   

 

 

 

 

 

 

   
   

 

 

 

 

   

 

758'8 Illlllllllllllllllllllllllllllllllllllllllllllllll

. “W, .n 3 1293 018209043

This is to certify that the

thesis entitled

IS THERE A RELATIONSHIP BETWEEN ATTACHMENTS T0
PARENTS, AND MALE ANTISOCIAL BEHAVIOR

presented by
James Earl Lyons II

has been accepted towards fulﬁllment
of the requirements for

M.S. ﬂegreein Criminal Justice- Urban
Studies

 

 

Major professor

Date 3/3 3/7?

0-7639 MS U is an Afﬁrmative Action/Equal Opportunity Institution

ﬁr v—T.~— _ '

H""V_V_" *-

 

 

LIBRARY

Michigan State
1 University

 

 

MACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINE return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE

DATE DUE

DATE DUE

 

AT? 0 1 AAA

M“; ”it

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1/93 WWW-9.14

IS THERE A RELATIONSHIP BETWEEN ATTACHMENTS TO
PARENTS, AND MALE ANTISOCIAL BEHAVIOR?

By

James Earl Lyons II

A THESIS

Submitted to
Michigan State University
in partial fulﬁllment of the requirements
for the degree of

MASTER OF SCIENCE
Department of Criminal Justice-Urban Affairs

1999

ABSTRACT

IS THERE A RELATIONSHIP BETWEEN ATTACHMENTS TO
PARENTS, AND MALE ANTISOCIAL BEHAVIOR?

By

James Earl Lyons II
The purpose of this study was to explore the connection between male attachments to
parents, and male antisocial behavior. A sample of 91 8 males ages 12-18 years from the
United States was used in the research. The study concluded that attachments to parents
were signiﬁcantly and negatively associated with antisocial behavior. Regression
analysis revealed; however, that Parental Attachment only explained 0% to 1% of the
variance in overall Antisocial Behavior. However, this study was a secondary analysis,
and validity of variables is a major limitation with reference to secondary analysis.
Therefore, additional research testing parental attachment and antisocial behavior

employing a primary data collection procedure is recommended.

Cepyright by
James Earl Lyons II
1999

For Tyler and Kyndal: Enjoy the little

things, for one day you may look back

and discover they were the big things.
Love Uncle J J

iv

ACKNOWLEDGMENTS

First, giving thanks to God for the blessing of waking up to “one more day”
everyday which allowed me to ﬁnish this project. I would like to thank my chair Dr.
Homer C. Hawkins for his patience when I would tell him day after day, “next week”
when he would ask me when I would be ﬁnishing. I would also like to thank the rest of
my committee, Dr. Sheila R. Maxwell and Dr. Charles Corley. Thank you for all of the
constructive criticism, which at times unnerved me but proved to be the tools I needed to
complete this thesis.

I want to say thank you to Alesha M. Price for taking the time to ensure that this
thesis was grammatically correct. To my fellow NCCU Eagle Jessica Davis: Thank you
for everything that you have done for me this year. I would still be almost in tears if it
were not for you.

Thank you Department of Urban Affairs for funding my education at Michigan
State University, and the School of Criminal Justice for supporting me. Temetra, I want
to especially thank you for the patience and dedication you have shown and given me
over the past two years. To my extended family James and Elizabeth Powell, thanks for
everything that you have done for me. Finally, I have to extend the ultimate thank you to
my parents James and Yvonne Lyons. Thank you for the you patience, love, and

understanding this last year. Ultimately this is for you.

TABLE OF CONTENTS

LIST OF TABLES .............................................................................. IX
KEY TO SYMBOLS AND ABBREVIATIONS ........................................... X
I. INTRODUCTION ...................................................................... l
A. The Problem
B. Research Objectives ............................................................... 2
C. Overview ............................................................................. 2
II. REVIEW OF RELATED LITERATURE ........................................... 4
A. Theoretical Framework-Introduction
to theory .............................................................................. 4
B. Social Control Theory ............................................................. 5
1. Elements of the Bond .......................................................... 5
2. Relations Among the Elements .............................................. 7
3. Attachments to Parents ........................................................ 8
C. Delinquency and the Failure of
Social Controls ..................................................................... 10
D. Social Control Theory: Gender
Differences and Delinquency ..................................................... 11
E. Social Control Theory: A Rural
and Urban Comparison ............................................................ 14

F. Social Control Theory: Speciﬁcations,
Replications, Suggestions, and

Criticisms ............................................................................. 15
1. A Longitudinal Study ......................................................... 15
2. An Empirical Test of Hirschi’s

Model ............................................................................ 17
3. Direct Parental Controls and Delinquency ................................. 18

4. A Partial Replication and Extension ....................................... 21

5. An Examination of the Elements

of the Bond ..................................................................... 23
6. An Elaboration of Hirschi’s Model .......................................... 25
7. Social Control Theory Revised ............................................. 29
8. Suggestion: A Synthesis of Social

Control Theory and Social Learning

Theory ........................................................................... 33
9. Criticisms of Social Control Theory ....................................... 35

G. The Continuity of the Parent- Child

Attachment ........................................................................ 37
1. The Continuity of Early Parent-

Child Relationships Into the Young
Adult’s College Years ......................................................... 38

vi

III.

IV.

2. The importance of Parental Attachments

Among a College and Inpatient Sample .................................... 40
3. Adolescent Attachment to Parents and
Peers .............................................................................. 41
H. Parental Management: A Summation of
Previous Studies ..................................................................... 43
1. Summary .............................................................................. 46
METHODS ............................................................................... 47
A. Sample .............................................................................. 47
B. Data Collection ...................................................................... 48
C. Data Analysis ........................................................................ 49
1. Univariate Analysis ............................................................ 50
2. Factor Analysis ................................................................. 50
3. Reliability Analysis ............................................................ 51
4. Statistical Signiﬁcance ...................................................... 51
D. Bivariate Analysis .................................................................. 51
l. Contingency Analysis ......................................................... 51
2. Chi-Square ...................................................................... 52
3. Pearson’s Product-moment Correlation
Coefﬁcient ( r) ............................................................... 52
E. Multivariate Analysis ............................................................ 53
1. Partial Correlation Analysis ................................................ 53
2. Multiple Regression ......................................................... 53
F. Conceptualization and Operationalization .................................... 54
1. Antisocial Behavior ......................................................... 54
2. Parental Attachment ......................................................... 54
G. Data Transformations ............................................................ 54
RESULTS .............................................................................. 58
A. General Description of the Sub-sample ....................................... 5 8
B. General Description of Antisocial Youth Behavior ............................ 59
C. General Description of Attached Youth ....................................... 61
D. The Computed Dependent Variable ............................................. 61
E. The Computed Independent Variable .......................................... 63
F. Contingency Analysis Results ................................................... 63
1. Total Attachment and Total Antisocial Behavior ........................ 63
2. Total Attachment and Acts of Illegal Drug ................................. 64
Distribution and Acts of Theft ............................................. 64
3. Total Attachment and Violent Antisocial Acts ........................... 64
4. Total Attachment and Acts of Robbery .................................... 64
5. Total Attachment and Assault Against
Authority Figures ................................................................ 64
6. Total Antisocial and the Various Demographic
Variables .......................................................................... 65
G. Pearson’s Product-moment Correlation and

vii

VI.
VII.
VIII.
IX.

Partial Correlation Coefﬁcient Analyses ......................................... 65

H. Regression Analysis Results ...................................................... 67
CONCLUSIONS ........................................................................ 70
A. Conclusions .......................................................................... 70
B. Limitations ........................................................................... 71
APPENDIX A ............................................................................. 75
APPENDIX B ............................................................................. 78
APPENDIX C ............................................................................. 82
REFERENCES ........................................................................... 84

viii

LIST OF TABLES

Table
1 Sold Illegal Drugs and Acts of Theft Index ............................................ 55
2 Robbery of Others (not student or parents ) Index .................................... 55
3 Violent Behavior Index ................................................................... 55
4 Assault on Authority Figures (parents and teachers .................................. 56
5 Parental Attachment Scale ............................................................... 5 6
6 Zero-order Correlation Matrix of Parental Attachments and Antisocial

Behavior .................................................................................... 82
7 Partial Correlations of Parental Attachment and Antisocial

Behavior .................................................................................... 66
8 Stepwise Regression Analysis ......................................................... 68
9 Regression Analysis ....................................................................... 68
10 Regression Analysis ....................................................................... 68

AIQ
ANOVA
ANTISOC

ASSAULT

ATTPAREN

ROBBOTH
SACQ
SOLDSTO
VIOLENT
DV

IV

SD

KEY TO SYMBOLS AND ABBREVIATIONS

Aspects of Identity Questionnaire
Analysis of Variance

Total Antisocial

Assault on Authority Figures (parents and teachers)
Attachment to Parents

Acts of Robbery

Student Adaptation to College Questionnaire
Sold Illegal Drugs and Acts of Theft
Violent Antisocial Acts

Dependent Variable

Independent Variable

Standard Deviation

Null Hypothesis

Hypothesis

Independent Variable

Dependent Variable

Explained Variation (Correlation Squared)
Attained Signiﬁcance Level

Alpha Level

Chi-Square

Correlation (Pearson)

Sample Size

t T-Statistic
81 g Signiﬁcance

,6 Standardized Regression Coefﬁcient

xi

INTRODUCTION

Will

The inquiry of j uvenile delinquency and the possible causes of this phenomenon
in America have evolved greatly. From casual assumptions to well organized scientiﬁc
studies, possible causes of delinquent behavior have been the interest of the
neighborhood community as well as the scientiﬁc community.

This topic was chosen because antisocial behavior appears to be a developmental
trait that begins early in life and often continues through adolescence, into adulthood
(Patterson, DeBaryshe, & Ramsey, 1989; Loeber, 1982; Olweus, 1979). For many
children, stable manifestations of antisocial behavior begin as early as the elementary
school grades (Fanington, Ohlin, & Wilson, 1986). There is a history of empirical
studies that have identiﬁed family variables as consistent indicators for early forms of
antisocial behavior (LeBlanc, 1994; Patterson et a1., 1989; Loeber, 1982; Loeber &
Dishion, 1983; McCord, McCord, & Howard, 1963; Glueck & Glueck, 1962). Through
studies, attachment or bonding to parents and parental variables are consistently found to
be related to delinquency, yet, the importance of parenting skills tends to be ignored in
most major theories of delinquency (Cemkovich & Giordano, 1987). This is best
characterized by the belief that parent-child interaction variables are not nearly as
important as peer interaction, school factors, and various structural factors such as intact
families (i.e. parents not divorced) in understanding delinquent behavior patterns
(Cemkovich & Giordano). Cemkovich and Giordano have asserted that family variables
such as parenting have been ignored because of the “guidelines” established by scholars

of the social sciences (1987: 296):

This has been attributed to a desire to a avoid “psychologizing.” That is, while
Criminologists include in their analyses such structural or social variables such as
social class, blocked opportunities, and peer relations, such variables as
personality characteristic and parent-child relationships are avoided because they
are “too psychological.” This is yet another example of the misguided loyalty to
artiﬁcial disciplinary boundaries which is far too characteristic of the social
solences.

This particular study is important because it will assess or determine the role
attachment to parents’ plays with reference to antisocial and pro-social characteristics in
male children. The potential signiﬁcance of this study is to show whether or not
attachments between parents and male children are major factors in predicting antisocial
behavior. If parental attachment is found to be related to antisocial behavior; then, a
more adequate knowledge of the possible causes of antisocial behavior would assist
policy makers and programmers in designing more effective policies and programs
regarding parent-child relations, and antisocial behavior.

Emamhﬂbjﬁsﬁxﬁ
The research objectives of this study are:
1. To determine whether an association exists between parental attachment and male
antisocial behavior.
2. To determine how much variance in antisocial behavior is attribute to parental
attachment.
919111931
Criminologists are concerned about family variables relative to the development
of early forms of antisocial behavior and later delinquency. This study of the American
youth population examines male behaviors as well as parental factors. Self-reported

delinquency and parent-child attachments were measured using the available variables in

the National Youth Survey (Wave II). The dataset contained information on male and

females in the American population; however, only the male cases were used. A measure
was made of the two variables self-reported delinquency and parent-child attachment.
This research is an important endeavor because follow up studies of antisocial
males show that as adults they ultimately contribute disproportionately t0 the incidence of
criminal activity, accidents, alcoholism, chronic unemployment, divorce, psychiatric
illness, and the demand on welfare services.
The following section explains social control theory in detail as well as numerous studies
that have utilized a social control approach. Also included in the literature review is a
section that discusses the importance of attachment and how it relates to children
becoming socially competent as young adults. The literature review also includes
previous work concerning parent-child interactions as dynamics of the family structure

and a section examining factors associated with being at risk for antisocial behavior.

CHAPTER TWO
REVIEW OF RELATED LITERATURE

Throughout the span of sociological research, several theories have emerged
regarding delinquency causation. In Robert Troj anowicz and Merry Morash’s Juvenile
W the authors discussed several theories of
delinquency causation. Travis Hirschi’s social control theory (1969) was found to have
the strongest empirical research (J unger & Marshall, 1997; Akers, 1994; Trojanowicz &
Morash, 1992; Agnew, 1985, 1991; Gardner & Shoemaker, 1989; Wiatrowski, Griswold,
& Roberts, 1981; Krohn & Massey, 1980; Johnson, 1979; Hepburn, 1977; Hindelang,
1973). Thus, it offers one of the most adequate explanations for delinquency causation.

This research project is centered on Travis Hirschi’s control theory. Social
control theory asks the question, “why do humans refrain from engaging in delinquent
behavior?” Social control theory incorporates four main elements that Hirschi theorized
to be the foundation of conventional behavior. Conventional behavior is conceptualized
as behavior that is socially acceptable and of high moral standards. Those elements are
attachment, commitment, involvement, and belief. Moreover, Hirschi theorized that
these elements work independently to socialize individuals in order to mold the
individuals obey the laws and norms of the society in which they live. This chapter is
dedicated to providing the reader with a complete overview of Hirschi’s social control
theory. This chapter also provides the reader with numerous past studies that have tested
the validity and efﬁcacy of social control theory. Towards the end of this chapter is a
section that examines the importance of early attachment to parents. This last section

demonstrates the effect of the early bond in later years.

4

WM

Travis Hirschi’s social control theory, also known as bonding theory, stresses
aspects of youths becoming bonded or socially integrated into the norms of society
(Champion, 1992). Control theory assumes that people will commit crimes if it is left to
their own will. The question that control theory asks then is, why do most people not
commit crimes. Control theory seeks to answer that question by focusing on elements
that Hirschi thought restrained individuals from committing crimes. These devices break
down in certain conditions that result in crime and uncontrollable behavior. Thus,
individuals, according to social control theory, commit crime because of the weakness of
devices restraining them from doing so, not because of the strength of devices driving
them to do so (V old, Bernard, & Snipes, 1988).

Control theory classiﬁes, describes, and proposes a process that orders and
speciﬁes the interrelationships among the elements of the bond (Akers, Krohn, Lon,
Lanza-Kaduce, & Radosevich, 1979).

ElememmftheBnnd

The element attachment is deﬁned as the effective bond that the youth forms with
signiﬁcant others. Social control theory proposes that if there is a lack of attachment to
signiﬁcant others, all characteristics that the deviant individual exhibits be attributed to
the lack of attachment. Norms and moral restraints are deﬁned and shared by members
of society; therefore, an individual who violates a societal norm and suffers from a lack
of a sufﬁcient bond to society is insensitive to the opinion of others and is not bound by
its norms (Hirschi, 1969). Thus, control theory suggests that attachment to any
conventional person, group, or institution outside of one’s self promotes moral behavior

(Durkheim, 1961).

Commitment is deﬁned as the individual’s actual or anticipated investment in
conventional activities. To the person committed to conventional lines of action, a risk
that exceeds the value of what is to be stolen is not worth stealing. The concepts of
commitment assumes that the organization of society is such that the interests of most
persons would be endangered if they were to engage in criminal acts (Hirschi, 1969).
Therefore, when an individual feels that his or her future status depends on present
actions, the individual becomes committed to conventional lines of action and is therefore
committed to conformity. Hirschi’s social control theory assumes that when an
individual is committed to conventional activities the individual chooses not to engage in
delinquent activities.

The element of involvement stresses the notion that juveniles of any social or
economic background can become susceptible to committing delinquent acts if leisure
time outweighs conventional recreational involvement (Matza & Sykes, 1961).
Involvement refers to involvement in conventional activities, which lead toward socially
valued success and status objectives. The quality of a youth’s participation in
conventional activities and his or her relationship to future goals and objectives is
important in preventing delinquency (Wiatrowski et a1., 1981). Time spent on homework
or other school related activities, for example, is viewed as future predictors to success in
attaining educational goals which are prerequisites to hi gh-status occupations. The
assumption that is linked with the element of involvement is that a person may be too
preoccupied engaging in conventional activities to ﬁnd time to engage in deviant
behavior (Hirschi, 1969). The element of belief is based on the assumption that al
individuals are socialized (perhaps imperfectly). This particular element asks the

question, “why do men violate the rules in which they believe?” Control theory

assumes that individuals do not develop systems of rationalization in order to justify
commission of acts they want to commit. Instead, the beliefs allow an individual to
commit deviant acts without motivation from a system of rationalization that explains and
justiﬁes the attachment of illegal means. Control theory also postulates that “there is
variation in the extent to which people believe they should obey the rules of society”
(Hirschi, 1969: 26). Therefor, the less a person believes he or she should obey the rules,
then the more likely he or she will violate the rules.

WWW

According to Hirschi (1969), any individual who has bonded with conventional
society through any one particular of the four aforementioned elements has probably
bonded similarly with reference to the other three elements as well. Control theory
assumes that an attachment to a conventional other and a commitment to achievement
and prosperity are expected to relate in a positive manner (Hirschi, 1969).

Commitment and involvement’s relationship is important to understand because
delinquent acts are events that take place at speciﬁc points in time. If an individual is
involved in a conventional activity or committed t lines of conformity, it is less likely that
her or she will engage in acts of delinquency. In other words, it is not deﬁnite that a
conventionally attached individual will conform. It is only assumed that the individual is
less likely to engage in delinquent behavior. Speciﬁcation of some of the conditions
necessary for delinquent behavior to occur oﬁen produces indeterminacy. Indeterminacy
is speciﬁed as lacking the ability to be measured. It is not correct to imply that an
individual will necessarily commit an act of deviance because of a lack of bonds to
conventional society. Therefore, it is more appropriate to conclude that an individual

weakly attached to conventional society is more likely to commit delinquent acts than

an individual with strong attachments to conventional society.
Attachmenttoﬁarems

Control theory assumes that crime is deterred greatly by the bond of affection for
conventional others. The stronger the bond is between parents and children, the more the
adolescent contemplates not committing the delinquent act. Parental ideals are passed to
their children through the emotional bond (McKinley, 1964). “In control theory, then,
attachment to parents becomes a central variable, and many of the variations in
explanations of this relationship are the basis of control theory” (Hirschi, 1969: 86). The
link between socialization and internalization of norms and attachment is difﬁcult to
determine; therefore, control theory assumes that the moral element is within the
attachment itself. The element attachment is easily seen to vary with the same person
over time. For example, if the attachment or bond between parent and child is weakened,
the probability of delinquent behavior will increase. If the bond becomes stronger the
probability for delinquency declines (Hirschi).

Attachment to conventional parents is assumed to work against the commission of
delinquent acts through direct control and psychological presence. The child who spends
a great deal of time in his or her parents’ presence is less likely to be faced with
opportunities to commit delinquent acts for one obvious reason. The reason being that
the child is constantly with or near the parent. However, the time it takes to commit a
delinquent act is miniscule, and direct control, according to Hirschi holds little or no
water in delinquency theory (Hirschi, 1969). Therefore, if a child is confronted with a
delinquent opportunity and declines to commit the illegal act because of internal
questions concerning parental approval or reactions, that parent is present

psychologically. Psychological presence is based on how much time parents spend

with their children or the extent to which the child interacts with the parent on a personal
level (Hirschi: 89-90):

The child is less likely to commit delinquent acts not because his parents actually

restrict his activities, but because he shares his activities with them; not because

his parents actually know where he is, but because he perceives them as aware of
his location. Following this line of reasoning, we can say that the more the child
is accustomed to sharing his mental life with his parent, the more he is
accustomed to seeking or getting their opinion about his activities. We can also
say that he is more likely to perceive them as part of his social and psychological
ﬁeld, and the less likely he would be to neglect their opinion when considering an
act contrary to law-which is, after all, potential source of embarrassment and/or
inconvenience to them.

Hirschi (1969) sampled 17,500 students entering 11 public junior and senior high
schools of that area in the fall of 1964. The sample was drawn as part of the Richmond
Youth Project. Hirschi hypothesized that the social control elements would correlate
negatively to delinquency.

Hirschi conducted a study integrating the social control principles. Self-reported
delinquency by the mother’s virtual supervision indicated that among the children with
low supervision, 55% reported committing two or more acts of delinquency, while 12%
of children with high supervision reported committing two or more delinquent acts. The
results for self-reported delinquency by intimacy of communication with the father
indicated that of the children, who reported the least intimate communication with their
father, 43% reported committing two or more delinquent acts. Five percent of the
children who reported committing two or more delinquent acts reported more intimate
communication with the father. Affectionate identiﬁcation with the father as it relates to
self-reported delinquency was measured by asking children the question, “Would you
like to be the kind of person your father is?” Of those children who answered, “In every

way,” 16% admitted to committing two or more delinquent acts. Of those who answered

“Not at all,” 38% admitted to committing two or more acts of delinquency. Attachments

9

to peers were measured by asking the children, “Have any of your close friends ever been
picked up by the police?” Of those who answered “No,” 7% reported committing two or
more delinquent acts. Of those who answered “Four or more,” 45% reported committing
two or more acts.

The section of Hirschi’s study that tested the commitment to conventional actions
indicated the following. When the question, “Do you smoke cigarettes?” was asked,
among those who answered, “Yes, I began before 13,” 48% reported committing at least
two acts of delinquent behavior. Of those who answered “No,” 12% admitted to
committing two or more delinquent acts.

The section of the study based on involvement indicated that 34% of the of the
children who reported spending one and a half hours or more devoted to homework
reported committing one or more delinquent acts. Sixty-four percent who reported
spending less than one half hour on homework reported committing one or more
delinquent acts.

The belief element’s results indicated that of those who answered, “Strongly
agree” to the question “I have a lot of respect for the Richmond police,” 12% reported
committing two or more acts of delinquency. Of those who answered, “Strongly agree,”
45% reported two or more acts of delinquency.

'l i ° v

Delinquency results when there is a breakdown or absence of internalized norms
and rules set by society. Reiss (1951) conducted a study in which he isolated a set of
social controls he thought to be related to delinquent recidivism. Upon isolation of the
social controls, Reiss sought to evaluate them as prognosticators of delinquent

recidivism. The researcher chose from each juvenile’s personal folder items that were

10

appropriate measures of social control in primary groups and social control by
community and institutions.

Measures of social control from the primary group included economic status of
the family and parental conﬂict. The economic status of the family as discussed by Reiss
(1951) is the ability of the family to provide for the needs of the child. The researcher
conceptualized that the economically dependent family has a lack of ﬁnancial stability,
which often leads to lack of respect for the supposed provider (usually the father).
Dependent economic status usually results in diminished parental authority over family
members. Loss of parental authority (especially among fathers) is greatest over the
adolescents of the family.

For Reiss’ study, parental conﬂict was hypothesized to have an effect on
delinquency. Parental conﬂict, as measured in this study involved parents of juvenile
delinquents who were together structurally but had conﬂicts within their structural unit.
Absence of a formal break does not necessarily mean that the social-psychological unit of
the family is intact.

The results indicated a statistically signiﬁcant association between economic
status of family and outcome on probation. There was a 27.5% failure rate for children
with parents who earned over $270 per month and a 53.8% failure rate for children whose
parents were on public assistance. The parental conﬂict measure indicated that children
who had parents that got along well had a 26.3% failure rate and those couples with gross
incompatibility had children that indicated a 42% failure rate.

S '1C 1 III '5 I n'm lDl'
This thesis deals with patterns concerning male antisocial behavior. Therefore it

is appropriate to include studies pertaining to social control theory and gender. Jill

ll

Rosenbaurn (1987) conducted an empirical examination of social control theory as
postulated by Hirschi (1969). The analyses focused on male and female delinquent
behavior by social class and three types of delinquent behavior; drug offending, property
offending and violent offenses, which served as the measures of delinquency. The
independent measures that were used are attachment to parents, peers, and school; class,
commitment to conventional activities, involvement in conventional activities, delinquent
friends, dating, and belief in the law (Rosenbaum).

The data for Rosenbaum’s study were derived from a self-reported questionnaire
administered to a sample of male and female adolescents in Seattle between 1977 and
1979. A stepwise multiple regression analysis technique employed in the study.

The results for property offending indicated that the explained variance for
females was 13% and 19% for the lower and upper classes respectively. For their male
counterparts the explained variation for the lower and upper class was 7% and 8%
respectively. The stepwise analysis for drug offending yielded stunning results for
females, especially the upper class females. The explained variation for drug offending
females was 47% for upper class females and 17% for lower class females. The
explained variation was 12% and 15% for their lower and upper class male counterparts.
The stepwise multiple regression table for violent behavior indicated slightly different
results from property offending and drug offending. The results for the lower class males
and females were slightly greater for the lower class males. The results were 6% and 5%
for males and females respectively. For the upper class males and females the results
were more consistent wit the preceding results, indicating 5% and 7% of the variance for
males and females (Rosenbaum, 1987).

The correlation for property offending and the independent measures indicated

12

negative associations between property offending and all of the independent measures
with the exceptions of delinquent friends and dating. This held true for both social
groups of males and the upper class females (attachment to peers did not enter the
equation for upper class females). The same held true for lower class females with the
exception of attachment to parent, which was found to correlate positively and highly
with property offending (Rosenbaum, 1987).

The correlation for drug offending and the independent measures indicated
negative associations between drug offending and all of the independent measures except
delinquent friends and dating. This held true for both social groups of females and the
upper class males (attachment to parents and peers did not enter the equation for upper
class females). There was an exception, however regarding the lower class males. The
same held true for the lower class males with the exception of dating, which was sound to
correlate negatively with the drug offending (Rosenbaum, 1987).

Finally, the correlation between violent behavior and the independent measures
did not enter the equation. However, positive correlations were found between drug
offending and number of delinquent friends for both social groups of males and the upper
class females. The upper class females also indicated a positive correlation between
violent behavior and dating (Rosenbaum, 1987).

Overall, this examination of social control theory’s relationship with different
types of delinquency indicates important differences. The most obvious result is that the
theory accounts for more explained variation in female than male involvement for both
drug and property offending, and a negligible amount of explanation of violent behavior
for both sexes. This study gives an important look at gender differences and how social

control theory explains those differences. However, several independent variables did

13

not enter the equation and the correlation between violent behavior and the independent

measures were unclear.

 

This section of the literature review entails how efﬁcient social control theory
predicts behavior in urban and rural locations. Gardner and Shoemaker (1989) conducted
a study entertaining this type of comparison. The sample consisted of seven hundred
thirty three 8“1 —12th grade students living in a predominantly rural and urban locale in
Virginia. A cross-sectional multivariate analysis technique was employed. The
operational measures consisted of a self-report survey (Gardner & Shoemaker). The
researchers hypothesized that there would be a signiﬁcant inverse relationship between
components of the social bond and delinquency, with one exception, attachment to peers,
which is expected to be positive. Gardner and Shoemaker also hypothesized that the
associations between social bond factors and delinquency would be stronger in the rural
locale instead of the urban locale. He combined explanatory power of the social bond
components would be greater within the rural sample than the urban sample (Gardner &
Shoemaker),

The measures of delinquency for Gardner and Shoemaker’s study were property
offenses, violent behavior, drug possession, and the less serious juvenile misbehaviors.
The social bonding measures were involvement in school and community, attachments to
peers, parents, teachers and church; commitment to church, school, parents, and peers;
and beliefs (Gardner &Shoemaker, 1989).

The zero-order correlation for the total sample indicated negative correlations
between the social bonding measures and the delinquency measures with the exception

being, positive relationships between attachment to peers, drug possession, juvenile

14

misbehaviors, and total delinquency. The individual indexes indicated that for the total
sample, the social bonding measures explained 41% of the variation in total delinquency.
The index for the urban sample indicated that the independent measures explained 27 of
the variance in total delinquency. The rural sample indicated that 47% of the explained
variation was attributable to the social bonding measures. The results found in this study
were concurrent with the hypotheses of the study (Gardner & Shoemaker, 1988).

o . o o . o o ‘ o 000
_.E .I. [‘0‘ _.. 71.71 \‘0 1 '1, -,.'., 'I 5|. H

Social control theory has become one of the most dominating theories of
delinquency because the theory has received much empirical support. However, may
researchers have contended that Hirschi’s (1969) social control theory does not explain
why the more serious acts of delinquency occur. According to Robert Agnew (1985)
social control theory, as proposed by Hirschi, as only explained 1% to 2% of the variance
in ﬂiture delinquency when a longitudinal study was applied. This small variance is due
to the assumption that the social control variables have a causal impact on delinquency,
and the majority of the studies regarding social control theory have been cross-sectional
studies. Robert Agnew stated that “since social control variables only explained 15 to
2% of the variance in delinquency, cross-sectional studies gave greatly exaggerated the
importance of Hirschi’s theory” (Agnew, 1985: 47).

Robert Agnew (1985) conducted a longitudinal study testing the predictive power
of social control theory. Agnew wanted to test the hypotheses that control theory only
predicts delinquent acts that are minor in nature, and those longitudinal results will
indicate less variance between the social control variables and delinquency. The

argument that Agnew was making was that longitudinal studies are more accurate

15

than cross-sectional studies when engaging in research concerning social control
variables and the variables’ effects on delinquency.

Agnew (1985) employed a panel design and collected the ﬁrst set of data from
2,213 boys beginning the 10th grade and collected the second set of data from 1,886 of
the members of the original sample, which accounted for 85% of the original sample. In
order to test the previously stated hypotheses, Agnew (1985) used total delinquency and
seriousness of delinquency as measures of delinquency. When the social control
variables from the Time 1 section were regressed on total delinquency and seriousness of
delinquency, the cross-sectional regression results indicated that the social control
variables explained almost twice as much variance in total delinquency. The cross-
sectional study also revealed that only four of the eight control variables used has a
signiﬁcant effect on seriousness of delinquency while six of the eight control variables
had a signiﬁcant effect on total delinquency. These results conﬁrm that control theory is
best able to explain minor delinquency.

The longitudinal regression results indicated that the longitudinal effects of the
control variables are lesser than the cross-sectional effects. Only three of the eight
control variables had a signiﬁcant effect on Time 2 seriousness of delinquency, and
likewise, three of the eight control variables had a signiﬁcant effect on Time 2 total
delinquency. The amount of variance explained by the control variables over and above
Time 1 delinquency, indicated that the control variables accounted for 1.5% of the
variance in Time 2 seriousness of delinquency and 1.8% of the variance in Time 2 total
delinquency. The regression results suggested that the social control variables as a whole
has a very small impact on Time 2 delinquency (Agnew, 1985).

Although the majority of the results supporting control theory’s ability to explain

l6

minor delinquency was obtained using cross-sectional techniques, Agnew (1985)
explains that this is because “prior researchers have usually tested control theory with
scales based toward minor offenses. For that reason, they have greatly exaggerated the
importance of this theory” (Agnew, 1985: 53).

Thompson et a1. (1984) empirically tested the validity of Hirschi’s control theory
(1969). These researchers sought to test Hirschi’s approach through multiple regression
analysis in the form of path diagrams. The diagrams were derived through a path
analysis, which assumed that all variables included in the model were measured
independently of each other. This type of analysis goes beyond indicating relationships
between two pairs of variables (i. e. correlation procedures). The sample size was 724
juveniles with varied self-reported delinquency experiences. The juveniles were ﬁom a
number of high schools and correctional institutions in a metropolitan area in the
southwest. The study included males and females.

The ﬁndings of this study were derived from correlation matrices, and two-path
analysis. The results of the correlation matrices indicated that for males, the associations
between delinquent companions and delinquent acts was .71. The association between
attachment and delinquent acts was -. 29. Conventional attitudes and delinquent acts
indicated an association of -. 36. In addition, delinquent companions indicated a negative
and moderate association with attachment and delinquent acts (-. 27 and -. 33
respectively). For females, the association between delinquent companions and
delinquent acts was strong but not as pronounced as the association among males. For
females, the correlation coefﬁcient was .48. Among females, attachment and

conventional attitudes related negatively and moderately to delinquent acts indicating

17

correlations of -. 26 and -. 30 respectively. There was a weak but negative association
between delinquent companions and attachment (-. l3), and delinquent companions and
conventional attitudes indicated a correlation of -. 14 (Thompson et a1., 1984).

The path analysis of Hirschi’s model indicated ﬁndings consistent with control
theory; however, the path coefﬁcients were signiﬁcantly weaker than the original
correlation. For males and females, the path coefﬁcients between attachment and
delinquent acts were -. 12 and -. 13 respectively. At the same time, a negative and low
path coefﬁcient was indicated between conventional attitudes and delinquent acts (-. 28
and -. 23). Furthermore, attachment and conventional behavior combined, explained only
13% of the variation in delinquent acts among males and 10% among females.

These ﬁndings are inconsistent with Hirschi’s proposition of the bonding
elements interrelating (Thompson et a1., 1984). In addition, the explained variation,
when delinquent companions are added to the path analysis, increases to 52% and 30%
for males and females respectively. Hirschi’s original model did not include delinquent
companions. Hirschi suggested that any relationship between delinquent companions and
delinquency was spurious and would disappear when controlling for other factors. These
data do not support Hirschi’s original model and the researchers suggest that Hirschi
revise the theory to not only include conventional peers, but include delinquent peer
association as well (Thompson et al.).

E . E l :2 l 1 I: 1'

Wells and Rankin (1988) undertook a project that sought to examine direct
parental controls. The researchers asserted that Hirschi’s (1969) theory was not inclusive
of all parental controls. In particular, Wells and Rankin suggested that Hirschi’s (1969)

model was not complete because of his lack of discussion about direct parental controls.

18

Direct controls are those dimensions such as parental regulation, parental monitoring, and
punishment. Direct controls are concerned with the physical restriction and surveillance
of behaviors. Indirect controls are dimensions such as attachment, involvement, belief,
and commitment. These four components are conducive to Hirschi’s model. The two
elements that are most closely related to parental controls are attachment and
involvement. These elements are said to be indirect controls because they are essentially
concepts that do not involve one-on-one parent-child interactions. Indirect controls are
based primarily on affectionate attachment to or identiﬁcation with conventional persons
(especially parents). Hirschi (: 88) argued that direct parent controls offer little effect in
controlling delinquent behaviors beyond that already offered through attachments or
indirect controls:

Since most delinquent acts require little time, and since most adolescents are

frequently exposed to situations potentially deﬁnable as opportunities for

delinquency, the amount of time spent with parents would probably be only a

minor factor in delinquency prevention. So-called “direct control” is not, except

as a limiting case, of much substantive or theoretical importance. The important
consideration is whether the parent is psychologically present when temptation to
commit a crime occurs.

The aim of Wells and Rankin’s (1988) research project was to indicate
conceptually what the relevant content of direct parental control variables might be and to
examine empirically what relationships direct parental controls have with delinquency.
The study analyzed the Youth in Transition Panel data. The data set included four waves
of surveys administered to a nationwide sample of boys who were in the 10th grade in
1966. A total of 2,277 boys were selected into the sample list. Of the sample, 97%
(2,213) of the boys agreed to participate in the initial survey. Eighty-ﬁve percent of the

original sample was present in the second wave, and 73% of the original sample were

included in the sample at the fourth wave. However, for the purpose of their study, Wells

19

and Rankin used only the ﬁrst and second waves of the Youth in Transition Panel study.
The researchers focused on the ﬁrst and second waves because the respondents were
sophomores and juniors at the time (15 to 17 years old). These ages correspond to the
point in adolescence at which delinquent activities are most frequent and problematic
(Wells and Rankin, 1988).

The multidimensional content of direct parental control was measured by four
variables. Those four variables were chosen to correspond closely to the dimensions of
direct control of children by parents who were described in the previous paragraph.
Those dimensions were regulation, monitoring, and punishment. The four variables used
were ﬁrst, regulation/restriction (a summary index of ﬁve items indicating the degree to
which parents decided their sons’ friends and activities). Second was strictness, (an
index of how strict respondents rated their parents to be). Third was punishment
contingency (a measure of how frequently parents ignored rather than punished
wrongdoing). The last variable was punitiveness (a summary index if items rating how
vigorously and frequently parents punished their sons, ranging from yelling to hitting)
(Wells and Rankin, 1988).

Six variables were used to measure involvement in juvenile delinquency. Those
variables were total delinquency, theft/vandalism, assault/threat, and trouble with parents,
school delinquency, and trouble with police.

All of the direct control measures were coded as three category variables. Mean
levels of delinquency were compared across categories using analysis of variance
(ANOVA) procedures. This type of statistical procedure provides both an estimate of the
strength of association between parental control and delinquency variables as well as a

test of the linearity of their correlation. Nonlinear associations were indicated to exist

20

between perceived strictness of parents and delinquent behaviors. The other three direct
control measures indicated generally linear associations with delinquent behaviors.
However, the differences for regulation/restriction were somewhat weak across all of the
measures of delinquency. Contingency and punitiveness both show consistent
associations with all forms of delinquent behavior (Wells and Rankin, 1988).

The correlation coefﬁcients indicate that for three of the four direct control
variables (strictness, punishment contingency, and punitiveness), the coefﬁcients are
moderate (mostly between .1 and .2) and statistically signiﬁcant. Punitiveness was the
strongest and most consistently related to delinquency with correlation ranging ﬁ'om .18
to .11. A multiple regression model combining all of the direct parental controls
explained 22%, 16%, 19%, 21%, 23%, and 13% of the variance in total delinquency.
The delinquency measures were theft/vandalism, assault/threat, trouble with parents,
school delinquency, and troubles with police respectively (Wells and Rankin, 1988).

The results of Wells and Rankin’s (1988) study indicated that measures of direct
parental control relate in ways consistent to the occurrence of delinquent behavior. The
researchers argued that direct controls are signiﬁcantly related to various measures of
delinquency and suggest that direct parental controls be researched further and should not
be dismissed as empirically and theoretically irrelevant, as suggested by Hirschi (1969).

Michael Hindelang (1973) sought to examine the extent to which social control
theory could be replicated. Hirschi (1969) focused primarily on urban males, so
Hindelang focused primarily on rural males and females. Hindelang’s major dependent
variable was a recency index. A recency index is a simple sum across all acts. Hirschi

also used a recency index, which was composed of several questions regarding

21

involvement in delinquent behaviors.

Hindelang (1973) used the same independent variables as Hirschi (1969). The
independent variables were attachment to parents, attachment to school, attachment to
peers, involvement in conventional activities, and belief. The dependent variables were
delinquent involvement and the recency index.

Attachment to parents, as Hirschi (1969) conceptualized, was that when parental
attachments are strong, parental values (presumed to be anti-criminal) are more readily
accepted. The question that Hirschi asked was, “Would you like to be the person your
father is?” In the recency index, Hirschi found that 64% of those who answered, “In
every way” compared to 41% of those who answered “Not at all” scored low on the
recency index. Hindelang’s results for the rural males were 61% and 36%. The results
for the rural females were weaker with a score of 65% and 48%, which was low on the
recency index. Among the rural groups, attachment to parents was related to reported
delinquent involvement, as Hirschi predicted (Hindelang, 1973).

Control theory postulates that weak attachments to school reduce the interest in
school, which allows the child freedom to engage in delinquent behavior. For this aspect
of the study the children were asked, “Do you care about what you teachers think of
you?” Among the rural males who reported that they cared “A lot,” 70% reported 10w
delinquency involvement. The results were the same in 67% of the rural females studied.
Of those rural children who reported “Not much,” 35% and 42% were the males and
females respectively reported low delinquency involvement. Hirschi’s (1969) results
indicated 66% and 36% respectively.

The variable attachment to peers was the ﬁrst variable that was found by

Hindelang (1973) to diverge 'from Hirschi’s (1969) results. Hirschi hypothesized that the

22

greater the peer attachment (conventional peers), then the lower the reported delinquent
involvement (Hirschi failed to include unconventional peers in his study). Hindelang
found that among the rural children there was a generally positive relationship between
the extent of identiﬁcation with ﬁiends and the extent of reported delinquency. The
compared results were as follows: Hirschi fund that among the urban males, 64% who
reported identifying most closely with friends reported low delinquency involvement
while 47% who reported “Not at all” (close ties with ﬁiends) reported low delinquency
involvement. Among the rural sample of males who reported to have close ties with their
friends, 48% reported low delinquency, and 67% of those who identiﬁed the least with
friends, reported low delinquency involvement.

The ﬁnal two variables, involvement in conventional activities and belief, as studied by
Hindelang (1973) were found to be similar to the results provided by Hirschi (1969). The
overall correspondence between Hirschi’s results and Hindelang’s is substantial with the
exception of peer attachments. Hindelang stated that “this may indicate that Hirschi’s
control theory will have to be more speciﬁc about attachment to peers. The theory may
need to be reconceptualized in terms of attachment to conventional and unconventional
peers” (Hindelang: 487).

1E"E1El [15.115]
Krohn and Massey (1980) examined the overall effects of the elements of Hirschi’s social
control theory on four separate measures of deviance using data drawn from a sample of
3,065 adolescents. A self-report questionnaire was administered to male and female
adolescents in six communities in three Midwest states. The researchers investigated the
individual effectiveness of the elements, distinguished differences between minor and

major delinquency, and whether social control theory is more effective

23

in explaining deviant behavior among females than it is for males.

The results addressing the individual effectiveness of the elements were produced
by bivariate regression. The correlation matrix indicated moderate inverse association
between 7 of the 10 bonding measures and the four measures of delinquency. The 10
bonding measures were maternal and paternal attachments, peer attachments,
commitment, educational and career aspirations, grade point average, parents morals,
moral duty to obey the law, and school learning helps ﬁnd job. The three bonding
measure that were not consistent with Hirschi’s original theory were peer attachments,
educational aspirations, and career aspirations. The second question, which pertains to
seriousness of crime, is answered by multiple regression analyses. Alcohol and
marijuana conceptualized soft drugs and hard drugs were deﬁned as barbiturates,
hallucinogens, amphetamines, and opiates. For alcohol and marijuana use and abuse, the
theory accounted for 29% of the variance. Twenty-four percent of the variance with
reference to status offenses and minor delinquent acts was attributed to the theory. The
theory accounted for 16% of the variance in hard drug use and 18% of the variance in
serious delinquent acts.

Multiple regression analyses were also conducted to establish how much the
theory contributed to gender differences. The overall effect of the commitment items is
greater for females than for males, with the opposite being true for the attachment items.
The partial regression coefﬁcient results indicated that 13% and 11% of the variance in
alcohol and marijuana use was explained by attachment for males and females
respectively; with 8% and 6% of hard drug usage attributed to attachments. For the
commitment item 13% and 17% of the variance in alcohol and marijuana use was

attributed to attachments; and for hard drug use, the numbers were 7% and 14% for males

24

and females respectively.

The theory continued to indicate its explanatory power with relation to females in
the belief item. In the belief item for alcohol and marijuana usage, 19% of the variance
was explained by belief with reference to males. It was indicated that 23% of the
variance was explained by belief with reference to females. With reference to hard drug
usage among males, 6% of the variance was attributed to belief. The respective
percentage was 13% for females. For all variables representing the seriousness of
delinquency scale, 24% of the variance was attributed to the bonding variables for males
and 25% for females. For all variables representing the seriousness of delinquency scale,
18% of the variance was attributed to the bonding variables for males and 19% for
females (Krohn & Massey, 1980).

In conclusion, the study of Hirschi’s social control theory (1969) suggests that
there is sufﬁcient support for the theory’s major hypotheses to continue the task of
clariﬁcation and modiﬁcation. Krohn and Massey stated “to the extent that this process
continues, the theory will remain an important focus in criminological thinking” (Krohn
& Massey, 1980: 42).

Cemkovich and Giordano (1987) conducted a study that built directly upon the
work of Hirschi (1969). Their research project was an extension of Hirschi’s model
because of questions concerning the dimension (5) of attachment. According to
Cemkovich and Giordano (1987) Hirschi was not speciﬁc with regard to what constitutes
attachment. The researchers asserted that Hirschi was somewhat ambiguous when it
comes to whether or not attachment is one dimensional or multidimensional in nature.

For example, Hirschi gives the impression that the element attachment is a one

25

dimensional construct by asserting that attachment is one of four ways in which
individuals are bound to the social order (i.e. attachment, involvement, commitment, and
belief). However, at other times Hirschi concentrates on the communication origins of
attachment or on the feelings of affection which characterize attachment. When this is
the case, Hirschi is giving the impression that attachment is multidimensional in nature.

The purpose of Cemkovich and Giordano’s (1987) research was to examine the
several dimensions of family attachment and interaction, and the dimensions’ impact on
delinquency. In addition to this elaboration of Hirschi’s (1969) model, the researchers
are also interested in whether the speciﬁed variables are consistent across different types
of family structure, or whether there are difference in interaction dynamics by family type
(Cemkovich & Giordano). For example, if a child is attached to both parents, are his or
her chances of being non-delinquent double those of a child attached to a single parent?
Finally the research will examine whether or not here are race or sex differences in the
inﬂuence of family interaction variables.

The data used for the study was based on a sample of adolescents living in private
households in a North Central Standard Metropolitan Statistical Area. A cross-section of
youth between the ages of 12 and 19 years was obtained through a multistage probability
sample design. The number of successfully conducted interviews was 942. With
reference to gender, 51% and 49% were adolescent females and males respectively. With
reference to ethnicity, 45% were white and 50% were nonwhite.

The dependent variables used in the analysis were delinquency involvement, sex,
race, and home status. Delinquency involvement was broken down into ﬁve degrees of
frequency and seriousness that ranged from non-offender to high-frequency major

offenders. Home status was broken down into three categories, which were, both parents,

26

mother only, and mother/stepfather. The independent variables were ﬁrst, caring and
trust (degree of intimacy with parents). The second variable was, control and supervision
(extent of parental monitoring). Third, intimate communication (sharing of private
thoughts with parents). Next, identify support (belief that parents accept character of
child). Fourth, parental disapproval of peers (what do parents think of child’ friends).
Next, conﬂict (extent of parent-child arguments). Finally, instrumental communication
(parent-child communication of important topics such as problems a school).

The results of a four-way (ANOVA) with age controlled as a covariant indicated
that males, surprisingly, had a higher mean level of caring and intimacy than females.
Also unexpected was the indication that females engaged in more arguments with their
parents than males. However, females are subjected to greater control and supervision
and are therefore more likely to engage in more conﬂicting conversations with their
parents than males. Between sex and race for the control and supervision dimension,
nonwhite females reported the greatest level of supervision and white males the least.
With the exception of intimate communication, each of the family relationship
dimensions is signiﬁcantly related to delinquency involvement. What was assumed by
the researchers to be the most important result was the indication that none of the seven
family factors differed by home status. The authors contend that (1987: 299):

This clearly weakens the traditional argument that broken homes, female-headed

households, or those with stepfathers are necessarily negative and/or

disadvantageous socialization environments.

A correlation matrix of delinquency, age, and the seven family interaction factors
indicate that all of the family dimensions are moderately associated with delinquency at

the zero-order, and the dimensions are all related in the hypothesized directions. That is,

delinquency is inversely associated with all of the family factors with the exception being

27

conﬂict and parental disapproval of peers.

A regression analysis by home statues indicates that family variables explain only
small amounts of the variance in delinquency. The results indicate that 15%, 17%, and
16% of the variance in delinquency is attributed to the seven factors as they ﬁt, with
reference to the total sample, both natural parents, mother only, and mother/stepfather
respectively.

An unstandardized regression coefﬁcient table explained by family factors and
age revealed that the amount of variance in delinquency explained by the model ranges
from a high 23% for white males and a low 8% for nonwhite males. The model also
indicates that the family factors explain more variance in female delinquency than males.
The explained variation is 14% and 12% for females respectively. When race is
considered, the model explains more variation in delinquency for whites (18%) than for
nonwhites (11%) (Cemkovich & Giordano, 1987).

In summation, the ﬁndings indicate that there are important sex and race
differences in family interaction dynamics. However, there are no signiﬁcant variation
by family structure. Signiﬁcant relationships between all but one of the family
interaction dimensions and delinquency, along with the lack of any signiﬁcant
relationship between family structure and delinquency suggest that internal family
dynamics affect delinquency more so than family structure. The researchers concluded
that instead of thinking of attachment as a dichotomous conception (attached/unattached),
it is more appropriate to think of attachment as more multidimensional in nature. In other
words, a more detailed elaboration of the elements constituting attachment provides a
more rigorous test of exactly how family processes affect delinquency (Cemkovich &

Giordano, 1987).

28

Socialﬁontmllhmiiexised

Travis Hirschi’s social control theory (1969) assumes that criminality is the
essential nature of humans. If this is believed to be true, then the element of conformity
must be addressed in order to explain why some people choose not to engage in
criminalistic activities. Conformity is achieved through socialization, which is the
formation of a bond between society and an individual. According to the social control
theory, socialization is the formation of the bond comprised of four major elements,
which are attachment, involvement, commitment, and belief. Hirschi contends that the
stronger each element, then the less likely the occurrence of delinquent behavior.

It has been implied by various researchers that Hirschi (1969), only concerned
with the individual variance of each element of the theory, has failed to consider how the
four elements might act simultaneously to affect the likelihood of delinquency behavior.
Also, with respect to the relations of the elements of the bond, social control theory falls
short of complete explanation because the association between the elements of the theory
were only hypothesized by Hirschi and not empirically tested.

A study undertaken by Wiatrowski et a1. (1981) sought to address he
aforementioned concerns and develop an extension of social control theory. Because of
the importance of this study in pointing out the shortcomings of Hirschi’s control theory
an extensive analysis of this study will be documented in this literature review.

The researchers addressed these shortcomings of control theory by raising three
questions that Hirschi’s theory presented. The three questions raised by Hirschi’s (1969)
theory were modiﬁed and tested by the present researchers to address the shortcomings
charged. Construction and data analysis is as follows, along with the present researchers’

method of testing these developments.

29

While analyzing Hirschi’s (1969) theory, the researchers asked, “To what extent
does any of Hirschi’s four elements represent empirically distinct components of
socialization?” (Wiatrowski et a1., 1981: 526). The researchers of this study implied that
Hirschi was contradictory in his discussion of each element’s individual predictive
power. If the association between the elements of the bond explain the most variance,
then the elements independently would not constitute analytically distinct elements of
the bond. The researchers sought to test these assumptions, so measures were created of
the bonds to estimate how much each element contributes to the others to the explanation
of delinquent behavior.

The second question raised by the researchers was with reference to the number
of elements of the bond. The issue is since the predictive power of the bond is so modest,
then additional elements of the bond should be considered. The researchers conducted a
factor analysis of the bond’s structure to discover and ﬁnd new elements. The ﬁnal
question raised by the researchers discussed the constructs that Hirschi chose to include
and those he chose to delete which were considered to be important. Research prior to
Hirschi’s theory construction has asserted that socioeconomic status, personal ability, and
signiﬁcant others play a major part in educational and occupational aspirations. Hirschi
stresses the importance of educational and occupational aspirations but fails to
incorporate these constructs into his theory. The researchers of this study conducted a
pair of multivariate analyses incorporating educational and occupational aspirations along
with the elements of attachment, involvement, commitment, and belief. The ﬁrst analysis
was intended to replicate Hirschi’s ﬁndings. The second analysis was based on the factor
analysis which used new and deﬁned elements which lead to a revised model which the

researchers contended was more complete than Hirschi’s (Wiatrowski et a1., 1981).

30

The study was longitudinal in nature and the reconstruction of the scales and indices
demonstrated high alpha coefﬁcients, which demonstrated their internal reliability.

The ﬁrst analysis addressed the issue of the extent to which the elements make
unique contributions to predicting delinquency and to what extent the variables
simultaneously make on predicting delinquency. Hirschi’s results (1969), which were
intentionally replicated, indicated that all four of the elements of the bond with their
respective measures explain 32% of the variance in delinquency. The individual
elements of the bond; however, indicated that each individual element independent
accounted or only minute variations in the dependent variable. The element attachment
accounted for 7% of the change in predicting delinquency, commitment contributed 11%
of the change, involvement accounted for 1% of change, and ﬁnally, belief accounted for
2% of the variance in predicting delinquency.

The researchers of this project conducted a second analysis using the elements of
the bond that they deleted from the factor analysis. Only a few measures were selected
and the results indicated a 17% variation collectively among the variables. The bond
elements independently accounted for 3%, 0%, 1%, and 3% of the variance in predicting
delinquency. The elements of the bond were attachment, commitment, involvement, and
belief respectively. These results show that incrementally each element of the bond adds
little explanation of delinquency in relation to the total amount of variance explained.
The second question involved the discussion of the structure of the bond with reference to
the elements selected to represent the bond. A factor analysis done by the researchers
included the original elements as selected by Hirschi (1969), and the elements were
expected to appear as separate factors. However, the researchers found that belief did not

appear a factor, and several of the original elements appeared to have

31

overlapped one another in the attempt to stand on its own as an adequate predictor of
delinquency. These results suggest that a more complex interpretation of the social bond
needs to be done. The researchers’ summarized this phenomenon by stating “some of
Hirschi’s postulated dimensions emerge as distinct factors, but the general picture of the
components of the bond is altered. What this suggests is that it may be more appropriate
to reconceptualize the nature of the bond” (Wiatrowski et a1., 1981: 532). Finally, the
multivariate analyses conducting corporated measures of socioeconomic status, ability,
and grades along with the original elements developed by Hirschi (1969) and the
elements deﬁned by the factor and communality analysis are discussed brieﬂy. The
multivariate analyses’ results were represented in the form of simple and complex
models of the social bond. The ﬁrst multivariate analysis included the corporated
measures of socioeconomic status, ability, and grades, which were controlled, as well as
the original elements of the bond developed by Hirschi. A path analysis was used to
examine the effects of the model on delinquency outcomes. The end result indicated a
19% proportion of the variance in delinquency.

The researchers contend that the R square (R2) value of this model is low because
the scales constructed to measure the predictive power of the variables were ravaged
because they were based on Hirschi’s (1969) theoretical statement. Finally, the complex
model of social control theory constructed after factor analysis variables were added
explained 33% of the variance in delinquency. In contrast to the ﬁrst model, the revised
model implies a moderate positive direct effect of ability and socioeconomic status.
Grades make a moderate negative direct contribution to delinquency.

The next section of the literature review is dedicated to looking at a synthesized

model of delinquency control.

32

.99: '01' ,5 u 1" o .0 ' 010. 1‘0; 1-10. 0 '. yon-rt: 1";

The bulk of this literature review has concentrated on the shortcomings,
replications, and uniqueness of Hirschi’s (1969) social control theory. The ﬁnal section
of this literature review takes a look at how a synthesized view of delinquency theory can
better predict delinquency behavior. This study’s hypothesis was that a combination of
the social bonding notion from control theory along with speciﬁc principles of social
learning theory will lay a groundwork for a more comprehensive theory of delinquent
behavior than either perspective alone. This study is included in this paper because it is
the opinion of the writer that if a complete picture of social control theory is to be
addressed, then the notion that the theory can be expanded to be more predictive of
delinquency is important.

Whatever shortcomings may exists in the study of delinquency, a lack of
theoretical perspective is not one of them. Through this literature review, social control
theory has been collectively discussed in terms of its internal structure, its basic
principles, its social-psychological nature, empirical consistency, and empirical
inconsistencies. The conclusion of this discussion of social control theory will take a
ﬁnal route towards social control theory’s aspects but also look at how the theory can be
expanded to have more predictive power.

Social control theory has been extensively covered, so before the forthcoming
study is discussed, a short discussion of social learning theory’s basic components will be
discussed. Social learning theory as deﬁned by Bandura (1977) emphasizes the effect of
a model’s behavior on an individual who is capable of emitting imitative responses.
Social learning theories are primarily psychological theories and are derived inductively.

The differences among social control theory and social learning theory are

33

detailed in the following paragraph.

Social control theory (1969) asserts that human beings are prone to delinquency
from the outset but are restrained ﬁ'om delinquency due to bonds they attain with
conventional others and institutions. Control theory asks the question, “why do people
refrain from committing delinquent acts?” It does not ask, “why do people engage in
delinquent acts?” Social learning theory (1977) asserts that human beings’ behavior is
acquired through direct conditioning and imitation of others’ behavior (Akers et a1.,
1979). The control theory perspective implies that attraction or attachment to deviant
others is not a cause of delinquency. The social learning perspective asserts that
attachments to deviant others increases the likelihood of delinquency if the rewards for
engaging in delinquent acts outweigh the rewards for conforming to conventional
behavior. This is commonly known as the “matching law.” A study conducted by Rand
Conger (1976) examined the implications of both the control and the learning model in
order to produce a more adequate predictor of delinquency.

The ﬁndings were broken down into two interpersonal categories: attachment to
peers and attachment to parents. The ﬁrst set of results, which tested the theories on
attachment to peers indicated that attachment to peers (conventional), as control theory
postulates, were not always inversely associated (Conger, 1976). The results indicated a
random ﬂuctuation in the relationship between attachment to peers and delinquent
behavior, which is contradictory to control theory. The social learning position was
supported which predicts either a negative relationship between delinquency and
attachment to conventional peers or a positive relationship between delinquency and
attachment to deviant peers.

The second set of ﬁndings which tested the theory’s predictive power on

34

attachments to parents indicated that actual parental behavior, not just the simple
identiﬁcation with parents, is the most important part of the bond between parents and
child, which protects against delinquency involvement. The social control model predicts
that communication between parent and child will decrease the likelihood of delinquent
behavior. The social learning model predicts that communication between parent and
child will be negatively associated with delinquency only if his or her parents reinforce
initiations from the child.

This study argues that social control theory is more incomplete that incorrect. By
combining the notion that individuals are bonded to others with certain principles from
social learning theory, the groundwork for an expanded social control theory will be
developed. Conger argues that (1976: 37):

The social learning model clariﬁes how attachments will inﬂuence deviant

behavior by including to whom one is attached in the analysis. Social control is

brought about through differential reinforcement and punishment in many subtle
ways that are far ﬁ'om being completely understood.

2 . . . ES . l 2 l I]
Social control theory, like any other theory has weaknesses. Empirical research
from testing the consistency of the theory have unearthed that social control theory, as
Hirschi formulated does not predict major types of delinquency as well as it explains
trivial types of delinquency. Agnew (1985) suggested that this phenomenon is from
Hirschi’s testing of control theory with scales biased toward less serious forms of
delinquency. Social control theory has been credited with “being able to explain some of
the differences between children and levels of delinquency, but failing to predict why
certain adolescents exhibit particular patterns of delinquency” (Troj anowicz & Morash,
1992: 68). For example, why do some adolescents become habitual offenders, but others

participate occasionally in minor delinquency?

35

Hirschi’s social bonding theory (as labeled by Akers, 1973) has also been
criticized for overestimating the signiﬁcance of involvement in conventional activities
and giving little credence to the role played by unconventional peer attachments. It has
been suggested through empirical studies that peer attachments be divided into
conventional and unconventional peer associations and examined independent of each
other in relationship to delinquency (Thompson, Mitchell, & Dodder, 1984; Mastueda,
1982; Mastueda & Heimer, 1987; Patterson & Dishion, 1985; Mitchell & Dodder, 1980;
Elliot & Ageton, 1980; Poole & Regoli, 1979; Linden, 1978; Conger, 1976; Silberrnan,
1976; Hindelang, 1973; Jensen, 1972). The theory also meets strong opposition from
differential association theorists from the suggestion that parental attachment and peer
attachments are positively correlated (Bandura, 1977; Sutherland & Cressey, 1974;
Burgess & Akers, 1966; Cloward & Ohlin, 1960; Cohen, 1955). Differential association
theorists suggest that delinquent behavior or conventional behavior depends on the
proportion of the reinforcement offered by the peer or the parent. Whoever offers the
most in the form of a reward gets the desired behavior. Differential association theorists
argue that it is either the parent or the peer that gets the desired behavior from the child or
friend, not the parent and the peer.

Social control theory, as it is proposed by Hirschi (1969) is also criticized for
failing to incorporate direct parental controls (Wells & Rankin, 1988). Wells and Rankin
argue that indirect parental controls (i.e., attachments) are no more important than direct
parental controls (i.e., parental monitoring). Wells and Rankin insisted that Hirschi’s
dismissal of direct parental control was not based on an empirical demonstration of its
unimportance, but rater in an ideological preference for other forms of social control

(Wells and Rankin).

36

Hirschi is criticized for not being speciﬁc with regard to exactly what constitutes the
dimension of attachment (Cemkovich & Giordano, 1987). At times, Hirschi speaks of
attachment as a one-dimensional construct belonging to one of four elements of social
control (attachment, involvement, commitment, and belief). At other times Hirschi
concentrates on the communication origins of attachment, or on the feelings of affection
which characterize attachment. In this instance, Hirschi is giving the impression that
attachment is more multidimensional in nature. Cemkovich and Giordano suggest that
Hirschi systematically identify the dimensions of attachment. Finally, criticisms of the
theory come in the form of questions that were unanswered by Hirschi’s research such as:
1. How do the bonds become weakened or broken?

2. How much do the bonds need to be weakened for delinquency to occur?

3. How does weakening of one bond affect others?

This section of the literature review addresses the attachment between parent and
child during the adolescent’s young adult years. The purpose of this section is to present
literature that details the continuity of the attachment. Frequently asked questions
concerning parent-child attachments are those such as, “Does the attached child continue
to exhibit pro-social characteristics once he/she is independent of their parents, or how
important is the attachment once the child has physically separated from the parent?”
The ﬁrst question is a valid question to as because general belief it that departure requires
a loosening of family ties. Furthermore, failure in loosening ties is expected to result in
the difﬁculty in the development of intimate relationships outside the family (Haley,
1980; Freud, 1969; B105, 1967).

The next few sections discuss how the early attachment develops in the youth.

37

Empirical evidence (Sullivan & Sullivan, 1980) has suggested that feelings of parental

closeness actually increase following departure from the family. Sullivan and Sullivan
found that adolescent males reported increased feeling of affection and communication
towards parents after leaving home to attend a residential college.

- o o . - ' .- ' ' - , --
I. .l I.\. 2., ,1-I'IQ,\ 19'00001 ...I'$.- 9..

Kim

Kenny (1987) conducted a study that focused on the nature and ﬁmction of
parent-child adolescent attachment and separation during the leaving home process. The
study attempted to explore the relationship between characteristics of the parental
attachment in late adolescence and indices of social competence during the same period.
Characteristics of secure attachment are expected to be associated with self-reports of
social competence.

The subjects for this study were 173 ﬁrst year residential college students at the
University of Pennsylvania. The subjects were predominantly white (90% of the males
and 81% of the females). The parental attachment measure consisted of a 70 item
parental questionnaire designed to adapt Ainsworth, Blehar, Walters, & Wally’s 1978
conceptualization of attachment for use with college students in a self-report format.
Ainsworth et al’s., conceptualization of attachment is that secure attachment fosters
feelings of conﬁdence in expressing one’s needs and feelings with the expectation that
one will inﬂuence and be accepted by others. Social competence was measured by
assertion and dating competence, which were operationalized by the Dating and
Assertion Self-report Questionnaire (Levenson & Gottman, 1978).

A factor analysis was executed and extracted separate factors for male and
females. In the female category the factors were Quality of Relationship with Parent,

Parental Role in Providing, Emotional Support of Parent, Parental Fostering, Autonomy,

38

and Adjustment to Separation. The factors extracted in the male category were General
Attachment, which includes everything but Adjustment to Separation with reference to
the female factors. The multiple regression analyses were completed separately for the
male and female samples. The results indicated that for both men and the women, the
Adjustment to Separation factor was signiﬁcantly associated with self-ratings of assertion
and dating competence. The results were .33 and .51 for prediction of assertion with
reference to males and females respectively; and .33 and .25 for prediction of dating
competence for males and females respectively. Among the females, the Quality of the
Relationship with Parents accounted for 19% of the variance in the prediction of assertion
variable, while Adjustment to Separation accounted for 26% of the variance in the
prediction of assertion variable. In the male sample, 11% of the variance in prediction of
assertion was attributed to Adjustment to Separation. With reference to the dating
competence variables; 11% and 6% of the variance was attributed to Adjustment to
Separation for males and females respectively.

These ﬁndings have suggested that parental attachment often continue through
adolescence into young adulthood. These young adults appear to look for their parents or
support while at the same time asserting their independence. Although correlation
ﬁndings cannot be interpreted as causal understanding the relationship between parental
attachment and assertion is reasonable. Since the majority of this sample came from
afﬂuent families, assertion amongst other behaviors was probably valued and expected.
There has been little research conducted in this area. According to these ﬁndings, ﬁrst-
year college students often continue to view their parents as a secure base of support and
caring (Kenny, 1987). Because the continuity of the parent-child attachment is so

important in understanding the effects of the initial attachment, more studies concerning

39

this aspect of the attachment is much needed.
1‘ 1111111: 1 '-_.‘1.. 1.11111'1 31101? .. o 1.11 111:1 1.111 ‘

Pro-social behaviors can be deﬁned as anything accepted and prescribed by a
particular society. It is often hypothesized that parent-child positive attachments mold
the child into an individual who exhibits characteristics of a socially acceptable
individual. Throughout this literature review, it has shown that there are many different
forms of antisocial behavior, or behavior not accepted as conventional or socially
acceptable. This study conducted by Kenny and Hart (1992) looks at a different type of
antisocial behavior. Kenny and Hart (1992) investigated the relationship of parental
attachment to eating disorder symptoms among late adolescent and young adult omen in
inpatient treatment for eating disorders among a non-clinical sample of college women.
These researchers hypothesized that characteristics of secure attachment, including
positive affect, use of parents as a source of support, and view of parents facilitating
autonomy would be inversely associated to eating disorder symptoms for the combined
sample of eating disorder and college women.

The inpatient sample consisted of 68 females diagnosed as eating disordered. The
non-clinical college sample consisted of 162 ﬁrst-year college women enrolled at a
private, urban, coeducational Jesuit university in the northeast. Parental attachment was
broken down by factor analysis to include Affective Quality of Attachment, Parental
Fostering of Autonomy, and Parental Role in Providing Emotional Support. Attachment
was measured by a 55-item questionnaire. Five scales of the Eating Disorder Inventory
(Garner, 1991) were used to assess eating disorder symptoms. The ﬁve scales were the

drive for thinness, body dissatisfaction, bulimia, ineffectiveness, and the maturity fears

scales.

40

Correlations indicated among the sample that the inpatient sample was less
attached to their parents; and therefore were more susceptible to eating disorders. For the
Bulimia and Ineffectiveness scale the inpatient sample indicated correlations of -.29 and
-.29, both being signiﬁcant at the .05 level. The college sample indicated correlations of
.01 and -.20, with only the latter being statistically signiﬁcant at the .05 level (correlated
with Affective Quality of Attachment). Among women with eating disorders, affective
closeness to parents and reaching out to available family support in conjunction with
parental support for autonomy appear to be adaptive responses to difﬁculty. Eating
disorder women described themselves as less securely attached to their parents when the
college women. Hence the higher degree of eating disorders, as indicated previously.
More speciﬁcally, the inpatient sample characterized their parents as less supportive of
their independent strivings and described themselves as less likely to seek out and receive
comfort from their parents in times of stress (-.15, -.O6, -.11, -.23, and -.10 when parents
fostering of autonomy was correlated with the eating disorders). The college sample,
whose parents showed more interests in their independent strivings indicated correlations
of -.02, -.06, -.03, -.28, and -.22; with the latter two correlations being signiﬁcant at the
.001 level (p< .001).

These results indicate that women who have never had an early attachment to
their parents have grown up without an effective attachment. Therefore, they view their
parents as interfering in their strives to become independent, and are likely to experience
anxiety and this case obsessions with weight
AdolescanAnachmentsjoBamMReers

Lapsley, Rice and Fitzgerald (1990) undertook a study to determine whether early

bonding between parents and children remain important throughout the life span. Prior

41

research has in indicated that positive attachments to parents convey a number of
advantages for adolescents’ social and personal development(Armsden & Greenberg,
1987; Greenberg, Siegel, & Leitch, 1983). The researchers hypothesized that early
attachment to parents would account for a signiﬁcant amount of variation in both
personal and social identity. The study also hypothesized that early attachment to parents
would account for personal and social identity in various college aspects (peers).

The participants totaled 130 males and females that were enrolled as psychology
major freshmen at a small Catholic university. Also, 123 juniors and seniors who were
enrolled in a child development course volunteered to participate. Attachment was
assessed with the Inventory of Parent and Peer Attachment (Armsden & Greenberg).
Factors extracted from attachment were Trust, Communication, and Alienation. Identity
was used to assess personal and social identity. Identity was assessed by the Aspects of
Identity Questionnaire (AIQ) developed by Cheek and Briggs (1982). College
adjustment was assessed by the Student Adaptation to College Questionnaire (SACQ)
developed by Baker and Siryk (1984). The SACQ is a 67-item self-report measure that
assesses four feature of college adjustment. Namely the adjustments were academic,
social, personal-emotional, and goal commitment.

Regression analyses indicated that the block of the attachment variables (Trust,
Communication, and Alienation) accounted or a signiﬁcant amount of variation in the
personal identity scores of both freshmen and upperclassmen (9% and 7% respectively).
The addition of the peer attachment variables to the regression indicated an 11% change
(positive) for freshmen, and a 10% change (positive) for the upperclassmen. Hence, the
attachment variables appear to account for approximately 20% of the variance in the

personal identity score of freshmen and about 17% of the variance in upperclassmen.

42

The block of parental attachment variables also accounted for signiﬁcant
proportions of the variance in social identity scores for both freshmen and upperclassmen
(8% and 9% respectively). The addition of the peer variables did not improve the
prediction of social identity for freshmen. However the peer variables did improve the
prediction of social identity for upperclassmen (from 9% to 16%). The measures of
parent and peer attachment were also entered as predictors of adjustment to college. In
the freshman group the block of parent attachment variables accounted for a signiﬁcant
amount of variation only in academic adjustment scores (8%). The peer attachment
variables did not improve the prediction. With reference to the personal-emotional
adjustment and the goal commitment; the parental attachment variable accounted for 16%
and 9% of the variance respectively. The addition of the peer attachment variable
improved the prediction of personal-emotional adjustment by 5% and goal-commitment
by 11%.

In conclusion, it was indicated by the data that freshmen did not differ
signiﬁcantly from upperclassmen in their felt attachment to parents or peers. The
principle aim of this study was to chart the relationship between attachment and several
indices of adolescent adaptation. Regarding the social identity variable, the ﬁndings
suggest that adolescent attachments to parents and to a lesser extent peers signiﬁcantly

predicts social identity among the freshmen and upperclassmen.

 

The disruptions or omissions in the parents’ application of parental monitoring and
parental authoritarianism or effective discipline have consistently been found to
correlated with preadolescents’ antisocial behavior (Patterson & Stouthamer-Loeber,

1984). The emphasis upon monitoring and discipline rests upon the fact that when

43

counseling the parents of antisocial children, it becomes necessary to teach the majority
of the parents to monitory their child and to discipline him or her, for antisocial behavior
whenever antisocial behavior occurred (Patterson & Dishion, 1985). Patterson and
Dishion conducted a study that tested the hypothesis that a parental failure to monitor
their children effectively contributes directly and indirectly to delinquent activity. The
model assumed that ineffective or a lack of parental monitoring contributed indirectly to
delinquency by increasing the chance that the child will become involved with a deviant
peer group which, in turn, contributes directly to delinquent activity. The direct
assumption is that when there is a disruption in parental monitoring, the child will have
more unsupervised “street time” which means more time to engage in antisocial or
delinquent activity.

The results of the study proved the hypotheses of indirect and direct delinquency
involvement. The results of the standardized structural coefﬁcients and factor loadings
for a revised two-stage model of delinquency indicated that social skills and parental
monitoring are both found to be related to deviant peers. The results also indicated that
80% of the variance in delinquency in delinquency was explained by parental monitoring
and social skills.

Family management variables have emerged as the result of years of clinical work
with pre-adolescent, antisocial children. In the majority of these studies, the parent was
the agent of change for altering the behavior of the problem child. Patterson and
Stouthamer-Loeber (1984) developed speciﬁc family variables used in a study correlating
family management practices for delinquency as a result of these studies. Patterson and
Stouthamer-Loeber hypothesized that certain familial variables would be found to

correlate signiﬁcantly with criterion of delinquency. The familial variables were parental

44

monitoring, parental discipline, problem solving, and reinforcement. The criterion
measures for delinquency were police contacts that were deﬁned as boys who have had a
police contact resulting in a juvenile court record and a self-reported delinquency
measure of delinquent lifestyle. The grades of the boys who were studied, were inclusive
of and between grades 7 and 10.

The correlation of the family management variables with measures of delinquency
indicated a signiﬁcant negative correlation of monitoring and discipline with both
measures of delinquency, which gave strong support to the researchers’ hypotheses. The
correlation for the monitoring variable and the discipline variable accounted for two and
a half as much variance as did any other measure of parental skill. When correlated with
the delinquency measure, delinquent lifestyle, problem solving, and reinforcement
indicated correlations of .04 and —.24 respectively. It was necessary to include the results
of the familial variables problem solving and reinforcement to show that the family
variables, parental monitoring; when disrupted, and parental discipline when
inconsistently, severely, and nonchalantly applied account for more signiﬁcant variations
in delinquency than most other familial variables (Reiss, 1951).

The data results indicate that familial variables provide strong support for the
hypothesis that disruptions in certain family management practices are associated with
delinquent behavior for young adolescent boys. In this process, parental monitoring
plays the central role. Lack of parental monitoring that increases with age insures the
possibility of several unpunished trials of more or less delinquency (Patterson &
Stouthamer-Loeber, 1984). This phenomenon may also be dichotomous in nature. In the
beginning, it may determine which youths become engaged in delinquent activities, and

consequently, it may determine which youths become recidivists.

4S

Summary

This literature review was designed to inform the reader of the nature of parental
attachments to children and how levels of attachments relate to delinquency. The review
also serves to broaden the reader’s knowledge of Hirschi’s social control theory and to
show through previous studies that the theory is effective in describing and predicting
delinquency although internally the theory has some ﬂaws and shortcomings.
Additionally, this literature review was designed to incorporate theory. Hence, an
overview of the elements of the social bond and the relationship between these bonds and
delinquency were reviewed. Studies examining the continuity of the bond were also
included.

The continuity of the bond or attachment section of the literature review was
included to brieﬂy give the reader an idea of how important the early parent-child
attachment is to the development of pro-social characteristics. Continuity studies have
indicated that if early parental attachments are not established the child is more likely to
have problems establishing social relationships, solving problems, and the inability to

assimilate into conventional social settings.

46

CHAPTER THREE
METHODS

The following section will outline the research strategy of the study in order to
carry out the objectives stated in chapter one. Included in this section is the research
design, the sample population, the data collection, and data analysis. A secondary
analysis of the National Youth Survey (Wave II, see Appendix B) collected by Delbert
Elliot for the Behavioral Institute in Boulder, Colorado will be carried out in order to
achieve the objectives stated in the problem statement. Wave II was selected because the
ages of the respondents ranged from 12 to 18 years. These ages correspond to the point
in adolescence at which antisocial activities are most frequent and problematic (Wells &
Rankin, 1988). The National Youth Survey is a longitudinal survey of delinquency and
drug use. The ﬁrst wave was conducted from January to March of 1977 and focused on
delinquency during the 1976 calendar year. Subsequent analyses were conducted on an
annual basis numbering seven waves to date.
53mm:

The data will be used were acquired in a national survey of 1,725 youths in the
United States. The ages of the youth ranged from 12 to 18 years. A total of 1,725
interviews were conducted from January to March of 1978 and focused o delinquency
during the 1977 calendar year. In the second wave 53% (n=918) of the respondents were
male and 47% (n=807) were female. Fifteen percent (n=252, n=257, n=258, n=253) of
the respondents were ages 12, 13, 15, and 16 respectively. Sixteen percent (n=269)
reported that they were 14 years old, which was the mode. Twelve and fourteen percent

(n=197 and n=239) indicated that they were 18 and 17 years old. The majority of the

47

youths studied were White (79%). Fifteen percent (n=260) were Black, 4% (n=76) were
Mexican-American, .5% (n=8) were American-Indian, 1% (n=17) were Asian-American,
and .2% (n=3) made up the category “Other”.

Since this study targets male antisocial behavior, a sub-sample was chosen from
the larger sample collected by Delbert Elliot. The sub-sample was chosen by deleting the
female respondents from the original sample. The characteristics of the sub-sample are
described in the beginning of Chapter Four.

Michigan

The questionnaire used in this study and the data collected were obtained (in part)
through the Inter-university Consortium for Political and Social Research (ICPSR) and
the Behavioral Research Institute, with Delbert Elliot as the principal investigator. The
primary data were obtained by area probability sampling of American youth. The
questionnaire utilized had 24 sections. The ﬁrst three sections of the questionnaire were
devoted to obtaining demographic information and included a respondent receipt form
(all interviewed youth received $5). The fourth section assessed general activities that
occurred the previous year (from Christmas 1976 to Christmas 1977). The ﬁfth section
of the questionnaire was devoted to assessing the importance of certain activities to the
respondents and how well the respondents were at doing those activities. The sixth
section was dedicated to obtaining information on the respondents’ feelings and beliefs
about ﬁiends, teachers, and family. Sections seven and eight assessed how the
respondent thought parents, friends, and teachers would describe him (respondent) and
how the respondents think his parents and close friends would react if he or she engaged
in certain behaviors. Section nine is devoted to assessing how parents and close ﬁiends

would act if the respondents were to get in trouble at school, with the police, or in the

48

neighborhood. Section 10 assesses the respondents’ perception of right and wrong and
how it applies to other children in his age group. Sections 11 and 14 are devoted to
assessing the behavior of the respondents’ close friends (from Christmas 1976 to
Christmas 1977). Sections 12 and 13 are devoted to assessing the respondents’ views on
the roles of men and women in American society and violence between people. Section
15 deals with assessing the extent that the respondents’ parents engaged in drug use.
Sections 16, 17, and 18 are devoted to the frequencies and rates the respondents’ have
participated in delinquent activities and drug consumption. Section 19 is used to
determine the impact that the respondents’ drug usage (if any) has affected their families.
Sections 20 and 21 seek to determine how many times (if any) that the respondent has
been the victim of a robbery. These two sections also assess how many times the
respondent (if any) has attacked someone sexually, or how many times he has been
attacked sexually by someone other than a parent, and the respondents’ knowledge of
handguns (how to use them, how to obtain them, ever used a handgun). Section 22 is
devoted to those respondents who indicated that they had sexually assaulted someone.
This section asks the respondents when the act took place, where, and the reaction of the
victim. Section 23 is an informal section that gives the respondent a chance to add
comments to any information that was given to the interviewer. The ﬁnal section is the
interviewer’s observations before and during the interviews.
W

The statistical techniques that will be utilized are the following:
1. Univariate Analysis

Frequency Distribution

Percentages

49

2. Factor Analysis
3. Reliability Analysis
4. Statistical Signiﬁcance
5. Bivariate Analysis

Contingency Tables

Chi-Square

Pearson’s Product-moment Correlation Coefﬁcient
6. Multivariate Analysis

Partial Correlation-Analysis

Multiple Regression

n' ' si

The types of Univariate analysis that will be utilized are frequency distributions
and percentages. The frequency distribution is a type of statistical analysis that involves
the measurements of each set of objects on a single variable.
W

The factor analysis is conducted during the very early stages of most statistical
investigations. Factor analysis is a data reducing technique that removes the redundancy
from a set of correlated variables and represents the variables with a smaller set of
derived variables, or factors. In other words, the factor analysis procedure groups similar
variables. The procedure can also be said to remove duplicated information among a set
of variables. A quantity called an eigenvalue, which corresponds to the equivalent
number of variables is used to extract variables. The eigenvalue’s concept is to extract
factors that account for the most variance in the data collection (Kachigan, 1991). The

intricate formulas that factor analysis uses to perform these operations are beyond the

50

scope of this study, therefore only its importance and what it produces is discussed.
E 1' l 'l' ! l .

Reliability analysis is a process in which several variables are measured for scale
appropriateness. Reliability tests produce and alpha coefﬁcient. If the coefﬁcient is over
.5, the scale is considered to be reliable. The reliability output gives an alpha coefﬁcient
(Cronbach’s Alpha) of all of the combined items and then gives individual would be
alphas, if the researcher chooses to delete certain items from the scale.

The statistical signiﬁcance of a relationship that is observed in a set of sample of
data is expressed in terms of probabilities. Statistical signiﬁcance is the outcome of
bivariate analysis and multivariate analysis. The particular probability that is used as the
criterion for an unlikely outcome, supposing that the working hypothesis (H0) is true, it is
referred to as the signiﬁcance level. The signiﬁcance level is designated by the Greek
letter alpha, a. Three common signiﬁcance levels (a=.01, a=.05, and a=.10) are used in
hypothesis testing. The level of signiﬁcance of this study is set at a .05 level (a: .05).
This means the probability of a relationship as strong as the observed one being
attributable to sampling error alone is no more than 5 in 100. Simply, the probability of
sampling error is discounted when the stated level of signiﬁcance is less than ﬁve

chances in one hundred.

B' . l ! I .
:2 . 5 l .

Contingency analysis will be used to conduct a certain type of bivariate analysis.
A contingency analysis tests whether or not the effect of the column variable (IV) on the

pattern of cell frequencies is independent (not contingent) or dependent (contingent)

51

upon the effect of the row variable (DV) upon the pattern of cell frequencies (Rosenberg,
1990). Contingency tables or crosstabs indicate in a broad way whether values on two
dichotomous variables are contingent or dependent on one another. Contingency tables
also show the joint frequency distribution of two independent variables. The relationship
between the two variables will be presented in the form of percentage distributions.
Chiﬁguare

The chi-square (f) test will be used to test for signiﬁcance with reference to the
contingency analysis. The calculated chi-square will be compared to the critical value set
for the .05 level of signiﬁcance coupled with the degrees of freedom of the table. If the
critical value is higher than the chi-square value the hypothesis of independence is
accepted. If the critical value is lower that the calculated chi-square value the hypothesis
of independence is rejected, and it is assumed that there is a less than 5% chance of
making a Type I error, which is incorrectly rejecting the null hypothesis. While
univariate analysis and sub-group comparisons focus on describing the units of analysis
under study, bivariate analysis focuses on the variables.
EsmnlhodmLMomenLConelationﬁoefﬁcienm)

Pearson’s Product-moment Correlation Coefﬁcient ( r) is a measure of the degree
of association between two variables (Graham, 1994). The correlation coefﬁcient r can
range in value from —1.00 to +1.00. A correlation coefﬁcient of r = -1.00 signiﬁes a
perfect negative linear relationship. A correlation coefﬁcient of r = +1.00 signiﬁes a
perfect positive linear relationship.

The Pearson’s Product-moment Correlation Coefﬁcient will be utilized to predict
the values of the independent variable, parental attachment, in order to see how its

increase or decrease in value affects the dependent variable, antisocial behavior. The

52

correlation coefﬁcient assumes that the relationship between two variables is linear.

MII' 'I!l'
13.]: l.,}.

A partial correlation analysis is the process that examines the effect of the
independent variable on the dependent variable when a third variable is held constant. If
the partial relationship is essentially the same as the original relationship (corresponding
zero-order) this outcome is called a replication. This means that the original relationship
has been replicated within smaller sub-groups and that the control variable has no
inﬂuence on the original relationship (Babbie, 1995).

I l l . l E .

Regression gives more information than the correlation analysis. It is necessary
to analyze the notion that antisocial behavior is affected simultaneously by several
elements of attachment. Multiple regression analysis provides a means to analyzing such
situations (Maxﬁeld & Babbie, 1995). Multiple regression uses several predictor
variables to estimate values on a criterion variable (Kachigan, 1991). In using many
predictor variables instead of j ust one, the aim is to reduce even further the errors of
prediction; or equivalently, to account for more of the variance of the criterion variable.

A stepwise procedure will be used in the regression analysis. Stepwise
procedures allow the selection of the one predictor variable that accounts for the most
variance in the criterion variable, and then one at a time, add the variables which account
for the most of the remaining or residual unexplained variance (Kachigan, 1991 ). The
predictor variables are continuously introduced until the resulting increase in R2
(explained variance) becomes insigniﬁcant. The R2 value is the squared correlation

produced by the Pearson test.

53

C I l' |' ID I' l’ I'

The dependent variable in this study is antisocial behavior. The independent
variable is parental attachment. The two variables are conceptualized and operationalized
in this study as follows:

Antisocial Behavior in this study is deﬁned by acts, the detection of which is

thought to result in punishment of the person committing them, by agents of the

larger society. Antisocial behavior is going against the conventional social order
because of lack of concern for the moral ethic set forth by the conventional
society (Hirschi, 1969). This variable will be measured using the responses of the

questions in Appendix B.

Parental Attachment is the psychological bond that a child shares with his

conventional parent/s. The bond represents the psychological presence of the

parent and the presence of this bond depends very much on the extent to which
the child interacts with the parent/s on a personal basis. It is thought that the
stronger this bond, the more strongly the child is bound to his parents’
expectations, and therefore the more strongly the child is bound to conformity

with the legal norms of the larger system (Hirschi, 1969). This variable will be
measured using the responses of the questions in Appendix B.

DataILansformations

This section describes the process that was undertaken to transform the primary
data to measurable form for the purpose of this study.

The dependent variable for this study is antisocial behavior. Sixteen separate acts
in the self-reported delinquency section were chosen to represent antisocial behavior. All
of the antisocial offenses were combined into a single index that represents general or
total participation in antisocial behaviors. Since individual antisocial items are rather
diverse, an attempt was made to reduce them to smaller more manageable sets of
indicators (sub-indexes). Four components emerged from a factor analysis indicating
four dimensions of antisocial involvement. The speciﬁc acts that correlated the highest
with each other were combined to form the speciﬁc antisocial indexes. The ﬁrst speciﬁc

antisocial index indicates acts of theft and illegal drug distribution (Table 1).

54

Table 1 Sold Illegal Drugs and Acts of Theft

 

 

n = 918 Mean SD Factor Loading
Sold Marijuana 1 .95 .674
Sold Hard Drugs 1 .41 .798
Stolen Motor Veh. 1 .31 .395
Stolen Items Worth >

$50 1 .41 .875
Stolen Items Worth <

$50 1 .85 .643
Stolen Items Worth

(SS-$50) l .51 .836

 

Cronbach’s Alpha = .75
The second antisocial index indicates participation in acts of robbery of fellow students

and others (not including parents and teachers, Table 2).

 

 

Table 2 Robbery of Others
11 = 918 Mean SD Factor Loading
Robbed Others (not

students or

teachers) 1 .34 .88
Robbed Students 1 .44 .88

 

Cronbach’s Alpha = .70
The third antisocial index includes participation in other violent antisocial acts (Table 3).

Table 3 Violent Behaviors

 

 

n = 918 Mean SD Factor Loading
Attacked Someone I .40 .72
Hit Other Students 2 1 .70
Carried Hidden

Weapon 1 1 .73
Been in Gang Fight 1 .66 .77

 

Cronbach’s Alpha = .63
The fourth factor of antisocial behavior includes attacks on authority ﬁgures such as

parents and teachers (Table 4).

55

Table 4 Assault on Authority Figures

 

 

n = 918 Mean SD Factor Loading
Hit Teacher 1 .48 .59
Hit Parent 1 .29 .78
Robbed Teacher 1 .13 .72

 

Cronbach’s Alpha = .40

The factor loadings obtained from the varimax rotation for the individual items were all
.46 or above. Refer to Appendix A for the speciﬁc items in each index. Demographic
variables are also used in the study. The demographic variables are ethnicity, age, grade
in school, and grade point average.

The independent variable is attachment to parents. After reviewing the codebook
for the National Youth Survey (Wave 11), six items were chosen to represent attachment
to parents. Factor Analysis revealed loadings on one factor. The extracted factor was
named attachment to parents. The combined items were used to form the Parental
Attachment scale (Table 5).

Table 5 Parental Attachment

 

 

n = 918 Mean SD Factor Loading
Inﬂuence of Parents 4 .98 .62
Parents Believe You

Do Things

Well 4 .99 .61
Parent Comforting

Important 2 .81 .71
Good Impression on

Parents 4 .75 -.43
Get Along With

Parents 4 1 .73
Importance of

Honesty 5 .74 .57

 

Cronbach’s Alpha = .52

The following Hypotheses are tested in this study:

56

H12

H22

H92

There is a negative association between the parental attachment variable
and the combined antisocial behavior variable.

There is a negative association between the parental attachment variable
and various acts of illegal drug distribution and acts of theft.

There is a negative association between the parental attachment variable
and violent antisocial acts.

There is a negative association between the parental attachment variable
and acts of robbery of fellow students and others (not parents or teachers).

There is a negative association between the parental attachment variable
and assault on authority ﬁgures.

There is a negative association between grade in school and all of the
antisocial measures.

There is a negative association between age and the all of the antisocial
measures.

There is a negative association between grade point average and all of the
antisocial measures.

There is no association between ethnicity and all of the antisocial
measures.

57

CHAPTER FOUR
RESULTS

g l D . I' [ II S I -S 1

This chapter will present the ﬁndings of this study. Through the use of univariate
analysis, summary statistics describe the sub-sample population. Clearly stated, the
summary statistics are obtained through the usage if frequency distribution outputs. The
sub-sample population consists of the entire male population available in the original
study (n=918).

From the sub-sample population the ages of the males range from 12 to 18 years.
The ages of the males in the sub-sample population are basically evenly distributed with
each age group containing between 12% and 16% of the total sub-sample population.
The median age of the males is 15 years. Eighty percent (n=101) of the 12 year olds in
the sub-sample were White, 13% (11: 16) were Black, 7% (n=9) were Mexican-
American, and .8% (n=1) were Asian-American. Forty seven percent (n=60) of the 12
year olds were in middle school, 45% (n=57) were in high school, 3% (n=4) were ﬁrst or
second year college students, and 5% (n=6) were not in school. Among the 13 year olds,
77% (n=98) and 17% (n=22) were White and Black respectively. Six percent (n=7) and
.8% (n=1) made up the Mexican-American and American-Indian population among 13
year olds. Thirty-four percent (n=44) and 60% (n=77) of the 13 year olds were in middle
school and high school, while 2% (n=3) and 3% (n=4) were either ﬁrst or second year
college students or those not in school respectively.

With reference to the 14 year olds, 74% (n=100) were White, 20% (n=27) were
Black, 4% (n=6) were Mexican-American, and .7% (n=1) were American-Indian and

Asian-American collectively. Thirty six percent (n=48) and 53% (n=71) were in middle

58

and high school, while 4% (n=6) and 7% (n=10) were ﬁrst or second year college
students and those not in school respectively.

The 15 year olds numbered 74% (n=109) and 20% (n=29) which represented the
White and Black populations respectively. The Mexican-American and Asian-American
population made up 5% (n=7) and 2% (n=3) of the 15 year old sub-sample population
respectively. Thirty-one percent (n=46) of the 15 year olds were in middle school while
59% (n=87) were in high school. The ﬁrst and second year college students made up 3%
(n=4) of the 15 year old population while 7% (n=11) were not in school.

The frequency outputs indicated that among the 16 year olds, 80% (n=110) were
White males and 15% (n=21) were Black males. The Mexican-American population
represented 4% of the 16 year olds, while both the American-Indian and Asian-American
population numbered .7% (n=1) collectively. Among the 16 year olds, 41% (n=57) were
in middle school and 45% (n=62) were in high school. In addition, 7% (n=9) and 7%
(n=10) were ﬁrst and second year college students and those not in school respectively.

Next, among the 17 year olds, 79% (n=104) and 15% (n=20) were White and
Black respectively, while 3% (n=4) and 7% (n=9) were ﬁrst and second year college
students and those not in school respectively.

This section will report the general descriptions of the youths’ antisocial behavior.
Frequencies were obtained for 16 antisocial behavior items. Those items are listed in
Appendix A. Almost 7% of the sub-sample indicate participation in all 16 of the
antisocial behaviors once or twice a year while 1% admit to participating in all 16
behaviors at least once a month. One and a half percent of the males indicate

participating in all of the antisocial behaviors once every two to three months with the

59

exception of sold hard drugs. The closer the responses are to a higher frequency of
participation, the smaller the percentages. However, the number of acts becomes more
frequent. For example, only 5% 0f the sub-sample report engaging in antisocial acts once
every 2-3 weeks. Those acts are (sold marijuana, carried hidden weapon, broken into
building, stole items worth less than $5, stole items worth between $5-$50, attacked
someone, been in gang ﬁghts, hit teacher, and hit other students). The same percentage
indicates those respondents who answered 2-3 times per week to participating in half of
the total available behaviors (same as the previous category with the addition of sold hard
drugs, stole items worth more than $50, and used force on other students). An even
smaller percentage (.4%) indicates those who answered once a week to being involved in
eight select antisocial behaviors (sold marijuana, carried hidden weapon, broken into
building, stolen items worth between $5-$50, attacked someone, and used force on
students). The same percentage of .4% represents those youths who responded to
participating in six of the 16 antisocial behaviors once a day (same as above category
with the addition of been in gang ﬁghts, and hit other students).

Among the median age of 15, the most “popular” antisocial act is hit other
students. The percentage is 41% (n=5 8), and the rate is once or twice a year, which is the
second most “popular” act among the median age group. The most “popular” antisocial
act among the high school and middle school youths is hit other students once or twice a
year. The percentages are 37% (n=164) and 35% (n=118) respectively. Hit other
students once or twice a year is also the act most participated in by the Black and White
sub-sample population (32%, n=45; 38% (n=25 8) respectively. The next section of the
results will discuss the descriptive characteristics that univariate analysis produced for

attached youth.

60

E III 'l' [ill I I]: II

This section will reveal the general characteristics of the attached youth. With
keeping in mind the univariate results of the antisocial youth, it is probable that the
number of youths responding to the attachment items in a way leaning towards high
parental attachment will be high. The frequency output indicates that ahnost 2% (n=17)
of the sub—sample’s parents inﬂuence them a little while an expected high percentage of
43% (n=390) indicates those respondents who felt their parents inﬂuenced them a great
deal. The frequency statistics indicate that 69% (n=634) of the males think that it is very
important that their parents think they do things exceptionally well while only 2% (n=l6)
of the sub-sample thinks that it is not important at all. The frequencies also indicate that
well over half of the youths think it is very important that their parents comfort them
when they are feeling depressed or stressed (61%, n=560). An exceptionally high 82%
(n=750) of the youths indicate that it is very important to get along with their parents.
The important to be honest with parents category is one of only two categories where the
majority of the sub-sample did not answer to the extreme response with reference to high
positive attachments. Importance of parental impression is the other category. Although
57% (n=521) answered agree to the honest with parents category, 26% (n=240) strongly
agreed. The ﬁnal category; with parents, good impression important even if it means
being dishonest drew a percentage of 62% (n=567) to the disagree response while 18%
(n=169) strongly disagreed. For the strongly agree response to the same question, .9%
(n=8) of the sub-sample answered strongly disagreed.
W

A factor analysis was conducted on all of the diverse antisocial acts. Factor

analysis revealed four components indicating four dimensions of antisocial behavior.

61

this allowed the researcher to compute the distinct components into four categories or
factors (Tables 1-5). These categories are SOLDSTO, which measures participation in
acts of theft and illegal drug distribution. The next category is ROBBOTH, which
measures participation in acts of robbery of fellow students or other (not including
parents and teachers). The VIOLENT category measures violent antisocial acts, and
ASSAULT measures attacks on authority ﬁgures (parents and teachers). The ANTISOC
measure represents all 16 antisocial items. The items were combined into a single index
that represents general participation in antisocial activities. For the ANTISOC variable,
the responses were recoded into low, medium, and high antisocial involvement. The four
separate indices of antisocial behavior were recoded from having seven
responses(Appendix B) to four responses. The responses were recoded to “almost never
or seldom”, “once or twice a month”, “once or twice a week”, and “once or twice a day”.
For the computed variable SOLDSTO, 95% (n=830) of the sub-sample fall into the
almost never or seldom category while .9% (n=8) are included in the once or twice a day
category. Percentages of 3% (n=30) and 1% (n=9) fell into the once or twice a month
and once or twice a week categories respectively. The results of the frequency output for
the ROBBOTH variable show that 97% (n=851) of the sub-sample are a part of the
almost never or seldom category while on the extreme .7% (n=6) fall into the once or
twice a day categories respectively. The variable VIOLENT includes 75% (n=603) of the
sub-sample which falls into the once or twice a month category while 21% (n=169) and
5% (n=3 6) are a part of the once or twice a week and once a day categories respectively.
Eighty-eight percent (n=765) of the sample population are a part of the almost never or

seldom category while a percentage of .1% (n=1) is indicated for those who fall into the

extreme category of once or twice a day. The remaining percentages of 11% (n=93) and

62

1% (n=11) are inclusive of the once or twice a month and once or twice a day categories
respectively. Finally, the univariate statistics indicate that 94% (n=815) of the
respondents indicate low antisocial involvement, 5% (n=42) indicate middle antisocial
involvement, and 2% (n=l4) indicate high rates of antisocial involvement.
Will:

The independent variable parental attachment is represented by the single index
ATTPAREN (Table 5). The six attachment indicators were combined to accomplish this.
Factor analysis revealed no signiﬁcant differences between the indictors of attachment of
the subjects of this study. The attachment index was recoded into low, medium, and high
parental attachment. The univariate analysis revealed that 5 7% (n=520) of the
respondents have a high attachment, 37% (n=36) have a medium attachment, and 7%
(n=61) indicate having a low attachment to their parents.
W11:

This section will present the ﬁndings of the contingency analysis. The results of
the contingency analysis indicate that total attachment and total antisocial behavior are
related statistically and negatively in the population; however, the individual variables
that make up the total antisocial variables are not statistically signiﬁcant when crossed
with total attachment. Therefore, it can be concluded that combined, the antisocial items
are more readily related to total attachment than they would be individually. The results
of the contingency analysis are below.

The contingency analysis results indicate that the two variables are related. A chi-
square of 11.53 and a critical value of 9.48 indicate that the two variables are statistically

signiﬁcant in the population. Since the probability is smaller than .05 ( p=.021) the

63

hypothesis of independence between total attachment and total antisocial behavior is
rejected and a statistically signiﬁcant relationship is indicated. Therefore, the more that a
child is attached to is parent, the less likely he will participate in antisocial acts. It can
also be assumed that the higher the attachment the fewer the number of respondents in
the antisocial categories as the categories go from low to high involvement. This is
evident where 96% of the highly attached youths are in the low antisocial category, 3%
are in the middle antisocial category, and 1% are in the high antisocial category.
1.. 311111'1 1.11; 1 '_°-. 31.?! 11,111,115- 0 Ir'i

According to the contingency analysis output, the two variables are not
statistically signiﬁcant (chi-square = 3.86; critical value = 9.48). The probability is over
the accepted value of .05 (p = .145) which also allows the null hypothesis of
independence to be accepted.
I l ! l 1 If 1 i . . l 5

The probability is larger than .05 (p = .114). The hypothesis of independence
between total attachment and violent antisocial acts is accepted. A chi-square of 4.346
and a critical value of 6 indicate that the variables are not statistically signiﬁcant.

_ O . . - -
0 a a- 1 L" 1 11¢ a .0 \000 o .1 1 11! C 1 It! 2 1 11!

199911115)

Again, the probability is larger than .05 (p = .471). The hypothesis of
independence between total attachment and acts of robbery is accepted (chi-square =
1.505; critical value = 6).

01.312111711103 -. 3211 6.10. ‘1 2.11 ‘1 1‘
The two variables total attachment and assault against authority ﬁgures are also
found to be independent in the population (chi-square = .674; critical value = 6). The

probability value is again larger than .05 (_ p = .714).

64

115.1 llII'E 1.1,.“

Contingency analyses were conducted with the antisocial measures and the
various demographic variables. When ethnicity was crosstabulated with total antisocial
behavior, a chi-square value of 38.27 and a critical value of 18.307 allow the hypothesis
of independence to be rejected. The probability is small (p = .000) therefore there is a
less than 5% chance of incorrectly rejecting the hypothesis of no difference. Another
weak statistically signiﬁcant relationship is indicated for the SOLDSTO variable and
ethnicity. The hypothesis of independence is rejected (chi-square = 24.84; critical value
= 11.07) and the chance of making a Type I error is less than 5% (p = .000). Ethnicity
and acts of robbery are indicated as being dependent in the population (chi-square =
61.45; critical value 11.07). The probability of p = .000 allows the hypotheses of
independence to be rejected. The ﬁnal two measures of assault against authority ﬁgures
and various acts are independent of ethnicity in the population (chi-square = 1.41, critical
value = 11.07; chi-square = 5.6, critical value = 11.07 respectively). For the remainder of
the demographic variables there are no relationships indicated by the contingency
analysis with exception of total antisocial behavior and age. The hypothesis of
independence is rejected in this case because the chi-square was indicated as 13.78 and
the critical value is indicated as 12.59. There was a less than 5% chance of making a
Type I error (p = .032).

' 11"11.--11_111'1 1 ‘2'11111’11'; 1 '1'11 1'1 "1

The results indicate considerable support for the hypothesis of this study. The
antisocial indices are inversely associated to parental attachment and are statistically
signiﬁcant in the population with the exception of assault on authority ﬁgures

(ASSAULT), which is inversely associated to parental attachment, but not

65

statistically signiﬁcant (r = .022, p = .269). The correlation matrix indicates that the
strength of the associations are; however, weak ( Table 6 Appendix C).

A partial correlation coefﬁcient analysis was conducted in order to examine the
effect of parental attachment on antisocial behavior with grade, ethnicity, grade point
average, and age held constant in four separate instances. The goal was to test the
independent effect of parental attachment. There were no signiﬁcant differences from the
partial relationships and the original relationship. A replication was achieved when
grade, ethnicity, grade point average, and age were controlled (Table 7).

Table 7 Partial Correlations of Parental Attachment and Antisocial Behavior

 

 

Constant Partial
Grade -.092**
Ethnicity -.093**
Grade Point Avg. -.091**
Age -.090**

 

Original Relationship = -.095**

Concerning the demographic variable grade in school, a predicted inverse
association is indicated for all of the antisocial measures with the exception of the
robbery measure, which indicates a very weak positive association (Table 6; r = .017).
Although the association is not statistically signiﬁcant, the bivariate result, which are
supported by the frequencies suggest that middle school students participate in antisocial
behaviors more that high school students and more than 1St and 2nd year college students.
The hypothesized inverse associations with reference to age and the antisocial measure
indicate that there are more positive associations than negative association. This could be
misleading when considering that middle school students participate in more antisocial

activities; however, it must be remembered that almost 40% of the 18 year olds were in

66

middle school, 41% of the 17 and 16 year olds, and 31% of the 15 year olds were in
middle school grades. Therefore, the ages of the middle school population should be
taken into consideration. This is especially true when we look at the statistically
signiﬁcant associations between age and the SOLDSTO measure (r = .072*).

The hypothesized inverse association between grade point average and antisocial
behavior was indicated for acts of robbery and the violent acts measure (Table 6), but not
signiﬁcantly. There were no signiﬁcant associations between ethnicity and antisocial
behavior. The bivariate associations between the parental attachment variable and the
measures of antisocial behavior serve as support, albeit weak, for the viability of the
research hypotheses. The question remains; however, of the overall impact of the
attachment variable on the explanation of the range of antisocial behaviors. To address
this question, the results of the multiple regression will be discussed.
WWI:

Stepwise regression analysis was conducted to determine which independent
variables offered the best explanation of antisocial behavior. Based on the stepwise
regression, when parental attachment and age were regressed on acts of theft and illegal
drug distribution it was indicated that parental attachment and age explained only 1% of
the variation in the theft and illegal drug distribution measures (Table 8). Although the
explained variation is low, the slope of the regression line is statistically signiﬁcant. The
signiﬁcant t level is .039. The standardized slope or the Beta weight for parental
attachment is -.071 and for age the Beta weight is .071 which signiﬁes very weak
negative and positive relationships with acts of theft and illegal drug distribution.

The next two regression analyses were enter method analyses, instead of stepwise

because there is only one predictor variable, parental attachment.

67

Table 8 Stepwise Regression Analysis

 

 

 

SOLDSTO
Standardized t sig.
Parental Attach. -.071 -2.07* .039
Age .071 2.06* .039
R = .01
*p S .05

When parental attachment was regressed on antisocial acts, the explained
variation was less than 1% (Table 9). Less than 1% explained variation was also
indicated when parental attachment was regressed on violent antisocial acts (Table 10).
The slope for parental attachment and antisocial acts was found to be signiﬁcant and the
Beta weight was indicated as weak and negative ( t = .000; ,6 = -.O9). The indicated slope
of the regression for parental attachment and violent antisocial acts was found to be
signiﬁcant ( t = .000) and the relationship id indicated as a weak negative relationship (,6
= -.O9).

Table 9 Regression Analysis

 

 

 

 

 

AN T ISOC

Standardized t sig.
Parental Attach. -.09 -2.69* .000
R2.- .008
*ps .05
Table 10 Regression Analysis

VIOLENT

Standardized t sig.

Parental Attach. -.09 -2.65* .000

R2: .008

 

68

69

CHAPTER FIVE

CONCLUSIONS
Conclusions

To recap, this thesis sought to determine whether or not there was an association

between parental attachment and male antisocial behavior. Travis Hirschi’s social
control theory was used as the theoretical framework for this project. Control theory’s
element of attachment was used to pursue the research objectives of this project. As
hypothesized, a negative association was found to exist between parental and male
antisocial behavior. Speciﬁcally, the antisocial behaviors that were found to be
negatively associated with parental attachment were the violent acts and illegal drug and
theft measures (see Table 6). Although these associations were consistent with the
hypotheses of the study, the regression analysis revealed that the parental attachment
variable accounted for only 0% to 1% of the variance in antisocial behavior. These
ﬁndings are consistent with Agnew’s 1985 study, which suggested that social control
variables such as parental attachment have a very small impact on antisocial behavior.
Although the regression analyses were damaging to the theory used in this study, the
regressions do not prove that social control theory is unimportant in explaining future
antisocial behavior. The antisocial acts used in this study were for the most part
considered serious by the public at large. According to studies (see criticisms in Chapter
2) control theory is best able to explain antisocial acts that are minor in nature. Also, this
study employed a cross-sectional approach and researchers have suggested that cross-
sectional studies are less efﬁcient in explaining antisocial behavior than longitudinal
studies.

This study assessed whether or not there was an association between parental

70

attachment and antisocial behavior. It was thought that social control theory, as proposed
by Hirschi would account for more of the variance in antisocial behavior that was
indicated in this study. However, it must be noted that many of the variables had to be
computed using available data, and in some instances, what was available was not
thought 0 be perfect enough measures for the target variable. It is important to note that
the original investigator’s purpose in doing his survey was to collect information on
speciﬁcs not conducive to this studies goals and objectives. This is why many of the
variables under consideration for this study were not adequately measurable.

There is a great deal of information conceming antisocial behavior among male
children; consequently, this is an open ﬁeld for further research in the area.
Recommendations for further study in this area would be to undertake the problem using
a primary data collection. Primary data collection’s purpose would be to design an exact
instrument for data collection that would consists of more accurate measures of the
variables used in this study.

The limitations of this study are associated with the use of secondary analyses and
cross-sectional studies. First, the question of validity surfaces when conducting a
secondary analysis. The secondary analysis is limited to using variables possibly
operationalized differently than what the researcher is interested in. Thus, it is
questionable whether secondary data provides valid measures of the variable of interest
(Babbie, 1995). Another limitation of this study is with reference to the indices and
scales used in the original study. If there were problems with the original scales and
indices, the secondary analysts must deal with those problems accordingly (Davis, 1996).

The third limitation of this study is the usage of cross-sectional data. Although

71

the original data used in this study is a part of a longitudinal study only one wave will be
used therefore making the analysis cross-sectional. Agnew (1991) has asserted that
Hirschi’s control theory (1969) must be studied under longitudinal conditions because
longitudinal studies have found that several of Hirschi’s social bond variables explain
only 1% to 0% of the variance in delinquency. Several researchers have asserted that
longitudinal studies have indicated that Hirschi’s social control theory explains only
trivial types of delinquency and not major types of delinquency (Agnew, 1985, 1991;
Paternoster & Iovanni, 1986; Paternoster, 1988; Massey & Krohn, 1986; Elliot,
Huizinga, & Ageton, 1985; Paternoster, Saltzman, & Waldo, 1983; Wells & Rankin,
1983). The longitudinal data in other words, suggest that cross-sectional studies have

greatly exaggerated the importance of Hirschi’s theory.

72

APPENDICES

73

APPENDIX A
ITEMS USED IN THE STUDY

74

APPENDIX A

 

Items

Items as Worded

 

ATTPAREN-attachment to parents

SOLDSTO—selling illegal drugs and acts of theft

75

How much have your parents
inﬂuenced what you’ve
thought and done? How
important is it for you to have
parents who comfort you
when you’re unhappy about
something? How important
is it for you to have your
parents think you do things
well? How important is it for
you to get along well with
your parents? How much do
you agree or disagree that it
is important to be honest with
your parents, even if they
become upset or you get
punished? How much do you
agree or disagree that making
a good impression is more
important than telling the
truth about parents?

How many times in the last
year have you sold marijuana
or hashish (“pot”, “grass”,
“hash”)? How many times in
the last year have you sold
hard drugs such as heroin,
cocaine, and LSD? How
many times in the last year
have you stolen (or tried to
steal) a motor vehicle, such
as a car or motorcycle? How
many times in the last year
have you stolen (or tried to
steal) something worth more
than $50? How many times
in the last year have you
stolen (or tried to steal)
something worth less than
$5? How many times in the
last year have you stolen (or

VIOLENT-violent antisocial acts

ROBBERY—acts of robbery

ASSAULT-assault on authority ﬁgures

ANTISOC-combined antisocial

76

tried to steal) things worth
between $5 and $50?

How many times in the last
year have you attacked
someone? How many times
in the last year have you hit
(or threatened to hit) other
students? How many times
in the last year have you been
involved in gang ﬁghts?
How many times in the last
year have you carried a
hidden weapon other than a
plain pocket knife?

How many times in the last
year have you used force
(strong-ann methods) to get
money or things from other
people (not students or
teachers)? How many times
in the last year have you used
force (strong-anti methods)
to get money ﬁom other
students?

How many times in the last
year have you hit (or
threatened to it) one of your
parents? How many times in
the last year have you hit (or
threatened to hit) your
teacher? How many times in
the last year have you used
force (strong-ann methods)
to get money or things from a
teacher or other adult at
school?

All items in VIOLENT,
ROBBERY, ASSAULT, and
SOLDSTO are combined
together to form a single
index.

APPENDIX B
NATIONAL YOUTH SURVEY WAVE 11
(IN PART)

77

APPENDIX B

This section deals with your own behavior. I’d like to remind you that all your answers
are conﬁdential. I’ll read a series of behaviors to you. Please give me your best estimate
of the EXACT NUMBER of times you’ve engaged in each behavior during the last year.
(FOR ANY BEHAVIOR THAT THE RESPONDENT HAS ENGAGED IN 10 OR
MORE IN THE LAST YEAR, RECORD THE RESPONSES TO THE SECOND
COLUMN SAYING “Please look at the responses on the orange card and select the one
which best describes how often you are involved in this behavior.”)

Once Once Once 2-3 Once 2-3
A Every A times a times
Month 2-3 Week A Day A Day
Weeks Week
How many times in the LAST

YEAR have you:

1. Stolen (or tried to
steal) a motor
vehicle, such as l 2 3 4 5 6
a car or
motorcycle?

2. Stolen (or tried to
steal) something 1 2 3 4 5 6

worth more than
$50?

3. carried a hidden 1 2 3 4 5 6
weapon other than
a plain pocket
knife?

4. Stolen (or tried
steal) things worth 1 2 3 4 5 6
$5 or less?

5. Attacked someone? I 2 3 4 5 6

6. Been involved in
gang ﬁghts? l 2 3 4 5 6

7. Sold marijuana or
hashish? 1 2 3 4 5 6

8. Hit students? 1 2 3 4 5 6

78

Hit a parent? 1 2 3 4 5 6

10. Sold hard drugs
such as heroin, 1 2 3 4 5 6
cocaine, and LSD?

11. Used force (strong-arm)
to get things from
other students? 1 2 3 4 5 6

12. Used force to get
things from a teacher
or other adult? 1 2 3 4 5 6

13. Used force to get
things from people
(not students or 1 2 3 4 5 6
teachers)?

14. Stolen (or tried
to steal) things
worth between 1 2 3 4 5 6
$5-$50?

15. Broken into a
building or vehicle
(or tried) to steal 1 2 3 4 5 6
something or just
to look around?

16. Hit a teacher? 1 2 3 4 5 6
In this section, I would like to ask you some questions concerning how you feel about
your parents. Please answer as accurately as possible. Thank you.

1. How much have your parents inﬂuenced what you’ve thought and done?

A Great Deal Quite a Bit Some Not too Much Very Little
5 4 3 2 1
2. Its important to be honest with your parents, even if they become upset or you get
punished.
A Great Deal Quite a Bit Some Not too Much Very Little
5 4 3 2 1
3. Making a good impression is more important than telling parents the truth.
A Great Deal Quite a Bit Some Not too Much Very Little
5 4 3 2 1

79

How important is it to you...

4. to have parents who comfort you when you are unhappy about something?

Very Important Somewhat Important Not Important
5 3 1
5. to have your parents think you do things well?
Very Important Somewhat Important Not Important
5 3 1
6. to get along well with your parents?
Very Important Somewhat Important Not Important
5 3 1

80

APPENDIX C
ZERO-ORDER CORRELATION COEFFICIENT MATRIX
OF PARENTAL ATTACHMENTS AND
ANTISOCIAL BEHAVIOR

81

93 M Z 835m:
65. mo. 2: an “585:3... ﬂ noun—9.30 4
6.6. S. 2: 3 2.35:»? m_ «.33—0......03

 

c2; “.5. main Ewe:— Eom u m> own-52% 2:...— 0620 n X
.9233— ..c 33.. u Sr «3.. n .X

394 3982.3 use—c; u 9% .Eﬂﬁzm u nX
mouswmm 522.23 E. “—533. u ~> Begum E 2:20 n «X
3683.2 3:. .H u Cw EoEﬁaﬁ< .555..— n .X

I «sham. «rwNm. ermMN. ecbbc. mac. aﬂhc. 25. mncf *4.ch.1 m?
11 imam. :3... team. ms: 9....- ns. 2... 3.. c»

1|... :8... 3.2:. :8: .2. cm... E..- tS..- or

II 3.9%. mg. E...- «m.... an..- «8.- 9

i 2... mm... «8. 2...- 33...- C»

II :2..- ..«...- E..- m8.- mx

1: E...- c«... 3......- .x

m2..- NS. mX

mac-c.- «X

.X
m? Cr 9 Cr c» mX vX AX «X .X
.8323— 39832 2:. 358.383. .852..— .«o x232 «8:29.30 couLOéLoN e 935,—.

82

LITERATURE CITED

83

REFERENCES

Dataﬁollection

Elliot, Delbert. NATIONAL YOUTH SURVEY [UNITED STATES]: WAVE
II, 1977 [Computer ﬁle]. ICPSR version. Boulder, CO: University of Colorado,
Behavioral Research Institute [producer], 1978. Ann Arbor, MI: Inter-university
Consortium for Political and Social Research [distributor], 1994.

Books

Ainsworth, M. D. S., Blehar, M. C. Walters, E., &Wally, S. (1978). Battemsof

WW Erlbaum Associates,
Hillsdale, NJ.

Aker81R-L- (1973). DexiantheleciaUcamingmmar-h- Belmont,
CA: Wadsworth Publishing.

Akers,R- L- (1994). Criminologicaljheoticszmmgnandmluation. Los
Angeles: Roxbury.

American Psychological Association (1995). PublicationJAanuaL (4th ed.).
Washington, DC: American Psychological Association.

Bandura, A. (1977). Socialleamjngﬂeom Englewood Cliffs, New Jersey:
Prentice-Hall, Inc.

Champion, J. D. (1992). Wm New York: Macmillan
Publishing Company.

Cloward, R. A. & Ohlin, L. E. (1960). MW New York:

Free Press.

Cohen,A. (1955). Qelirrquemhoxsnlheculmmfrhegang. Glencoe, IL: The

Free Press.

Durkheim, E. (1961). MoraLeducation, New York: The Free Press.

Elliot, D. S., Huizinga, D, & Ageton, S. (1985). WWW
durgjrse, Beverly Hills, CA: SAGE.

84

Farrington, D. P., Ohlin, L. E. &Wilson, J. Q. (1986). Unde_s_tandr_ng_and
99W New York: Springer-Veriag.

Fowler, F. J. (1993). W Newbury, Park: SAGE.

Glueck, S., & Glueck, E. (1962). WWW Boston:
Houghton Mifﬂin Company.

Graham, A. (1994). W Chicago: NTC/ Contemporary
Publishing Company.

Haley, J. (1980). LoayinghomoNew York: McGraw-Hill.
Hirschi, T. (1969). MW Berkeley & Los Angeles: University

of California Press.

Johnson-R E- (1979). luxemmdehnquencxanirtsongms- Cambridge:

Cambridge University Press.

Kachigan, S- K. (1991). WWW
introduction, New York: Radius Press.

Maxﬁeld, M. G., &Babbie, E. (1995). Researchmethodsfomﬁminaljustise
a_d_c_[imino1ogy, Belmont, California: Wadsworth Publishing Company.

Mckinley, D. G. (1964). Sooialolasiandfamilylife, New York: The

Free Press.

Norusis, M. J. (1996). WWW Englewood Cliffs, New

Jersey: Prentice-Hall, Inc.

Norusis, M. J. (1997). W Upper Saddle River,

New Jersey: Prentice-Hall, Inc.

Reckless, W., & Smith, M. (1932). ,Inyenilooeiingnonoy, New York: McGraw—
Hill Book Company.

Rosenberg, K. M. (1990). WWW Dubuque, IA:
Wm. C. Brown Publishers.

Sutherland, E. H., & Cressey, D. R. (1974). Criminology, (9th ed.).
Philadelphia: Lippincott Press.

Trojanowicz, R., & Morash, M. (1992). W
control, (5th ed.). Englewood Cliffs, New Jersey. Prentice-Hall, Inc.

Vold, G. B., Bernard, T. J ., & Snipes, J. B. (1998). flfhggmjjga] 91211111anng
(4th ed.). Oxford: Oxford University Press.

85

P'l'l

Achenbach, T. M. ,& Edelbrock, C. S. (1979). The classiﬁcation of child

psychopathology: A review and analysis of empirical efforts. Esyohologioalﬁniloiin,
85, 1275-1301.

Agnew, R. (1985). Social control theory and delinquency: A longitudinal test.
Criminolmll. 47-61-

Agnew, R. (1991). A longitudinal test of social control theory and delinquency.
IoumalofResr-atchmﬂnmoandﬂehnquencxﬁ 126- 156

Akers, R. L., Krohn, M. D. HLanza-Kaduce L. &Radosevich, M. (1979). Social
learning and deviant behavior: A speciﬁc test of a general theory. muonoan

Sooiologioal Boxing, 44, 636- 655.

Arrnsden, G. C., & Greenberg, M. T. (1987). The inventory of parent and peer
attachment: Individual differences and their relationship to psychological well-being in

adolescence. loumalofiomhandAdolescenceJo. 427-453.

Blos, P. (1967). The second individuation process of adolescence.

BsrmhoanalxsrsShrdLChilMZ. 162- 186

Burgess, R. L., & Akers, R. L. (1966). A differential association - reinforcement
theory of criminal behavior. Sociaiﬁrohlomoﬁ 128-147.

Cemkovich, J. D., & Giordano, P. C. (1987). Family relationships and
delinquency. Cn'minology._25, 295-321.

Cheek, J. M. ,&Briggs, S. R. (1982). Self-consciousness and aspects of 1dent1ty
Ioumalofkesearchinﬁersonahm 401-408

Conger, R. (1976). Social control and social learning models of delinquent

behavior. Criminology,_1_4__l 7-40.

Elliot, D. S., &Ageton, S. S. (1980). Reconciling race class differences in self-

reported and ofﬁcial estimates of delinquency. AmonoanSocioogioaLRﬂiLAL 95-
110.

Forgatch, M. S., Chamberlain, P., & Gabrielson, P. (1982). A video training tape.
Eugene, Oregon: Castalia.

Gardner, L. & Shoemaker, D. J. (1989). Socialbonding and delinquency: A
comparative analysis. WM 481- 500.

Greenberg, M. T., Siegel, J. M., & Leitch, C. J. (1990). The nature and
importance of attachment relationships to parents and peers during adolescence. Jonmai

oonuthandAdolescence-LLZ. 401-408.
86

Grove, W., & Crutchﬁeld (1982). The family and juvenile delinquency. Ibo
Sociclcgmalﬂuanﬂlm 301 -319

Hepburn, J. R. (1977). Testing alternative models of delinquent causation.

IoumalofELimimlemdﬁnminologxﬂ 450- 460

Hindelang, M. J. (1973). Causes of delinquency: A partial replication and
extension. W 471 -.487

Jensen, G. F. (1972). Parents, peers, and delinquent action: A test of the

differential association perspective. W562 575.

Junger, M. & Marshall, I. H. (1997). The interethnic generalizability of social

control theory: An empirical test. Ioumalofkeseamhjnﬁnmeandﬂelinqmli
79- 112.

Kenny, M. E. (1987). The extent and function of parental attachment among
ﬁrst-year college students. LoumﬂoﬂoujhoniAdolosomJo, 17-29.

Kenny, M. E., & Hart, K. (1992). Relationship between parental attachment and

eating disorders 1n an inpatient sample and a college sample. loumolofﬂounsoling
W521- 526

Krohn, M. D., & Massey, J. L. (1980). Social control and del1nquent behav1or

An examination of the elements of the social bond. Ibo Sociological Qoort oily, 2 l , 529-
543.

Lapsley, D. K., Rice, K. G., &FitzGerald, D. P. (1990). Adolescent attachment,
identity, and adjustment to college: Implications for the continuity of adaptation

hypothesis WWW 561-565

Larzelere, R. E., & Patterson, G. R. (1990). Parental management: Mediator of
the effect of socioeconomic status on early delinquency. W 301-.323

Le Blanc, M. (1994) Family, school, delinquency, and criminality, the
predictive power of an elaborated theory for males. CriminalﬂohaxionjnsLMemol
11mm 101- 117

Levenson, R. W., & Gottman, J. M. (1978). Toward the assessment of soc1al
competence. WWW 453- 468

Linden, R. (1978). Myths of middle-class delinquency: A test of the
generalizability of social control theory. Y_o_u_th_ano_Soojo1y,_2, 407-432.

Loeber, R. (1982). The stability of antisocial and delinquent child behavior: A
review Quinta/elem 1431- 1446

87

Massey, J. L. (1986). A longitudinal examination of an integrated process model

of deviant behavior. Sooialﬁomogﬁi, 106-134.

Matsueda, R. L. (1982). Testing control theory and differential association. A
causal modeling approach. WWW 489- 504.

Mastueda, R. L. ,& Heimer, K. (1987). Race, family structure, and delinquency:
A test of differential association and social control theories. Amorioanﬁooiologioal

Roviow, 52, 826- 840.

Matza, D., & Sykes, M. (1961). Juvenile delinquency and subterranean values.

WWW 712- 719

McCord, W, McCord, J. &Howard, A. (1963). Fam111al correlates of
aggression in nondelinquent male children. JoumaloﬂAonormolmdﬁooiol
Esyohology, 62.79- 93.

Mitchell, J, & Dodder, R. A. (1980). An examination of types of delinquency
through path analysis. W 239-248.

Olweus, D. (1979) Stability of aggressive reaction patterns in males: A review.

BsychologicaLBJuIetin._&§. 852- 875

Patterson, G. R, DeBaryshe, B. D., & Ramsey, E. (1989). A developmental
perspective on antisocial behavior. MW 329- 335.

Patterson, G. R., & Dishion, T. J. (1985). Contributions of families and peers to
delinquency. Co'minoiogy, 23, 63-79.

Patterson, G. R., & Stouthamer-Loeber, M. (1984). The correlation of family
management practices and delinquency. Chiliﬂexolopmonij, 1299-1307.

Paternoster, R., Saltzman, L. E. & Waldo, G. P. (1983). Perceived risk and social
control: Do sanctions really deter? Law and Sooioiy Royiow w, 1 Z, 457- 479.

Paternoster, R. & Iovanni, L. (1986). The deterrent effect of perceived severity:

A reexamination. W 751-777.

Poole, E. D. ,&Regoli, R. M. (1979). Parental support, delinquent friends, and

delinquency: A test of interaction effects. WWW
]_Q, 180- 193.

Reiss, A. J. (1951). Delinquency as the failure of personal and social controls.

W196- 207

Rosen, L. (1985). Family and delinquency: Structure or function. Criminology,
23, 553-573.

88

Rosenbaum, J. L. (1987). Social control, gender, and delinquency: An analysis
of drug, property, and violent offenders. W1 17-132.

Silberman, M. (1976). Toward a theory of criminal deterrence. Amorioao
SEW 442- 461

Thompson, W. E., Mitchell, J ., & Dodder, R. A. (1984). An empirical test of
hirschi’s control theory of delinquency. mm 11-22.

Wells, L. E., & Rankin, J. H. (1983). Self-concept as a mediating factor in
delinquency. SooiglﬂlchologLQuanoﬂon, 11-22.

Wells, L. E., & Rankin, J. H. (1988). Direct parental controls and delinquency.
Criminclcgm 263-285-

Wiatrowski, M. D. ,Griswold, D. B. ,,&Roberts M. K. (1981). Social control
theory and delinquency. MW 525- 541.

Scales
Garner, D. (1991). W Odessa, FL: Psychological

Assessment Resources.

Brown A M (1991) WWW

andJuLenﬂLdelinguencx. Unpublished master’s thesis, Michigan State University, East
Lansing.

DaVis, J- S- (1996). AnassessmenmfhelpzseekingbehaxioLamonwctimmf
Mam. Unpublished master’s thesis, North Carolina Central University,

Durham.

89

 

"111111111111111111111111111“