<%BANNER%>

The Year of the Ox

Permanent Link: http://ufdc.ufl.edu/UFE0044423/00001

Material Information

Title: The Year of the Ox Social Bonding Theory and Juvenile Delinquency in a 1973 Chinese Birth Cohort
Physical Description: 1 online resource (96 p.)
Language: english
Creator: Timbs, Allison L
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2012

Subjects

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

Notes

Abstract: CHINAThis dissertation uses data collected from a birth cohort born in 1973 in the Wuchange district of China.  Elements of Hirschi’s social bonding theory and Burgess and Akers’ social learning theory (specifically the presence of delinquent peers) were used as independent variables and offender status and deviance status were used as the dependent variables.  In this cohort there were only 81individuals who were identified as delinquent.  The sample used in this analysis consists of the offenders and 81 other individuals who were matched based on a short list of social criteria. Analysis show support for most social bonding elements.  Variables relating to the presence of delinquent peers produced results that are surprising based on the previous literature.  Some support is found for social bonding theory and some support was also found for the effects of peers.  But the support for social bonding theory is more mixed and weaker than typically found in previous research.  The same is true for peer variables, and there is one finding on peer effects, rarely seen in previous research, that indicates that differential association with deviant peers is related to the respondents more conforming behavior.  This is opposite from expectation based on social learning theory.  The relevance of these findings for the theories and future research is discussed.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Allison L Timbs.
Thesis: Thesis (Ph.D.)--University of Florida, 2012.
Local: Adviser: Akers, Ronald L.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2012
System ID: UFE0044423:00001

Permanent Link: http://ufdc.ufl.edu/UFE0044423/00001

Material Information

Title: The Year of the Ox Social Bonding Theory and Juvenile Delinquency in a 1973 Chinese Birth Cohort
Physical Description: 1 online resource (96 p.)
Language: english
Creator: Timbs, Allison L
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2012

Subjects

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

Notes

Abstract: CHINAThis dissertation uses data collected from a birth cohort born in 1973 in the Wuchange district of China.  Elements of Hirschi’s social bonding theory and Burgess and Akers’ social learning theory (specifically the presence of delinquent peers) were used as independent variables and offender status and deviance status were used as the dependent variables.  In this cohort there were only 81individuals who were identified as delinquent.  The sample used in this analysis consists of the offenders and 81 other individuals who were matched based on a short list of social criteria. Analysis show support for most social bonding elements.  Variables relating to the presence of delinquent peers produced results that are surprising based on the previous literature.  Some support is found for social bonding theory and some support was also found for the effects of peers.  But the support for social bonding theory is more mixed and weaker than typically found in previous research.  The same is true for peer variables, and there is one finding on peer effects, rarely seen in previous research, that indicates that differential association with deviant peers is related to the respondents more conforming behavior.  This is opposite from expectation based on social learning theory.  The relevance of these findings for the theories and future research is discussed.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Allison L Timbs.
Thesis: Thesis (Ph.D.)--University of Florida, 2012.
Local: Adviser: Akers, Ronald L.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2012
System ID: UFE0044423:00001


This item has the following downloads:


Full Text

PAGE 1

1 THE YEAR OF THE OX: SOCIAL BONDING THEORY AND JUVENILE DELINQUENCY IN A 1973 CHINESE BIRTH COHORT By ALLISON L TIMBS 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 2012

PAGE 2

2 2012 Allison L Timbs

PAGE 3

3 To my parents for all their love and support

PAGE 4

4 ACKNOWLEDGMENTS I would like to thank my committee chair, Ron Akers, for being an ou tstanding mentor and teacher since I began my graduate studies at UF. In addition to Ron Akers, I also want to thank Lonn Lanza Kaduce, Charles Frazier, and Marion Borg for being a truly wonderful (and patient) committee. I would like to thank my CU famil y for all their help and support along the way. I will always be grateful for the help Jackie Sandifer, Craig Rogers, and Frank Cheatham lent to this effort. Also, Mackenzie and Cici deserve a special acknowledgement for everything they have done since b efore the beginning.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 7 LIST OF ABBREVIATION S ................................ ................................ ........................... 11 ABSTRACT ................................ ................................ ................................ ................... 12 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .... 13 2 REVIEW OF THE LITERATURE ................................ ................................ ............ 24 Empirical Support for Social Bonding Theory ................................ ......................... 24 Bonding Theory and Asian Samples ................................ ................................ ....... 27 3 METHODS ................................ ................................ ................................ .............. 31 Hypotheses ................................ ................................ ................................ ............. 31 The Sample ................................ ................................ ................................ ............ 31 The Variables ................................ ................................ ................................ .......... 33 Parental Attachment ................................ ................................ ......................... 35 Educational Attachment ................................ ................................ .................... 37 Attachment to Peers ................................ ................................ ......................... 38 Differential Peer Association ................................ ................................ ............ 38 Commitment to Education ................................ ................................ ................ 39 Involvement ................................ ................................ ................................ ...... 40 Deviance ................................ ................................ ................................ .......... 41 4 ANALYSIS ................................ ................................ ................................ .............. 52 Attachment and Offender Status ................................ ................................ ............. 53 Parental Attachment ................................ ................................ ......................... 53 Educational Attachment ................................ ................................ .................... 53 Attachment to Peers ................................ ................................ ......................... 54 Commitment ................................ ................................ ................................ ..... 54 Involvement ................................ ................................ ................................ ...... 54 Differential Peer Association ................................ ................................ ............ 55 Analysis of All Independent Variables on Offender Status ............................... 56 Deviance and Attachment ................................ ................................ ....................... 57 Deviance and Pa rental Attachment ................................ ................................ .. 58 Deviance and Educational Attachment ................................ ............................. 58 Deviance and Peer Attachment ................................ ................................ ........ 59

PAGE 6

6 Deviance and Commitment to Education ................................ ......................... 59 Deviance and Social Involvement. ................................ ................................ ... 60 Deviance and Differential Peer Associat ion ................................ ...................... 61 All Independent Variables and Deviance Status ................................ ............... 62 5 SUMMARY AND CONCLUSION ................................ ................................ ............ 79 APPENDIX A MEASURES OF BONDING ................................ ................................ .................... 89 Attachment: ................................ ................................ ................................ ...... 89 Educational ................................ ................................ ................................ 89 Parental ................................ ................................ ................................ ...... 89 Peers ................................ ................................ ................................ ......... 90 Commitment ................................ ................................ ................................ ..... 90 Involvement ................................ ................................ ................................ ...... 90 B MEASURES OF DIFFERENTIAL PEER ASSOCIATION ................................ ....... 92 LIST OF REFERENCES ................................ ................................ ............................... 93 BIOGRAPHICAL SKETCH ................................ ................................ ............................ 96

PAGE 7

7 LIST OF TABLES Table page 3 1 Parental attachment reliability statistics ................................ .............................. 43 3 2 Parental att achment item total statistics ................................ ............................. 43 3 3 Parental Attachment/offender status Crosstabulation ................................ ......... 43 3 4 Educational Attachment Reliability Sta tistics ................................ ...................... 44 3 5 Educational Attachment Item Total ................................ ................................ ..... 44 3 6 Educational Attachment/offender status Crosstabulation ................................ ... 44 3 7 Offender Status/Attachment to Peers Crosstabulation ................................ ....... 44 3 8 Friends Absent From Work/offender status Crosstabulation .............................. 45 3 9 Friend Truant/offender status Crosstabulation ................................ .................... 45 3 10 Friends abide by law/offender status Crosstabulation ................................ ........ 45 3 11 Friends Steal/offender status Crosstabulation ................................ .................... 45 3 12 Friends Detained/offender status Crosstabulation ................................ .............. 45 3 13 Friends Compulsory Labor/offender status Crosstabulation ............................... 46 3 14 Commitment to Education Reliability Statistics ................................ ................... 46 3 15 Commitment to Educatio n Item Total Statistics ................................ .................. 46 3 16 Commitment to Education/ Offender Crosstabulation ................................ ......... 46 3 17 Visits/offender status Crosstabulation ................................ ................................ 46 3 18 Attend Lectures/offender status Crosstabulation ................................ ................ 47 3 19 Public Security/offender status Crosstabulation ................................ ................. 47 3 20 Take Part in Other/offender status Crosstabulation ................................ ............ 47 3 21 Deviance Reliability Statistics ................................ ................................ ............. 47 3 22 Deviance Item Total Statistics ................................ ................................ ............ 47 3 23 Deviance/offender status Crosstabulation ................................ .......................... 48

PAGE 8

8 3 24 Parental Attachment/Deviance Crosstabul ation ................................ ................. 48 3 25 Commitment/Deviance Crosstabulation ................................ ............................. 48 3 26 Educational Attachment/Deviance Crosstabulation ................................ ............ 48 3 27 Friends Absent Work/Deviance Crosstabulation ................................ ................ 49 3 28 Friends Truant/Deviance Crosstabulation ................................ .......................... 49 3 29 Friends abide by law/Deviance Crosstabulation ................................ ................. 49 3 30 Friends Steal/Deviance Crosstabulation ................................ ............................. 49 3 31 Friend Detained/Deviance Crosstabulation ................................ ........................ 50 3 32 Friends Compulsory Labor/Deviance Crosstabulation ................................ ........ 50 3 33 Take Part In Visits/Deviance Crosstabulation ................................ ..................... 50 3 34 Attend Lectures/Deviance Crosstabulation ................................ ......................... 50 3 35 Take part public security/ deviance crosstabulation ................................ ........... 51 3 36 Take part OTHER/deviance crosstabulation ................................ ...................... 51 4 1 Parent attachment/offender status model summary ................................ ........... 64 4 2 Parent attachment/offender status ANOVA ................................ ........................ 64 4 3 Parent attachment/offender status coefficients ................................ ................... 64 4 4 Education attachment/offend er status model summary ................................ ...... 64 4 5 Education attachment/offender status ANOVA ................................ ................... 64 4 6 Education attachment/offender status coefficients ................................ ............. 64 4 7 Get along with schoolmates/offender status model summary ............................. 65 4 8 Get along with schoolmates/offender status ANOVA ................................ ......... 65 4 9 Get along with schoolmates/offender status coefficients ................................ .... 65 4 10 Commitment/offender status model summary ................................ .................... 65 4 11 Commitment/offender status ANOVA ................................ ................................ 65 4 12 Commitment/offender status coefficients ................................ ............................ 65

PAGE 9

9 4 13 Involvement/off ender status model summary ................................ ..................... 66 4 14 Involvement/offender status ANOVA ................................ ................................ .. 66 4 15 Involvement/offender status Coefficients ................................ ............................ 67 4 16 Peer Association /offender status Model summary ................................ ............. 67 4 17 Peer Association/offender status ANOVA ................................ .......................... 68 4 18 Peer Association/offender status coefficients ................................ ..................... 69 4 19 Parent attachment/deviance model summary ................................ ..................... 69 4 20 Parent At tachment/ Deviance ANOVA ................................ ............................... 70 4 21 Parent attachment/deviance Coefficients ................................ ........................... 70 4 22 Education Attachment/ Deviance Model summary ................................ ............. 70 4 23 Education/deviance ANOVA ................................ ................................ ............... 70 4 24 Education/deviance coefficients ................................ ................................ ......... 70 4 25 Get along with schoolmates/deviance Model summary ................................ ...... 70 4 26 Get along with schoolmates/deviance ANOVA ................................ ................... 70 4 27 Get along with schoolmates /deviance coefficients ................................ ............. 71 4 28 Commitment/deviance model summary ................................ ............................. 71 4 29 Commitment/deviance ANOVA ................................ ................................ .......... 71 4 30 Commitment/deviance coefficients ................................ ................................ ..... 71 4 31 Involvement/deviance model summary ................................ .............................. 72 4 32 Involvement/devi ance ANOVA ................................ ................................ ........... 72 4 33 Involvement/deviance coefficients ................................ ................................ ...... 72 4 34 Peer Association/deviance model summary ................................ ....................... 73 4 35 Peer Association/deviance ANOVA ................................ ................................ .... 74 4 36 Peer Association/deviance coefficients ................................ .............................. 75 4 37 All indepen dent variables/ offender status model summary ............................... 76

PAGE 10

10 4 38 All independent variables/offender status ANOVA ................................ ............. 76 4 39 All independent variable s/offender status coefficients ................................ ........ 76 4 40 All independent variables/deviance model summary ................................ .......... 77 4 41 All independent variables/deviance ANO VA ................................ ....................... 77 4 42 All independent variables/deviance coefficients ................................ ................. 78

PAGE 11

11 LIST OF ABBREVIATION S A NOVA Analysis of variance

PAGE 12

12 Abstract of Di ssertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy THE YEAR OF THE OX: SOCIAL BONDING THEORY AND JUVENILE DELINQUENCY IN A 1973 CHINESE BIRTH COHORT By Allison Leigh Timbs August 2012 Chair: Ronald Akers Major: Criminology, Law, and Society The following uses data collected from a birth cohort born in 1973 in the Wuchange district of China. y as well as variables pertaining to differential peer association were used as independent variables and offender status and deviance status were used as the dependent variables. In this cohort there were only 81 individuals who were identified as delinq uent. The sample used in this analysis consists of the offenders and 81 other individuals who were matched based on a short list of social criteria. Analysis show support for most social bonding elements. S upport is found for social bonding theory but it is more mixed and weaker than typically found in previous research done in American society There are also mixed findings for peer variables, and there is one finding on peer effects, rarely seen in previous research, that indicates that differential association with deviant peers is related to the respond ents more conforming behavior obviously counter to expectation The relevance of these findings for social bonding theory and future research is discussed.

PAGE 13

13 CHAPTER 1 INTRODUCTION T he purpose of my study is to examine juvenile delinquency in a non western used will come from an existing data set collected in 1991 92 and 2000 from a sample of Chinese youths living i n the Wuchange district of Wuhan, China. Originally began by Wolfgang, and modeled on the Philadel phia cohort study (Wolfgang, Figlio, & Sellin 1972) this project follows a birth cohort of all 5,341 persons who were born in 1973 in this district. (Friday Ren, & Weitekamp 2003) The timing of this project provides a rather unique perspective for researchers. At the time this group of children was coming of age, China was undergoing massive social and political change. (Friday 20 05) Beginning in the mi d 1900 s the Chinese government began instituting policies that went against traditional Chinese values. One such policy this law called for monogamy within marriages, free choice of spouse without parental interference, equal inheritance rights for male and female children, equal protection for children of both sexes (including illegitimate children), and stipula tions for divorce (Friday 2005). To westerners this law does not seem all that objectionable or shocking; however, in China these dictates were met with a great deal of resistance. Not only did these laws break from age old tradition; they allowed the government to take control of the institution of marriage. Before these laws, matters related to marriage and family and the elders in the family. child policy of 1972, which ordered that couples in urban areas could be licensed for o nly one child.

PAGE 14

14 s social benefits and parents we re forced to pay fines. (Friday, 2005) In 1980, the law was a pplied to all of China. (Friday, 2005) As a result of this, the you ths in this sample tended to come from smaller families. Also in 1980, China revised the marriage laws to raise the legal age of marriage to 20 (females) and 22 (males), regulations regarding divorce were also loosened, both males and females became respon sible for birth control, violence, bigamy, and desertion was outlawed, and stipulations regarding care of the elderly were outlined. The revised marriage laws also expressly forbade ar ranged marriages. (Friday 2005) For the youth in this sample the effe cts and ramifications of these laws were being felt all around them. Matters that were once entirely family domain were being snatched away and regulated by the government. Regardless of the official intent of these laws, the effect was to push Chinese s ociety towards a more western model. The significant and rapid change this society was experiencing makes it a unique opportunity for empirical examination. This cohort was three years old when the Cultural Revolution ended and the oming into being. This generation found themselves in a different being were causing shifts towards an open market model and young people were now competing for their place in the system. By no means did China morph into a capitalist society; but efforts were made to trade on a global level and that in turn caused some capitalistic traits became apparent. (Friday 2005) The data provided in this project provide insigh this time period. Other literature has captured Asian or Asian American samples; but

PAGE 15

15 this sample i s the only known sample that captures the experiences of the youth who grew up in the midst of these chaotic social and political changes that were occurring in (Friday, 2005) Although these political changes undoubtedly moved the operation and cultural traditions of the Chinese family and the relations hip of individuals to society toward western culture, there still remained differences in these areas compared to American and other western societies. Despite the changes, China remains one of the most structured communist regimes in the world. Becoming political structure of nation. The degree of this political control is something that cannot be forgotten when examining this data. For this nation, the political and social struct ures are one and the same. The overarching goal of society is the progression and survival of the political agenda. Survival of the political machine is at the heart of the social structure and conformity is given a level of importance that is simply not seen in Western society. Bonding theory places a great degree of importance on the maintenance of social ties and conformity and this analysis allows for the theory to be tested in a society that values conformity to the social norm above all else. The ch aracteristics of this sample and the nature of the data collected on it offer a unique opportunity to test the cross cultural generalizability of social bonding theory to youth in Chinese society. If measures of concepts and variables from other criminolo gical theories developed in America had been included, the same could be said a bout tests of those theories. Those data were not collected, but sufficient data were collected to allow a reasonable test of social bonding theory, and that is the focus of th is my study

PAGE 16

16 cohort study (Wolfgang, Figlio, & Sellin, 1972); however Wolfgang died before the project could be completed. Wolfgang did collect t he first wave of data in 1991 and 1992 when the juveniles were 18 years of age. (F riday, Ren, Weitekamp, & Leuven 2003) The second wave of data was collected by Friday, Ren, and Weitekamp in 2000 when the persons in the sample were 27 years of age. One of the most striking c haracteristics of the Wuhan data is the low delinquency rate among the juvenile population (1.5%). My study test s theory finds support with this particular sample in Chinese society. The proposed thrust of bonding theory is to account for conformity, and in a society with such a low rate of non conformity the theory should be supported as well if not better than has been found in the large number of studies testing the theory in American society. To date, social bonding theory has received a moderate amount of support from a very large body of research with western (primarily adolescent) samples. There is a growing segment of literature using various types of Asian samples to test social bonding theory; h owever, much of this research tends to focus exclusi vely on drug and/or alcohol use and resident Asian American populations in the United States are commonly used. In many of these cases, it was not possible to distinguish Chinese students from other Asia n students in the sample. Because of strict migration and immigration policies in China, the sample in this dataset consists s olely of Chinese students. My study not only add s to that growing body of literature, but it also shed s light on the dynamics of delinquency in a rapidly changing, but still traditional social climate like that of China in the 1980 s. Wolfgang also spearheaded another cohort study similar to the Philadelphia study; this

PAGE 17

17 study utilized a Puerto Rican birth cohort. The results from t his study yielded results that were in agreement with the finding of the Philadelphia study. (Nevares, Wolfgang, & Tracy, 1990) Social bonding (Hirschi, 1969) theory maintains that strong social bonds enhance conformity and control or restrict delinquent a nd deviance behavior. The theory purports to be uninterested in why individuals engage in delinquent and deviant behavior; instead the theory questions why people conform to social and legal norms and do not engage in delinquent and deviant behavior. Wi thin society there is a common value system that outlines the boundaries of acceptable behavior (1969, p.23). Hirschi argues that all creatures are inherently self serving; however in the case of humans this behavior is often labeled as delinquent and/or deviant and must be controlled (1969, p. 31). As such, acts that are considered to be delinquent and/or deviant are, in fact, natural and all would commit them unless constrained from doing so. What should interest theorists the most, then, are the mech anisms that are in place that stop delinquent and deviant behavior. Hirschi maintains that the answer to this questions lies within the concept of the social bond. Strong social bonds control delinquency and deviance and weak social bonds allow it to occu r. (Hirschi, 1969) According to Hirschi the social bond is comprised of four elements: attachment, commitment, involvement, and belief (1969, pp. 16 27). Taken together, these elements create a system that promotes conformity to the rules of society. In order to maintain good standing in this system, an individual must follow the rules. The first element of the bond, attachment to others, posits that being attached to others as individuals, in groups, and as participants in social institutions will serve to limit

PAGE 18

18 delinquent and deviant behavior. By and large, members of society believe in the relationship with others who believe in the rules. An individual who is n ot bonded to others is not at risk of jeopardizing social relationships by engaging in unacceptable behavior. (Hirschi 1969, pp. 18 22) The degree to which an individual desires to reach or maintain a certain social goal or status comprises the element o f commitment. Hirsch maintains that individuals who are committed to reaching their goals will not be willing to jeopardize them in order to engage in an unacceptable act with ephemeral rewards. These goals and 1969, p.21). The element of involvement speaks to the level of engagement individuals maintain within their social sphere. Legitimate social obligations create a framework around which individuals can center their life. The demands of these obligations will leave little room for deviant or delinquent behavior. (Hirschi 1969, pp.21 23) or endorsement of the law or rules of society, with the assumption that there is a common set of consensually agreed upon and supported cultural values for society as a whole. The reference in the theory is not to belief in subcultural or other values that run counter to general culture; rath er it is to the stronger or weaker acceptance of the general societal norms. If a person does not feel social rules or laws are valid, then they will be less likely to abide by them. Individuals who follow the rules have a respect for the common value sy stem and feel bound by it. Hirschi argues that an individual does not need to construct an elaborate system of justifications or excuses in order to

PAGE 19

19 engage in deviant or delinquent behavior; nor does he need to adopt a set of beliefs that run counter to those of conventional society. (1969, pp. 23 30) Justifications may be used, but underneath it all, justifications for breaking rules simply indicate weaker belief in conventional values. child relationship. If the bond to the parent is weak, then the chances of delinquent and deviant behavior are increased (Hirschi 1969, p.23). Hirschi outlines how strong bonds to parents deter delinquency and deviance in several ways. First, children who are well bonded to parents will not have as much time for deviance or delinquency since they delinquent and deviant activities take little time; so the amount of time s pent in the direct presence of parents is most likely a minor deterrent to deviance and delinquency. The this, he means whether or not the child will consider the react ions of the parents upon learning of the deviant or delinquent sical location of the parents. (Hirschi 1969, pp. 87 88) Psychological pre sence is heightened in children who believe their parents know where they are and what they are doing. Even though these children may not be directly supervised by their parents at all times, they do include their parents in their social lives by keeping them aware of their activities. This in turn serves to extend the in fluence of parental deterrence. (Hirschi 1969, pp. 88 91) Children who maintain active and open communication with parents also tend to exhibit lower levels of

PAGE 20

20 deviance and higher level s of conformity. Hirschi argues that children who can communicate their thoughts and ideas with their parents will be more likely to consider their reaction and opinion when deciding whether or not to engage in delinquent or deviant behavior. Hirschi als o noted that children of parents who provide an explanation of the reasons behind rules and dictates are less likely to violate those rules. (Hirschi 1969, pp. 90 92) In predicting the effect of the parental bond Hirschi points out that there must also b e a certain level of affection for the parents present. In order for the parents influence to be positive the child has to be concerned about maintaining the relationship. (Hirschi 1969, pp. 93 94) The bond to the parent is critical because it is one of the first bonds formed and will influence forming, strengthening, or weakening other bonds in the life and facilitate an overall weaker social bond (Hirschi 1977 p.333). The next type of bond Hirschi considers significant is that of the bond to school. The school occupies such an important role because it is the dominant social institution in the transition from childhood to adulthood. The strength of this atta chment is firstly the skills or opportunity to perform well in school are not as likely to feel a strong attachment to the institution. As such, the student will not feel compelled to follow the rules and norms put forth by the institution. Students who perceive an inability to compete in school will most likely lose interest and enjoyment in school. When this occurs, the chances for delinquency and deviance are further exacerbated since the attachment is weakened yet again. Poor performance and lack of interest also serves

PAGE 21

21 to reduce the level of involvement in school related activities. This lack of involvement is yet another way that the o verall social bond is weakene d. (Hirschi 1968, pp. 110 125) Students who underperform and feel no attachment to the school will most likely reject any coercive power the institution may hold. At this point, the student does not feel compelled to follow the rules of the school. (Hir schi 1969, p132) For the weakly attached student, the negative consequences of his or her present behavior is somewhat disassociated from the future. Students with higher social goals will see education as a necessary tool for achieving their goals and will not be willing to risk this potential by engaging in delinquent or deviant behavior. The weaker student is not likely to have such goals (or if they do have such goals they will in some way be cognizant of the futility of these aspirations) and will n ot view the school as a social bond worth maintaining. (Hirschi 1977, 334 336) This focus on attachment to the family and school raises the question of how the theory deals with attachment to friends and others of similar age, that is attachment to peers, and in particular the role of attachment or friendship with peers who engage in delinquency or deviance I will come back to this question below. Hirschi argues that delinquent or deviant peers come along after social bonds have been weakened and delinq uency and deviance is not only imminent but already engaged in as a behavior 1977, p. 336). Individuals choose their friends based on similar interests and backgrounds. It is only natural that children with weak social bonds and tendencies towards delinquency and deviance

PAGE 22

22 would seek out friends with si milar qualities. Therefore, such friendships have no i ndependen behavior. Attachment to peers is significant regardless of the delinquency or deviance status of the peers (1969 p. 229). Hirschi did concede (1969 p.230). As noted above much of the previous research on social bonding theory has found differential peer association to have significant effects, often stronger than the effects of bonding variables, on offending behavior. D ifferential peer ass ociations is usually understood in the context of s ocial l earning t heory (Akers, 1985; 1998) Briefly, social learning theory relies on four main sets of concepts/variables differential association, definitions, differential reinforcement and imitation. Differential association pertains to the groups that one associates with ( not just peers and friends but also family, secondary groups, and so on); these groups expose one to both behavioral and normative patterns that is expected to have effects on one Differential reinforcement pertains to whether or not behaviors are rewarded or punished, the amount, frequency, consistency, and severity with which the behaviors are rewarded or punished (Akers 1998) Imitation occurs when an individual engages in a certain behavior after seeing someone do it. Definitions are the attitudes, beliefs, and values about desirable/non desirable, acceptable/unacceptable behavior. (Akers & Sellers, 2009, pp.89 90) According to social learning theory, behavior (conforming and non conforming) is the result of the social learning process involving mainly these variables. While this dataset does have items allowing measures of p eer associations, it does not have

PAGE 23

23 sufficient other measures to allow consideration of other variables from general social learning theory. The refore, the current data do no t allow a test of a full social learning model. Although expectations regarding th e effects of differential association with delinquent peers derive from, and have direct relevance for, social learning theory, no effort is made here to present or test social learning theory as a competing perspective to social bonding theory. Rather the focus is on testing the validity of Social Bonding theory in a different societal and cultural context. Therefore, differential peer association is included in the analysis as an important variable based on prior research that should be included as a pos sible intervening or interacting variable with the bonding variables. The political str ucture of China during the 1970 s and 80s was not conduci ve to social science research. The sample used in my study analysis is one of the few snap shots we have into this culture during such a tum ultuous time. My study will contribute to current criminological knowledge by expanding the scope of inquiry beyond western samples and go beyond any previou s research on Chinese samples. Social bonding theory has been used with Asian American and Asian samples, but studies using Chinese samples are rare. This project is unique in that it uses a birth cohort fro m one Wolfgang and his colleagues administered extensive questionnaires which will a llow for a test of social bonding theory on three elements of the social bond: attachment, involvemen t, and commitment. This sample provide s a s social ity (or lack thereof) to a non western sample.

PAGE 24

24 CHAPTER 2 REVIEW OF THE LITERA TURE Empirical Support for Social Bonding Theory Social bonding theory has received a great deal of empirical attention from scholars. In 1993 Kempf conducted a meta analysis o bonding theory. She identified 71 studies that had been conducted from 1970 1991. Only 17 of the studies identified by Kempt used all four elements of the bond. Of the four, attachment was most commonly measured. Fifty studies measured attachment to parents in some way (p. 154). Thirty nine measured attachment to peers (p. 157). Twenty nine studies measured attachment to school; but none in the way Hirschi originally did (p.158). The element of involvement was only me asured in 25 studies (p. 159); however this element was cited by several to be rather ambiguous and was either skipped completely or combined with another element, principally commitment (p.153). Thirty seven studies included a measurement for belief; many of these did stick with (p. 161). Ultimately, Kempf found that the empirical support for bonding theory was inconsistent, and she was not able to draw any definitive c onclusions about the empirical status of the theory. She did identify three areas of concern regarding bonding theory: the overall generalizability of the theory, the theories inability to explain certain forms delinquency, and the applicability of the th eory across various age, race, and gender groups. However, she did argue that these empirical weaknesses could have more to do with the way the theory has been tested and the wide range of operationalizations that she found. (Kempf, 1993) Akers and Selle rs (2009) reviewed the state of empirical knowledge regarding social bonding theory and came to the

PAGE 25

25 overall conclusion that the theory has received some degree of support; however, this support was in the low to moderate range. Hirschi did not discuss the role of religion; however several studies of bonding theory have included religious involvement as a measure of social bonding. Baier and Wright (2001) found in a meta analysis of 60 studies that the presence of a religious faith and involvement in churc h or other religious activities served to insulate juveniles from delinquency in every study examined. Similar results were also found in a meta analysis conducted by Johnson, Spencer, Larson, and McCullough (2000). Agnew and Peterson (1989) examined the i mpact of the social bond within the context of leisure activities and found that most types of organized activities were protective of delinquency. Unorganized activities, such as spending time with friends unsupervised, appeared to facilitate delinquent activities. The presence of parents during these activities did not appear to have much impact, except in cases where the juvenile disliked the activity (in these cases rates of delinquency were higher). In other tests of bonding theory examining the par ental bond, Alarid and Cullen, (2000) used a sample of adult inmates and found that the parental bond was only a predictor of violent offending in females. Using the Youth in Transition Data, Rankin and Wells (1990) examined indirect and direct parental c ontrol and found that both types of control are protective against delinquency. However, this relationship changes when parental control becomes excessive. When this happens, juvenile delinquency is more likely to occur. Rankin and Kern (1994) found tha t parental attachment was most predictive in cases of families where both parents were present in the household.

PAGE 26

26 In other tests of bonding theory the impact of the attachment to school has been highlighted. Gottfredson (1986) examined the effectiveness o f a school based delinquency prevention program in the context of social bonding theory. This program used various methods to improve the level of attachment that students felt towards the school. Some initiatives included efforts to increase teacher invo lvement, bring community resources into the school and provide specialized services for high risk youths. He found that the program was useful for the general student population; however, high risk children were not significantly affected by the program. Rosenbaum and Lasley (1990) examined the relationship between attitudes toward school and found that a positive attitude towards school reduc ed delinquency in males only. In an effort to explain the nature of the relationship between parental and educati onal bonds with delinquency, Liska and Reed (1985) explored both types of bonds and found that a weak parental bond is predictive of delinquent behavior. However, delinquency is predictive of weak educational bonds. The theory contends that weak education al bonds will lead to delinquency, but Liska and Reed argue this is not the case. The possible explanation they offer is that delinquency causes a negative reaction from school officials and that serves to weaken the bond. The strength of the parental re lationship is not predicated on the behavior of the child like the educational relationship is. In essence, a weak parental bond will lead to deviance, which in turn will lead to negative reactions from the school and that will serve to weaken the attac hm ent to the school. (Liska & Reed1985) Krohn and Massey (1980) found that belief, attachment, and a combined element of commitment/involvement were able to expla in minor forms of delinquency. H owever,

PAGE 27

27 more serious forms of delinquency could not be explaine d with the s ame variables. Wiatrowski, Griswold, and Roberts, (1981) used one wave of the Youth in Transition data and found that some bonding variables were applicable in explaining delinquency. However, Agnew (1985) conducted a similar test of bonding t heory using the first and second waves of the Youth in Transition data and found that bonding variables were not significant in explaining the changes that occurred between waves one and two. Agnew found that children who were delinquent at wave two were likely delinquent at wave one as well. Overall, social bonding theory has been tested many times and almost always receives some degree of support. Bonding Theory and Asian Samples Social bonding theory was created in the American academic setting drawi ng on ideas that had developed over many years in Europe and the United States, and all initial tests of the theory were conducted with American samples and later samples from other western societies. There have been some tests of social bonding in other societies that indicate that social bonding theory has some cross cultural applicability. (Ching Mei, 1989; Junger & Marshall, 1997; Ellickson, Collins,& Bell, 1999; Ozbay & Ozcan, 2007) More recently there has been an emergence of research with a focu s on samples of Asian respondents. This has been done in one of two ways; samples using Asian American subjects and samples using Asian subjects residing in Asian countries. Much of the available literature relies on the first method and studies using Chi nese samples are exceedingly rare. The first part of this section will review some studies testing bonding theory with Asian American samples. Nagasawa, Qian, and Wong (2000) examined drug use among Asian A merican high school students. The sample broke d own as follows: 296 Japanese, 2,168

PAGE 28

28 Chinese, 669 Korean, 1,295 Filipino, 924 Southeast Asian, 525 Asian Indians, and 112 Pacific Islander. The total sample size was 5,989. Of this sample, Pacific Islander and Filipino students reported the highest drug u se and Chinese students reported the lowest drug use. For Chinese and Asian Indians familiar bonds are important; however, for Japanese, Korean, Filipino, and Pacific Islanders familial bond did not show to be all that important. For Southeast Asians, te achers were a very important sources of control. However, the effect of association with delinquent peers overshadowed all the social bonding variables for all groups, a finding that runs counter to social bonding or non effect on delinquency and that association with delinquent peers is the result of weakened social bonds. In a community based research study conducted by Le, Monfared, and Stockdale (2005) similar results were foun d. This particular project compared delinquency rates of 329 Chinese, Cambodian, Laotian, Mien, and Vietnamese youths. School attachment was found to be highly predictive for Chinese and Vietnamese youths. Parental attachment did not appear to be as imp ortant as social bonding theory would suggest. Delinquency of peers was found to be a strong predictor; however, its power varied among the various groups. Ultimately, social bonding theory received some support; but the effect of delinquent peers oversha dowed those variables (contrary to the theory). Using the same data, Chang and Le (2005) examined the effect of social bonds on academic achievement. They found that having delinquent peers did not directly predict poor grades. However, those with delin quent peers did tend to have negative attitudes towards school and this in turn lead to poor grades. Students with

PAGE 29

29 positive attitudes toward school tended to have higher grades in spite of negative peer associations. Attachment to parents did not appear to have any substantial effect on academic performance. Go and Le (2005) conducted a study of 1st and 2nd generation Cambodian students in Oakland, Ca lifornia They found that the presence of delinquent peers was predictive of delinquent behavior. The po wer of this association was further strengthened when age was taken into account. They found no significant differences between males and females. While these studies are suggestive of the extent to which social bonding explanations of delinquency are ge neralizable across Asian ethnic groups in the cultural context of the U.S., they tell us nothing about generalizing to delinquency in Asian countries with a different cultural context. One of the earliest attempts using an Asian sample was by Zhang and M essner (1996) who used a small delinquent sample of 57 and a control group of 86. These were students aged 15 18 years residing in Tianjin, China. The data were collected by the Nankai University and the Tianjin Communist Youth League in 1988. They foun d that juveniles who are strongly attached to family tend be strongly attached to school as well. These closer ties to school in turn yield lower delinquency rates. Students who reported a high degree of respect for school also tended to report lower del inquency. However, support for family and school attachment was mediated by the presence of delinquent peers. That is, having delinquent peers was more predictive of delinquency than any of the bonding measures tested and the direct effects of the bonding variables were reduced in the presence of the peer association variable.

PAGE 30

30 Davis, Tang, and Ko (2004) conducted a similar study with at risk youths in Hong Kong The sample consisted of 398 at risk and 320 control students aged 12 18. The juveniles in th e at risk group tended to be older than the control group; but the control group had more education and performed better in school. The at risk participants were found in group homes and probation/social work records. This study consisted of semi structu red interviews. Like previous studies, social bonding variables were moderately supported; but as found in previous research they were overshadowed by the presence of delinquent peer associations A strong familial bond was found to be somewhat protectiv e against delinquency, but its strength is diminished by the i nfluence of delinquent peers. When peer association, family bonds, and school attachment was combined over 50% of the variance was accounted for, mainly due to the effects of peers. Hwang and Akers (2006) examined the effects of peer and parental influences on drug use in South Korean youths and found that parental bonds could offer some degree of explanation for juvenile drug use. However, the presence of drug using peers offered a more power ful explanation. Tanioka and Glaser (1991) used a sample of Japanese youths and found that students with better attitudes towards school tended to have lower delinquency rates. However, they also found that teacher opinion only matter s in cases of minor delinquency.

PAGE 31

31 CHAPTER 3 METHODS Hypotheses Based on the propositions of the theory and the review of previous tests of social bonding theory with western and Asian samples the following hypotheses were used: Based on the propositions of the theory and the review of previous tests of social bonding theory with western and Asian samples the following hypotheses were used: H1: Non offenders will have stronger parental bonds than offenders H2: Non offenders will have a stronger bond to school than offenders H3: Non offenders will have stronger bonds with peers than offenders H4: Non offenders will demonstrate a stronger commitment to education than offenders H5: Non offenders will have a lower level of social involvement than offenders H6 : Non offenders wil l be less likely to associate with delinquent peers than offenders. H7: H8: H9 H10: H11: H12: quent peers than The Sample The Delinquency in a Chinese Birth Cohort dataset is a restricted dataset which was obtained through ICPSR. Marvin Wolfgang began the project in the Wuchange district of Wuhan, China which was intended to be a repli cation of his Philadelphia birth

PAGE 32

32 cohort study (Wolfgang, Figlio, & Sellin, 1972). All 5,341 children born in Wuchange in 1973 were included in the study. In 1990 Wolfgang began searching official data to identify juvenile offenders. Once juvenile offend ers were identified, a control sample was matched using head to head matching. The pairs were matched on characteristics occupation, and school district. Because the rate of juve nile delinquency was low (1.5%) within this group, the matched samples totaled 162. The final sample consisted of 152 males and 10 females. This makes it essentially a male sample and does not allow for consideration of gender issues. Researchers then con ducted extensive (Friday, Ren, & Weitekamp 2003) The interviews were given around the respondents 18th birthday and covered a broad range of topics. The rol e and influence of family, school, and peers were thoroughly examined. A team was sent back in 1995 to follow up; but the team was dismissed due to lack of funding and any data they collected were never recovered. (Friday Ren, & Weitekamp 2003) At this point, the project stagnated and the existing data were scattered among researchers. When the project was revived, a new research team consisting of Professors Friday, Ren, Weitekamp, Kerner, and Taylor was put t ogether and its first however identifying information was not available to follow up wi th the samples. Researchers did conduct a second wave of data which relied on public records

PAGE 33

33 stage another 41 offenders were identified which brought the total up to 122. Little is known about how these 41 new offenders fit into the overall picture. It can be determined that these new offenders are not part of the original 81 offenders used in this study. However, it is impossible to know if any of the new off enders are part of the control group. The new offenders are from the entire birth cohort, not just the sample used in this dataset I nterestingly, recidivism rates were virtually non existent with this sample. One of the seminal findings from the Philade lphia study (Wolfgang Figlio, & Sellin, 1972) is that a small group of high frequency repeat offenders commit a majority of crime. With the Chinese cohort, reoffending was extremely rare. Only 14 out of the final 122 offenders were arrested for a second offen s e. There is no record of any offender being arrested more than twice. (Friday 2005) When comparing offenders and non offenders there are similarities in marital status (49.8% and 55.2% respectively) and la ck of religion affiliation (only 2.3% of non offenders report any religious affiliation and no offenders reported a religious affiliation). Repeat offenders, however, are less likely to be married (35.3%) and more likely to have been divorced (5.9% compar ed to 1% and 2.9% for non offenders and offenders). Among non offenders, 19.4% are unemployed, compared to 41% of offenders and 70.6% of repeat offenders. Non offenders are also more likely to be employed in a professional type career. ( Friday, Ren & Weitekamp, 2003) The Variables This dataset is comprised of data gathered during in depth interviews with the matched samples in 1990. The survey given was very lengthy and captured numerous variables appropriate for t esting social bonding theory. Interviews were conducted by Wolfgang and his team on site in the Wuchang District. Criminal histories of the

PAGE 34

34 respondents were checked and verified through local police records. The surveys were done in person and the resu the data from these surveys was coded and electronically recorded. The records that the respondents; however, th e Chinese government required that copies of the original interviews be turned into the police bureau for analysis. The police analysis has never been revealed and it is unknown if the information submitted to them was de identified as well. Based on the information available about the provenance of this dataset it is unclear if respondents were granted any kind of anonymity during the interview process. The data allow for the concepts of attachment, commitment, and involvement to be measured. Unfortuna tely, there were no variables included which could be reasonably analysis, it is not uncommon for empirical tests of social bonding to exclude one or more elements of the so cial bond. (Kempf, 1993) For reasons noted above variables regarding delinquent peer associations will also be included. Because of the homogeneity of the sample, variables such as age, gender, race, and religion will not be considered in analysis. In most hypotheses tested offender/non offender status as measured by official record was used as the dependent variable. To obtain information about delinquency, police records were reviewed to identify individuals with a j uvenile delinquency record who met the demographic requirement of the study. Dividing offenders by seriousness of offense was initially considered; however the homogeneity of the offenses committed was not conducive to this type of analysis. Most of the

PAGE 35

35 variables used are similar to the ones Hirschi used in the Richmond Youth Project (1969 p. 35). T he data create some limitations on what items can be used. The most significant of these limitations is the lack of variance and missing data on several items. Because of this, it was necessary to use substitute variables to measure the desired theoretical constructs. The wording of some of these items do not always make perfect sense in the English wording, but are still serviceable for purposes of testi ng relationships. Some examples of this include the attachment to peers items, various involvement items, and some of the deviance items. Another point to consider with this ass umptions regarding the ord i nality of the data. While the answer choices are not ideal, it is important to consider that the questions and their answers have been translated from English to Chinese and back to English again and that may have the unintended effect of making the ord i nality less clear. Despite this, one is able to see a t least an implicit progression from the least desirable or pleasant to the most desirable or pleasant in the answer choices and this progression does lend ord i nality to these items This analysis relies on the ordnality implied by the original researchers. The original translations of the questions and answer choices are included in the appendices Parental A ttachment The concept of attachment was measured using several variabl es available from interviews with the matched samples (Tables 3 1, 3 2, and 3 3) original conception of social bonding theory, attachment to parents is the most critical bond (Hirschi 1969, p.85). Most of the attachment variables used were similar to those Hirschi used. Variables pertaining to parental attachment where scaled together and a

PAGE 36

36 Alpha was chosen because it provides an assessment of how closely related a set of variables are as a group. in any of the scales was 0.7. The final scale for parental attachment had a Cronbac h Alpha of .953 and consisted of 14 items. During the reliability analysis, the variable to not be a reliable indicator of parental attachment; however, perceived attitude of the father was found to a reliable indicator of parental at tachment. Responses regarding attitudes towards parents and perceived parental attitudes towards the child have been included in previous studies of social bonding theory. (Hirschi 1969, pp. 91 94). Variables pertaining to level of parental help with ho mework, communication with parents about daily events, school events, and the future in general were shown to be reliable indicators of parental attachment. Once the scale for parental attachment was computed, the scores were then categorized into four di fferent groups. Twelve of the variables used in the scale we re 3= Cares for everything and things must be done as said, 2= way to deal with you, 1= Cares only for my livelihood and seldom anything else, 1= Cares but always consults you. the lowest possible total was 15 and highest possible total was 29. This gav e a total range of 47 points. Scores of less than

PAGE 37

37 Between 3 and 8 were cod 20 was coded as Analysis will also be done which will use index scores for parental attachment instead of these categorizations. It is important to note that the coding for these, and several other variables used, ranged from negative to positive numbers. To make sure that the inclusion of negative numbers did not affect the reliability analysis the scales for commitment and educatio nal attachment were recoded using only positive numbers and the reliabil ity analysis was redone with identical results The original research team assigned these values and based on the checks of the reliability analysis it was not necessary to deviate fr om the original coding. Educational A ttachment school. relationships with teachers (1969, pp. 121 126). In this data, respondents were also asked to provide an overall rating of the school they attended. In his own research Hirschi found attitude towards school to be an important indicator of attachment (1969, pp. 121 123). Several variables were identified a s being potential indicators of .735 (Tables 3 4, 3 5, and 3 6) These variables pertained to the participants feeling about school, relationships with teachers, and their o verall opinion of their school. The about school ded as follows: 1= Monotonous and flat, and has no great pleasure, 1= highly significant, but has great pressure, 2= rich and colorful, and lations with 2=Do not care, 1= The relationships with quite a few of them are not good, 1= very ordinary with all

PAGE 38

38 middle school 2=did not go, 1 A relatively inferior school, 1= an ordinary school, and 2= A key 5 through 6. Scores of 5 through 1 and 2 were Analysis will also be done which will use index scores for educational attachment instead of the three listed categories. Attachment to P eers Measuring attachment to peers proved to be rather problematic with this dataset. In the original surveys several questions were asked regarding attachment to peers. However, when analysis began it became evident that data was completely missing for most of these variables. It is not clear if all respond ents simply refused to answer these questions or if the data was lost or removed somewhere in t his datasets complex history. Attachment to peers was measured using one variable. While this is not ideal, it is unavoidable with this data. The variable use concept of attachment to peers (1969, p.259). Responses to this variable were coded 1= Not in good terms with quite a few of them, 1=All are very ordinary or most of Differential Peer Association not important The variables he choose to measure attachment to peers did not measure the type of peers; rather, they simply measured the level of attachment only (1969, p.259). Past research indicates that the presence of delinquent peers is an

PAGE 39

39 important factor to consider when analyzing causative factors to delinquency and deviance. The presence of delinquent peers was measured through a number of questions respondents were asked about the behavior of their peers as indicated by iant behavior (such as stealing, playing truant, and being absent from work) or (detained by the police or being assigned to compulsory labor) Positive behaviors such as abiding the law were also included on the list. All of these variables, taken together, did not create a relia ble scale (Tables 3 8 through 3 13) In light of this, 6 variables relating to peer behavior were used individually in the analysis. Commitment to E ducation To measure the respon pertaining to the type of student the respondent desires to be and the desired educational level will be used. These items are very similar to those used by Hirschi to measure commitment in the Richmond Y outh Project (1969, p.171 and p. 179). Respondents were also asked about their current educational status and the reasons why they are no longer in school. These questions are also consistent with the items used by Hirschi in the Richmond Youth Project (1 969, p 172). N umerous variables were identified as being potential measures of commitment; however, reliability analysis .909. 1=Never thought of, 1= Middle pertained to the stage of education the respondent hoped to attain was coded as: 1=Jr Middle School, 2= Technical School, 3= Sr. Middle School, 4=Specialized Secondary

PAGE 40

40 School and 5= University. The third variable asked respondents how they would feel if 2=Happy and willing to leave school to earn, 1=Do not c are, to go to school or to leave school, 1= Disappointed, and 2= Very sorry and surely try every way to continue These variables yielded a score range between 2 and 10. The scale range with then divided into the categories. Scores of 2 through 1 were co (Tables 3 14, 3 15, and 3 16) Analysis will also be done which will use index scores for educational commitment ins tead of these categorizations. Involvement The third element of the social bond, involvement, also proved a challenge to measure. Initially, involvement was to be measured with a series of items regarding time spent doing homework, time spent reviewing lessons, the types and quantity of non curricular activities they engage in, their feelings about these activities, and the importance of maintaining these activities and other social contact. Unfortunately, the data for all of these variables was missing Again, it is unclear why the data was missing. However, other variables were available which could be used as indicators of The items used to measure level of involvement are in keeping with those used by Hirschi (196 9, pp. 189 195). In total, four variables were selected to determine how involved each respondent was in their social structure ( Tables 3 17 through 3 20) The five variables asked respondents if they participated in the following events: take part i n visits, public security activities, public lectures, and 1=No and 1= Yes= format.

PAGE 41

41 of these variables. For anal ysis, these variables will be used individually. Deviance In addition to dividing the groups by offender/non offender status, an index was created which measured deviant behavior that is not considered criminal or did not lead to the arrest and eventual p this index, non delinquent respondents were categorized based on their deviance status. Items on this scale included questions about punishment in school, fighting, truancy, stealing money from home, conning others out of money, stealing money from strangers, substance use, sexual relationships, g ambling, lying, and cheating. In total, 12 items were used to measure deviance. The respo 2= Yes, Rel iability analysis revealed four variables that are reliable The possible score range for deviancy was 8 through 4. For descriptive purposes, these items were made into an index and respondents were be categorized based on their deviance scores. Categories will include 8 through 4=very deviant 3 through 0= deviant a nd 1 through 4= not deviant which is presented in the reliability analysis presented in Tables 3 21, 3 22, and 3 23. Deviance status was used as the dependent variable and levels of social bonds were us ed as the independent variables (Tables 3 24 through 3 36) The face validity of the variables that were excluded by the reliability analysis is very strong. To ensure that important variables concerning deviance were not excluded another analysis was conducted and there were no significant changes from the initial

PAGE 42

42 devian ce findings. When considering this data, it is important to remember that the delinquency and deviance in this sample is relatively minor. The deviance variables that the reliability analysis excluded showed little to no variance. This lack of variance made these variables of very little use when explaining the deviance committed by the members of this sample. Another point that was considered was the difference in the delinquency status of the two halves of this sample. The non offe nders were selecte d out and the devian t behavior they engaged in was analyzed separately. The non offenders were similar in the regard that they tended to engage in minor forms of deviance.

PAGE 43

43 Table 3 1. Parental attachment reliability statistics Cronbach's Alpha Cronba ch's Alpha Based on Standardized Items N of Items .958 .962 13 Table 3 2. Parental a ttachment i tem t otal s tatistics Scale Mean if Item Deleted Scale Variance if Item Deleted Corrected Item Total Correlation Squared Multiple Correlatio n Cronbach's Alpha if Item Deleted Father help with study 9.69 116.674 .712 .669 .956 Mother help with study 10.56 121.167 .719 .699 .956 Talk to parents about school 7.26 119.150 .911 .881 .952 Parent help learn to read 9.22 110.149 .841 .772 .953 Pa rents help learn to write 9.02 113.850 .771 .920 .955 Parents help read books 9.33 111.774 .816 .806 .954 Parents help counting 9.94 114.543 .850 .881 .952 Parents help recite poems 9.06 112.729 .796 .921 .954 Parents help drawing 9.97 114.838 .857 .85 1 .952 Attitude f ather 10.33 124.482 .726 .613 .956 Discuss study 9.41 120.057 .759 .738 .955 Discuss occupation 8.95 119.314 .829 .765 .953 Discuss marriage 9.56 121.689 .734 .829 .956 Table 3 3. Parental Attachment /offender status Crosstabulation Offender Status Total Non Offender Offender Parent Attachment Very Weak Parental Attachment 13 22 35 Weak Parental Attachment 7 27 34 Moderate Parental Attachment 26 23 49 Strong Parental Attachment 35 9 44 Total 81 81 162

PAGE 44

44 Table 3 4. Education al Attachment Reliability Statistics Cronbach's Alpha Cronbach's Alpha Based on Standardized Items N of Items .735 .741 3 Table 3 5. Educational Attachment Item Total Scale Mean if Item Deleted Scale Variance if Item Deleted Corrected Item To tal Correlation Squared Multiple Correlation Cronbach's Alpha if Item Deleted Feel about school life 1.40 4.079 .576 .340 .637 Relations w/ teachers 1.22 3.429 .535 .286 .692 My middle school is 1.12 3.823 .581 .346 .624 Table 3 6. Education al Attachment /offender status Crosstabulation Offender Status Total Non Offender Offender Educational Attachment Scale Weak Education Bond 0 23 23 Moderate Education Bond 19 49 68 Strong Education Bond 62 9 71 Total 81 81 162 Table 3 7. Off ender Status/Attachment to Peers Crosstabulation Get Along with Schoolmates Total Not on good terms with quite a few of them All are very ordinary or most of them are ordinary All are very friendly or most of them are very friendly Offender Status Non Offender Offender 0 12 50 61 31 8 81 81 Total 12 111 39 162

PAGE 45

45 Table 3 8 Friends Absent From Work /offender status Crosstabulation Offender Status Total Non Offender Offender Friends absent From work No 76 56 132 Occasionally 4 7 11 Frequently 1 18 19 Total 81 81 162 Table 3 9 Friend Truant /offender status Crosstabulation Offender Status Total Non Offender Offender Friends play truant No 76 56 132 Occasionally 1 8 9 Frequently 4 17 21 Total 81 81 162 Table 3 10 Friends abid e by law /offender status Crosstabulation Offender Status Total Non Offender Offender Friends abide by law No 78 67 145 Occasionally 3 0 3 Frequently 0 14 14 Total 81 81 162 Table 3 11 Friends Steal /offender status Crosstabulation Offender St atus Total Non Offender Offender No 23 69 92 Occasionally 0 6 6 Frequently 58 6 64 Total 81 81 162 Table 3 1 2 Friends Detained /offender status Crosstabulation Offender Status Total Non Offender Offender friends detained No 81 66 147 Occas ionally 0 9 9 Frequently 0 6 6 Total 81 81 162

PAGE 46

46 Table 3 1 3 Friends Compulsory Labor /offender status Crosstabulation Offender Status Total Non Offender Offender Friends compulsory labor No 81 64 145 Occasionally 0 15 15 Frequently 0 2 2 Tota l 81 81 162 Table 3 1 4 Commitment to Education Reliability Statistics Cronbach's Alpha Cronbach's Alpha Based on Standardized Items N of Items .909 .910 3 Table 3 15 Commitment to Education Item Total Statistics Scale Mean if Item Deleted Scale V ariance if Item Deleted Corrected Item Total Correlation Squared Multiple Correlatio n Cronbach's Alpha if Item Deleted Student Becoming In School 2.30 6.682 .798 .637 .886 Stage of edu .33 5.988 .839 .706 .853 Not continue school 3.42 6.742 .822 .680 .8 68 Table 3 16 Commitment to Education / Offender Crosstabulation Offender Status Total Non Offender Offender Commitment Weak Commitment 15 53 68 Moderate Commitment 33 12 45 Strong Commitment 33 16 49 Total 81 81 162 Table 3 17 Visits /of fender status Crosstabulation Offender Status Total Non Offender Offender Take part in visits No 3 13 16 Yes 78 68 146 Total 81 81 162

PAGE 47

47 Table 3 18 Attend Lectures /offender status Crosstabulation Offender Status Total Non Offender Offender Attend lectures No 0 13 13 Yes 81 68 149 Total 81 81 162 Table 3 19 Public Security /offender status Crosstabulation Offender Status Total Non Offender Offender Take part public security No 32 81 113 Yes 49 0 49 Total 81 81 162 Table 3 20 Take Part in Other /offender status Crosstabulation Offender Status Total Non Offender Offender Take part OTHER No 61 80 141 Yes 20 1 21 Total 81 81 162 Table 3 21 Deviance Reliability Statistics Cronbach's Alpha Cronbach's Alpha Based on Stan dardized Items N of Items .840 .844 4 Table 3 2 2 Deviance Item Total Statistics Scale Mean if Item Deleted Scale Variance if Item Deleted Corrected Item Total Correlation Squared Multiple Correlation Cronbach's Alpha if Item Deleted Smoke and drink .79 6.080 .722 .527 .775 G ambling 1.11 5.714 .765 .601 .754 Fighting and mischief 1.71 8.083 .626 .430 .827 Talking dirt 1.37 6.893 .628 .405 .816

PAGE 48

48 Table 3 2 3 Deviance /offender status Crosstabulation Offender Status Total Non Offender Offend er Deviance Very Deviant 2 54 56 Moderately Deviant 43 27 70 Not Deviant 36 0 36 Total 81 81 162 Table 3 2 4 Parental Attachment/Deviance Crosstabulation Parent al Attachment Total Very Weak Parental Attachment Weak Parental Attachment Moderat e Parental Attachment Strong Parental Attachment Deviance Not Deviant 0 3 9 24 36 Deviant 17 6 27 20 70 Very Deviant 18 25 13 0 56 Total 35 34 49 44 162 Table 3 25 Commitment/Deviance Crosstabulation Commitment Scale Total Weak Commitment Mo derate Commitment Strong Commitment Deviance Not Deviant 0 9 27 36 Deviant 23 30 17 70 Very Deviant 45 6 5 56 Total 68 45 49 162 Table 3 26 Educational Attachment/Deviance Crosstabulation Education Attachment Scale Total Weak Education Bond M oderate Education Bond Strong Education Bond Deviance Not Deviant 0 0 36 36 Deviant 3 34 33 70 Very Deviant 20 34 2 56 Total 23 68 71 162

PAGE 49

49 Table 3 27 Friends Absent Work/Deviance Crosstabulation Friends absent from work Total No Occasionally F requently Deviance Not Deviant 36 0 0 36 Deviant 60 7 3 70 Very Deviant 36 4 16 56 Total 132 11 19 162 Table 3 28 Friends Truant/Deviance Crosstabulation Friends play truant Total No Occasionally Frequently Deviance Not Deviant 36 0 0 36 Deviant 62 3 5 70 Very Deviant 34 6 16 56 Total 132 9 21 162 Table 3 29 Friends abide by law /Deviance Crosstabulation Friends abide by law Total No Occasionally Frequently Deviance Not Deviant 36 0 0 36 Deviant 66 3 1 70 Very Deviant 43 0 13 56 Total 145 3 14 162 Table 3 30 Friends Steal/Deviance Crosstabulation Friends steal Total No Occasionally Frequently Deviance Light Not Deviant 2 0 34 36 Deviant 41 4 25 70 Very Deviant 49 2 5 56 Total 92 6 64 162

PAGE 50

50 Table 3 31 Fri end Detained/Deviance Crosstabulation Friends detained Total No Occasionally Frequently Deviance Light Not Deviant 36 0 0 36 Deviant 60 6 4 70 Very Deviant 51 3 2 56 Total 147 9 6 162 Table 3 3 2 Friends Compulsory Labor/Deviance Crosstabulati on Friends compulsory labor Total No Occasionally Frequently Deviance Not Deviant 36 0 0 36 Deviant 70 0 0 70 Very Deviant 39 15 2 56 Total 145 15 2 162 Table 3 33 Take Part In Visits/Deviance Crosstabulation Take part in visits Total No Yes Deviance Not Deviant 1 35 36 Deviant 4 66 70 Very Deviant 11 45 56 Total 16 146 162 Table 3 34 Attend Lectures/Deviance Crosstabulation Attend lectures Total No Yes Deviance Not Deviant 0 36 36 D eviant 1 69 70 Very Deviant 12 44 56 Total 13 149 162

PAGE 51

51 Table 3 35 Take part public security/ deviance crosstabulation Take part public security Total No Yes Deviance Light Not Deviant 8 28 36 Deviant 49 21 70 Very Deviant 56 0 56 Total 1 13 49 162 Table 3 36 Take part OTHER/deviance crosstabulation Take part OTHER Total No Yes Deviance Light Not Deviant 23 13 36 Deviant 63 7 70 Very Deviant 55 1 56 Total 141 21 162

PAGE 52

52 CHAPTER 4 ANALYSIS In order to assess the relative strengt h of the independent variables on the dependent variables with the groupings of offender/non offender and deviance status hierarchical linear modeling analysis was conducted Hierarchical linear modeling typically assumes that data are normally distribute d However, this data distribution does have homoscedasticity violations. Hierarchical linear modeling is used in spite of this violation because the power of the statistical technique reduces the potential for an inflated error rate. The data is so skewe d toward the lower zero frequency that the only meaningful measure of offending is to dichotomize the first dependent variable into 0/1 n on offender/offender. The natures of the responses to the independent variables also cause the data to be irregularl y distributed. Hierarchal linear modeling is advantageous in that it is built on the general linear model which uses an F test to test for significance of coefficients. This in turn allows for the interpretation of the partial correlations to determine p ercent variance of each independent variable on the dependent variable. Using percent variance allows for a more causal explanation than the odds ratio that would ultimately be produced using multiple logistic regressions In this sample the non offender sample was chosen based on their shared similarities with their counter parts in the offender group. Hierarchal linear modeling is able to adjust to the fact that the control group is not a truly random sample drawn from the entire cohort. While technical ly the independent variables are ordinal, they were treated as interval variables. This is not an uncommon practice because treating ordinal data as interval allows for the use of more sophisticated

PAGE 53

53 statistical methods and is advantageous in data sets wit h several variables. (Agresti & Finlay 1997, p. 15). Attachment and Offender Status Parental Attachment Hypothesis 1 holds that non offenders will have stronger parental bonds than offenders. In order to test this hypothesis hierarchical linear modelin g was used. When using offender status as the dependent variable and level of parental bonding as the independent variable a significance level of .000 was obtained. For the purposes of analysis, offender was coded as 1 and non offender was coded as 0. The Beta of .359 presented in Table 4 3 indicates that the relationship between the variables is negative. Meaning, offenders tend to have lower levels of parental bonds than non offenders. Based on this a nalysis Hypothesis 1 is supported. When the par ental attachment scale was deconstructed and index scores were used there was no statistically significant difference from the categorized parental attachment scale. Educational A ttachment Hypothesis 2 holds that non offenders will have a stronger bond t o school than offenders. The analysis using offender status as the dependent variable and level of educational attachment as the independent variable are presented in Tables 4 4, 4 5, and 4 6. This analysis produced a significance level of .000 was obtai ned A Beta of .717 indicated that the relationship between the variables is negative. Or, offenders tend to have lower levels of educational attachment than non offe nders. Based on this analysis H y pothesis 2 is supported. While performing this analysi s the educational attachment scale was deconstructed and index scores were used and there was no statistically significant difference in the findings.

PAGE 54

54 Attachment to P eers Hypothesis 3 holds that non offenders will have a stronger bond to peer than offende rs. When using offender status as the dependent variable and level of peer attachment as the independent variable a significance level of .000 was obtained To measure peer attachment a scale could not be created and a single item asking respondents to r ate the importance of getting along with peers was used. Table 4 9 presents a Beta of .399 which indicates that the relationship between the variables is negative. Meaning, offenders tend to have lower levels of peer attachment than non offenders. Base d on this analysis Hypothesis 3 is supported. Commitment Hypothesis 4 states that non offenders will demonstrate a stronger commitment to edu cation that offenders. Tables 4 10, 4 11, and 4 12 use offender status as the dependent variable and commitment to education as the independent variable These analysis yield a signifi cance level of .000 A Beta of .448 indicates that the relationship between the variables is negative. Meaning, offenders tend to have lower levels of commitment to education than n on offenders. The commitment to education scale was also deconstructed and index scores were used and no statistically significant difference was found. Based on this analysis Hypothesis 4 is supported. Involvement Hypothesis 5 states that non offenders will have a lower level of social involvement than offenders. To measure involvement a reliable scale could not be created. Instead, four individual variables were used. The item regard ing the tendance at public lectures was found to not be significantly related to offender status (Table 4 15) However, when items regarding the respondents

PAGE 55

55 participation in park visits was used a significance level of .005 and a Beta of .167 was produced. This indicates that non offenders are more likel y to engage in park visits than offenders. The final model presented in Table 4 15 revealed three variables that produced significant results. When the items pertaining to attendance at events concerning public security was analyzed, a significance level of .000 and a Beta of .591 resulted. This indicated that non offenders are more likely to attend these events than offenders. par t .013 and a Beta of .154 was produced. This would indi cate that non offenders are results, there is partial support for Hypothesis 5 Differential Peer Association Hypothesis 6 holds that offenders will be more likely to asso ciate with delinquen t peers than non offenders. The analysis presented in Tables 4 16, 4 17, and 4 18 suggests that the variable riends a bide t he is not significant The item regarding friends being detained by police produc ed a significance level of 0.008 and a Beta of 0.16 which indicates that offenders are more likely to have friends who have been detained. Three other variables produced a significance level that was greater than 0.05, however they were included in the final model because the B eta is significantly f riends play absent from and compulsory The item concerning friends punished to compulsory labor yielded a significance level of 0.11 and a Beta of 172. The item concerning friends committing truancy yiel ded a significance level of .055 and a Beta of 0.121 Friends missing work produ ced a significance level of .065 and a Beta of 0.13 8

PAGE 56

56 The analysis of these three variables indicates that offenders a re more likely to have friends who engage in this type of behavior. Having friends who steal produced a significance level of 0.000 and a Beta of .56 9 This indicates that non offenders are more likely to have friends who steal. This analysis runs c ounter to previous literature examining the effect of deviant peers on delinquent behavior. This finding does contradict Hypothesis 6 ; but is not necessarily contradictory to Social Bonding Theory. Interestingly enough, variables concerning positive behav ior o f peers did not produce a significant relationship with offender/non offender status. This analysis would suggest weak to partial support for Hypothesis 6 Analysis of All Independent Variables on Offender Status Individually, analyses of the indepen dent variables have shown varying degrees of support. In the final step of analysis, all independent variables were entered into a (Tables 4 37, 4 38, and 4 39 ) In order t o determine the fit of the independent variable in the model the partial correlations were used. Partial correlations indicate the variable when the linear effect of anot (Bachman and Paternoster, 1997, p.522) The partial correlation for each variable was squared and the resulting number was converted in to a percentage by multiplying by 100. This percentage was then used to determi ne the amount of the variance that each independent variable accounts for. In the final analysis, there were four variables that, together, accounted for more than 50% of the variance. These variables were: educational attachment, having friends who ste al, parental attachment, and taking part in public security events. Educational

PAGE 57

57 attachment yielded a significance level of .001 and a partial correlation of .271, which squared equals 0.0734. This means that educational attachment accounted for 7.3% of t he variance in offender status. Friends who steal produced a significance level of .000 and a partial correlation of 0.287. This accounted for 8.24% of the variance in offender status. Parental attachment accounted for 16.6% of the variance in offender status with a significance level of .000 and a partial correlation of 0.407 Taking part in public security events accounted for 24.8% of the variance in offender status with a significance level of .000 and a partial correlation of 0.498 Friends who h ave been detained by the police produced a significance level of .024 and accounted for 3.53% percent of the variance in offender status. When all of the independent variables were analyzed together, the other independent variables did not yield statistic ally significant results or account for a significant amount of the variance. Deviance and Attachment Using reliability analysis variables with strong face validity were eliminated. Variables pertaining to punishment in school, truancy, stealing from h ome, conning money from others, rob b ing others for money and sexual relationships were eliminated in the reliability analysis. To make sure this did not affect the analysis, these variables were examined further. The variables pertaining to stealing mon ey from home, conning money from others, robbing others for money, and engaging in sexual relationships showed very little or no variance. The items re garding punishment in school, truancy in school and lying did have vari ance. However, it should be noted the items asking about truancy in school and being punished in school had 28 and 51 missing answers respectively. These variables were included with the other deviance variables that were

PAGE 58

58 not eliminated during the reliability analysis. In the overall mod el, including these variables did not change the findings. Deviance a nd Parental Attachment Hypothesis 7 deviants will have stronger parental attachment that In order to test this hypothesis multiple hierarchical linear modeli ng was used and the results of this analysis are presented in Table s 4 19, 4 20, and 4 21. When using level of deviance as the dependent variable and level of parental attachment as the independent variable a significance level of .000 was obtained. For t he purposes of analysis, offender was coded as 1 and non offender was coded as 0. A Beta of 0.563 indicated that the relationship between the variables is negative. This means that higher deviance scores indicate lower levels of parental attachment. Bas ed on this analysis Hypothesis 7 is supported. Both deviance and parental attachment were converted in to categorized variables for analysis. In order to be sure that these categorizations di d not affect the analysis, both of the variables were deconstruct ed and index scores based on 0 1 coding summing the 1 scores across the different items to produce an index score, were used in the analysis. No statistically significant differences were found in these analyses and the analysis with categorized variabl es was retained Deviance and Educational Attachment Hypothesis 8 holds that non deviants will have a stronger attachment to school than deviants. Table 4 24 presents a significance level of 0.000 and a Beta of 0.755 were obtained. This indicates that the relationship between the variables is negative and significant. Or, higher levels of deviance are associated with lower levels of attachment to school. Based on these findings, Hypothesis 8 is supported. Both

PAGE 59

59 deviance and educational attachment were made into categorized variables for this analysis. To make sure these categorization did not produce erroneous results, both of the scales were deconstructed and index scores were used in the analysis. No statistically significant differences were found in these analyses. Deviance and Peer Attachment Hypothesis 9 holds that non deviants will have a stronger bond to peers than deviants. When using deviance as the dependent variable and level of peer attachment as the independent variable a significance l evel of .000 was obtained (Table 4 27) To measure peer attachment a scale could not be created and a single item asking respondents to rate the importance of getting along with peers was used. A Beta of 0.410 indicated that the relationship between the variables is negative and significant. Higher levels of deviance are associated with lowers levels of peer attachment than non deviance. Based on this analysis Hypothesis 9 is supported. Again, the deviance scale was deconstructed and index scores were used and no statistically significant changes occurred. Deviance and Commitment to Education Hypothesis 9 t to Tables 4 28, 4 29, and 4 30 use level of delinquenc y as the dependent variable and commitment to educatio n as the independent variable and obtain a significance level of .000 This, along with a Beta of 0.638 indicates that the relationship between the variables is negative and significant. Meaning, h igher levels of deviance is associated with lower levels of educational commitment. Based on this analysis Hypothesis 10 is supported. Both deviance and commitment to education were converted into categorized variables for analysis. In order to be sure t hat these

PAGE 60

60 categorizations did not affect the analysis, both of the variables were deconstructed and index scores were used in the analysis. No statistically significant differences were found in these analyses. Deviance and Social Involvement. Hypothesis 11 of social involvement as the independent variable ( Tables 4 31, 4 32, and 4 33) Since it was not possible to create a scale me asuring social involvement, four separate variables were used. The variable regarding attendance of public lectures was not found to be significantly related to deviance levels. However, analysis of the other variables does indicate a significant relationship. Analysis of the variable concerning taking part in visits produced a significance level of 0.001 and a Beta of 0.209. This indicates that lower levels of deviance are associated with taking part in visits. Taking pa rt in public security produced a significance level of 0.000 when analyzed with deviance levels. A Beta of 0.552 indicated that the relationship between participation in public security events and levels of deviance is negative. Or, higher levels of del inquency are associated with lower levels of attendance at these types of events significance level of 0.011 and a Beta of 0.164. This would indicate higher levels of dev iance and lower rates of participation in these events. Based on the significance of four out of five involvement indicators, Hypothesis 11 is supported In the analysis of involvement variables the deviance scale was deconstructed and index scores were used and there was no statistically significant difference observed.

PAGE 61

61 Deviance and D ifferential Peer Association Hypothesis 12 create a reliable scale for differential peer associations individual indicato rs were used. Table 4 36 presents the seven indicators that were used in this analysis However, four of the variables were found to be not significant and the final model con sisted of thre e variables. The absent from abide the and The item concerning friends sentenced to compulsory labor yielded a significance level of 0.003 and a Beta 0.17 3 This suggests that lower levels of deviance are related to lower rates of compulsory labor sentences for friends. The item regarding truancy levels of friends produced a significance level of 0.010 and a Beta of 0.149. This indicates that lower devia nce levels are associated with lower rates of truancy among friends. The significance level of the analysis of friends who steal and levels of deviance is 0.000 and the Beta is 0.413. The negative Beta would suggest that lower levels of deviance ar e associated with higher rates of friends who steal. This is a finding that runs counter to Hypothesis 12 and previous literature regarding the role of peers. Out of the seven variables used to measure the presence of delinquent peers four of the variabl es did not suggest a sig nificant relationship. Of the three variables that did have an acceptabl e significance level, only two variables produced relationships in the direction indicated by the hypothesis. Based on these findings, null Hypothesis 12 cann ot be rejected.

PAGE 62

62 All Independent Variables and Deviance Status The final stage of analysis for the dependent variable of deviance status was to include all the independent variables in a Hierar chical Linear Modeling analysis ( Table 4 42 ) Again, the parti al correlations will be used to determine how well each individual variable fits into the model. The partial correlations were squared in order to determine the percentage of the variance that each variable accounted for. In this model, 5 variables were found to be significant. Those variables were: attendance at events pertaining to public security, educational attachment, having friends who are truant, Attendance at public events pert aining to public security yielded a significance level of .000 and accounted for 18.4% of the variance in the deviance model. Educational attachment accounted for 6.4% of the variance and produced a significance level of .002. Having friends who play trua nt produced a significance level of .008 and accounted for 4.9% of the variance in the deviance status model. Attendance at public the deviance model and produce a signific ance level of .008 Commitment to education produced a significance level of .05 and each accounted for 2.6% of the variance in the deviance model. The other independent variables, when analyzed together, did not prove to be statistically significant, nor did they account for a significant amount of the variance. Using deviance as an index score did not produce any significantly different findings. When examining the deviance variables the entire sample was used. To determine if offender s would differ fr om non offenders in terms of deviance, a similar analysis was conducted using on ly non offenders Overall, this analysis was very

PAGE 63

63 similar to the analysis which used the entire sample. In this analysis, educational attachment and commitment to education we re the only variables that produced an the expected direction and based on the partial correlations educational attachment and commitment to education accounted for 6. 3% and 7.8% of the variance respectively. The only noteworthy departure in this analysis was the change in the variable pertaining to attendance at events related to public security. In the overall sample, this variable was one of the most significant va riables. However, for the non offenders attendance at events related to public security only accounts for 3.16% of the variance. This is a drastic change from the 18.4% percent variance found in the analysis of the overall sample This role of the varia ble pertaining to attendance at public security events will be discussed in the final chapter.

PAGE 64

64 Table 4 1. Parent attachment/offender status model summary Model R R Square Adjusted R Square Std. Error of the Estimate 1 .359 a .129 .124 .46950 a. Predi ctors: (Constant), Parent Attachment Table 4 2. Parent attachment/offender status ANOVA b Model Sum of Squares df Mean Square F Sig. 1 Regression 5.230 1 5.230 23.728 .000 a Residual 35.270 160 .220 Total 40.500 161 a. Predictors: (Constant), P arent Attachment b. Dependent Variable: Offender Status Table 4 3. Parent attachment/offender status coefficients a Model Unstandardized Coefficients Standardized Coefficients t Sig. Correlations B Std. Error Beta Zero order Partial Part 1 (Constant ) .930 .096 9.722 .000 ParentAttachment1 .163 .034 .359 4.871 .000 .359 .359 .359 a. Dependent Variable: Offender Status Table 4 4. Education attachment/offender status model summary Model R R Square Adjusted R Square Std. Error of the Esti mate 1 .717 a .513 .510 .35096 a. Predictors: (Constant), Education Attachment Table 4 5. Education attachment/offender status ANOVA b Model Sum of Squares df Mean Square F Sig. 1 Regression 20.792 1 20.792 168.797 .000 a Residual 19.708 160 .123 Total 40.500 161 a. Predictors: (Constant), Education Attachment b. Dependent Variable: Offender Status Table 4 6. Education attachment/offender status coefficients a Model Unstandardized Coefficients Standardized Coefficients t Sig. Correlations B Std. Error Beta Zero order Partia l Part 1 (Constant) .745 .033 22.302 .000 Education Attachment .131 .010 .717 12.992 .000 .717 .717 .717 a. Dependent Variable: Offender Status

PAGE 65

65 Table 4 7. Get along with schoolmates/offender status mod el summary Model R R Square Adjusted R Square Std. Error of the Estimate 1 .399 a .159 .154 .46131 a. Predictors: (Constant), Get along with schoolmates Table 4 8. Get along with schoolmates/offender status ANOVA b Model Sum of Squares df Mean Square F Sig. 1 Regression 6.451 1 6.451 30.312 .000 a Residual 34.049 160 .213 Total 40.500 161 a. Predictors: (Constant), Get along with schoolmates b. Dependent Variable: Offender Status Table 4 9. Get along with schoolmates/offender status coefficients a Model Unstandardized Coefficients Standardized Coefficients t Sig. Correlations B Std. Error Beta Zero order Partial Part 1 (Constant) .800 .065 12.226 .000 Get along with schoolmates .274 .050 .399 5.506 .000 .399 .399 .399 a. Dependent Variable: Offender Status Table 4 10. Commitment/offender status model summary Model R R Square Adjusted R Square Std. Error of the Estimate 1 .448 a .201 .196 .44980 a. Predictors: (Constant), Commitment Table 4 11. Commitment/offender status ANOVA b Model Sum of Squares df Mean Square F Sig. 1 Regression 8.129 1 8.129 40.181 .000 a Residual 32.371 160 .202 Total 40.500 161 a. Predictors: (Constant), Commitment b. Dependent Variable: Offender Status Table 4 12. Commitment/offender status c oefficients a Model Unstandardized Coefficients Standardized Coefficients t Sig. Correlations B Std. Error Beta Zero order Partial Part 1 (Constant) .682 .046 14.977 .000 Commitment .060 .009 .448 6.339 .000 .448 .448 .448 a Dependent Var iable: Offender Status

PAGE 66

66 Table 4 13. Involvement/offender status model summary Model R R Square Adjusted R Square Std. Error of the Estimate 1 .694 a .482 .465 .36686 2 .693 b .480 .467 .36615 3 .691 c .478 .468 .36589 a. Predictors: (Constant), take p art OTHER, Take part in visits take part public security, attend lectures b. Predictors: (Constant), take part OTHER, Take part in visits take part public security c. Predictors: (Constant), take part OTHER, Take part in visits, take part public security Table 4 14. Involvement/offender status ANOVA d Model Sum of Squares df Mean Square F Sig. 1 Regression 19.505 5 3.901 28.986 .000 a Residual 20.995 156 .135 Total 40.500 161 2 Regression 19.452 4 4.863 36.274 .000 b Residual 21.048 157 .134 Total 40.500 161 3 Regression 19.347 3 6.449 48.171 .000 c Residual 21.153 158 .134 Total 40.500 161 a. Predictors: (Constant), take part OTHER, Take part in visits take part public security, attend lectures b. Predictors: (Constant), tak e part OTHER, Take part in visits take part public security c. Predictors: (Constant), take part OTHER, Take part in visits, take part public security d. Dependent Variable: Offender Status

PAGE 67

67 Table 4 15 Involvement/offender status Coefficients a Model Un standardized Coefficients Standardized Coefficients t Sig. Correlations B Std. Error Beta Zero order Partial Part 1 (Constant) .336 .151 2.228 .027 Take part in visits .097 .091 .116 1.073 .285 .207 .086 .062 Attend lectures .063 .100 .068 .628 .531 .295 .050 .036 Take part public security .322 .034 .591 9.359 .000 .659 .600 .539 Take part OTHER .103 .047 .138 2.181 .031 .349 .172 .126 2 (Constant) .293 .134 2.185 .030 .118 .133 .052 .884 .378 .000 .070 .051 Take part in visits .145 .049 .174 2.967 .003 .207 .230 .171 Take part public security .325 .034 .597 9.625 .000 .659 .609 .554 Take part OTHER .109 .046 .146 2.343 .020 .349 .184 .135 3 (Constant) .400 .057 7.044 .000 T ake part in visits .140 .049 .167 2.877 .005 .207 .223 .165 Take part public security .322 .034 .591 9.593 .000 .659 .607 .552 Take part OTHER .115 .046 .154 2.503 .013 .349 .195 .144 a. Dependent Variable: Offender Status Table 4 16. Peer Association /offender status Model summary Model R R Square Adjusted R Square Std. Error of the Estimate 1 .705 a .497 .460 .36847 2 .705 b .497 .464 .36725 3 .704 c .496 .466 .36654 4 .701 d .492 .465 .36681 5 .697 e .486 .463 .36759 6 .697 f .4 86 .466 .36661 a. Predictors: (Constant),, friends detained, friends abide by law friends compulsory labor, friends play truant, friends absent from work, friends steal b. Predictors: (Constant) friends detained, friends abide by law friends compulsor y labor, friends play truant, friends absent from work, friends steal c. Predictors: (Constant) friends detained friends compulsory labor, friends play truant, friends absent from work, friends steal d. Predictors: (Constant) friends detained, frien ds compulsory labor, friends play truant, friends absent from work, friends steal e. Predictors: (Constant), friends detained, friends compulsory labor, friends play truant, friends absent from work, friends steal f. Predictors: (Constant), friends det ained, friends compulsory labor, friends play truant, friends absent from work, friends steal,

PAGE 68

68 Table 4 17. Peer Association /offender status ANOVA g Model Sum of Squares D f Mean Square F Sig. 1 Regression 20.134 11 1.830 13.481 .000 a Residual 20.366 150 .136 Total 40.500 161 2 Regression 20.134 10 2.013 14.928 .000 b Residual 20.366 151 .135 Total 40.500 161 3 Regression 20.079 9 2.231 16.606 .000 c Residual 20.421 152 .134 Total 40.500 161 4 Regression 19.914 8 2.489 18.500 .000 d Residual 20.586 153 .135 Total 40.500 161 5 Regression 19.691 7 2.813 20.817 .000 e Residual 20.809 154 .135 Total 40.500 161 6 Regression 19.667 6 3.278 24.388 .000 f Residual 20.833 155 .134 Total 40.500 161 a. Predictor s: (Constant) friends detained, friends abide by law friends compulsory labor, friends play truant, friends absent from work, friends steal b. Predictors: (Constant) friends detained, friends abide by law friends compulsory labor, friends play truan t, friends absent from work, friends steal c. Predictors: (Constant) friends detained, friends compulsory labor, friends play truant, friends absent from work, friends steal d. Predictors: (Constant) friends detained, friends compulsory labor, friend s play truant, friends absent from work, friends steal e. Predictors: (Constant), friends detained, friends compulsory labor, friends play truant, friends absent from work, friends steal f. Predictors: (Constant),friends detained, friends compulsory la bor, friends play truant, friends absent from work, friends steal, g. Dependent Variable: Offender Status

PAGE 69

69 Table 4 18. Peer Association /offender status coefficients a Model Unstandardized Coefficients Standardized Coefficients t Sig. Co rrelations B Std. Error Beta Zero order Partial Part 1 (Constant) .723 .087 8.357 .000 Friends compulsory labor .214 .096 .153 2.221 .028 .328 .178 .129 Friends detained .230 .076 .199 3.013 .003 .300 .239 .174 Friends play truant .069 .049 .095 1.416 .159 .296 .115 .082 Friends abide by law .038 .059 .043 .636 .526 .270 .052 .037 Friends steal .283 .044 .546 6.492 .000 .626 .468 .376 Friends absent from work .093 .057 .124 1.630 .105 .342 .132 .094 2 (Constant) .723 .080 8.999 .000 Friends compulsory labor .214 .095 .153 2.252 .026 .328 .180 .130 Friends detained .230 .076 .199 3.029 .003 .300 .239 .175 Friends play truant .069 .048 .095 1.433 .154 .296 .116 .083 Friends abide by law .038 .059 .043 .639 .524 .270 .052 .037 Friends steal .283 .043 .546 6.595 .000 .626 .473 .381 Friends absent from work .093 .057 .124 1.635 .104 .342 .132 .094 3 (Constant) .679 .070 9.676 .000 Friends compulsory labor .233 .093 .167 2.506 .013 .328 .199 .144 Friends de tained .205 .073 .177 2.806 .006 .300 .221 .162 Friends play truant .088 .045 .122 1.944 .054 .296 .155 .112 Friends steal .289 .043 .558 6.788 .000 .626 .481 .391 Friends absent from work .106 .055 .142 1.917 .057 .342 .153 .111 4 (Constant) .683 .070 9.722 .000 Friends compulsory labor .240 .093 .172 2.579 .011 .328 .204 .149 Friends detained .194 .073 .168 2.672 .008 .300 .210 .154 Friends play truant .088 .045 .121 1.935 .055 .296 .154 .112 Friends steal .294 .042 .569 6.941 .000 .626 .488 .401 Friends absent from work .103 .055 .138 1.861 .065 .342 .148 .107 a. Dependent Variable: Offender Status Table 4 19. Parent attachment/deviance model summary Model R R Square Adjusted R Square Std. Error of the Estimate 1 .563 a 317 .313 .61801 a. Predictors: (Constant), Parent Attachment

PAGE 70

70 Table 4 20. Parent Attachment/ Deviance ANOVA b Model Sum of Squares df Mean Square F Sig. 1 Regression 28.420 1 28.420 74.410 .000 a Residual 61.111 160 .382 Total 89.531 161 a. P redictors: (Constant), Parent Attachment b. Dependent Variable: Deviance Table 4 21. Parent attachment/deviance Coefficients a Model Unstandardized Coefficients Standardized Coefficients t Sig. Correlations B Std. Error Beta Zero order Partial Part 1 (Constant) 3.125 .126 24.826 .000 Parent Attachment .381 .044 .563 8.626 .000 .563 .563 .563 a. Dependent Variable: Deviance Table 4 22. Education Attachment/ Deviance Model summary Model R R Square Adjusted R Square Std. Error of the Est imate 1 .755 a .570 .568 .49029 a. Predictors: (Constant), Education Attachment Table 4 23. Education/deviance ANOVA b Model Sum of Squares df Mean Square F Sig. 1 Regression 51.070 1 51.070 212.451 .000 a Residual 38.461 160 .240 Total 89.531 161 a. Predictors: (Constant), Education Attachment b. Dependent Variable: Deviance Table 4 24. Education/deviance coefficients a Model Unstandardized Coefficients Standardized Coefficients t Sig. Correlations B Std. Error Beta Zero order Partial Par t 1 (Constant) 2.508 .047 53.717 .000 Education Attachment .206 .014 .755 14.576 .000 .755 .755 .755 a. Dependent Variable: Deviance Table 4 25. Get along with schoolmates/deviance Model summary Model R R Square Adjusted R Square Std. Error o f the Estimate 1 .410 a .168 .162 .68244 a. Predictors: (Constant), Get along with schoolmates Table 4 26. Get along with schoolmates/deviance ANOVA b Model Sum of Squares df Mean Square F Sig. 1 Regression 15.014 1 15.014 32.237 .000 a

PAGE 71

71 Residual 74.517 160 .466 Total 89.531 161 a. Predictors: (Constant), Get along with schoolmates b. De pendent Variable: Deviance Table 4 27. Get along with schoolmates/deviance coefficients a Model Unstandardized Coefficients Standardized Coefficients t Sig. Co rrelations B Std. Error Beta Zero order Partial Part 1 (Constant) 2.581 .097 26.665 .000 Get along with schoolmates .419 .074 .410 5.678 .000 .410 .410 .410 a. De pendent Variable: Deviance Table 4 28. Commitment/deviance model summary Model R R Square Adjusted R Square Std. Error of the Estimate 1 .638 a .407 .403 .57614 a. Predict ors: (Constant), Commitment Table 4 29. Commitment/deviance ANOVA b Model Sum of Squares df Mean Square F Sig. 1 Regression 36.422 1 36.422 109.726 .000 a Residual 53.109 160 .332 Total 89.531 161 a. Predictors: (Constant), Commitment b. Dependent Variable: Deviance Table 4 30. Commitment/deviance coefficients a Model Unstandardized Coefficients Standardized Coefficients t Sig. Correlations B Std. Error Beta Zero order Parti al Part 1 (Constant) 3.184 .111 28.709 .000 Commitment .563 .054 .638 10.475 .000 .638 .638 .638 a. Dependent Variable: Deviance

PAGE 72

72 Table 4 31 Involvement/deviance model summary Model R R Square Adjusted R Square Std. Error of the Estimate 1 .678 a .460 .443 .55658 2 .678 b .459 .446 .55529 a. Predictors: (Constant), take part OTHER, Take part in visits take part public security, attend lectures b. Predictors: (Constant), take part OTHER, Take part in vis its take part public security Table 4 32. Involvement/deviance ANOVA c Model Sum of Squares df Mean Square F Sig. 1 Regression 41.206 5 8.241 26.603 .000 a Residual 48.325 156 .310 Total 89.531 161 2 Regression 41.120 4 10.280 33.338 .000 b Re sidual 48.411 157 .308 Total 89.531 161 a. Predictors: (Constant), take p art OTHER, Take part in visits, take part public security, attend lectures b. Predictors: (Constant), take part OTHER, Take part in visits take part public security c. Depend ent Variable: Deviance Table 4 33. Involvement/deviance coefficients a Model Unstandardized Coefficients Standardized Coefficients t Sig. Correlations B Std. Error Beta Zero order Parti al Part 1 (Constant) 1.598 .229 6.990 .000 Take part in vi sits .199 .138 .160 1.445 .151 .223 .115 .085 attend lectures .080 .152 .059 .527 .599 .318 .042 .031 take part public security .442 .052 .547 8.486 .000 .616 .562 .499 take part OTHER .174 .072 .157 2.426 .016 .361 .191 .143 2 (Constant) 1.543 .203 7.594 .000 Take part in visits .260 .074 .209 3.499 .001 .223 .269 .205 take part public security .447 .051 .552 8.723 .000 .616 .571 .512 take part OTHER .181 .070 .164 2.575 .011 .361 .201 .151 a. De p endent Variable: Deviance

PAGE 73

73 Table 4 34. Peer Association /deviance model summary Model R R Square Adjusted R Square Std. Error of the Estimate 1 .755 a .570 .539 .50647 2 .755 b .569 .541 .50529 3 .754 c .568 .543 .50419 4 .753 d .567 .544 .50349 5 .752 e .565 .546 .50273 6 .750 f .563 .546 .50233 7 .747 g .558 .544 .50342 a. Predictors: (Constant), friends detained, friends abide by law friends compulsory labor, friends play truant, friends absent from work, friends steal b. Predictors: (Constant) fr iends detained, friends abide by law friends compulsory labor, friends play truant, friends absent from work, friends steal, c. Predictors: (Constant) friends detained, friends abide by law friends compulsory labor, frie nds play truant, friends steal d. Predictors: ( Constant) friends detained friends compulsory labor, friends play truant, friends steal e. Predictors: (Constant) friends compulsory labor, friends play truant, friends steal f. Predictors: (Constant) friends compulsory labor, frie n ds play truant, friends steal, g Predictors: (Constant) friends compulsory labor, friends play truant, friends steal,

PAGE 74

74 Table 4 35. Peer Association /deviance ANOVA h Model Sum of Squares df Mean Square F Sig. 1 Regression 51.054 11 4.641 18.094 .000 a Residual 38.477 150 0 .257 Total 89.531 161 2 Regression 50.978 10 5.098 19.967 .000 b Residual 38.553 151 0 .255 Total 89.531 161 3 Regression 50.892 9 5.655 22.245 .000 c Residual 38.639 152 0 .254 Total 89.531 161 4 Regression 50 .744 8 6.343 25.021 .000 d Residual 38.786 153 0 .254 Total 89.531 161 5 Regression 50.609 7 7.230 28.606 .000 e Residual 38.922 154 0 .253 Total 89.531 161 6 Regression 50.419 6 8.403 33.302 .000 f Residual 39.112 155 0 .252 Total 89.5 31 161 7 Regression 49.996 5 9.999 39.456 .000 g Residual 39.535 156 0 .253 Total 89.531 161 a. Predictors: (Constant) friends detained, friends abide by law friends compulsory labor, friends play truant, friends absent from work, friends ste al b. Predictors: (Constant) friends detained, friends abide by law friends compulsory labor, friends play truant, friends absent from work, friends steal c. Predictors: (Constant) friends detained, friends abide by law friends compulsory labor, frie nds play truant, friends steal d. Predictors: (Constant) friends detained friends compulsory labor, friends play truant, friends steal e. Predictors: (Constant) friends compulsory labor, friends play truant, friends steal f. Predictors: (Constant) friends compulsory labor, friends play truant, friends steal g. Predictors: (Constant) friends compulsory labor, friends play truant, friends steal, h. De pendent Variable: Deviance

PAGE 75

75 Table 4 36. Peer Association /deviance coefficients a Model Unstand ardized Coefficients Standardized Coefficients t Sig. Correlations B Std. Error Beta Zero order Parti al Part 1 (Constant) 2.190 .119 18.417 .000 Friends compulsory labor 0.272 .132 .131 2.052 .042 .386 0.165 .110 Friends detained 0.067 .105 .039 .638 .525 .085 0.052 .034 Friends play truant 0.130 .067 .120 1.945 .054 .382 0.157 .104 Friends abide by law 0.053 .081 .041 .646 .519 .322 .053 .035 Friends steal .260 .060 .338 4.350 .000 .624 .335 .233 Friends absent from work 0.0 43 .078 .039 .556 .579 .373 .045 .030 2 (Constant) 2.187 .118 18.457 .000 Friends compulsory labor 0.266 .132 .128 2.020 .045 .386 .162 .108 Friends detained 0.072 .104 .042 .690 .491 .085 .056 .037 Friends play truant 0.131 .067 .121 1.967 .051 .382 .158 .105 Friends abide by law 0.054 .081 .042 .664 .508 .322 .054 .035 Friends steal .260 .060 .338 4.357 .000 .624 .334 .233 Friends absent from work 0.045 .078 .041 .582 .562 .373 .047 .031 3 (Constant) 2.191 .118 18.560 .000 Friends compulsory labor 0.285 .127 .137 2.236 .027 .386 .178 .119 Friends detained 0.079 .103 .046 .768 .443 .085 .062 .041 Friends play truant 0.125 .066 .116 1.903 .059 .382 .153 .101 Friends abide by law 0.061 .080 .047 .762 .448 .322 .0 62 .041 Friends steal .259 .060 .336 4.347 .000 .624 .333 .232 4 (Constant) 2.192 .118 18.603 .000 Friends compulsory labor 0.308 .124 .148 2.488 .014 .386 .197 .132 Friends detained 0.075 .103 .044 .731 .466 .085 .059 .039 Friends pl ay truant 0.135 .064 .125 2.109 .037 .382 .168 .112 Friends steal .261 .059 .340 4.406 .000 .624 .336 .234 5 (Constant) 2.166 .101 21.399 .000 Friends compulsory labor 0.319 .122 .154 2.622 .010 .386 .206 .139 Friends play truant 0.155 .062 .143 2.509 .013 .382 .198 .133 Friends steal .279 .055 .363 5.110 .000 .624 .380 .271 a. Dependent Variable: Deviance

PAGE 76

76 Table 4 37. All independent variables/ offender status model summary Model R R Square Adjusted R Square Std. Error of the Estimate 1 .851 a .725 .688 .28021 a. Predictors: (Constant), friends compulsory labor Take part OTHER, friends detained, Take part in visits friends abide by law friends play truant, Take part public security, friends absent from work, friends ste al, getting along with schoolmates Parent Attachment, Commitment, Education attachment, Attend lectures, Table 4 38. All independent variables/offender status ANOVA b Model Sum of Squares df Mean Square F Sig. 1 Regression 29.351 19 1.545 19.674 .000 a Residual 11.149 142 0 .079 Total 40.500 161 a. Predictors: (Constant), friends compulsory labor Take part OTHER, friends detained, Take part in visits, friends abide by law friends play truant, Take part public security, friends absent from wo rk, friends steal, getting along with schoolmates Parent Attachment, Commitment, Education attachment, Attend lectures, b. Dependent Variable: Offender Status T able 4 39. All independent variables/offender status coefficients a Model Unstandardized Coef ficients Standardized Coefficients T Sig. Correlations B Std. Error Beta Zero order Partial Part 1 (Constant) .476 .205 2.327 .021 Parent Attachment .174 .033 .382 5.314 .000 .304 .407 .234 Commitment .040 .048 .067 .821 .413 .403 .069 .036 Education attachment .199 .059 .279 3.360 .001 .669 .271 .148 Getting along with schoolmates .028 .040 .058 .702 .484 .503 .059 .031 Take part in visits .091 .074 .108 1.227 .222 .207 .102 .054 Attend lectures .093 .082 .10 1 1.131 .260 .295 .094 .050 Take part public security .222 .033 .408 6.835 .000 .659 .498 .301 Take part OTHER .069 .040 .093 1.736 .085 .349 .144 .076 friends absent from work .065 .046 .086 1.405 .162 .342 .117 .062 friends play t ruant .057 .039 .078 1.461 .146 .296 .122 .064 friends abide by law .030 .046 .034 .649 .517 .270 .054 .029 friends steal .138 .039 .266 3.573 .000 .626 .287 .157 friends detained .145 .063 .125 2.282 .024 .300 .188 .100 friends compulsory la bor .074 .079 .053 .932 .353 .328 .078 .041 a. Dependent Variable: Offender Status

PAGE 77

77 Table 4 40. All independent variables/deviance model summary Model R R Square Adjusted R Square Std. Error of the Estimate 1 .895 a .801 .774 1.59638 a. Predictors: (Cons tant), friends compulsory labor Take part OTHER, friends detained, Take part in visits friends abide by law friends play truant, Take part public security, friends absent from work, friends steal, getting along with schoolmates Parent Attachment, Commi tment, Education attachment, Attend lectures, Table 4 41. All independent variables/deviance ANOVA b Model Sum of Squares df Mean Square F Sig. 1 Regression 1456.451 19 76.655 30.079 .000 a Residual 361.877 142 2.548 Total 1818.327 161 a. Predi ctors: (Constant), friends compulsory labor Take part OTHER, friends detained, Take part in visits, friends abide by law friends play truant, Take part public security, friends absent from work, friends steal, Getting along with schoolmates Parent Atta chment, Commitment, Education attachment, Attend lectures, b. D ependent Variable: Deviance

PAGE 78

78 Table 4 42 All independent variables/deviance coefficients a Model Unstandardized Coefficients Standardized Coefficients t Sig. Correlations B Std. Error Beta Zero order Partial Part 1 (Constant) 2.554 1.166 2.191 .030 Parent Attachment .172 .187 .056 .921 .359 .597 .077 .034 Commitment .542 .274 .136 1.975 .050 .658 .164 .074 Education attachment 1.050 .337 .220 3.112 .002 .742 .253 .116 G etting along with schoolmates .146 .230 .044 .633 .528 .649 .053 .024 Take part in visits .384 .421 .068 .912 .363 .182 .076 .034 Attend lectures .127 .468 .021 .271 .787 .294 .023 .010 Take part public security 1.051 .185 .288 5.671 .000 .687 430 .212 Take part OTHER .609 .226 .122 2.691 .008 .422 .220 .101 friends absent from work .092 .262 .018 .351 .726 .427 .029 .013 friends play truant .598 .222 .123 2.695 .008 .405 .221 .101 friends abide by law .169 .261 .029 .647 .518 .346 .054 .024 friends steal .410 .219 .118 1.870 .064 .682 .155 .070 friends detained .230 .362 .030 .635 .527 .107 .053 .024 friends compulsory labor .792 .453 .085 1.748 .083 .414 .145 .065 a. Dependent Variable: Deviance

PAGE 79

79 CHAPTER 5 SUMMARY AND CONCLUSION My study western hiladelphia Birth Cohort Study and is made more unique by the fact that it reveals dramatic differences from the Philadelphia study. There is a limited amount of empirical research into bonding theory in Asian cultures and this dataset provides a fantasti c insight into Modern China that was closed to social researchers. The cohort in this study came of age during a time when China was changing and evolving very quickly and this dataset helps paint a picture of the state of mind of Chinese youths and their families during this time. Aside tested in a communist society where little is known about the effectiveness or dimensions of social bonds. Bonding theory places a high value on societal conformity and this analysis offers a unique chance to test the theory in a society were conformity is given the utmost importance. Contextualizing the meaning and significance of these findings in a society that prides itself on both its political prowess and secretiveness proves quite difficult. Many of the facts that would normally be useful in making sense of research findings are not available and in some cases are considered to be secrets of the state. The birth cohort t hat was used in this dataset was born in 1973 and by the time of their 18th birthday the delinquency rate was an astonishing 1.5%. This low rate of delinquency led the researchers to match each delinquent respondent to a non delinquent. These matches wer e based on gender, parental economic status,

PAGE 80

80 of the sample that w as used in this analysis was 162 While this small of a number only captures a small percentage of the cohort, it did allow for researchers to conduct in depth interviews with each respondent. The premise behind my study was that children who were classified as delinquent would have lower levels of social bonding as outlined ory (1969) and they would be more likely to have friends who engaged in deviant or delinquent behavior To further understand the mechanics of the relationship between socia l bonding theory and peer association respondents were also categorized by the le vel of deviance. All the respondents were asked questions about behaviors that were either deviant or illegal but not detected by law enforcement. Responses to these questions were used to create a deviance scale and hen analyzed with the independent variables concernin g social bonding and peer association When the dependent variable of offender status and the independent variables pertaining to social bonding and peer association were analyzed individually there di d appear to be support for social bonding theory and most of the variables pertaining to peer association were not proven to be significant and two variables went opposite of the expected direction. However, when all the independent variables were analyzed together a different picture emerged. The independent variable that accounted for most of the variance in the offender status was attendance at events pertaining to public security. Parental attachment and educational attachment were also significant in this analysis. The scale that was created for educational attachment and the indicator that

PAGE 81

81 was used for peer attachment did not prove to be significant when all the variables were taken together. When the dependent variable of deviance status and the independent variables concernin g social bonding and peer association were analyzed together there was also a change from the individual analysis In the individual analyses of these variables there was support for most social bonding variables and most va riables relating to peer association were found to have little to no significance or significance in the opposite of the expected direction. H owever, when analyzed together 5 variables were found to be significant. Attendance at events related to public security accounted for the greatest amount of the variance in the deviance model. Educational attachment and commitment to education also were found to be significant. The involvement variables per taining to attendance at if icant results. H aving friends who play ed truant in school was significant as well When looking at the variables individually there was support for hypotheses 1 5 and hypotheses 7 11 However, when models for offender status and deviance status are use d with all the independent variables this changes quite a bit. Hypotheses 1, 2, 8, and 10 found support and some of the variables used to measure hypotheses 5, 6, 11, and 12 found support. The results, when compared to previous tests of socia l bonding t heory and peer association are different than what the literature would suggest. Previous literature would suggest that the delinquent behavior or status of peers would be an important predictor or the most important predictor of offender or deviance sta tus. In the individual analysis, some indicators of de linquency status of peers produced results that

PAGE 82

82 run counter to previous research. This analysis would suggest that not offending and low levels of deviance are more predictive of having friends who st eal. The reason for this drastic departure from existing research is not immediately clear. One explanation may be a technical problem with the coding of the data. When the dataset was reassembled it is possible that these variables may have been coding incorrectly. Unfortunately the original data is not available to determine if this is the case or not. It is also possible that this finding is the result of a social and political structure in which the influence of peers is weaker than that of a western society. While there was s ome support for differential peer association as a predictor here, it is nowhere near the levels that have been seen in previous studies done in the United States and elsewhere. Previous studies have also tended to show a moder ate level of support for bonding theory. It is true that some variables proved significant, but the level of support is not quite what the hypotheses or previous literature suggested they may be. When analyzing offender status and deviance taking part i n events related to public security accounted for more of the variance than any other variable in the analysis. The reason why this variable produced such strong results may say more about the societal reaction to deviance by Chinese authorities at the ti me than the desires of the individual. T he exact disciplinary measures taken by the Chinese government in regards to juvenile delinquency are not known. It is entirely possible that youths who have been caught offending would be barred from attending eve nts related to public security. The data does not allow one to interpret if attendance at the events occurred before delinquency status or not. I t could be that the offender status was known by authorities and offenders were not allo wed to attend (seen as a privilege to attend or as a possible

PAGE 83

83 security issue if allowed to attend) and therefore produced the finding rather than the chance of attendance or non attendance at such meetings occurred prior to the offender status designation. If the opportunity to attend such meetings did occur before the In this latter case the finding would indicate that a gener al bond to society might be more relevant (compared to bonds to family, friends, and school ) in the kind of society characterizing China of that era than would be true in American society. Parental attachment came in next with 16.6% of the variance. Socia l bonding theory puts a great deal of emphasis on the parental bond and in this analysis this particular bond did not dominate. However, the analysis does show that parent bond is important. This finding should be expected in a society in which the famil y is still very important to the individual. The finding that attendance at meetings on public security suggests that the social system at the time of the study had a political system and public priorities of the state that had become paramount even in the lives of adolescents. When using deviance status as the dependent variable, parental bonding did not produce a statistically significant result. This finding regarding deviance status and f Social Bonding Theory and the majority of previous research. Educational attachment did produce significant result in both offender status and deviance status which does bode well for Social Bonding Theory. In each model some of the variables pertainin g to involvement and peer association showed significance The inability to create a reliable scale for

PAGE 84

84 involvement and delinquency status of peers may explain the disjuncture between hypothesized expectations and the results of the analysis. The majori ty of previous research has shown that social bonding variables receive moderate support, but that support is essentially undone when the presence of deviant peers is added to the analysis. That was not the case with my study A few of the variables for peer association proved to be significant, but they accounted for little of the variance in either the offender status or deviance status models. This departure from previous research is rather puzzling; however, the explanation for this might be found in the overall composition of this cohort. Only 81 or 1.5% of this cohort had been processed by the police and were considered delinquent. Wuhan is also a rather large urban area. The reason why delinquent peer groups are not showing up in this sample may be because there are not enough delinquent children living in each neighborhood and attending each school for these peer groups to form In the sample itself, 126 or 78% of respondents fell into the very or moderately deviant category. In the final anal ysis for the deviance model the differential peer association variables did not fare any better than they did in the offender status model. However, it is important to remember that the variables used to measure deviance consisted of mostly minor deviant behaviors such as swearing and these behaviors are treated differently than the ones used to measure differential peer association For example, in the deviance scale fighting and troublemaking were paired together into one question and in the differentia l peer association variable list these two things were asked separately. With this dataset, this disparity between the deviance scale and variables used to test differential peer association is unavoidable and the potential effects of this disparity must be recognized.

PAGE 85

85 If it is possible to conduct valid research in China today (which of course has undergone very significant economic but little political change since the time the data here were collected) this issue should be directly addressed in future r esearch. The assumption would be that the level of delinquency and deviance among Chinese adolescents in the present time would be greater and provide a wider range of variance. But that remains to be seen. Future research may reveal a more substantive explanation. As suggested, the findings may indicate there is something about the social/family system and political system and culture in China at the time of study (and perhaps even in contemporary time s ) that was sufficiently different than the societi es in which social bonding theory and the effects of delinquent/deviant peers had been tested previously. Bonding variables and a few peer variables do have effects that are in line with previous research, but neither has as strong effects as would be exp ected. Again, if possible, future research in China should include more direct indicators of political, family, and peer differences and similarities with those variables in Western Society. Even though this dataset is unique, it is not without its flaws The lack of variance in the dependent variables of delinquent and deviant behavior has already been discussed. Lack of variance also necessitated the substitution of some of the independent variables as well. The Chinese government confiscated the orig inal whole dataset and it has not been seen since. The dataset that was used in this analysis was put back together again when researchers picked up the project several years later. It is clear from examination of the data that some items did not make it into the final copy intact. The exact reason for this is not known. This missing data meant that some variables that would have most likely been stronger indicators could not be used.

PAGE 86

86 Despite missing a few variables, it was possible to test the hypothe ses outlined Another issue that must be considered when using this dataset is that the questions were originally written in English, translated into Chinese, and then later translated back to English. The original English versions of the questions are n o longer available and it is possible that these questions and the list of responses may have suffered during this process. Some of the items have awkward and not entirely clear wording in the current English version available for the dataset Additionall y the bonding model had to be excluded from this study because of the political nature of the community in which the research was carried out. Asking questions about belief in the rule of law and the legitimacy of the socia l structure would most likely not have yielded accurate answers due to fear of sanction. It is also important to note that much of the data regarding the behavior of friends and the respondents own deviance and delinquency were self report data. Self rep ort data in the best of circumstances can have issues concerning response validity. Respondents may lie about, over report under report or mistake the facts they are reporting to for various reason and this can degrade the integrity of the data In this case though, the respondents lived in a society with rigid state control and it is not likely that they had any expectation that answers would remain anonymous. One must nswers given to these questions. This fear could account for the lack of variance and missing data on some items. It could also have affected the measures of differential peer association. Worry that police or other authorities would have access to the d ata and identify not only themselves but also their friends could have a chilling effect on

PAGE 87

87 behavior Despite the actual and potential problem s, this dataset and my subs equent study is a valuable tool to the criminological field in that it presents sociological and criminological glimpses into one of the most rigidly controlled and the largest closed communist society on the planet In the past decade criminology and oth er social sciences have begun to explore transnational and cross cultural study rather aggressively, but China has continued to remain a source of mystery for researchers. There is plenty of research testing Social Bonding Theory, but the amount of resear ch using social bonding theory in a non western sample is limited and research using social bonding theory with a sample of Chinese respondents living in China is virtually non existent. Understanding how western theories fit into Chinese culture is somet hing that the current criminological knowledge base so rely lacking. My study is one of the first steps in filling that void. Future research using this cohort has already proven to be impracticable, based on the attempts of researchers to carry out a fol low up with this sample; however, as suggested above research in the power of western theories with Asian samples needs to be continued. Asia, China in particular, is quickly becoming the key player in a globalized world and more social science research i s needed to understand this growing and expanding culture. Conducting social science research in a communist country has, and will most likely continue, to be a challenge for researchers. However, in recent years researchers in other fields have been abl e to gain access and it is not outside the realm of possibilities that doors will open for the social science researcher as well. In

PAGE 88

88 the meantime, continuing on the growing trend of research in other Asian countries is essential to furthering the knowledg e base.

PAGE 89

89 APPENDIX A MEASURES OF BONDING Attachment: Educational 1. How do you feel about school life? a. Rich and colorful, and highly significant b. Monotonous and flat, and has no pleasure c. Highly significant but has great pressure d. Other 2. How do you feel abou t your relations with your teachers? a. Very good all of them b. Very ordinary with all of them c. Very good, but the relations with individual teachers are strained d. The relations with most of them are ordinary, but good with individual teachers e. The relations with quite a few of them are not good f. Do not care 3. The middle school you studied at is: a. A key school b. An ordinary school c. A relatively inferior school d. I did not go to middle school Parental 4. Is there anybody in your family who helps you in your study? (check all t hat apply) a. Father b. Mother c. Private teacher d. Relatives or friends e. None 5. Do you talk with your parents about things in the school? a. Almost every day b. Once or twice a week c. Once or twice a month d. Seldom or never 6. What is the usual attitude of your parents towards you? (Choose answer for mother and father) a. Cares for everything and things must be done as said b. you c. Cares only for my livelihood and seldom for anything else d. Cares for everything, but always consult you when anything crops up e. Cares for nothing and I can do what I like

PAGE 90

90 7. The topics discussed often among your family members are: 1 frequently 2 Occasionally, and 3 nil a. Family life b. Interesting things in the news c. d. Children e. Recreation and amusement f. Issue about marriage g. Current affairs and politics h. Literature and art i. Building up family fortunes j. Other Peers 8. How do you feel about getting along with your schoolmates? a. All are very friendly b. All are very ordinary c. Most of them are very good and individuals are ordinary d. Most of them are ordinary and individuals are very good e. Not in good terms with quite a few of them f. Commitment 1. What kind of student d id you think of becoming in school? a. Middle level b. Above the middle c. The best d. Never thought of it 2. Which stage of education do you hope to complete finally? a. Junior middle school b. Technical school c. Senior middle school d. Specialized secondary school e. A majoring fiel d in college f. University g. Graduate school 3. If you cannot continue schooling because of something unexpected, what kind of feeling will come over you? a. Happy, and willing to leave school to earn money b. Do not care, to go to school and to school are the same c. Dis appointed d. Very sorry and surely try every way to continue schooling Involvement Which of the following activities of social practice do you take part in while in school?

PAGE 91

91 1) Social survey __ 2) Looking after elderly persons of no family __ 3) Doing good d eeds for handicapped schoolmates or neighbors __ 4) Social activities of public interest (for example, donations of money and goods, support for renovation of the Great Wall, Rescue of pandas, assistance for the construction of former Chinese Soviet areas, relief to the people in disaster areas) __ 5) Visits (visits to models and war heroes, historical relics and heritage, exhibition on legal education, display on scientific and technological achievements, etc.)__ 6) Lectures (such as lectures on "study for the rejuvenation of the Chinese nation", "with motherland is my heart", etc.)__ 7) Activities on maintaining public security (such as assisting in maintaining traffic order, fighting against evildoers and evil deeds in disruption of social security, takin g part in volunteer publicity activities on streets, etc.)__ 8) Others___________________________________

PAGE 92

92 APPENDIX B MEASURES OF DIFFERENTIAL PEER ASSOCIATION Something about your companions who often play with you: Frequently Occasionally Nil Fighting Trouble making Absent from work Play truant Abide by law and behave oneself Acts of stealing and pickpocket Flirtations with the opposite sex Having been detained Compulsory labor one__twice__thrice___

PAGE 93

93 LIST OF REFERENCES Agnew, R. (1985). Social control theory and delinquency: A longitudinal test. Criminology, 23 (1), 47 61. doi:10.1111/j.1745 9125.1985.tb00325.x Agnew, R., & Petersen, D. (1989). Leisure and delinquency. Social Problems, 36 (4), 332 350. Agresti, A., & Finlay, B. (1997). St atistical methods for the social sciences Upper Saddle River, NJ: Prentice Hall. Akers, R., (1985). Deviant Behavior: A Social Learning Approach Belmont, CA: Wadsworth Akers, R. (1998). Social Learning and Social Structure: A General Theory of Crime a nd Deviance. Boston: Northeastern University Press. Akers, R., & Sellers, C. (2009). Criminological theories: Introduction, evaluation, and application (5th ed.). Los Angeles: Roxbury. Alarid, L., Burton, V., & Cullen, F. (2000). Gender and crime among felony offenders: Assessing the generality of social control and differential association theories. Journal of Research in Crime and Delinquency, 37 (2), 171 199. doi:10.1177/0022427800037002002 Bachman, R., & Paternoster, R. (1997). Statistics for crim in ology and criminal justice (2 nd ed.). Boston: McGraw Hill. Baier, C., & Wright, B. (2001). "If you love me, keep my commandments": A meta analysis of the effect of religion on crime. Journal of Research in Crime and Delinquency, 38 (1), 3 21. doi:10.1177/00 22427801038001001 Chang, J. (2005). The influence of parents, peer delinquency, and school attitudes on academic achievement in Chinese, Cambodian, Laotian or M ien, and Vietnamese youth. Crime & Delinquency, 51 (2), 238 264. doi:10.1177/0011128704273469 Ching Mei, L. (1989). The study of social learning and social bonding variables as predictors of cigarette smoking behavior among ninth grade male students in Taipei, Taiwan, the R epublic of C hina University of Oregon. Davis, C. (2004). The impact of peer family and school on delinquency: A study of at risk Chinese adolescents in Hong Kong International Social Work, 47 (4), 489 502. doi:10.1177/0020872804046255 Ellickson, P. L., Collins, R. L., & Bell, R. M. (1999). Adolescent use of illicit drugs other t han marijuana: How important is social bonding and for which ethnic groups? Subst use Misuse, 34 (3), 317 346. doi:10.3109/10826089909035649

PAGE 94

94 Friday, P. C. (2005). A C hinese birth cohort: Theoretical implications. Journal of Research in Crime and Delinquenc y, 42 (2), 123 146. d oi:10.1177/0022427804266561 Friday, P., Ren, X., & Weitekamp, E. (2003). Delinquency in a Birth Cohort in Wuchang District, Wuhan China, 1973 2000. (ICPSR version) [data file and codebook]. Ann Arbor, MI: Inter university Consortium fo r Political and Social Research [distributor] Doi: 10.3886/ICPSR03751 Friday, P. C., Ren, X., Weitekamp, E., & Leuven, K. U. (2003). Delinquency in a C hinese birth cohort, final report. Washington, DC: Go, C., & Le, T. (2005). Gender differences in Cambo dian delinquency: The role of ethnic identity, parental discipline, and peer delinquency. Crime & Delinquency, 51 (2), 220 237. doi:10.1177/0011128704273466 Gottfredson, D. (1986). An empirical test of school based environmental and individual intervention s to reduce the risk of delinquent behavior. Criminology, 24 (4), 705 731. Hirschi, T. (1969). Causes of delinquency Berkeley,CA: University of California Press. Hirschi, T. (1977). Causes and prevention of juvenile delinquency. Sociological Inquiry, 47 ( 3 4), 322 341. doi:10.1111/j.1475 682X.1977.tb00804.x Hwang, S., & Akers, R. L. (2006; 2006). Parental and peer influences on adolescent drug use in K orea. Asian Journal of Criminology, 1 (1), 51 69. doi:10.1007/s11417 006 9009 5 Johnson, B., De Li, S., L arson, D., & McCullough, M. (2000). A systematic review of the religiosity and delinquency literature: A research note. Journal of Contemporary Criminal Justice, 16 (1), 32 52. doi:10.1177/1043986200016001003 Junger, M., & Marshall I. H. (1997). The inter ethnic generalizability of social control theory: An empirical test. Journal of Research in Crime and Delinquency, 34 (1), 79 112. doi:10.1177/0022427897034001005 McCor d & F. Adler (Eds.), Advances in criminological theory (pp. 143 182). New Brunswick, NJ: Transaction Publishers. Krohn, M. D., & Massey, J. L. (1980). Social control and delinquent behavior: An examination of the elements of the social bond. The Sociologi cal Quarterly, 21 (4), 529 544. Le, T., Monfared, G., & Stockdale, G. (2005). The relationship of school, parent, and peer contextual factors with self reported delinquency for C hinese, Ca mbodian, L aotian or M ien, and V ietnamese youth. Crime & Delinquency, 51 (2), 192 219. doi:10.1177/0011128704273472

PAGE 95

95 Liska, A., & Reed, M. (1985). Ties to conventional institutions and delinquency: Estimating reciprocal effects. American Sociological Review, 50 (4), 547 560. Nagasawa, R., Qian, Z., & Wong, P. (2000). Social control theory as a theory of conformity: The case of Asian/Pacific drug and alcohol nonuse. Sociological Perspectives, 43 (4), 581 603. Nevares, D., Wolfgang, M., & Tracy, P. (1990). Delinquency in Puerto R ico: The 1970 birth cohort study Westport, CT: G reenwood Press. delinquency in the high schools of A nkara, T urkey International Journal of Offender Therapy and Comparative Criminology, 50 (6), 711 726. doi:10.1177/0306 624X05283525 International Journal of Offender Therapy and Comparative Criminology, 52 (2), 134 157. doi:10.1177/0306624X07309182 Rankin, J., & Wells, E. (1990). The effect of pa rental attachments and direct controls on delinquency. Journal of Research in Crime and Delinquency, 27 140 165. Rankin, J., & Kern, R. (1994). Parental attachments and delinquency. Criminology, 32 (4), 495 515. Rosenbaum, J., & Lasley, J. (1990). School community context, and delinquency: Rethinking the gender gap. Justice Quarterly, 7 (3), 493 513. Tanioka, I., & Glaser, D. (1991). School uniforms, routine activities, and the social control of delinquency in J apan. Youth & Society, 23 (1), 50 75. doi:1 0.1177/0044118X91023001003 Wiatrowski, M. D., Griswold, D. B., & Roberts, M. K. (1981). Social control theory and delinquency. American Sociological Review, 46 (5), 525 541. Wolfgang, M. E., Figlio, R. M., Figlio, R. M., Figlio, R. M., & Sellin, J. T. (19 72). Delinquency in a birth cohort Chicago: University of Chicago Press. Zhang, L., & Messner, S. F. (1996). School attachment and official delinquency status in the P eople's R epublic of C hina Sociological Forum, 11 (2), 285 303. doi:10.1007/BF02408368

PAGE 96

96 BIOGRAPHICAL SKETCH Allison Leigh Timbs completed a Bachelor of Science in Criminology at Arkansas Status University. At Florida State University she completed a Master of Science in Criminology in 2003. In 2006, she completed a Master of Arts in Poli tical Science at Arkansas State University.