 88

conditions. In this condition, PTFC had a significant and negative effect on ER but CE had a significant and positive effect on ER. The data suggested a different effect hierarchy (Vakratsas & Ambler, 1999) and Hypothesis 5 was not supported in the negative issue condition for the Opponent. 88

PAGE 89

PIFC PTFC ER MR CE Figure 4-1. Full path diagram of PIFC, PTFC, MR (message reall), ER (emotional response), and CE (candidate electability) ER CE PC P A D VI IC Figure 4-2. Measurement model for CE indicated with IC (issue competence), PC (personal characteristics), and VI (voting intention) and ER indicated with P (pleasure), A (arousal), and D (dominance) 89

PAGE 90

ER CE PC P A D VI IC .50* .89* .66* .77* .61* .59* Figure 4-3. Factor loadings in positive image condition p < .05 PIFC PTFC ER MR -.03 .63* .29* .34* .16* .08 .24* .56* CE .06 Figure 4-4. Path coefficients in positive image condition p < .05 ER CE PC P A D VI IC .59* .76* .64* .83* .71* .21* Figure 4-5. Factor loadings in positive issue condition p < .05 90

PAGE 91

PIFC PTFC ER MR .19* .55* .21* .22* .05 .16* .32* .67* CE .11 Figure 4-6. Path coefficients in positive issue condition p < .05 ER CE PC P A D VI IC .47* .88* .64* .69* .34* .37* Figure 4-7. Factor loadings in negative image condition for the candidate p < .05 91

PAGE 92

PIFC PTFC ER MR .17* .28* .22* .25* .00 .05 .42* .60* CE .21* Figure 4-8. Path coefficients in negative image condition for the candidate p < .05 ER CE PC P A D VI IC .59* .65* .48* .51* .07 .42* Figure 4-9. Factor loadings in negative image condition for the opponent p < .05 92

PAGE 93

PIFC PTFC ER MR .15 .19 .09 -.11 .07 -.18* .60* .60* CE -.54* Figure 4-10. Path coefficients in negative image condition for the opponent p < .05 ER CE PC P A D VI IC .58* .94* .64* .78* .29* .33* Figure 4-11. Factor loadings in negative issue condition for the candidate p < .05 93

PAGE 94

PIFC PTFC ER MR .32* .42* .27* .29* -.21* -.08 .16 .62* CE .14 Figure 4-12. Path coefficients in negative issue condition for the candidate p < .05 ER CE PC P A D ET IC .40* .93* .52* .50* .36* .57* Figure 4-13. Factor loadings in negative issue condition for the opponent p < .05 94

PAGE 95

PIFC PTFC ER MR -.05 -.29* .10 -.01 .09 -.06 .18* .64* CE .08 Figure 4-14. Path coefficients in negative issue condition for the opponent p < .05 95

PAGE 96

Table 4-1. Personal qualities and political issues Personal qualities Political issues Paired samples test Mean SD Mean SD t (df = 43) Positive image 5.61 .97 1.39 .69 24.20* Positive issue 2.09 1.14 5.73 .87 -17.24* Negative image candidate 4.30 1.58 1.41 .79 11.74* Negative image opponent 5.20 1.37 1.84 1.29 13.69* Negative issue candidate 1.57 .97 2.27 1.48 -3.63* Negative issue opponent 2.41 1.53 5.59 1.23 -11.67* p < .05 Table 4-2. Familiarity with candidates Candidate Opponent Mean SD Mean SD Positive image 1.10 .37 Positive issue 1.05 .22 Negative image 1.17 .44 1.22 .48 Negative issue 1.12 .33 1.19 .40 Table 4-3. Information and transformation Information Transformation Paired Samples Test Mean SD Mean SD t (df = 249) Positive image 3.41 1.09 4.20 1.14 -10.70* Positive issue 4.46 1.20 3.12 1.10 21.19* Negative image 3.06 .99 4.05 1.09 -18.26* Negative issue 4.33 1.11 3.10 1.07 16.98* p < .05 Table 4-4. Reliability coefficients of informational/transformational scale Information Transformation Positive image .81 .93 Positive issue .81 .91 Negative image .78 .87 Negative issue .79 .89 96

PAGE 97

Table 4-5. Message recall Image recall Issue recall Total recall Mean SD Mean SD Mean SD Positive image 1.83 1.03 .05 .21 1.88 1.11 Positive issue 1.22 1.21 1.48 1.05 2.70 1.48 Negative image 3.08 2.13 .52 .81 3.60 2.03 Negative issue 2.28 2.39 2.23 2.81 4.52 4.09 Table 4-6. Candidate issue competence Candidate Opponent Paired samples test Mean S D Mean SD t (df = 249) Positive image 3.98 .79 Positive issue 4.85 1.14 Negative image 4.71 1.15 3.08 1.17 17.07* Negative issue 4.02 1.21 4.15 1.49 -1.11 p < .05 Table 4-7. Candidate characteristics Candidate Opponent Paired samples test Mean SD Mean SD t (df = 249) Positive image 4.33 .76 Positive issue 4.70 1.20 Negative image 3.63 1.03 3.31 .79 4.79* Negative issue 3.22 1.09 3.77 1.06 -6.98* p < .05 Table 4-8. Reliability coefficients of candidate characteristics scale Candidate Opponent Positive image .76 Positive issue .93 Negative image .85 .83 Negative issue .87 .93 97

PAGE 98

Table 4-9. Voting intention Candidate Opponent Paired samples test Mean SD Mean SD t (df = 249) Positive image 2.49 .72 Positive issue 2.98 .84 Negative image 2.44 .90 2.36 .80 1.27 Negative issue 2.04 .87 2.40 .98 -4.75* p < .05 Table 4-10. Pleasure Candidate Opponent Paired samples test Mean SD Mean SD t (df = 249) Positive image 5.54 1.31 Positive issue 5.98 .78 Negative image 4.82 1.34 3.75 1.48 8.94* Negative issue 4.75 1.17 3.85 1.37 8.40* p < .05 Table 4-11. Arousal Candidate Opponent Paired samples test Mean SD Mean SD t (df = 249) Positive image 3.71 1.91 Positive issue 4.85 1.12 Negative image 4.39 1.75 4.68 1.87 -2.76* Negative issue 4.13 1.72 4.15 1.69 -.18 p < .05 Table 4-12. Dominance Candidate Opponent Paired samples test Mean SD Mean SD t (df = 249) Positive image 5.65 1.80 Positive issue 5.20 1.07 Negative image 4.99 1.70 4.93 1.72 .58 Negative issue 5.28 1.59 5.12 1.72 1.71 98

PAGE 99

Table 4-13. Fit indices in experimental conditions 2 df GFI NFI CFI RMSEA SRMR Positive image 66.69 22 .94 .94 .95 .10 .05 Positive issue 79.85 22 .94 .93 .94 .10 .07 Negative image candidate 32.33 22 .97 .95 .98 .04 .05 Negative image opponent 64.20 22 .95 .82 .87 .09 .07 Negative issue candidate 52.62 22 .96 .93 .96 .07 .05 Negative issue opponent 51.76 22 .95 .84 .89 .08 .07 99

PAGE 100

CHAPTER 5 DISCUSSION Overview This study was designed to investigate the effects of issue/image and positive/negative political advertisements on message processing and voting behavior with a new measurement tool of message learning, i.e. the informational/transformational scale (Puto & Wells, 1984). The constructs in this study were selected based on a chain reaction model of advertising effects (Lavidge & Steiner, 1961) and the relationships between them were analyzed with t-Tests, ANOVA, MANOVA, and structural equation modeling. This study generated very interesting results regarding the mechanism through which different types of political advertisements were processed. In the following sections, findings related to the hypotheses and research questions are presented. Perception of Advertising Content Hypothesis 1 proposed that issue ads are perceived as more informational than transformational whereas image ads are perceived as more transformational than informational. This was supported, as the PIFC and PTFC scores was significantly different for each ad and the difference was in line with the primary content focus of the ad. Both the positive and negative image ads were perceived as more transformational than informational, and the transformational scores were above the scale mid-point whereas the informational scores were below the scale mid-point. Both the positive and negative issue ads were perceived as more informational than transformational, and the informational scores were above the scale mid-point whereas the transformational scores were below the scale mid-point. The findings confirmed the conceptual similarity between information/transformation and issue/image advocated in previous studies (Shen, 2007; Shen & Kim, 2006). 100

PAGE 101

Memory Hypothesis 2 proposed that issue ads are bettered remembered than image ads. This was supported, as the F-test statistic for emphasis (image vs. issue) was significant and the total recall of candidate characteristics and candidate issue stands in the two issue ad conditions was higher than that in the two image ad conditions. The result confirmed previous findings about the superiority of issue ads in generating higher recall than image ads (Kaid et al., 1992). Research Question 1 asked how valence (positive vs. negative) influences recall. The F-test statistic for valence was also significant and the total recall of candidate characteristics and candidate issue stands in the two negative ad conditions was higher than that in the two positive ad conditions. Therefore, the answer was that negative ads were better remembered than positive ads. Research Question 2 asked how valence interacts with emphasis on recall. The F-test statistic for the interaction term was not significant. Therefore, the answer was that valence did not interact with emphasis on recall. The results about recall should be interpreted with caution. One possible methodological flaw in this study was that it remained unchecked whether the four ads had the same amount of message units for recall. Issue ads might be better recalled than image ads because they had more message units than image ads. Negative ads might be better recalled than positive ads because they had more candidates (the Candidate and the Opponent) than positive ads (the Candidate only). It is highly worthwhile to calibrate the message units and keep the amount of the units constant across experimental conditions in future research on recall of political ads. 101

PAGE 102

Candidate Electability Issue, Characteristics, and Voting Intention Candidate electability was measured with issue competence, candidate characteristics, and voting intention. The results did not provide consistent answer or support to Research Question 3 and Hypotheses 3 and 4. In fact, the results largely depended on the focus of each measurement and the status of the candidates. For example, in terms of issue competence, issue ads generated more favorable candidate attitude than image ads, but positive ads did not differ from negative ads. In terms of candidate characteristics and voting intention, positive ads generated more favorable candidate attitude than negative ads, but there was no difference between issue and image ads. Additionally, the effects of two negative ads varied between the two candidates. For the Candidate, the negative issue ad led to less favorable candidate electability than the negative image ad. By sharp contrast, for the Opponent, the negative issue ad led to more favorable candidate electabililty than the negative image ad. The negative issue ad backfired on the Candidate in terms of candidate characteristics because the Candidate was less likable than the Opponent in that condition, which was inconsistent with previous findings about the effectiveness of negative issue ads to overcome the backlash syndrome (Johnson-Cartee & Copeland, 1989; Kahn & Geer, 1994; Sonner, 1998). In future research, it is highly worthwhile to reexamine the relationship between ad content and the backlash effect. Cognition and Emotion Hypothesis 5 proposed that emotional response has a larger impact on candidate electability than cognitive response measured with message recall. This was analyzed through path analysis, and the hypothesis was supported in most conditions except the negative issue condition. 102

PAGE 103

In the positive image condition, emotional response had a significant effect on candidate electability whereas cognitive response had an insignificant effect on candidate electability. In the positive issue condition, both emotional response and cognitive response had a significant positive impact on candidate electability, but the emotional response appeared to have a stronger effect on candidate electability. In the negative image for the Candidate condition, emotional response had a significant effect on candidate electability whereas cognitive response had an insignificant effect on candidate electability. In the negative image for the Opponent condition, both emotional response and cognitive response had a significant positive impact on candidate electability, but the emotional response appeared to have a stronger effect on candidate electability. In the negative issue for the Candidate condition, both emotional and cognitive responses had an insignificant effect on candidate electability. In the negative issue for the Opponent condition, again both emotional and cognitive responses had an insignificant effect on candidate electability. In fact, in that condition, there was a significant effect from candidate electability to emotional response. The results in the positive ad condition further confirmed the conceptual similarity among image, transformation and emotion. Image is analogous to transformation (Shen, 2007) and transformation focuses on the emotional aspect of advertising (Puto & Wells, 1984); therefore, it is not surprising that emotional response was much more powerful than cognitive response in determining candidate electability in the positive image condition. Similarly, issue is analogous to information (Shen, 2007) and information focuses on the cognitive aspect of advertising (Puto & Wells, 1984); therefore, cognitive response gained significance in the positive issue condition. 103

PAGE 104

The results in the negative ad condition are also very interesting. In the negative image for the Candidate condition, the effects of emotional and cognitive response were identical to those in the positive issue condition. In the negative image for the Opponent condition, the effects were identical to those in the positive image condition. A close look into the message recall for this condition indicated that the participants inferred unmentioned issues, such as campaign finance and tax, from the negative image ad. It was likely that those implicitly inferred issues influenced the effects of emotional and cognitive responses in a way similar to the explicitly mentioned tax issue in the positive issue condition. Furthermore, the insignificant effects of both emotional and cognitive response in the negative issue for the Candidate condition may be due to limited message about the candidate. The only message about the Candidate was a still image of the Candidate and a voice-over Im Chris Choloca. I approve this message. The opposite direct effect from candidate electability to emotional response in the negative issue for the Opponent condition might mean that the participants had some existing attitude about candidates who advocated higher tax, and this existing attitude influenced emotional response to the Opponent who also supported higher tax. Memory, Emotion, and Candidate Electability Research Questions 4 and 5 investigated the complete chain reaction process from message learning, message recall, emotion, and electability in all the experimental conditions. This was also analyzed through path analysis. In the positive image condition, PTFC had a significant impact on message recall whereas PIFC had an insignificant impact on message recall. This pattern made good sense because image and transformational cues were the primary content of the positive image ad. PTFC also had a significant impact on emotional response whereas PIFC still had an insignificant impact on emotional response. The relationship between PTFC and emotional response was well 104

PAGE 105

understandable because transformational cues were the emotional message of an ad (Puto & Wells, 1984). Additionally, both PIFC and PTFC had a significant and direct impact on candidate electability. In the positive issue condition, PTFC had an insignificant impact on message recall whereas PIFC had a significant impact on message recall. This pattern made good sense because issue and informational cues were the primary content of the positive issue ad. Both PTFC and PIFC had a significant impact on emotional response. In other words, besides the conceptually determined relationship between PTFC and emotional response (Puto & Wells, 1984), PIFC, the primary content in the positive issue ad, served as another important source of emotional response. Additionally, both PIFC and PTFC had a significant and direct impact on candidate electability. In the negative image for the Candidate condition, PTFC had an insignificant impact on message recall whereas PIFC had a significant impact on message recall. Both PTFC and PIFC had a significant impact on emotional response. The patterns in this condition were opposite to those in the positive image condition but were identical to that in the positive issue condition for the same candidate. There might be inferred issue learning, such as campaign finance and tax, from the negative image condition, which led to issue-induced effects as recorded in the positive issue condition. Again, both PIFC and PTFC had a significant and direct impact on candidate electability in this condition. In the negative image for the Opponent condition, neither PTFC nor PIFC had a significant impact on message recall. PTFC had a significant impact on emotional response, which could be explained with the conceptual similarity between transformation and emotion (Puto & Wells, 1984). PIFC also had a significant impact on emotional response, which was also found in the 105

PAGE 106

positive issue condition. It was likely that the inferred issue learning became a significant source of emotional response. The impact from PIFC to emotional response was negative, which was in line with the valence of the ad. Finally, neither PIFC nor PTFC had a significant and direct impact on candidate electability in this condition. In the negative issue for the Candidate condition, both PIFC and PTFC had a significant impact on message recall, and the impact from PTFC was negative. PTFC also had a significant impact on emotional response, whereas PIFC had an insignificant impact on emotional response. The patterns in this condition seemed to suggest that transformational cues or the secondary content in this condition had a stronger impact on message recall and emotional response than informational cues or the primary content. Finally, both PIFC and PTFC had a significant and direct impact on candidate electability in this condition. In the negative issue for the Opponent condition, neither PTFC nor PIFC had a significant impact on message recall. PTFC had a significant impact on emotional response, whereas PIFC had an insignificant impact on emotional response. Neither PIFC nor PTFC had a significant and direct impact on candidate electability in this condition. Finally, unlike the other conditions, the causal arrow between emotional response and candidate attitude was from candidate electability to emotional response. Limitations and Future Research There are a few methodological and theoretical limitations in this study. They should be well addressed in future research. First, message recall was a problematic construct in the study. There was no check on the amount of message units in each ad, and it is impossible to rule out that some ads were better recalled because there were more message units in them. It is also questionable to add message recall into the original chain reaction model. College students are known for low involvement 106

PAGE 107

with political campaigns (Yoon et al., 2005) and low involvement makes recall an inadequate predictor of memory-based effects (Petty, Cacioppo, & Goldman, 1981). In future research, more efforts should be spent on equivalent amount of message units for recall and enough involvement to generate adequate impact from recall to other behavioral variables. Second, the Scale might not fit negative advertising. Unlike most product advertising and positive political advertising which features one product or one candidate, there are two candidates in a negative ad and the perception of informational and transformational cues could be very different. It is possible that the Scale might not adequately meet this increased complexity of message content. The participants might experience difficulty in choosing the right answers to the Scale items. The increased measurement error, as indicated by lower reliability coefficients in the negative ad condition, led to a bad fit of the model to the data. This problem was alleviated with the latent PIFC and PTFC scores from IRT because they were assumed to have zero measurement error. In future research, it is very important to develop new PIFC and PTFC items for negative advertising. The use of college students was another limitation. College students generally have low interests in politics (Yoon et al., 2005). This characteristic is a possible impediment to the generalizability of the study results in regard to the broader voter population. In future research, it is very important to include non-student population to test the proposed model. A further limitation may be inherent in the limited number of ads for this study. In future research, it is very worthwhile to including a broader spectrum of issues and image at different levels of political races. 107

PAGE 108

CHAPTER 6 IMPLICATIONS This study provides various theoretical and managerial implications to political campaigning practitioners and researchers. In order to investigate the underlying mechanism of candidate attitude formulation in response to political advertising, this study was among the first to empirically test the chain reaction theory of advertising effects (Lavidge & Steiner, 1961). In addition, this study used the informational/transformational scale (Puto & Wells, 1984) to measure message learning, which has never been used for political advertisements, in spite of its theoretical applicability to issue and image political advertisements (Shen, 2007). Furthermore, this study balanced both emotional and cognitive antecedent to candidate electability with the use of a non-verbal emotional measure known as AdSAM (Morris, 1995) and message recall (Petty et al., 1981). This study also balanced the emotional and cognitive perspectives of candidate electability with a tripartite structure composed of issue competence, candidate characteristics, and voting intention (Kahn & Geer, 1994). The attempt to link information/transformation, issue/image, and cognition/emotion proved a legitimate approach to examine the dynamics of issue/image and positive/negative advertisements in voting behavior. Finally, besides structural equation modeling, item response theory was used to estimate latent response to issue and image content in political advertisement and improve model fitting because of reduced measurement error. This approach is well applicable to scale-based research. The managerial implications are particularly evident from a copy-testing standpoint for political campaigning. With some modifications, the informational/transformational scale can be a valid measure of message learning for issue and image in political advertisements, especially for a new and unfamiliar candidate. The enhanced calibration of advertising content will eventually help identify a manner in which issue and image can be managed that will lead to 108

PAGE 109

optimal mental processing and favorable candidate evaluation. The real-life complexity of using positive or negative issue and image to support or attack a candidate was also investigated in this study, and the four message conditions in this research can be used as the base category for a broader range of message strategies. Furthermore, emotional response had a stronger impact than cognitive response on candidate electability, or it was what the respondents felt about the candidates rather than what they remembered about the candidates from the political advertisements that shaped their judgment on the candidate. This finding will provide directions to create effective advertisements that serve to create the right emotion for the right candidate. Finally, given the enormous amount of investment in political advertising, this line of research will become increasing critical in light of the value of democracy and informed voting decision. 109

PAGE 110

APPENDIX A QUESTIONNAIRES Pretest 1. How much did you learn about the government policies/political issues related to the candidate from the ad you just watched? Very little 1 2 3 4 5 6 7 A great deal 2. How much did you learn about the personal qualities related to the candidate from the ad you just watched? Very little 1 2 3 4 5 6 7 A great deal 3. What political issues were mainly associated the candidate in the ad? _____________________________________________________ 4. What characteristics of the candidate stand out in your mind? Describe the candidate in adjective forms _____________________________________________________ 110

PAGE 111

Main Survey Sample (Negative Advertisement) Instructions (continued) Q1. Have you ever watched this advertisement before? <1> Never <2> A few times <3> Many times Q2. Please use the scale below to indicate your familiarity with Chris Chocola, the candidate. <1> Very unfamiliar <2> Unfamiliar <3> Neither unfamiliar nor familiar <4> Familiar <5> Very familiar Q3. Please use the scale below to indicate your familiarity with Joe Donnelly, the candidates opponent. <1> Very unfamiliar <2> Unfamiliar <3> Neither unfamiliar nor familiar <4> Familiar <5> Very familiar Please use the scale to indicate how you feel about Chris Chocola after watching the ad. HAPPY or ELATED UNHAPPY or SAD STIMULATED or INVOLVED CALM or BORED NOT IN CONTROL IN-CONTROL or EMPOWERED 111

PAGE 112

Please use the scale to indicate how you feel about Joe Donnelly after watching the ad. HAPPY or ELATED UNHAPPY or SAD STIMULATED or INVOLVED CALM or BORED NOT IN CONTROL IN-CONTROL or EMPOWERED Instructions (Continued) Please circle the number that best indicates the degree of your agreement or disagreement with each of the statements listed below. There are no correct or incorrect answers to those statements. Questions Q1A. I learned something from this advertisement that I didnt know before about Chris Chocola, the candidate. -4 -3 -2 -1 0 1 2 3 4 very strong disagreement very strong agreement Q1B. I learned something from this advertisement that I didnt know before about Joe Donnelly, the candidates opponent. -4 -3 -2 -1 0 1 2 3 4 very strong disagreement very strong agreement Q2. I would like to have an experience like the one shown in the advertisement. -4 -3 -2 -1 0 1 2 3 4 very strong disagreement very strong agreement Q3. The advertisement did not seem to be speaking directly to me. -4 -3 -2 -1 0 1 2 3 4 very strong disagreement very strong agreement 112

PAGE 113

Q4. There is nothing special about Chris Chocola, the candidate, that makes him different from Joe Donnelly, the candidates opponent. -4 -3 -2 -1 0 1 2 3 4 very strong disagreement very strong agreement Q5. While I watched this advertisement, I thought how Chris Chocola, the candidate, might be useful to me. -4 -3 -2 -1 0 1 2 3 4 very strong disagreement very strong agreement Q6. The advertisement did not teach me what to look for when choosing Chris Chocola, the candidate. -4 -3 -2 -1 0 1 2 3 4 very strong disagreement very strong agreement Q7. This advertisement was meaningful to me. -4 -3 -2 -1 0 1 2 3 4 very strong disagreement very strong agreement Q8. This advertisement was very uninformative. -4 -3 -2 -1 0 1 2 3 4 very strong disagreement very strong agreement Q9. Chris Chocola, the candidate, fits my lifestyle very well. -4 -3 -2 -1 0 1 2 3 4 very strong disagreement very strong agreement Q10. I could really relate to this advertisement. -4 -3 -2 -1 0 1 2 3 4 very strong disagreement very strong agreement Q11. Choosing Chris Chocola, the candidate, makes me feel good about myself. -4 -3 -2 -1 0 1 2 3 4 very strong disagreement very strong agreement 113

PAGE 114

Q12. If they had to, Chris Chocolas party could provide evidence to support the claims made in this advertisement. -4 -3 -2 -1 0 1 2 3 4 very strong disagreement very strong agreement Q13. Its hard to give a specific reason, but somehow Chris Chocola, the candidate, is not really for me. -4 -3 -2 -1 0 1 2 3 4 very strong disagreement very strong agreement Q14. This advertisement did not really hold my attention. -4 -3 -2 -1 0 1 2 3 4 very strong disagreement very strong agreement Q15A. This advertisement reminded me of some important facts about Chris Chocola, the candidate, which I already knew. -4 -3 -2 -1 0 1 2 3 4 very strong disagreement very strong agreement Q15B. This advertisement reminded me of some important facts about Joe Donnelly, the candidates opponent, which I already knew. -4 -3 -2 -1 0 1 2 3 4 very strong disagreement very strong agreement Q16. If I could change my lifestyle, I would make it less like the people who choose Chris Chocola, the candidate. -4 -3 -2 -1 0 1 2 3 4 very strong disagreement very strong agreement Q17A. When I think of Chris Chocola, the candidate, I think of this advertisement. -4 -3 -2 -1 0 1 2 3 4 very strong disagreement very strong agreement Q17B. When I think of Joe Donnelly, the candidates opponent, I think of this advertisement. -4 -3 -2 -1 0 1 2 3 4 very strong disagreement very strong agreement 114

PAGE 115

Q18. I felt as though I were right there in the advertisement, experiencing the same thing. -4 -3 -2 -1 0 1 2 3 4 very strong disagreement very strong agreement Q19. I can now accurately compare Chris Chocola, the candidate, with Joe Donnelly, the candidates opponent, on matters that are important to me. -4 -3 -2 -1 0 1 2 3 4 very strong disagreement very strong agreement Q20. This advertisement did not remind me of any experience or feelings Ive had in my own life. -4 -3 -2 -1 0 1 2 3 4 very strong disagreement very strong agreement Q21. I would have more confidence in my ability to judge the merits of choosing Chris Chocola, the candidate, now than before I saw this advertisement. -4 -3 -2 -1 0 1 2 3 4 very strong disagreement very strong agreement Q22. It is the kind of advertisement that keeps running through your head after youve seen it. -4 -3 -2 -1 0 1 2 3 4 very strong disagreement very strong agreement Q23. Its hard to put into words, but this advertisement leaves me with a good feeling about Chris Chocola, the candidate. -4 -3 -2 -1 0 1 2 3 4 very strong disagreement very strong agreement Q24. What personal qualities/characteristics of Chris Chocola, the candidate, stand out in your mind? Describe the candidate in adjective forms _____________________________________________________________________________ Q25. What personal qualities/characteristics of Joe Donnelly, the candidates opponent, stand out in your mind? Describe the candidate in adjective forms ______________________________________________________________________________ 115

PAGE 116

Q26. What government policies/political issues were mainly associated with Chris Chocola, the candidate, in the ad? ______________________________________________________________________________ Q27. What government policies/political issues were mainly associated with Joe Donnelly, the candidates opponent, in the ad? ______________________________________________________________________________ Q28. What is your best guess about Chris Chocolas competence in dealing with issues related to tax? 1 2 3 4 5 6 7 very competent very incompetent Q29. What is your best guess about Joe Donnellys competence in dealing with issues related to tax? 1 2 3 4 5 6 7 very competent very incompetent Q30. How much would you agree that trustworthy fits your impression of Chris Chocola? 1 2 3 4 5 6 7 strongly disagree strongly agree Q31. How much would you agree that provides strong leadership fits your impression of Chris Chocola? 1 2 3 4 5 6 7 strongly disagree strongly agree Q32. How much would you agree that hardworking fits your impression of Chris Chocola? 1 2 3 4 5 6 7 strongly disagree strongly agree 116

PAGE 117

Q33. How much would you agree that knowledgeable fits your impression of Chris Chocola? 1 2 3 4 5 6 7 strongly disagree strongly agree Q34. How much would you agree that trustworthy fits your impression of Joe Donnelly? 1 2 3 4 5 6 7 strongly disagree strongly agree Q35. How much would you agree that provides strong leadership fits your impression of Joe Donnelly? 1 2 3 4 5 6 7 strongly disagree strongly agree Q36. How much would you agree that hardworking fits your impression of Joe Donnelly? 1 2 3 4 5 6 7 strongly disagree strongly agree Q37. How much would you agree that knowledgeable fits your impression of Joe Donnelly? 1 2 3 4 5 6 7 strongly disagree strongly agree Q38. How likely would you vote for Chris Chocola? 1 2 3 4 5 very unlikely unlikely have no opinion likely very likely Q39. How likely would you vote for Joe Donnelly? 1 2 3 4 5 very unlikely unlikely have no opinion likely very likely 117

PAGE 118

APPENDIX B CORRELATION MATRICES Table B-1. Positive image condition Variable 1 2 3 4 5 6 7 8 9 1. PIFC 1.00 2. PTFC 0.56 1.00 3. Message Recall 0.06 0.14 1.00 4. Issue Competency 0.34 0.27 0.02 1.00 5. Personal Characteristics 0.50 0.57 0.17 0.43 1.00 6. Vote Intention 0.44 0.58 0.03 0.37 0.57 1.00 7. Pleasure 0.36 0.58 0.04 0.29 0.44 0.45 1.00 8. Arousal 0.29 0.55 0.09 0.16 0.34 0.32 0.56 1.00 9. Dominance 0.26 0.29 0.03 0.29 0.25 0.22 0.55 0.34 1.00 Table B-2. Positive issue condition Variable 1 2 3 4 5 6 7 8 9 1. PIFC 1.00 2. PTFC 0.67 1.00 3. Message Recall 0.16 0.08 1.00 4. Issue Competency 0.28 0.29 0.23 1.00 5. Personal Characteristics 0.49 0.52 0.21 0.47 1.00 6. Vote Intention 0.42 0.52 0.20 0.35 0.47 1.00 7. Pleasure 0.42 0.55 0.05 0.32 0.44 0.44 1.00 8. Arousal 0.34 0.43 0.14 0.28 0.39 0.30 0.58 1.00 9. Dominance -0.04 -0.01 -0.14 0.15 -0.02 0.05 0.20 0.15 1.00 Table B-3. Negative image condition for the candidate Variable 1 2 3 4 5 6 7 8 9 1. PIFC 1.00 2. PTFC 0.60 1.00 3. Message Recall 0.17 0.10 1.00 4. Issue Competency 0.32 0.26 0.15 1.00 5. Personal Characteristics 0.48 0.50 0.09 0.42 1.00 6. Vote Intention 0.37 0.43 0.08 0.25 0.57 1.00 7. Pleasure 0.24 0.28 -0.08 0.25 0.37 0.29 1.00 8. Arousal 0.22 0.22 0.06 0.11 0.26 0.22 0.19 1.00 9. Dominance 0.12 0.08 -0.01 0.13 0.11 0.15 0.30 0.13 1.00 118

PAGE 119

Table B-4. Negative image condition for the opponent Variable 1 2 3 4 5 6 7 8 9 1.00 2. PTFC 0.60 1.00 3. Message Recall 0.19 0.16 1.00 4. Issue Competency -0.18 -0.20 -0.17 1.00 5. Personal Characteristics -0.17 -0.04 -0.21 0.38 1.00 6. Vote Intention -0.07 -0.02 -0.02 0.23 0.36 1.00 7. Pleasure -0.30 -0.10 -0.18 0.31 0.15 0.16 1.00 8. Arousal 0.18 0.20 0.09 0.08 0.01 0.07 0.03 1.00 9. Dominance -0.07 -0.01 -0.01 0.17 0.20 0.13 0.22 -0.01 1.00 Table B-5. Negative issue condition for the candidate Variable 1 2 3 4 5 6 7 8 9 1. PIFC 1.00 2. PTFC 0.62 1.00 3. Message Recall 0.19 -0.01 1.00 4. Issue Competency 0.29 0.38 -0.02 1.00 5. Personal Characteristics 0.46 0.48 -0.02 0.54 1.00 6. Vote Intention 0.38 0.49 -0.12 0.42 0.59 1.00 7. Pleasure 0.31 0.38 -0.05 0.17 0.29 0.24 1.00 8. Arousal 0.16 0.26 0.11 0.03 0.13 0.08 0.22 1.00 9. Dominance 0.04 0.05 0.02 -0.04 0.12 0.17 0.26 0.11 1.00 Table B-6. Negative issue condition for the opponent Variable 1 2 3 4 5 6 7 8 9 1. PIFC 1.00 2. PTFC 0.64 1.00 3. Message Recall 0.01 0.06 1.00 4. Issue Competency -0.09 -0.04 -0.06 1.00 5. Personal Characteristics 0.11 0.07 -0.05 0.36 1.00 6. Vote Intention -0.06 -0.12 -0.03 0.35 0.47 1.00 7. Pleasure -0.09 -0.21 -0.12 0.12 0.12 0.22 1.00 8. Arousal 0.08 0.09 0.04 -0.01 0.13 0.02 0.14 1.00 9. Dominance 0.04 -0.04 -0.03 -0.04 -0.03 0.01 0.29 0.22 1.00 119

PAGE 120

LIST OF REFERENCES Aaker, D. A., & Bruzzone, D. E. (1985). Causes of irritation in advertising. Journal of Marketing, 49, 47-57. Aaker, D. A., & Norris, D. (1982). Characteristics of TV commercials percieved as informative. Journal of Advertising Research, 22, 61-70. Aaker, D. A., & Stayman, D. M. (1992). Implementing the concept of transformational advertising. Psychology & Marketing, 9, 237-253. Aaker, D. A., Stayman, D. M., & Hagerty, M. R. (1986). Warmth in advertising: Measurement, impact, and sequence effects. Journal of Consumer Research, 12, 365-381. Aaker, J. L. (1997). Dimensions of brand personality. Journal of Marketing Research, 34, 347-356. Anderson, L. R., & McGuire, W. J. (1965). Prior reassurance of group consensus as a factor in producing resistance to persuasion. Sociometry, 28, 44-56. Andreasen, N. C. (1997). Linking mind and brain in the study of mental illnesses: A project for a scientific psychopathology. Science, 275, 1586-1593. Averill, J. R. (1975). A semantic atlas of emotional concepts. JSAS Catalogue of Selected Documents in Psychology, 5, 330. Averill, J. R. (1982). Anger and aggression: An essay on emotion. New York: Springer. Baines, P. R., Worcester, R. M., Jarrett, D., & Mortimore, R. (2005). Product attribute-based voter segmentation and resource advantage theory. Journal of Marketing Management, 21, 1079-1115. Belk, R. W., Balm, K. D., & Mayer, R. N. (1982). Development recognition of consumption symbolism. Journal of Consumer Research, 9, 4-16. Ben-Ur, J., & Newman, B. I. (2002). Motives, perceptions and voting intention of voters in the 2000 U.S. presidential election. Psychology & Marketing, 19, 1047-1065. Berelson, B. (1966). Democratic theory and public opinion. In B. Berelson & M. Janowitz (Eds.), Reader in public opinion and communication (pp. 489-540). New York: Free Press. Blumenthal, S. (1980). The permanent campaign: Inside the world of elite political operatives. Boston: Beacon. Boster, F. J., & Mongeau, P. (1984). Fear-arousing persuasive messages. In R. N. Bostrom & B. H. Westley (Eds.), Communication Yearbook (Vol. 8, pp. 330-375). Newbury Park, Ca: Sage. 120

PAGE 121

Bradley, M. M., Greenwald, M. K., & Hamm, A. O. (1993). Affective picture processing. In N.Birbaumer & A. Ohman (Eds.), The structure of emotion: Psychophysiological, cognitive and clinical aspects. Toronto: Hogrefe & Huber. Brooker, G. (1981). A comparison of the persuasive effects of mild humor and mild fear appeals. Journal of Advertising, 10, 29-40. Browne, M. W., & Cudeck, R. (1993). Alternative ways of assessing model fit. In K. A. Bollen & J. S. Long (Eds.), Testing structural equation models (pp. 136-162). Newbury Park, CA: Sage. Bulter, P., & Collins, N. (1994). Political marketing: Structure and process. European Journal of Marketing, 28, 19-34. Bush, L. E., II. (1973). Individual differences in multidimensional scaling of adjectives denoting feelings. Journal of Personality and Social Psychology, 25, 50-57. Cacioppo, J. T., & Petty, R. E. (1989). The Elaboration Likelihood Model: The role of affect and affect-laden information processing in persuasion. In P. Cafferata & A. Tybout (Eds.), Cognitive and Affective Responses to Advertising (pp. 69-89). Lexington, MA: Lexington Books. Campbell, A., Converse, P., Miller, W. E., & Stokes, D. E. (1960). The American voter. New York: John Wiley and Sons. Campbell, A., Gurin, G., & Miller, W. E. (1954). The voter decides. Evanston, IL: Row, Peterson and Co. Chaiken, S. (1980). Heuristic versus systematic information processing and the use of source versus message cues in persuasion. Journal of Personality and Social Psychology, 39, 752-766. Chang, C. (2001). The impacts of emotion elicited by print political advertising on candidate evaluation. Media Psychology, 3, 91-119. Christ, W. G., Thorson, E., & Caywood, C. (1994). Do attitudes toward political advertising affect information processing of televised political commercials? Journal of Broadcasting & Electronic Media, 38, 251-270. Clark, M. S. (1982). A role for arousal in the link between feeling states, judgments, and behavior. In M. S. Clark & S. T. Fiske (Eds.). Hillsdale, NJ: Erlbaum. Cohen, J., & Basu, K. (1987). Alternative models of categorization: Toward a contingent processing framework. Journal of Consumer Research, 13, 455-472. Darke, P. R., Freedman, J. L., & Chaiken, S. (1995). Percentage discounts, initial price, and bargain hunting: A heuristic-systematic approach to price search behavior. Journal of Applied Psychology, 80, 580-586. 121

PAGE 122

Davis, J., J. (1997). Advertising Research: Theory and Practice. Upper Saddle River, NJ.: Prentice Hall. Decety, J., & Chaminade, T. (2003). Neural correlates of feeling sympathy. Neuropsychologia, 41, 127-138. Deighton, J., Romer, D., & McQueen, J. (1989). Using drama to persuade. Journal of Consumer Research, 16, 335-343. Dermody, J., & Scullion, R. (2001). An exploration of the advertising ambitions and strategies of the 2001 British General Election. Journal of Marketing Management, 17, 969-987. Devlin, L. P. (1989). Contrasts in presidential campaign commercials of 1988. American Behavioral Scientist, 32, 389-414. Devlin, L. P. (1993). Contrasts in presidential campaign commercials of 1992. American Behavioral Scientist, 37, 272-290. Devlin, L. P. (1997). Contrast in presidential campaign commercials of 1996. American Behavioral Scientist, 40, 1058-1084. Devlin, L. P. (2001). Contrast in presidential campaign commercials of 2000. American Behavioral Scientist, 44, 2338-2369. Devlin, L. P. (2005). Contrasts in presidential campaign commercials of 2004. American Behavioral Scientist, 49, 279-313. Duncan, C. P., & Nelson, J. E. (1985). Effects of humor in a radio advertising experiment. Journal of Advertising, 14, 33-40. Duncan, T. (2002). IMC: Using advertising and promotion to build brands. New York: McGraw-Hill. Edell, J. A., & Burke, M. C. (1987). The power of feelings in understanding advertising effects. Journal of Consumer Research, 14, 421-433. Ekman, P., & Friesen, W. V. (1978). Facial action coding system (FACS): A technique for the measurement of facial action. Palo Alto, CA: Consulting Psychologists Press. Embretson, S. E., & Reise, S. P. (2000). Item response theory for psychologists. Mahwah, NJ: Lawrence Erlbaum Associates. Ewing, M. T., Salzberger, T., & Sinkovics, R. R. (2005). An alternate approach to assessing cross-cultural measurement equivalence in advertising research. Journal of Advertising, 34, 17-36. 122

PAGE 123

Faber, R. J., Tims, A. R., & Schmitt, K. G. (1993). Negative political advertising and voting intent: The role of involvement and alternative information sources. Journal of Advertising, 22, 67-76. Feldman, L. A., Barrett, L., & Russell, J. A. (1998). Independence and bipolarity in the structure of affect. Journal of Personality and Social Psychology, 74, 967-984. Fishbein, M., & Ajzen, I. (1975). Belief, attitude, intention and behavior: An introduction to theory and research. Reading, MA: Addison-Wesley. Fishbein, M., & Ajzen, I. (1980). Predicting and understanding consumer behavior: Attitude-behavior correspondence. In I. Ajzen & M. Fishbein (Eds.), Understanding attitudes and predicting social behavior (pp. 148-172). Eaglewood Cliffs, NJ: Prentice-Hall. Fiske, S. T., & Taylor, S. E. (1984). Social cognition. New York: Random House. Frijda, N. H. (1986). The emotions. New York: Cambridge University Press. Geer, J. G., & Geer, J. H. (2003). Remembering attack ads: An experimental investigation of radio. Political Behavior, 25, 69-95. Gelb, B. D., & Pickett, C. M. (1983). Attitude-toward-the-ad: Links to humor and to advertising effectiveness. Journal of Advertising, 12, 34-42. Graber, D. A. (2004). Methodological developments in political communication research. In L. L. Kaid (Ed.), Handbook of political communication research (pp. 45-67). Mahwah, New Jersey: Lawrence Erlbaum Associates. Greenwald, A. G. (1989). Why are attitudes important? In A. G. Greenwald (Ed.), Attitude structure and function (pp. 1-10). Hillsdale, New Jersey: Lawrence Erlbaum Associates. Hair, J. F. J., Anderson, R. E., Tatham, R. L., & Black, W. C. (1995). Multivariate data analysis. Upper Saddle River, New Jersey: Prentice-Hall. Hambleton, R. K., & Swaminathan, H. (1985). Item response theory: Principles and applications. Boston: Kluwer-Nijhoff. Henneberg, S. C. (2004). The views of an advocatus dei: Political marketing and its critics. Journal of Public Affairs, 4, 225-243. Hill, R. P. (1989). An exploration of voter responses to political advertisements. Journal of Advertising, 18, 14-22. Hoyer, W. D., & Maclnnis, D. J. (2004). Consumer behavior (3rd ed.). Boston, MA: Houghton Mifflin. Izard, C. E. (1977). The emotions in life and science. In C. E. Izard (Ed.), Human emotions (pp. 1-18). New York: Plenum Press. 123

PAGE 124

Izard, C. E. (1992). Basic emotions, relations among emotions, and emotion-cognition relations. Psychological Review, 99, 561-565. Izard, C. E. (1993). Four systems for emotion activation: Cognitive and non-cognitive processes. Psychology Review, 100, 68-90. Izard, C. E., & Bartlett, E. S. (1972). Patterns of emotions: A new analysis of anxiety and depression. Oxford, England: Academic Press, 1972. James, W. (1884). What is an emotion? Mind, 9, 188-205. Johnson-Cartee, K. S., & Copeland, G. A. (1989). Southern voters' reaction to negative political ads in 1986 election. Journalism and Mass Communication Quarterly, 66, 888-893, 986. Johnston, A., & Kaid, L. L. (2002). Image ads and issue ads in U.S. presidential advertising: Using videostyle to explore stylistic differences in televised political ads from 1952 to 2000. Journal of Communication, 52, 281-300. Joreskog, K. G., & Sorbom, D. (2005). Lisrel 8.72. Lincolnwood, IL: Scientific Software International, Inc. Joslyn, R. A. (1980). The content of political spot ads. Journalism Quarterly, 57, 92-98. Kahn, K. F., & Geer, J. G. (1994). Creating impressions: An experimental investigation of political advertising on television. Political Behavior, 16, 93-116. Kahn, K. F., & Kenney, P. J. (2000). How negative campaigning enhances knowledge of Senate elections. In J. A. Thurber, C. J. Nelson & D. A. Dulio (Eds.), Crowded airwaves: Campaign advertising in elections (pp. 65-95). Washtington, DC: Brooking Institution. Kaid, L. L. (2001). Technodistortions and effects of the 2000 political advertising. American Behavioral Scientist, 44, 2370-2378. Kaid, L. L. (2002). Videostyle and political advertising effects in the 2000 presidential campaign. In R. E. J. Denton (Ed.), The 2000 presidential campaign: A communication perspective (pp. 183-197). Westport, CT: Praeger. Kaid, L. L. (2004). Political advertising. In L. L. Kaid (Ed.), Handbook of political communication research (pp. 155-202). Mahwah, NJ: Lawrence Erlbaum Associates. Kaid, L. L. (2005). Videostyle in the 2004 political advertising. In R. E. J. Denton (Ed.), The 2004 presidential campaign: A communication perspective (pp. 283-299). Lanham, MD: Rowman & Littlefield. Kaid, L. L., Chanslor, M., & Hovind, M. (1992). The influence of program and commercial type on political advertising effectiveness. Journal of Broadcasting & Electronic Media, 36, 303-320. 124

PAGE 125

Kaid, L. L., & Johnston, A. (2001). Videostyle in presidential campaigns: Style and content of televised political advertising. Westport, CT: Praeger. Kaid, L. L., & Postelnicu, M. (2005). Political advertising in the 2004 Election: Comparison of traditional television and Internet message. American Behavioral Scientist, 49, 265-278. Kaid, L. L., & Sanders, K. R. (1978). Political television commercials: An experimental study of type and length. Communication Research, 5, 57-70. Kaid, L. L., & Tecesco, J. C. (1999). Tracking voter reactions to television advertising. In L. L. Kaid & D. G. Bystrom (Eds.), The electronic election: Perspectives on the 1996 campaign communication (pp. 233-246). Mahwah, NJ: Lawrence Earlbaum. King, L. R. (1965). The marketing concept. In G. Schwartz (Ed.), Science in marketing. New York: John Wiley and Sons. Kitayama, S., Markus, H. R., Tummala, P., Kurokawa, M., & Kato, K. (1990). Culture and self-cognition.Unpublished manuscript. Kotler, P. (1975). Overview of political candidate marketing. Advances in Consumer Research, 2, 761-769. Lang, A. (1991). Emotion, formal features, and memory for televised political advertisements. In F. Biocca (Ed.), Television and political advertising: Psychological processes (Vol. 1, pp. 221-243). Hillsdale, New Jersey: Lawrence Erlbaum Associates. Lang, P. J. (1980). Behavioral treatment and bio-behavioral assessment: Computer applications b. In J. B. Sidowski, J. H. Johnson & T. A. Williams (Eds.), Technology in mental health care delivery systems (pp. 119-137). Norwood, NJ: Ablex. Larsen, R. J., & Diener, E. (1992). Promises and problems with the circumplex model of emotion. In M. S. Clark (Ed.), Review of personality and social psychology: Emotion (Vol. 13, pp. 25-59). Laskey, H. A., Day, E., & Crask, M. R. (1989). Typology of main message strategies for television commercials. Journal of Advertising, 18, 36-41. Lavidge, R. J., & Steiner, G. A. (1961). A model for predictive measurements of advertising effectiveness. Journal of Marketing, 25, 59-62. Lazarus, R. S. (1966). Psychological stress and the coping process. New York: McGraw-Hill. Lazarus, R. S. (1984). On the primacy of cognition. American Psychologist, 39, 124-129. Lazarus, R. S. (1991a). Emotion and adaptation. New York: Oxford University Press. Lazarus, R. S. (1991b). Progress on a cognitive-motivational-relational theory of emotion. American Psychologist, 46, 819-834. 125

PAGE 126

LeDoux, J. E. (1995). Emotion: Clues from the brain. Annual Review of Psychology, 46, 209-235. LeDoux, J. E., Iwata, J., Cicchetti, P., & Reis, D. J. (1988). Different projections of the central amygdaloid nucleus mediate autonomic and behavioral correlates of conditioned fear. Journal of Neuroscience, 8, 2517-2529. LeDoux, J. E., Romanski, L., & Xagoraris, A. (1989). Indelibility of subcortical emotional memories. Journal of Cognitive Neuroscience, 1, 238-243. Linden, W. J. v. d., & Hambleton, R. K. (1996). Handbook of moderen item response theory. New York: Springer-Verlag. Lock, A., & Harris, P. (1996). Political marketing vive la difference! European Journal of Marketing, 30, 21-31. Maarek, P. J. (1995). Political marketing and communication. London: John Libby. MacInnis, D. J., & de Mello, G. E. (2005). The concept of hope and its relevance to product evaluation and choice. Journal of Marketing, 69, 1-14. Mackay, C. J. (1980). The measurement of mood and psychophysiological activity using self-report techniques. In I. Marting & P. Venebles (Eds.), Techniques in psychophysiology (pp. 501-562). New York: Wiley. Mandler, G. (1962). Emotion. In R. Brown, E. Galanter, E. Hess & G. Mandler (Eds.), New directions in psychology (pp. 267-339). New York: Holt, Rinehart & Winston. Mandler, G. (1984). Mind and body. New York: Norton. Mano, H. (1997). Affect and persuasion: The influence of pleasantness and arousal on attitude formation and message elaboration. Psychology and Marketing, 14, 315-335. Marshall, S. W. (2006). Advertising message strategies and executional devices in television commercials from award-winning "effective" campaigns from 1999 to 2004.Unpublished manuscript, University of Florida, Gainesville, Florida. Martinelli, K. A., & Chaffee, S. H. (1995). Measuring new-voter learning via three channels of political information. Journalism and Mass Communication Quarterly, 72, 18-32. McGuire, W. J. (1989). Why are attitudes important? In A. R. Pratkanis, S. J. Breckler & A. G. Greenwald (Eds.), Attitude structure and function (pp. 37-69). Hillsdale, New Jersey: Lawrence Erlbaum Associates. Mehrabian, A. (1977). A questionnaire measure of individual differences in stimulus screening and associated differences in arousability. Environmental Psychology and Nonverbal Behavior, 1, 89-103. 126

PAGE 127

Mehrabian, A. (1995). Framework for a comprehensive description and measurement of emotional states. Genetic, Social, and General Psychology Monographs, 121, 339-361. Mehrabian, A., & Russell, J. A. (1974). An approach to environmental psychology b. Cambridge, MA: Oelgesschlager, Gunn. and Hain. Menon, M. K., Goodnight, J. M., & Wayne, R. J. (2006). Assessing advertising content in a hospital advertising campaign: An application of Puto and Wells (1984) measure of informational and transformational advertising content. Journal of Hospital Marketing & Public Relations, 17, 21-44. Merritt, S. (1984). Negative political advertising: Some empirical findings. Journal of Advertising, 13, 27-38. Mewborn, R. C., & Rogers, R. W. (1979). Effects of threatening and reassuring components of fear appeals and the physiological and verbal measures of emotions and attitudes. Journal of Experimental and Social Psychology, 15, 242-253. Mitchell, R. L. C., Elliott, R., Barry, M., Cruttenden, A., & Woodruff, P. W. R. (2003). The neural response to emotional prosody, as revealed by functional magnetic resonance imaging. Neuropsychologia, 41, 1410-1421. Mittal, B. (1999). The advertising of service: Meeting the challenge of intangibility. Journal of Service Research, 2, 98-116. Morris, J. D. (1995). Observations: SAM: The Self-Assessment Manikin: An efficient cross-cultural measurement of emotional response. Journal of Advertising Research, 35, 63-68. Morris, J. D., Bradley, M. M., Sutherland, J., & Wei, L. (1993). Assessing cross cultural transferability of standardized global advertising: An emotional response approach. Paper presented at the Annual Conference of the Association of Education in Journalism and Mass Communications, Kansas City. Morris, J. D., Bradley, M. M., Waine, C. A., & Lang, J. B. (1992). Assessing affective reactions to advertisements with the self-assessment manikin (SAM). Paper presented at the Southern Marketing Association Conference. Morris, J. D., Klahr, N. J., Shen, F., Villegas, J., Wright, P., He, A. G., et al. (2008). fMRI for emotional responses to TV commercials: A dimensional approach for locating emotion processors in human brain. Human Brain Mapping. Morris, J. D., Roberts, M. S., & Baker, G. F. (1998). The effects of music on emotional response, brand attitude, and purchase intent in an emotional advertising condition. In J. W. Alba & J. W. Hutchinson (Eds.), Advances in consumer research (Vol. 25, pp. 518-526). Provo, UT: Association for Consumer Research. 127

PAGE 128

Morris, J. D., Strausbaugh, K. L., & Nthangeni, M. (1996). Emotional response to advertisements (or commercials) across culture. In G. B. Wilcox (Ed.), Proceedings of the 1996 Conference of the American Academy of Advertising (pp. 47-54). Austin, TX: American Academy of Advertising. Morris, J. D., & Waine, C. A. (1993). Managing the creative effort: Preproduction and post-production measures of emotional responses. In E. Thorson (Ed.), Proceedings of the 1993 Conference of the American Academy of Advertising (pp. 158-176). Columbia, MO. Morris, J. D., Waine, C. A., & Lang, J. B. (1993). Assessing emotional responses to advertisements with (SAM) the self-assessment manikin. In Proceedings of the Allied Southern Business Association 1993 Annual Meeting: Allied Southern Business Association. Morris, J. D., Woo, C., Geason, J. A., & Kim, J. (2002). The power of affect: Predicting intention. Journal of Advertising Research, 42, 7-17. Morris, J. D., Woo, C., & Singh, A. J. (2005). Elaboration likelihood model: A missing intrinsic emotional implication. Journal of Targeting, Measurement & Analysis for Marketing, 14, 79-98. Natchez, P. B., & Bupp, I. C. (1968). Candidates, issues, and voters. Public Policy, 17, 409-437. Newhagen, J. E., & Reeves, B. (1991). Emotion and memory responses for negative political advertising: A study of television commercials used in the 1988 presidential election. In F. Biocca (Ed.), Television and political advertising: Psychological processes (Vol. 1, pp. 197-220). Hillsdale, New Jersey: Lawrence Erlbaum Associates. Newman, B. I. (2001). An assessment of the 2000 US presidential election: A set of political marketing guidelines. Journal of Public Affairs, 1, 210-216. Niffenegger, P. B. (1989). Strategies for success from the politcal marketers. Journal of Consumer Marketing, 6, 45-51. Nimmo, D. (1999). The permanent campaign: Marketing as a governing tool. In B. I. Newman (Ed.), Handbook of political marketing (pp. 73-86). Thousand Oaks, CA: Sage. O'Cass, A. (1996). Political marketing and the marketing concept. European Journal of Marketing, 30, 45-61. Oishi, S. (2006). The concept of life satisfaction across cultures: An IRT analysis. Journal of Research in Personality, 40, 347-462. O'Keefe, M. T., & Sheinkopf, K. G. (1974). The voter decides: Candidate image or campaign issue. Journal of Broadcasting, 18, 403-412. Omura, G. S. (1979). Role of attribute generality in cognitions of political candidates. Advances in Consumer Research, 6, 635-640. 128

PAGE 129

Ortony, A., & Turner, T. J. (1990). What's basic about basic emotions? Psychological Review, 97, 315-331. Osgood, C. E., Suci, G. J., & Tannenbaum, P. H. (1957). The measurement of meaning. Urbana, IL: University of Illinois Press. Panksepp, J. (1989). The neurobiology of emotions. In T. Manstead & H. Wagner (Eds.), Handbook of social psychophysiology (pp. 5-26). New York: Wiley. Petty, R. E., & Cacioppo, J. T. (1986). The elaboration likelihood model of persuasion. In L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 19, pp. 123-205). New York: Academic Press. Petty, R. E., Cacioppo, J. T., & Goldman, R. (1981). Personal involvement as a determinant of argument-based persuasion. Journal of Personality and Social Psychology, 41, 847-855. Plutchik, R. (1980). A general psychoevolutionary theory of emotion. In R. Plutchik & H. Kellerman (Eds.), Emotion:Theory, research, and experience (Vol. 1, pp. 3-33). New York: Academic. Plutchik, R. (1982). A psycho evolutionary theory of emotions. Social Science Information, 21, 529-553. Plutchik, R. (1984). Emotions: A general psychoevolutionary theory b. In K. R. Scherer & P. Ekman (Eds.), Approaches to emotion (pp. 197-219). Hillsdale, NJ: Lawrence Erlbaum Associates. Porter, M. E. (1985). Competitive advantage. New York: The Free Press. Puto, C. P., & Wells, W. D. (1984). Informational and transformational advertisng: The differential effects of time. Advances in Consumer Research, 11, 638-643. Quinn, P., & Kivijarv, L. (2005). US political media buying 2004. International Journal of Advertising, 24, 131-140. Resnik, A., & Stern, B. L. (1977). An analysis of information content in television advertising. Journal of Marketing, 41, 50-53. Richins, M. L. (1997). Measuring emotions in consumption experience. Journal of Consumer Research, 24, 127-146. Ridout, T. N., Shah, D. V., Goldstein, K. M., & Franz, M. M. (2004). Evaluating measures of campaign advertising exposure on political learning. Political Behavior, 26, 201-225. Roddy, B. L., & Garramone, G. M. (1988). Appeals and strategies of negative political advertising. Journal of Broadcasting & Electronic Media, 32, 415-427. 129

PAGE 130

Rogers, E. M. (1997). A history of communication study: A biographical approach. New York: Free Press. Roseman, I. J. (1984). Cognitive determinants of emotion : A structural theory. Review of Personality and Social Psychology, 5, 11-36. Roseman, I. J., Antoniou, A. A., & Jose, P. E. (1996). Appraisal determinants of emotions: Constructing a more accurate and comprehensive theory. Cognition & Emotion, 10, 241-277. Russell, J. A. (1978). Evidence of convergent validity on the dimension of affect. Journal of Personality and Social Psychology, 36, 1152-1168. Russell, J. A. (1980). A circumplex model of affect. Journal of Personality and Social Psychology, 42, 1161-1178. Russell, J. A., & Mehrabian, A. (1977). Evidence for a three-factor theory of emotions. Journal of Research in Personality, 11, 273-294. Samejima, F. (1969). Estimation of latent ability using a response pattern of graded scores. Psychometrika Monograph, 17, 1-100. Scherer, K. R., & Wallbott, H. G. (1994). Evidence for universality and cultural variation of differential emotion response patterning. Journal of Personality and Social Psychology, 66, 310-328. Schiffman, L. G., Sherman, E., & Kirpalani, N. (2002). Trusing souls: A segmentation of the voting public. Psychology & Marketing, 19, 993-1007. Schleuder, J. (1990). Effects of commercial complexity, the candidate, and issue vs. image strategies in political ads. Advances in Consumer Research, 17, 159-168. Schlinger, M. J. (1979). A profile of responses to commercials. Journal of Advertising Research, 19, 37-46. Schlosberg, H. (1952). The description of facial expression in terms of two dimensions. Journal of Experimental Psychology, 44, 229-237. Shama, A. (1975). Applications of marketing concepts to candidate marketing. Advances in Consumer Research, 2, 793-801. Shama, A. (1976). The marketing of political candidates. Journal of the Academy of Marketing Science, 4, 764-777. Shapiro, M. A., & Rieger, R. H. (1992). Comparing positive and negative political advertising on radio. Journalism Quarterly, 69, 135-145. 130

PAGE 131

Shaver, P., Schwartz, J., Kirson, D., & O' Connor, C. (1987). Emotion knowledge: Further exploration of a prototype approach. Journal of Personality and Social Psychology, 52, 1061-1086. Shen, F. (2007). Measuring issue and image in political advertising: An informational/transformational approach. Paper presented at the Annual Convention of Association for Education in Journalism and Mass Communication, Washintgton, D. C. Shen, F., & Kim, K. (2006). Videostyle in Presidential Campaigns 2000 & 2004: An informational versus transformational approach. Paper presented at the Association for Education in Journalism and Mass Communication, San Francisco, California. Simon, H. A. (1982). Comments. In M. S. Clarke & S. T. Fiske (Eds.), Affect and cognition: The seventeenth annual Carnegie symposium on cognition (pp. 333-342). Hillsdale, NJ: Erlbaum. Singh, J. (2004). Tackling measurement problems with Item Response Theory: Principles, characteristics, and assessment, with an illustrative example. Journal of Business Research, 57, 184-208. Singh, J., Howell, R. D., & Rhoads, G. K. (1990). Adaptive designs for Likert-type data: An approach for implementing marketing surveys. Journal of Marketing Research, 27, 304-321. Sonner, B. S. (1998). The effectiveness of negative political advertising: A case study. Journal of Advertising Research, 38, 37-42. Swaminathan, V., Zinkhan, G. M., & Reddy, S. K. (1996). The evolution and antecedents of transformational adverrtising: A conceptual model. Advances in Consumer Research, 23, 49-55. Teasdale, J. D. (1993). Emotion and two kinds of meaning: Cognitive therapy and applied cognitive science. Behavior Research Theory, 31, 339-354. Teasdale, J. D., & Barnard, P. J. (1993). Affect, cognition, and change: Remodelling depressive thought. Hove, England: Lawrence Erlbaum Associates. Teasdale, J. D., Howard, R. J., Cox, S. G., Ha, Y., Brammer, M. J., Williams, S. C. R., et al. (1999). Functional MRI study of the cognitive generation of affect. American Journal of Psychology, 156, 209-215. Thorson, E., Christ, W. G., & Caywood, C. (1991). Effects of issue-image strategies, attack and support appeals, music, and visual content in political commercials. Journal of Broadcasting & Electronic Media, 35, 465-486. Thorson, E., Reeves, B., & Schleuder, J. (1985). Message complexity and attention to television. Communication Research, 12, 427-454. 131

PAGE 132

Thorson, E., Reeves, B., & Schleuder, J. (1987). Attention to local and global complexity in television messages. In M. McLaughlin (Ed.), Communication Yearbook 10. Beverly Hills, CA: Sage. Tomkins, S. S. (1980). Affect as amplification: Some modifications in theory. In R. Plutchik & H. Kellerman (Eds.), Emotion: Theory, research and experience (pp. 141-164). New York: Academic Press. Vakratsas, D., & Ambler, T. (1999). How advertising works: What do we really know? Journal of Marketing, 63, 26-43. Vavreck, L. (2001). The reasoning voter meets the strategic candidate: Signals and specificity in campaign advertising, 1998. American Politics Research, 29, 507-529. Watson, D., & Tellegen, A. (1985). Toward a consensual structure of mood. Psychological Bulletin, 98, 219-235. Watson, D., Wiese, D., Vaidya, J., & Tellegen, A. (1999). The two general activation systems of affect: Structural findings, evolutinory considerations, and psychological evidence. Journal of Personality and Social Psychology, 76, 820-838. Weaver-Lariscy, R. A., & Tinkham, S. F. (1996). Advertising message strategies in U.S. congressional campaigns: 1982, 1990. Journal of Current Issues and Research in Advertising, 18, 53-66. Weiner, B. (1985). An attributional theory of achievement motivation and emotion. Psychological Review, 92, 548-573. Wells, W. D. (1988). Lectures and dramas. In P. Cafferata & A. Tybout (Eds.), Cognitive and affective responses to advertising (pp. 13-20). Lexington, MA: D. C. Heath. West, D. M. (1993). Air wars: Television advertising in election campaigns, 1952-1992. Washington, DC: Congressional Quarterly. West, D. M. (1994). Political advertising and news coverage in the 1992 California U.S. Senate campaigns. The Journal of Politics, 56, 1053-1075. Witte, K. (1995). Generating effective risk messages: How scary should your risk communication be? In B. R. Burleson (Ed.), Communication Yearbook (Vol. 18, pp. 229-254). Thousand Oaks, CA: Sage. Wring, D. (1997). Reconciling marketing with political science: Theories of political marketing. Journal of Marketing Management, 13, 651-663. Yoon, K., Pinkleton, B., & Ko, W. (2005). Effects of negative political advertising on voting intention: An exploration of the roles of involvement and source credibility in the development of voter cynicism. Journal of Marketing Communications, 11, 95-112. 132

PAGE 133

Zajonc, R. B. (1984a). The interaction of affect and cognition. In K. R. Scherer & P. Ekman (Eds.), Approaches to emotion (pp. 239-246). Hillsdale, NJ: Lawrence Erlbaum. Zajonc, R. B. (1984b). On the primacy of affect. American Psychologist, 39, 117-123. Zeithaml, V. A., & Bitner, M. J. (1996). Service marketing. New York: McGraw-Hill. 133

PAGE 134

BIOGRAPHICAL SKETCH Feng Shen was born in China. He received his Bachelor of Arts degree in english at Shanghai International Studies University (Shanghai, China) in 1995. After five years of industry experience with public health and marketing campaigns in China, he came to the United States and completed a Master of Science degree in communication at Cornell University, Ithaca, New York in 2002. After spending one year in China managing international investment, he joined the Ph.D. program at the University of Florida, Gainesville, Florida in 2005. He has been collaborating primarily with Dr. Jon Morris on several research projects. In 2006 and 2007, he taught Advertising Research at the University of Florida. He has now accepted a position as an Assistant Professor of Marketing at Erivan K. Haub School of Business, Saint Josephs University, Philadelphia, Pennsylvania.