—
=
_
:
—
—
—
—
—
__.
—_
——
—
——
—

 

THS

 

Maegan... NIiliiiiifiiiiiifl‘w
Unlvorslty

 

 

 

This is to certify that the
thesis entitled
PARENTAL ATTACHMENT, PEER ATTACHMENT, STRAIN, AND VIOLENCE:
A STUDY OF THE INTERACTION BETWEEN COMPONENTS OF SOCIAL
CONTROL THEORY AND STRAIN THEORY

presented by

Megan Ann Euurma

has been accepted towards fulfillment
of the requirements for

Mdegree in .inminaLJustice

 

Date /l%191 43

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

 

 

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

 

DATE DUE

DATE DUE

DATE DUE

 

 

 

 

 

 

 

ovwm

 

 

 

Add 11 7 2008

 

 

 

 

 

 

 

 

 

 

1M W14

PARENTAL ATTACHMENT, PEER ATTACHMENT, STRAIN AND VIOLENCE:
A STUDY OF THE INTERACTION BETWEEN COMPONENTS OF SOCIAL
CONTROL THEORY AND STRAIN THEORY
By

Megan Ann Buurma

A THESIS

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

MASTER OF SCIENCE

Department of Criminal Justice

1998

ABSTRACT
PARENTAL ATTACHMENT. PEER ATTACHMENT, STRAIN, AND VIOLENCE:
A STUDY OF THE INTERACTION BETWEEN COMPONENTS OF SOCIAL
CONTROL THEORY AND STRAIN THEORY
By

Megan A. Buurma

Youth violence has become an alarming problem in the United States, and the
ability of current theories to explain youth violence is questionable. The purpose of this
study is to combine components of social control theory and strain theory to explore the
possible interaction that may exist between these two theories. Specifically, this study
seeks to examine the interaction between parental and peer attachment and the strain in
those relationships and how this affects violence in youth. This study uses data from
Wave I and H of the National Youth Survey. Simple linear regression is used to examine
the both the direct effects of the variables used in this model and the interactions. This
study finds that an interaction between attachment and stress does not exist, however,
results indicate that peer pressure, family income, and peer delinquency all are important

factors affecting violence in youth.

Copyright by
MEGAN ANN BUURMA
1998

ACKNOWLEDGMENTS

I would like to thank my committee members for helping me accomplish this
goal. I would also like to extend a special thanks to Dr. Christina DeJong who sacrificed
her time to direct and teach me the tools that I needed to complete this great task.

To my family and friends, thank you for giving me the support and
encouragement that I needed throughout the past year, I could not have been as
successful without you.

To Jacob, your support, persistence and love helped me realized that I can

accomplish anything. Thank you.

iv

TABLE OF CONTENTS

List of Tables ............................................................................................................ vi
Chapter 1: Introduction ............................................................................................ l - 3
Chapter 2: Theory .................................................................................................... 4 - 13
Youth Violence
Social Control Theory
Strain Theory
Combination Theory
Chapter 3: Goals and Objectives ............................................................................. 14 - 15
Chapter 4: Data ........................................................................................................ l6 - 17
Chapter 5: Methods ................................................................................................. 18 - 27
Scales
Analysis
Chapter 6: Results .................................................................................................... 28 - 39
Bivariate
Multivariate
Chapter 7: Discussion .............................................................................................. 40 - 45
Limitations and Recommendations
Implications
Appendix: Questions Included in the Variable Scales ............................................ 46 - 48
Bibliography ............................................................................................................. 49 - 52

Table 1:

Table 2:

Table 3:

Table 4:

Table 5:

Table 6:

Table 7:

Table 8:

Table 9:

LIST OF TABLES

Demographic Characteristics for Youth Surveyed in the National Youth
Survey, Wave 1, 1978 ................................................................................ 17
Factor Analysis Coefficient Scores for Parental Attachment .................... l9 '
Factor Analysis Coefficient Scores for Peer Attachment .......................... 21
Factor Analysis Coefficient Scores for Family Stress ............................... 22
Factor Analysis Coefficient Scores for Peer Stress ................................... 23
Factor Analysis Coefficient Scores for Delinquency ............................... 25
Factor Analysis Coefficient Scores for Delinquent Peers ......................... 26
Bivariate Regression Analyses for Continuous Variables on

Self-Reported Violent Behavior ................................................................ 29
Bivariate Analysis of Variance Results for the Categorical Variables

and Self—Reported Violent Behavior ......................................................... 31

Table 10: Summary of the Regression Analysis for Variables Predicting Violent

Behavior .................................................................................................... 35

Table 11: Interaction Results of the Regression Analysis for Parental Attachment

and Family Stress ..................................................................................... 38

Table 12: Interaction Results of the Regression Analysis for Peer Attachment

and Peer Stress ......................................................................................... 39

vi

Introduction

Youth violence has become one of the leading causes of death for teenage
adolescents (Blumstein, 1995). With the increase in youth violence occurring all over the
United States, many wonder why youth are becoming more violent. Even more
important, can previously supported theories be used to explain this phenomena? The
purpose of this study is to examine social control theory and strain theory to see if they
can explain why some youth commit violent acts. Specifically, this study will examine
parental and peer attachment and the strain within the family and peer relationships.
Although these parts of both theories have found empirical support in the past (Gardner &
Shoemaker, 1998; Paternoster & Mazerolle, 1994), neither aspect has been used to
explain specific delinquency. In addition to examining these variables separately, this
study will combine these theories to see to see if attachment to parents or peers becomes
contingent on the strain experienced in those relationships.

Hirschi (1969) first developed social control theory from the idea that there are
certain aspects of youths’ lives that make them conform to societal norms. Although
there were many factors that Hirschi proposed as increasing resiliency in youth,
attachment to parents and peers were found to be most influential (Gardner & Shoemaker,
1989; Junger-Tas, 1992; Patterson, Pryor, & Field, 1995; Paschal], Ennett, & Flewelling,
1996; Rankin & Wells, 1990; Warr, 1993). Parental attachment has been the most
supported factor affecting delinquency, with many findings that suggest that youth with
strong attachments to their parents are less likely to become involved with delinquent
peers (Gardner & Shoemaker, 1989). Peer attachment has also been supported as a

(Junger-Tas, 1992). However, the original relationship between peer attachment and

delinquency was restated when researchers found that a major determining factor was
whether or not the youth were attached to delinquent peers (Cressey, 1964; J unger-Tas,
1992). In fact, juveniles who associated with deviant or delinquent peers were found to
be more likely to imitate such behavior (Cressey, 1964).

Years later, Agnew (1989) developed a revised strain theory that examined the
relationships that youth had with parents and peers. This theory held that youth with
stress within their familial relationships often felt the need to retaliate to reduce the
feeling of injustice these relationships caused (Brezina, 1996). Youth were also found to
participate in violent activities because of peer pressure (Pearl, Bryan, & Herzog, 1990).

Yet, whether youth have strong or weak attachments or high amounts of strain
within their relationships the ability of these theories alone to adequately explain all
delinquent behaviors is questionable (Trojanowicz & Morash, 1992). In fact, some
researchers believe that one theory cannot explain all behaviors because individuals are so
different (Trojanowica & Morash, 1992). Therefore, theories should be combined to see
if together they could explain more adequately youth behavior (Trojanowicz & Morash,
1992).

This study uses secondary data collected for the National Youth Survey that
questioned parents and youth about their lives and activities. A total of seven waves were
completed, however, only the first and second waves are used for this study.

Because the purpose of this study is to see if parental and peer attachments
interact with the stress present in those relationships, multidimensional scales are created
by combining variables that represented different aspects of each theoretical component.

Once the scales are defined and statistically supported, the following methods are

conducted:
0 To examine the direct relationships, Analysis of Variance (ANOVA) and
simple linear regression are used.
0 To examine the direct relationships while controlling for other variables,
multivariate regression is used.
0 To examine the interaction between attachment and strain, multiple regression

analyses is conducted.

Youth Violence in the United States

In 1990, 2.9 million people died resulting from violence in the United States
(Roth, n.d.). Although the US. has seen higher violent rates in the past for adults, youth
violence has risen dramatically since 1967 (Blumstein, 1995; Chandler, Chapman, Rand,
& Taylor, 1998; Greenfield, 1996; Kelley, Huizinga, Thomberry, & Loeber, 1997;
Ollendick, 1996; Roth, n.d.;). In fact, there has been a 67% increase in violent activities
such as murder, manslaughter, rape, robbery, and aggravated assault committed by
juveniles (Greenfield, 1997). Comparisons of American youth and youth from other
industrialized countries has found that American youth are the most violent (Kelly et al.,
1997). In addition, youth in the US. are also more likely to be victims of violent crimes
(Greenfield, 1996; Kelly et al., 1997); with the probability of death resulting from violent
acts being the highest for teenaged adolescents (Goldstein & Conoley, I997). The
previous findings (Blumstein, 1995; Chandler et al., 1998; Greenfield, 1996; Kelly et al.,
1997) and the increase in the media’s portrayals of youth violence occurring in the streets,
schools, and homes, has researchers and practitioners wondering: Why are American
youth so violent?

Delinquency Theories

Multiple researchers have developed theories used to explain why so many youth
have been involved in deviant and delinquent activities (Agnew, 1989; Hirschi, 1969).
Theories such as social control theory and strain theory have examined family and peer
relationships and the Stress or strain that juveniles have experienced (Agnew, 1991).
These theories have been used as ways to explain the behaviors of many juveniles

(Hirschi, 1969). Although these theories have limitations (Junger-Tas, 1992; Paternoster

& Mazerolle, 1994), both have been continually supported and reexamined to determine
how effectively these theories explain general delinquency. Recently researchers have
begun to combine these theories to explore the interaction that may exist between social
control and strain theory (Agnew, 1991: Paternoster & Mazerolle, 1994).
Social Control Theory

Hirschi's social control theory was first developed in the 1960's as a way to
rationally explain why some individuals engage in delinquent behavior (Hirschi, 1969).
Developed from the idea that researchers should be looking for reasons why some people
do not engage in criminal activities, Hirschi ( 1969) believed that individuals would
conform to social norms and uphold the beliefs of society if these individuals had strong
bonds to social institutions. Hirschi proposed four elements of social bonds: attachment,
commitment, involvement, and belief. This theory proposed that strong attachments to
parents and peers, commitment to society, involvement in non-delinquent activities, and
belief in conventional norms, would prevent juveniles from committing deviant or
delinquent activities (Hirschi, 1969). Research has empirically supported this theory; the
most notable relationship being the strength of attachment youth feel towards their
parents (Gardner & Shoemaker, 1989). In fact, numerous studies have found that youth
with strong attachments are less likely to be involved in delinquent activities (Gardner &
Shoemaker, 1989; Junger—Tas, 1992; Patterson, et al., 1995; Paschall, et al., 1996; Rankin
& Wells, 1990; Warr, 1993). In addition, involvement, commitment, and belief in
conventional activities have been found to be significantly related to non-delinquency in
youth (Gardner & Shoemaker, 1989).

Although all aspects of social control theory have been empirically supported

(Gardner & Shoemaker, 1989; Patterson et al., 1995), research has found that attachment
to parents and peers are the most significant predicting factors of later delinquency in
youth (Agnew, 1991; Akers & Cochran, 1998; Brownfield & Thompson, 1991; Knight &
Tripodi, 1996; Patterson et al., 1995; Thompson, Mitchell, & Dodder, 1984). And the
ability of parental and peer attachment to predict delinquency in youth suggests the
importance of this theoretical component. Because this study seeks to examine if youth
become violent behavior because of relationships in their lives, it is essential that we
investigate youth attachments.
M Attachment

Attachment to parents has frequently been measured by examining self-report data
that ask youth multiple questions regarding their family relationships (Gardner &
Shoemaker, 1989). Typically, parental attachment has been defined as the amount of
time youth spent with their parents, the feeling of closeness to their parents, and the
quality of the communication youth had with their parents (Gardner & Shoemaker, 1989).
Findings indicate that youth with strong attachments to their parents are less likely to be
involved in deviant or delinquent behaviors than their counterparts (Gardner &
Shoemaker, 1989; Junger-Tas, 1992; Le Blanc, 1991; Rankin & Wells, 1990; Sweet,
1991). Strong attachments to parents also resulted in having fewer friends that were
involved in deviant or delinquent activities, which suggests that high rates of family
attachment inhibited the youth from attaching with delinquent peers (W arr, 1993). One
reason could be that youth with Strong parental attachments spent more time at home, and
less time with their friends (W arr, 1993). Youth with strong attachments were also less

likely to be involved with delinquent peers because they reported fearing their parents'

disapproval (Warr, 1993). These youth consistently reported higher levels of
identification and acceptance of parental belief than their counterparts (J unger-Tas,
1992). Youth with high levels of identification and acceptance were also found to have
lower incidences of deviant or delinquent behaviors (J unger-Tas, 1992).

Yet, research that has examined social control theory cannot explain why some
juveniles with strong attachments to their parents still commit deviant or delinquent acts
(Junger-Tas, 1992). In addition, these studies have also been unable to explain why youth
with weak attachments to their parents are non delinquent (J unger-Tas, 1992); thus
suggesting that mediating factors may exist that enhance or remove the effects of parental
attachment on youth behavior (Agnew, 1991).

Peer Attachment

According to social control theory, strong attachments to peers may reduce the
probability of involvement in delinquent activities (Hirschi, 1969). However, other
researchers have not found support for the suggested direction of this relationship
(Cressey, 1964). Studies have indicated that the ability of peer attachments to impede
youth from committing crimes is contingent on the activities of those peers that the youth
is attached (Agnew, 1991; Akers & Cochran, 1985; Gardner & Shoemaker, 1989). In
fact, juveniles who associated with deviant or delinquent peers were found to be more
likely to imitate such behavior (Cressey, 1964). Association with peers has also been
shown to weaken parental attachments as youth move through adolescence (J unger-Tas,
1992). One reason could be that parental attachment was weaken because peers become
more important (Junger-Tas,1992; Pearl et al., 1990). Adolescent girls reported that they

engaged in delinquent behaviors because it was acceptable behavior according to their

friends (Pearl et al., 1990). This suggests that the original relationship proposed by
Hirschi (1969) was only partially correct. Instead, the relationship between peer
attachment and delinquency became contingent on whether or not the youth had
delinquent peers (Pearl et al., 1990); thus suggesting that peer relations could be positive
or negative influence, depending on the peers in which they associate (Cressey, 1964).

Yet, whether youth have strong or weak peer attachments, the ability of social
control theory to adequately explain delinquent behavior in youth is questionable
(Trojanowicz & Morash, 1992). Researchers who have tested social control theory
cannot always explain why youth want to be involved in delinquent activities (Junger-
Tas, 1992; Trojanowicz & Morash, 1992). Some may suggest that youth that report high
levels of attachment to delinquent peers are more likely to be involved in delinquent
activities (Junger-Tas, 1992). Few have explored this relationship in depth. In addition,
social control variables cannot be used to predict specific types of delinquency,
encompass all delinquent acts, or explain why some youth do become engaged in
delinquent activities while their counterparts do not (T rojanowicz & Morash, 1992).
Therefore, it is important to examine other theories.
Strain Theog

Robert Merton, who believed that juveniles became delinquent because they could
not obtain their personal goals, first developed strain theory. These juveniles were
believed to come from disadvantaged areas, which impeded their ability to adequately
provide the necessary tools to obtain expected goals (Burton & Cullen, 1992). The norms
of society were not followed by such youth because they did not provide pathways to

success. Therefore, the norms no longer restricted the youth from deviance or

delinquency but rather forced the adolescents to become involved in criminal activities
(Burton & Cullen, 1992; Burton, Cullen, Evans, & Dunaway, 1994). For instance, one
study found that youth that perceived fair or poor chances of graduating high school were
more likely to be involved in delinquent activities. And because the youth were not able
to obtain the expected goals of completion of high school, they turned toward
delinquency (Rankin, 1980). However, there was no proof that delinquency followed the
youth’s beliefs that they might not graduate from high school (Rankin, 1980). In fact, this
is a major problem with strain theory. Strain theorists have typically relied on self-report
data that measures youth’s perceptions to obtain future goals (Agnew, 1989). These
theorists have been unable to indicate causal direction, specifically whether delinquent
behavior caused youth to believe they are less able to obtain expected future goals or vice
versa (Agnew, 1989). In addition, strain theorists have tended to focus on future goals
and ignore the present goals of many adolescents. Family stress and peer pressure may be
immediate problems in the youth’s lives. And research indicates that youth tend to be
most concerned with immediate issues than future ones (Agnew, 1985, 1989; Paternoster
& Mazerolle, 1994).

Understanding the problems of Merton’s strain theory, Agnew (1985) developed a
revised theory that examined the strain experienced as a result of painful or negative
situations. Agnew stated that youth may use delinquency as a way to escape negative
situations. In addition, Agnew proposed that youth who cannot escape such situations
would become angry and aggressive. Although previous researchers held that youth tend
to leave when faced with aversive situations, Agnew (1989) believed that some youth

may be blocked from the ability to leave such Situations. This may be particularly true for

youth that live in stressful homes or have received a lot of peer pressure from their peers.
Generally, research empirically supports this new version of strain theory (Agnew, 1993;
Gottfredson, McNeil, & Gottfredson, 1991). In fact, neighborhood problems, negative
life events, negative relationships with adults, and strain from lack of goal attainment
were all related to general delinquency in youth (Paternoster & Mazerolle, 1994).
Changes in relationships were also found to cause stress in youth, who then turned to
delinquency as a way to release tension (Agnew, 1989); thus youth with high levels of
stress are likely to retaliate and use violence to reduce the feeling of injustice in their lives
(Gottfredson et al., 1991). Such youth are also more likely to report feeling cheated in
life and were more likely to believe they deserved to retaliate (Brezina, 1996).
Family Stress

The examination of the impact of external environmental factors in which families
operate has discovered that adolescents from highly disorganized areas were found to
have poor parental attachments and higher instances of involvement in delinquent
activities (Agnew, 1989; Gottfredson et al., 1991). These disorganized areas often
created high levels of stress in family relationships (Gottfredson et al., 1991). Parental
bonding was also weaker for adolescents from urban areas when compared to rural areas;
thus suggesting that external stress factors were mediating the effects of parental
influence, especially for youth living in the poorest and highly stressful neighborhoods
(Gardner & Shoemaker, 1989).

Mediating factors can also include internal environmental factors that decreased
the attachment youth have with their parents (Agnew, 1989). Internal stress factors could

include family socioeconomic status, marital discord, number of conflicts in the home,

10

and the movement of parents or adults in and out of the home (Agnew, 1993). These
factors were found to lead to stress, anger, and frustration in many youth (Agnew, 1993;
Gottfredson et al., 1991; Junger-Tas, 1992; Paternoster & Mazerolle, 1994). For
example, parents who reacted to the adolescents' delinquent activities with arguments and
conflict had teenagers that reported weaker bonds with their parents (J unger-Tas,l992).
Parents who have also experienced high levels of stress reported poor attachment with
their children; therefore suggesting that parental stress and youth stress reduced familial
bonding (Stern & Smith, 1995). Youth exposed to many family stress factors, such as
marital conflict and other problems with their parents, are also more likely to deal with
the stress in their lives by engaging in delinquent activities (Brezina, 1996; Paternoster &
Mazerolle, 1994). In fact, studies have found that youth with high amounts of stress in
their family lives used delinquent activities as a coping mechanism (Brezina, 1996;
Bruce, 1970; Windle, 1992).
Peer Pressure

During adolescence, peer relationships can be very important (Cressey, 1964;
Hirschi, 1969). Therefore, it is easy to believe that when youth are asked to commit
deviant or delinquent acts by their friends, it is hard for them to say no. Yet, some youth
seem not to be susceptible to such peer pressure (Pearl etal., 1990). Adolescents are
likely to engage in delinquent activities at the request of their peers if they perceived the
outcome to be a personal "payoff" or acceptable by their peers (Pearl et al., 1990).
Females are less likely to decline the invitation to be involved in delinquency because
they fear losing their friends. Females are also more likely to believe that they need to

satisfy their friends by their actions, whereas males engage in activities for personal gain

11

(Pearl et al., 1990). Other researchers have indicated that youth defined as "loners" were
less likely to be influenced by peer pressure. In fact, these youth were more likely to be
involved in conventional activities than youth who were highly involved with their
friends (Tolone & Tieman, 1990). Agnew ( 1991) found that the removal of juveniles
from their delinquent peers did not reduce the juveniles' delinquent behavior. Instead,
Agnew found that a reduction in the stress related to the peer relationship, such as the
perceived need to please or obtain acceptance from peers, was the main factor impacting
youth’s decisions. Thus suggesting that the stress in peer relationships was determining
the actions of youth (Agnew, 1991; Patternoster & Mazerolle, 1994).
Combination Approach

Research has suggested that criminal theories are not able to adequately explain
all delinquency in youth (Trojanowicz & Morash, 1992). Therefore, researchers have
begun to combine aspects of different theories to see if a combination approach will reach
better results (Paternoster & Mazerolle, 1994; Rankin & Wells, 1990). By combining
social control theory and strain theory, we can test whether or not the attachments to
family and peers have a direct effect on delinquency, or if they are contingent on the
stress that is prevalent in those relationships (Agnew, 1991). As previous research
suggested, the amount of stress in a relationship may mediate the relationship youth have
with their parents and peers (Agnew, 1991). In fact, some researchers found that general
strain theory and social control theory were complementary (Paternoster & Mazerolle,
1994).

Support for such an interaction has not been found (Agnew, 1991; Paternoster &

Mazerolle, 1994). Those studies that have concluded that there was no interaction used

12

unreliable scales whose alpha score was well below .6 (Paternoster & Mazerolle, 1994).
These studies also used questionable definitions of attachment and stress, and examined
the effects only for general delinquency (Paternoster & Mazerolle, 1994). Both unreliable
scales and the way in which stress was defined could have affected the final conclusions
that attachment and stress did not interact. In addition, support for the interaction may
not have occurred because of the dependent variable encompasses a whole spectrum of
delinquent activities. It may be possible that an interaction does occur for only some
delinquent activities. And based on other studies (Agnew, 1991; Paternoster &
Mazerolle, 1994), it is undetermined how well this mediating relationship will hold for

specific delinquent acts.

l3

Goals and Objectives

Similar to Agnew's (1991) study of peer attachment and stress and Paternoster and
Mazerolle’s ( 1994) study of the interaction between social control and general strain
theory, the purpose of this Study is to see if there is an interaction between social control
and strain theory that may explain why some juveniles engage in delinquent activities.
Specifically, this research will explore the relationship between family attachments,
family stress, peer attachments, and peer pressure and violent activity in youth. This
study hopes to find evidence that supports the belief that parental attachment and family
stress are negatively related and that peer attachment and peer pressure are positively
related to violent behaviors in youth. Based on previous research, the following general
hypothesized relationships will be examined:
Direct Effects

0 Youth with strong attachments to parents are less likely to be involved in
violent acts. Based on previous research, an direct causal relationship is
hypothesized.

0 Youth with weak attachments to peers are less likely to be involved in violent
acts. Because previous research suggests that youth with weak peer
attachment are less likely to be involved in violence, this study will
hypothesize a direct relationship.

0 Youth with high levels of family stress are more likely to be involved in
violent acts. This relationship is hypothesized as direct causation.

0 Youth with high levels of perceived peer pressure are more likely to be

involved in violent acts. This relationship is also hypothesized as direct

l4

causation.

Inttfltive/ModeLated Effects

Youth with weak attachments to parents and high levels of family stress are
more likely to be involved in violent crimes than youth with weak attachments
and low levels of family stress and youth who are strongly attached to their
parents.

Youth with strong attachments to peers and high levels of perceived peer
pressure are more likely to be involved in violent acts than youth with strong
attachments to peers and low levels of perceived peer pressure and youth that

reported weak attachments to peers.

15

Data

The data used in this research were collected for the National Youth Survey
(NYS) to learn about aspects of parents and youth's lives. The information was originally
collected to examine the lives of youth across the United States over a period of 11 years,
however only the first and second waves will be used. This survey collected information
on the demographics of respondents, family life, neighborhood descriptors, peer
interactions, school interactions, community involvement, and illegal activities of youth.
Some information was collected from their parents in Wave I. The subjects were selected
through area probability sampling. Once selected, the subjects were interviewed by
trained personnel (Elliot, 1977).
Youth Demographics

Youth demographics are an important component used for understanding the
population selected. Knowledge of the individuals selected for this survey defines the
ability of researchers to later make generalized conclusions (Maxfield & Babbie, 1995). A
total of 1726 youth and their parents were surveyed. The demographic characteristics for

Wave I and the control variables are listed in Table 1.

l6

Table 1
Demographic Characteristics for Youth Surveyed in the

National Youth Survey. Wave 1. 1978.

 

 

 

Characteristics Percentage
Sex
Male 53.2
Female 46.8
Ethnicity
White 79.0
Black 15.1
Other 5.9
Age" (years)
11 to 13 45.1
14 to 15 29.7
16 to 17 25.3
Grade
4th to 5th 5.6
61‘ to 3th 45.1
9th to 12‘h 47.2
Other 2.2
GPA
Mostly A's 17.0
Mostly B's 44.9
Mostly C's 32.8
Mostly D's and F's 5.2
Family Incomeb ($)
10,000 or Less 26.5
10,001 - 18,000 33.9
18,001 - 26,000 24.6
26,001 - More 15.0
Note. N = 1725.

“Average age was 13.87 years.
bAverage family income was $14,001 — 18,000.

17

Methods

Independent Vafiriables

LmentaIAHachment

Parental attachment has been operationalized in many different ways. For
instance, some researchers have defined attachment as the ability to control activities in
their children, the amount of closeness children feel towards their parents, amount of time
Spent with their parents, and the amount of communication parents have with their
children (Rankin & Wells, 1990). Use of secondary data has restricted the ability of this
study to operationalized parental attachment in a complete manner. However, the
National Youth Survey asked youth their perceptions concerning the feeling of closeness
to their parents and how much time they spent with their parents. Therefore, the parental
attachment variable scale will be based on these two theoretical components of
attachment (Gottfredson et al., 1991; Paschall, Ennett, & Flewelling, 1996). For
closeness, items will be combined that will consist of questions such as “Sometimes I feel
lonely when I’m with my family,” and “I feel close to my family (Elliot, 1977).” Amount
of time will be measured by including questions such as “On the weekends, how much
time have you generally spent talking, working, or playing with your family,” and “On the
average, how many afternoons during the school week . . . have you spent talking,
working, or playing with your family (Elliot, 1977)?” Questions regarding closeness are
asked in a five-point Likert-scale and youth are given choices for amount of time spent
with parents. The choices for amount of time range from zero days a week to five days a
week.

A correlation matrix was conducted to determine the reliability of combining

18

elements regarding closeness and amount of time. The results from the reliability
analysis indicated that the scale was highly reliable for the questions included (See the
Appendix). This was determined by examination of Cronbach’s alpha (.664). A factor
analysis was also conducted to determine if these variables are interrelated. The factor
score coefficients are listed in Table 2. Since all of the coefficient scores are above .5, all
of the variables selected can be used to create a scale, which will represent multiple

variables that are meaningful when combined.

 

Table 2

Factor Analysis Coefficient Scores for Parental Attachment

 

 

 

Variables §£o_re_
Aftemoons/ school week spent with family. .540
Evenings/School week spent with family. .578
Weekends spent with family. .632
Outsider with family. -.609
Family listens to problems. .554
Feel lonely with family. -.523
Feel close to family. .690
M2 or = .664.

 

l9

Peer Attachment

To be consistent with the operationalization of attachment used previously in this
study, and for the purposes of direct comparison between parental attachment and peer
attachment, peer attachment will also be operationalized as the reported feeling of
closeness to friends and amount of time the youth spent with these friends (Agnew,
1991). A scale will be constructed that will be inclusive of closeness to peers and time
spent with peers. Closeness to peers will be determined by combining items such as “I
don’t feel that I fit in very well with my friends,” and “I feel close to my friends (Elliot,
1977).” Time spent with peers will consist of such questions as “On the average, how
many evenings during the school week . . . have you spent with your friends,” and “On
the average, how many afternoons during the school week . . . have you spent with your
friends (Elliot, 1977)?” All questions are asked by using a five-point Likert-scale. As
with parental attachment, the higher youth score on this scale, the higher the level of
attachment was reported.

Based on the reliability analysis, the questions that represented the amount of time
the youth spend with their peers had to be eliminated from the scale. In order to make the
scale reliable above the .6 level, only questions that pertained to the closeness youth felt
toward their peers could be included. These questions combined resulted in a Cronbach’s
alpha of .638 (see Appendix). However, direct comparison of the scale to family
attachment will be difficult because the peer attachment scale could not include the
amount of time youth spent with their peers. A factor analysis was also conducted in
order to see if these variables can be combined into one factor. The results are listed in

Table 3. Since all of the variables are over .5, it was determined that the variables could

20

be combined into a scale.

 

Table 3

Factor Analysis Coefficient Scores for Peer Attachment

 

 

 

 

Variables Score
Don't fit well with friends. -.591
Feel close to friends. .694
Friends listen to problems. .699
Fee] lonely with friends. -.500
Friends don't take interest. -.730
Note. a: .638.

Family Stress

Family stress will be measured by combining elements relating to the marital
status of the parents, serious illnesses or accidents occurring during the year, other
children in the home with legal problems, employment status of parents, and the
movement of the family and its members. Previous research that has used the National
Youth Survey has indicated that the previous items may not be reliable when combined.
These studies continued to use the scale, regardless of the fact that the alpha was .5 or
lower.

Analysis of the reliability scale indicated that the reliability of the scale suffered

when the scale was inclusive of all items; therefore, multiple items had to be deleted to

21

get Cronbach’s alpha above the .6 level. Only questions that pertain to the martial status
and mobility of the family members were included in the final reliability analysis, which
resulted in an alpha of .641 (see Appendix). Items were excluded to see if the increase in
the reliability of the scale would change the interaction this scales has with the parental
attachment scale. This is particularly important since previous studies that used less
reliable scales found no interaction. A factor analysis was also conducted for these
variables and the results are listed in Table 4. All of the variables listed have a factor
coefficient score above .5, which suggests that these variables are interrelated and

therefore a scale could be created.

 

Table 4

Factor Analysis Coefficient Scores for Family Stress

 

 

 

Variables Scar;
Parents are divorced .620
Parents are separated. .802
Mother moves in and out of home. .602
Father moves in and out of home. .734
M; a = .641.

 

Peer Pressure
Peer pressure will be measured following Agnew’s (1991) concept of peer

pressure. Questions such as “if you found that your friends were leading you into trouble,

22

would you still run around with them,” and “It's okay to lie to keep your friends out of
trouble (Elliot, 1977),” will be used for the construction of the peer pressure scale (see
Appendix). These elements were found to be highly reliable in Agnew’s study. The
reliability analysis supported his findings, with a Cronbach’s alpha of .665. As stated
earlier, an alpha more than .6 is regarded as highly reliable; therefore, this scale not only
has theoretical but statistical support. In addition, the factor analysis revealed that these
variables had high factor loadings, which suggest that these variables are interrelated and

can be combined to make a multidimensional scale (See Table 5).

 

Table 5

F_actor Ayala/sis Coefficient Scores for Peer Pressure

 

 

 

Variables Sara
With friends, good impression is foremost. .563
Lying okay if keeps friends out of trouble. .636
Beat up kids to gain respect of friends. .652
Break rules to be popular. .732
Break parents' rules to keep friends. .696
Nate, a: .665

 

Demndent Variable

According to the National Victimization Survey, violent activities can be defined

as murder, manslaughter, forced rape, robbery, and aggravated assault (Greenfield, 1996;

23

Kelly et al., 1997). However, because this survey was originally developed to examine
the general delinquency of youth, specifically drug use and deviance, few indicators of
violent activity could be found. In addition, these respondents did not report violent
activity in large quantities. For most questions pertaining to violence, fewer than 10%
reported any violent activities. Therefore, to adequately represent violence in youth in
this sample, this study will construct scales for violence consists of variables that measure
the frequency of property damage, use of force, and assault. In order to show directional
cause, variables used in this scale were from Wave H. All other scales were conducted
from Wave I variables.

Analysis of the reliability scale for reported frequency of violence revealed that
the scale was slightly reliable. This was based on the fact that Cronbach’s alpha was only
.552, which is slightly below the .6 reliability level. However, the factor analysis showed
that the variables were interrelated and could be combined in a meaningful way (See

Table 6). Therefore, based on these analysis, the delinquency scale will be used.

24

 

Table 6

Factor Analysis Coefficient Scores for Delinquency

 

 

Variables §c_or§
Damaged family property. .523
Damaged other property. .617
Hit teacher. .743
Used force on students. .807
Used force on others. .793

 

Note. on = .552. All questions referred to the frequency of such acts.

 

Control Vag'ables

For this research five control variables will be used. Sex, race, age, family
income, and delinquency of peers were selected because prior research has indicated a
relationship between those variables and delinquency in youth (Brownfield & Thompson,
1991; Clark & Shields, 1996; Paschalleta1., 1996; Rankin & Wells, 1990; Sam, 1983;
Seydlitz, 1993; Stem & Smith, 1995). Percentages for each sex, race, age, and family
income are listed in Table 1. Reliability and factor analyses were conducted for the peer
delinquency variables. The results of these analyses are listed in Table 7 and the original
questions are listed in Appendix. Based on the reliability alpha, this scale is highly
reliable. In addition, the factor scores indicated that the variables are interrelated and can

be combined.

25

 

Table 7

factor Analysis Coefficient Scores for Peer Delinquency

 

 

 

 

Variables Saar;
Peer ever cheated on school tests. .566
Destroyed property. .647
Used marijuana. .789
Stolen something worth less than $5. .703
Used alcohol. .772
Broken into vehicle. .612
Stolen something worth more than $50. .608
Suggested that you break the law. .716
Gotten drunk. .754
Used prescription drugs. .624
Sold or given alcohol. .703
119$ or = .875.

Analysis

Two types of bivariate analysis will be conducted. In addition, two main types of
mulivariate analysis will be conducted: the direct effects of each independent variable,
controlling for sex, race, age, family income, and delinquency of peers, on violent activity

and the interactions between attachment and stress and their effects on violent activity.

26

Direct Effects

One-way Analysis of Variances (ANOVAs) and simple linear regressions will be
conducted to examine the bivariate relationship between the independent variables and
the dependent variable. These analyses will be used for a comparison of how the
relationships changed when other variables were controlled.
Direct Effects While Controlling for Other Variables

Multivariate regression will be used to simultaneously see how multiple
independent variables effect violent activity in youth. Regression will be used because
previous research suggests that the relationships between each of the independent
variables and delinquency are linear. Scatterplots of the data used in this study have
supported this finding. Other researchers examining these variables have used correlation
tables, bivariate analyses, or multiple analysis of covariance (MANCOVA). However,
many of these analyses use lower levels of measurement, which may affect the results of
the data. Since multivariate regression uses the highest levels of measurement (interval
and ratio) and can control for multiple factors, this analysis will be used to look at the
direct effect of the independent variables while controlling for other variables.
Interactions

Interactions will also be examined by conducting multiple regression. For this
analysis, the stress variables will be the moderators. Based on this assumption, we can
say that we are examining if the relationship between attachment to parents and peers and
violent youth behavior depends on whether or not the youth have stress within the

relationship.

27

Results
Bivariate Results

Independent Valiatifl

Table 8 contains data for the regression analyses of parental attachment, peer
attachment, family stress, peer pressure, age, and peer delinquency and their affects on
self-reported violent behavior. This table includes the constant, beta coefficient, t-score
and the Proportionate Reduction in Error statistic (r2) for each relationship.

The regression analysis (See Table 8) reveals that the relationship between
parental attachment and violent activity is significant (t = -2.542; p = .011). The analysis
also indicates that parental attachment only accounted for .7% of the explained variation
of the dependent variable, self-reported violent behavior. Based on these results, it can be
stated that the relationship between parental attachment and self-reported violence is

negative.

28

 

Table 8

Bivariate Regression Analyses for Continuous Variables on Self-Reported Violent

 

 

 

Behavior
Variable Constant B t 1’:

Parental Attachment .002 -.083 -2.542* .007
Peer Attachment .002 -.059 -l.818 .004
Family Stress .001 .012 .356 .000
Peer Pressure .006 .186 5.838** .035
Age .1 18 -.009 -.545 .000
Peer Delinquency .015 .272 7.362** .062
M.

*p < .05. **p < .01.

 

The analysis of the relationship between peer pressure and reported violent
behavior indicates that the relationship is significant (t = 5.838; p< .001). However, peer
pressure accounted for only 3.5% of the explained variation of the reported violent
behavior. These findings indicate that as the level of peer pressure increases, violent
activity reported by youth surveyed also increases (See Table 8).

Although other studies have found a relationship between peer attachment and
delinquency, results from the regression analysis for this study did not reveal a significant

relationship at the .05 level. The results for family stress indicates that there is not a

29

significant relationship between family stress and violent behavior reported by youth
surveyed.
Control Variables

The analysis for delinquency of peers and self-reported violent activity reveals
that there is a statistically significant relationship (I = 7.362; p < .001). These results
indicate that as the number of delinquent peers increases the number violent acts reported
also increases. According to these results, delinquency of peers only explained 6.2 % of
the explained variation in reported violent acts. Although the percentage of explained
variation is relatively low, the delinquent peer variable was the strongest predictor of self-
reported violent activity for this sample.

The regression analysis for age of respondent and reported violent acts indicates
that there is not a significant relationship between age and violent acts reported (t = -.545;
p = .586). Although this relationship is not significant statistically it is interesting that the
analysis indicates that as the age of the respondent increases the likelihood of violent
activity reported decreases. This is similar to what other studies have found (Menard &
Elliot, 1990).

Table 9 contains information for the AN OVA analyses conducted for all the
categorical level variables. Information in this table includes the mean and the F statistic.

For this study the dependent variable was recoded into z-scores. Those means that are
negative indicate that the group means were below the mean number of reported violent
acts for the whole sample. Positive means represent group means that were above the

mean number of reported violent acts for the whole sample.

30

Table 9

Analysis of Variance Results if Categorical Variablesa and Self-Reported Violent

 

 

Behind—or
Variable Mean E
Sex 33.987*
Male .120
Female -. 130
Grade 2.012
4th to 5‘h -. 127
6‘“ to 8th .084
9‘h to 12‘h -.057
Other -.162
Race .112
White -.008
Black .031
Other .030
Grade Point Average 3.827**
Mostly A's ‘ -.114
Mostly B's -.059
Mostly C's .087
Mostly D's .530
Mostly F's .329
Income 4.016**
10,000 or less .194
10,001 - 18,000 -.077
18,001 - 26,000 -.O21
26,001 - more -.085

 

Nate. a. The delinquency scale for this study was converted into z-scores. Those who
have a negative mean scored below the mean number of violent acts on the delinquency
scale and those with a positive mean scored above the mean number of violent acts on the
delinquency scale.

*p < .05. ** p < .01.

31

The distinction between the mean scores for males (. 120 violent acts) and females
(-. 130 violent acts) suggests that a difference exists between males and females (See
Table 9). This suspicion is supported by analysis of the F score (F = 33.987; p < .01).
This indicates that being male increases the likelihood of being above the mean number
of self-reported violent behaviors for this sample population.

For youth GPA, the AN OVA analysis reveals that the mean number of violent
acts reported by youth may differ between GPA groups (See Table 9); which was
supported by the F-score (F = 3.827; p = .004). The Bonferroni Post Hoc test reveals that
there is a statistically significant difference between those youth who reported receiving
Mostly D's (mean = .530 violent acts reported) and Mostly B's (mean = -.059 violent acts
reported; p = .015). The test also reveals that there is a significant difference between
those youth reporting Mostly D's (mean = .530 violent acts) and Mostly A's (mean = -.114
violent acts; p = .010). This suggests that youth who indicated that they were receiving
mostly D’s reported more violent acts than their counterparts who received better grades.

These are the only groups found to be statistically different.

For family income, the means for each group of income suggest that there might
be a difference between each group (See Table 9). The one-way AN OVA did indicate
that there is a difference between groups (F = 4.016; p = .007). A Bonferroni Post Hoc
test was conducted and revealed the following results: youth whose families reported
their family income at $10,000 or less were significantly different from youth whose
families reported income between $10,001 and $18,000 (p = .012); youth whose families
reported their income at $10,000 or less were significantly different from youth whose

families reported their income at $26,000 or more (p = .023). These results indicate that

32

youth from families with lower incomes, specifically income $10,000 or less, reported
more violent acts than youth from families with higher incomes.

Based on the analysis of the means for grade levels, the group means seem to
differ (See Table 9). However, the one-way ANOVA reveals that the mean number of
violent acts reported by youth did not differ between grade levels. This is based on the
significance of the F—score which is above .05 (F = 2.012; p = .11 1).

Based on the group means for each race, there seems to be only a difference
between white respondents and the other races ("White" mean: -.008 violent acts
reported; "Black" mean = .031 violent acts reported; "Other" mean = .030). However, the
results from the one-way AN OVA reveal that the mean number of violent acts reported
by youth is not different for the different groups of race.

Multivariate Results

Table 10 includes the summary of the regression results of four models that were
conducted. Model I includes data for parental attachment and seven control variables
(sex, race, grade, income, peer delinquency, GPA, and age) and their relationship with
self-reported violent behaviors. Model H includes peer attachment and the control
variables listed above, Model III includes family stress and the control variables, Model
IV includes peer pressure and the control variables and Model V includes all variables.
The table also includes the constant and r‘2 for each model.

The regression analysis of the multivariate results indicates that only two
variables, income and peer delinquency, consistently remained significant when
controlling for other variables (See Table 10). For income, the direction of the

relationship with self-reported violent behavior is negative for each model and the

33

constant remained Similar. Peer delinquency also continues to have a similar constant for
each model. The direction of the relationship between peer delinquency and violent
behavior reported continues to be positive. Two conclusions can be made from these
findings. First, income is an important predictor of violent behavior in youth. The results
indicate that youth from low-income families are more likely report using violence.
Second, the number of delinquent youth is the most important predicting factor of violent

activity.

34

 

Table 10

Summary of the Regression Analysis for Vambles Predicting Violent Behavior

 

 

Variable Mo_d§_l_1 54—04211]. Model HI Model IV Model V
Constant 1.315 1.297 1.319 1.190 1.178
Sex -l.689 -l.631 -l.673 -1.351 -1.308
Race .800 .852 .768 1.068 1.020
Grade -1.717 -1.642 -1.725 -1.674 -1.681
Income -2.263* -2.243* -2.356* -2.223* -2.277*
Peer Delinquency 7.578** 7.730** 7.717** 6.758** 6.576**
GPA -1.33O -l.356 -1.285 -1.112 -1.065
Age -.966 -.956 -.942 -.821 -.744
Parental Attachment -.234 -- -- -- .473
Peer Attachment -- -.477 -- -- -.038
Family Stress -- -- .430 -- .322
Peer Pressure -- -- -- -2.942** -2.856**
r2 .100 .100 .098 .109 .107

 

Note. Dashes indicate that those variables were not included in the model.

*p < .05. **p < .01.

 

Peer pressure was also found to be significant when controlling for other
variables. In fact, this variable was the second highest predictor of violence in youth.

These results are Similar to those found at the bivariate level, meaning that the direction

35

of the relationship continued to be negative. This indicates that as peer pressure increases
the number of violent offenses reported increases even while controlling for other
variables. Model IV, which included peer pressure and other control variables, accounted
for the most explained variation in the dependent variable (10.9%). This indicates that
the variables included in this model were the most significant predictors of violent
behavior reported by youth.

The relationship between parental attachment violent behavior did not remain
significant controlling for other independent variables. There are two possible
explanations for this finding. First, there is a possibility that a controlling variable is an
intervening variable. In this instance only income and peer delinquency could have been
the intervening variable because they were the only variables that remained significant
when controlling for other variables. Based on previous studies, there is a greater
possibility that peer delinquency was the intervening variable (J unger-Tas, 1992; Warr,
1993). However, this analysis does not indicate if this suspicion is correct. The second
possibility is that the bivariate relationship was spurious, meaning that income or peer
delinquency caused both parental attachment and violent behavior. Most studies would
suggest that income would be the most likely culprit of this type of conclusion because
most theorists would agree that parental attachment (or lack of) occurs before interactions
with peers (Gardner & Shoemaker, 1989; Junger-Tas, 1992; LeBlanc, 1991).

Sex of the respondent and GPA also did not remain significant when other
variables were included in the regression equation. For sex, this finding can be
contributed to an intervening variable because the sex of the respondent cannot be caused

by another variable. For example, sex can affect the number of delinquent peers one has,

36

which could then affect violent activity in youth. For GPA, this finding could be
contributed to either an intervening variable or a spurious relationship between GPA and
violent offenses reported. The remaining variables, peer attachment, family stress, age,
grade and race, continue not to be significant even when controlling for other variables.
Interaction

Table 11 and 12 contain the results for the interaction analyses. The constant, r2,
t-scores are all represented in these tables. Table 11 includes information from the
interaction between parental attachment and family stress and Table 12 includes the
results from the interaction between peer attachment and peer pressure.

Analysis of the regression analysis reveals that an interaction did not occur
between farme stress and parental attachment (See Table 11). This indicates that the
combination of parental attachment and family stress did not effect the probability of
reporting violent activity. Youth with strong attachments to parents and high levels of
family stress are not statistically different from youth with weak parental attachments and
high levels of family stress and youth with strong parental attachments. Therefore, the

proposed hypothesis is rejected.

37

 

Table 1 1

Interaction Results of the Regression Analysis for Parental Attachment and Family Stress

 

 

Variable B I [2.
Constant 1.297 .098
Parental Attachment -.054 -.139

Family Stress -.016 -.442

Parental Attachment . Family Stress .001 .035

Sex —. 125 - l .672

Race .008 .045

Grade -. 184 -1.725

Income -.085 -2.346*

Peer Delinquency .327 7.441**

GPA -.060 -1 .248

Age -.031 -.922

 

Note. The interaction is indicated by the (.) symbol.

*p< .05; **p < .01.

 

The results of the interaction between peer attachment and peer pressure also
reveal that an interaction did not occur (See Table 12). The combination of peer
attachment and peer pressure did not affect the relationship between the two independent

variables and the dependent variables. Based on this finding, the proposed hypothesis

38

was not supported. Therefore, there is no difference between youth with strong peer

attachments and high levels of peer pressure were more likely to be violent than youth

with strong peer attachments and low levels of peer pressure and those with weak peer

attachments.

 

Table 12

Integrction Results of the Regression Analysis for Peer Aflchment and Peer pressure

 

 

Variable B I i
Constant 1.099 .1 l 1
Peer Attachment -.001 -O.34

Peer Pressure .1 12 2.919**

Peer Attachment . Peer Pressure .037 1.169

Sex -.096 -1.274

Race .162 .631

Grade -. 184 -1 .695

Income -.078 -2.22 l *

Peer Delinquency .302 6.829**

GPA -.055 -1.142

Age -.028 -.837

 

Note. The interaction is indicated by the (r) symbol.

*p < .05; **p < .01.

 

39

Discussion

The purpose of this study was to explore the relationship between attachment and
stress and how it relates to violence in youth. Specifically, this study sought to examine
the possible interaction between attachment and stress. Past researchers believed that the
influence of attachment may be mediated by the stress within those relationships (Agnew,
1991; Paternoster & Mazerolle, 1994). However, researchers have not been able to
statistically support this assumption (Paternoster & Mazerolle, 1994). This study sought
to further the exploration of the belief that an interaction between attachment and stress
did exist and that previous studies failed to find such results because of poor scale
construction and the use nominal and ordinal level statistical methods. However, this
study found similar results, meaning that no interaction was indicated. Yet this study did
find some support for other theories that have been used to explain delinquent behavior.

One important finding was that peer pressure continued to be significant even
while controlling for other variables. This study found similar results to that proposed by
Agnew (1991) who believed that peer pressure would be a significant factor in youth's
lives. Although this theory has been typically used to examine general delinquency
(Agnew, 1991; Paternoster & Mazerolle, 1994), this study found that this relationship
was upheld when examining violence in youth. In addition, peer pressure continued to
remain significant even when controlling for the presence of delinquent peers, and
important finding that supports previous discoveries (Agnew, 1991). The results of this
study indicated peer pressure seemed to increase the likelihood of reporting violence,
which suggests that peer pressure may create anger and frustration in youth that propels

them toward using violence. This also suggests that peers are a major part of youth's lives

40

and that youth may feel the need to please their peers.

Parental attachment has been one of the most supported portions of the social
control theory (Gardner & Shoemaker, 1989; Junter-Tas, 1992; Paschall et al., 1996;
Patterson et al., 1995; Rankin & Wells, 1990; Warr, 1993). This study also found a
relationship between parental attachment and violence reported by youth. However, this
study did not totally support the theoretical assumption originally offered by Hirschi
(1996). Instead, this study found that the effects of parental attachment disappeared when
other variables were included in the model. Other studies have found similar results, and
have indicated that the number of delinquent peers played a role in the disappearance of
this relationship (J unger-Tas, 1992). Although statistically I can not make a sound
conclusion concerning this proposed relationship, previous research supports the
suspicion that this may have occurred (Junger-Tas, 1992). If this is true, it can be said
that parental attachment affected the number of delinquent peers a youth associated with
which in turn affected the amount of violent acts the youth reported.

Peer attachment, a component of social control theory, has also found statistical
support in the past (Akers & Cochran, 1985; Gardner & Shoemaker, 1989). Researchers
have suggested that youth with strong attachments to their peers are less likely to be
involved in delinquent activities (Gardner & Shoemaker, 1989). Theorists contend that
involvement with ones peers increases youths' identification and feeling of acceptance
(Gardner & Shoemaker,1989; Junger-Tas, 1992). These feelings in turn impede youth
from involvement in socially unacceptable behavior for fear of no belonging. However,
this study did not find such results. The results did indicate that those with stronger

attachments reported less violent activities, yet this relationship was not found to be

41

statistically significant. Instead, this study found that it was not how attached youth were
to their peers but to whom youth were attached. And this discovery is similar to what
others have concluded (Pearl et al., 1990), all of which support the differential association
theory.

Differential association theorists have often found support for their belief that the
number of delinquent peers a youth has affects the likelihood of being delinquent
(Cressey, 1964; Pearl et al., 1990). Even supporters of social control theory have found
that the type of friends with whom the youth associated mediated peer attachment (Pearl
et al., 1990). Although delinquent peers was not an independent variable in this study, it
was a control variable because of such conclusions. This theory has found support from
this study; a direct relationship between violence reported and the number of delinquent
peers did exist.

Family stress has been defined as both internal and external factors that create
stress within the family. Internal environmental factors have often been defined as
socioeconomic status, marital discord, number of conflicts in the home, and the
movement of parents or adults in and out of the home (Gottfredson et al., 1991). These
factors have been discovered to create high levels of stress, anger and frustration in youth
(Agnew, 1993; Gottfredson et al., 1991; Junger-Tas, 1992; Pattemoster & Mazerolle,
1994). This frustration is then directed to involvement in delinquent activities as a way to
cope with this stress (Agnew, 1993), which has been connected to violent behavior
exhibited by youth (Brezina, 1996). However, this study did not support such a
conclusion. Family stress was not found to increase the likelihood of being involved in

delinquent activities. Yet this study did find that family income was statistically related

42

to violent activity. Because of reliability issues stated earlier in Chapter 4, family income
could not be included in the family stress scale. However because there is ample
evidence that income did affect violent activity reported, there is a possibility that a
component of the definition of family stress did have some impact on violent activity. If
this is true, it could be concluded that family income created high levels of frustration and
stress in the youths' lives. This stress may have caused them to be involved in delinquent
activities.
Limitations and Recommendations

There were many limitations to this study that may have affected the relationships
found. Although every study has some statistical and methodological flaw, this study had
multiple issues that create concerns when making generalizations. First, because the
sample was surveyed in the late 1970's, the external validity of this study is questionable.
Youth and families in the past decade have changed dramatically. There are more single-
headed and double income families now than ever before. More children come from
marriages that have ended from divorce. Therefore, it is virtually impossible to compare
the youth of the 70's with those of today. These issues of today’s youth may influence
youths’ decisions to be involved in delinquent activities, especially those involving
violence. For instance, family stress may not be related to violent activity in youth during
the 1970's. But due to the weak external validity of this study it is difficult to conclude
that this is true for youth in the 90's because of the great differences between these youth.
Therefore, it is suggested a similar study be conducted with today's youth to see if the
increase in family and peer pressure felt by today's youth does violent behavior.

The second limitation was the ability to define family stress. Using secondary

43

data severely limited the ability to holistically encompass the family stress variable. The
questions we included in our scale only measured the status of the family (i.e., Are you
divorced?) Ideally, the scale would have included the feelings that accompanied this
change in status. For future studies, it is recommend that this variable encompasses more
in-depth definitions and questions that use higher levels of measurement.

Policy Implications

The main purpose of this study was to examine why youth commit violent acts.
Ideally, learning more about why individuals become involved in violence may help
create policies that may prevent or intervene with youth at risk for such behavior. Based
on the findings for this study, youth need help with dealing with stress caused by the
feeling to please their peers. Programs such as Drug Abuse Resistance Education
(D.A.R.E.) and Conflict Resolution/Mediation have been created to help youth say no to
negative behaviors and deal with stress. Yet many of these programs focus their energy
on younger children. More resources should delegated to programs that help adolescents
deal with peer pressure. And it is important that these programs are offered to
adolescents since adolescence is often the time when peers become most important and
influential.

Another important implication is the need for parents, teachers, and other
individuals working with youth to identify youths' friends. Programs may need to be
created that open communication lines between school officials, parents, and juvenile
justice programs. Programs may need to be developed to get troubled youth off the
streets, out of gangs or away from destructive friends. Parents may need to learn how to

increase the strength of their relationship with their child, and help them deal with the

stress of growing up. Whatever the method chosen, it is important for adults and positive
peer relationships are used to remove adolescents from unhealthy relationships with

delinquent youths.

45

APPENDIX

Questions Included in the Variable Scales.

Parental Attachment Scale

1.

I feel like an outsider in my family.
S—Strongly Agree, 4—Agree, 3—Agree Nor Disagree, 2—Disagree, l—Strongly
Disagree

My family is willing to listen ifI have a problem.
S—Strongly Agree, 4—Agree, 3—Agree Nor Disagree, 2—Disagree, l—Strongly
Disagree

Sometimes I feel lonely when I’m with my family.
5—Strongly Agree, 4—Agree, 3—Agree Nor Disagree, 2—Disagree, l—Strongly
Disagree

I feel close to my friends.
5—Strongly Agree, 4-—Agree, 3—Agree Nor Disagree, 2—Disagree, l—Strongly
Disagree

On the average, how many afternoons during the school week, from the end of school
or work to dinner, have you spent talking, working, or playing with your family?

On the average,-how many evenings during the school week, from dinnertime to
bedtime, have you spent talking, working, or playing with your family?

On the weekends, how much time have you generally spent talking, working, or
playing with your family?
5—A great deal, 4—Quite a bit, 3—Some, 2—Not too much

Peer Attachment Scfi

1.

Don’t feel that I fit in very well with my friends.
5—Strongly Agree, 4—Agree, 3—Agree Nor Disagree, 2—Disagree, l—Strongly
Disagree

My friends don’t take much interest in my problems.
5—Strongly Agree, 4—Agree, 3—Agree Nor Disagree, 2—Disagree, l—Strongly

Disagree

I feel close to my friends.

46

5—Strongly Agree, 4—Agree, 3—Agree Nor Disagree, 2—Disagree, l—Strongly
Disagree

4. My friends are willing to listen if I have a problem.
5—Strongly Agree, 4—Agree, 3—Agree Nor Disagree, 2—Disagree, l—Strongly
Disagree

5. Sometimes I feel lonely when I’m with my friends.
5—Strongly Agree, 4—Agree, 3—Agree Nor Disagree, 2—Disagree, l—Strongly
Disagree

Family Stress Scale

Which of the following events has occurred in your family in the past year?
Divorce

Separation

Mother moved in or out of home.
Father moved in or out of home.

9.93191"

Peer Pressure Scafi

1. With friends, making a good impression is foremost.
5—Strongly Agree, 4—Agree, 3—Agree Nor Disagree, 2—Disagree, l—Strongly
Disagree

2. Lying is okay if it keeps friends out of trouble.
5—Strongly Agree, 4—Agree, 3—Agree Nor Disagree, 2—Disagree, l—Strongly
Disagree

3. It is okay to beat up kids to gain respect of friends.
5—Strongly Agree, 4—Agree, 3—Agree Nor Disagree, 2—Disagree, l—Strongly
Disagree

4. You have to be willing to break some rules if you want to be popular with your
friends.
5—Strongly Agree, 4—Agree, 3—Agree Nor Disagree, 2—Disagree, l—Strongly
Disagree

5. It may be necessary to break some of your parent’s rules in order to keep some of your
friends.
S—Strongly Agree, 4—Agree. 3—Agree Nor Disagree, 2—Disagree, l—Strongly
Disagree

47

Delinquency Scale
How many times in the last year have you:

Damaged school property?
Damaged other property?
Used force on students?
Used force on others?

Hit a teacher?

MPPNr‘

Peer DelinquencyScale
How many of your friends have:

Cheated on school tests?

Destroyed property?

Used marijuana?

Stolen something worth less than $5?
Used alcohol?

Broken into a vehicle?

Stolen something worth more than $50?
Suggest you break the law?

. Gotten drunk?

10. Used prescription drugs?

11. Sold or given alcohol?

©WHQMPPP?

5—All of them, 4—Most of them, 4—Some of them, 2—Very few of them, l—None of
them

48

BIBLIOGRAPHY

Agnew, R. ( 1985). A revised strain theory of delinquency. Social Forces, 64, 151-167.

Agnew, R. (1989). A longitudinal test of the revised strain theory. Journal of
mantitative Criminology, 5, 373-387.

Agnew, R. (1991). The interactive effects of peer variables on delinquency.
Criminology, 29, 47-72.

Agnew, R. (1993). Why do they do it? An examination of the intervening mechanisms
between “social control” variables and delinquency. Journg of Research in Crime
and Delinquency. 30. 245-266.

Akers, R., & Cochran, J. (1985). Adolescent marijuana use: a test of three theories of
deviant behavior. Deviant Behavior. 6, 323-346.

Aultman, M. (1979). Delinquency causation: A typological comparison of path models.
Journal of Criminal Law and Criminology, 70,152-163.

Blumstein, A. (1995). Violence by young people: Why the deadly nexus? National
Institute of Justice Journal 229 1-59. Available
www.anrs.org/txtfileS/nijj_229.txt.

 

Brezina, T. (1996). Adapting to strain: An examination of delinquent coping responses.
Criminology, 34, 39-60.

Brownfield, D., & Thompson, K. (1991). Attachment to peers and delinquent behavior.
Canadian Journal of Criminology, 33, 45-60.

Bruce, N. (1970). Delinquent and non-delinquent reactions to parental deprivation.
British Journal of Criminology,10, 170-176.

Burton, V., & Cullen, F. (1992). The empirical status of strain theory. Joumg of Crime
and Justice, 15. 1-30.

Burton, V., Cullen, F., Evans, T., & Dunaway, R. (1994). Reconsidering Strain theory:
Operationalization, rival theories, and adult criminality. Joum_a1 of Quantitative

Criminology, 10, 213-239.

Chandler, D., Chapman, C., Rand, M., & Taylor, B. Students’ reports of school crime:
1989 and 1995. US Departments of Education and Justice. NCES 98-241/NCJ-

49

169607. Washington, DC: 1998.

Clark, R., & Shields, G. (1996). Families and delinquency: The interaction between
cohesion and adaptability. Journal of Crime and Justice. 19. 67-82.

Cressey, D. (1964). Delinquency, crime. gd differentialaassociation. Hague,
Netherlands: Martinus Nijhoff.

Elliott, D. National Youth Survey (United States): Wave 1, 1976(Computer file). ICPSR
version. Boulder, CO: University of Colorado, Behavioral Research Institute
(producer), 1977. Ann Arbor, MI: Inter-university Consortium for Political and
Social Research (distributor), 1994.

Fagan, J. (1998). Adolescent violence: A view from the street. NflResearch Preview,
1-8.

Felson, R. (1992). "Kick'em when they're down": Explanations of the relationship
between stress and interpersonal aggression and violence. Sociological Quarterly,
33, 1-16.

Gardner, L., & Shoemaker, D. (1989). Social bonding and delinquency: A comparative
analysis. The Sociological Quarterly. 30. 481-500.

Greenfield, L. (1996). Child victimizers: Violent offenders and their victims. Office of
Juvenile Justice Delinquency Prevention and Bureau of Justice Statistics, NCJ-
153258. Available: www.ncjrs,org/txtfiles/chldvrpy.txt.

Goldstein, A., & Conoley, J. (Eds.). (1997). School violence intervention: A practical
handbook. New York: Guilford Press.

Gottfredson, D., McNeil III, R., & Gottfredson, G. (1991). Social area influences on
delinquency: A multilevel analysis. Journal] of Research in Crime and
Delinquency. 28, 197-226.

J unger-Tas, J. (1992). An empirical test of social control theory. Journal of Quantitative
Criminology, 8, 9-28.

Kelly, B., Huizinga, D., Thomberry, T., & Loeber, R. (1997). Epidemiology of serious
violence. Office of Juvenile Justice Delinquency and Prevention. Available:
www.ncjrs.org/txtfiles 1/165 152.txt.

Kems, K., & Stevens, A. ( 1996). Parent-child attachment in late adolescence: Links to
social relations and personality. Journal of Youth and Adolescence 25 323-342.

 

Knight, K., & Tripodi, T. (1996). Societal bonding and delinquency: An empirical test

50

of Hirschi’s theory of control. J oumal of Offender Reaabilitation. 2; 117-129.

Le Blanc, M. (1991). Family, school, delinquency and criminality, the predictive power
of an elaborated social control theory for males. Criminal Behavior and Mengl
Health 4 101-117.

 

Maxfield, M., & Babbie, E. ( 1995). Research Methods for Criminal Justice and
Criminology. Belmont, CA: Wadsworth Publishing.

Menard, S. & Elliott, D. (1990). Self-reported offending, maturational reform, and the
easterlin hypothesis. Journal of Quantitative Criminology, 6, 237-267.

Ollendick, T. (1996). Violence in youth: where do we go from here? Behavior therapy’s
response. Behavior Therapy, 27. 485-514.

Paschall, M., Ennett, S., & Flewelling, R. (1996). Relationships among family
characteristics and violent behavior by black and white male adolescents. Journal
of Youth and Adolescence. 25. 177-197.

 

Paternoster, R., & Mazerolle, P. (1994). General strain theory and delinquency: A
replication and extension. Journal of Research in Crime and Delinquency, 31,
235-263.

Paterson, J ., Pryor, J ., & Field, J. (1995). Adolescent attachment to parents and friends
in relation to aspects of self-esteem. Journal of Youth and Adolescence 24 365-
376.

 

Pearl, R., Bryan, T., & Herzog, A. (1990). Resisting or acquiescing to peer pressure to
engage in misconduct: Adolescents' expectations of probable consequences.
Journal of Youth and Adolescence. 19. 43-55.

Rankin, J. ( 1980). School factors and delinquency: Interactions by age and sex.
Sociology and Social Resaarch. 64, 420-434.

Rankin, J ., & Wells, L. (1990). The effect of parental attachments and direct controls on
delinquency. Jamil of Research in Crime and Delinquency, 27, 140-165.

Roth, J. Understanding and preventing violence. Available:

www.ncjrs.orgltxtfiles/unprvtxt.

Sarri, R., (1983). Gender issues in juvenile justice. Crime and Delinquency, 29, 381-
397.

Seydlitz, R. (1993). Complexity in the relationships among direct and indirect parental
controls and delinquency. Youth and Society. 24, 243-275.

51

Stern, S., & Smith, C. (1995). Family processes and delinquency in an ecological
context. Social Service Review, 69, 703-731.

Sweet, R. (1991). Strengthening our families, fortifying our nation. Juvenile & Family
Court Journal,42, 1-1 1.

Thompson, W., Mitchell, J ., & Dodder, R. (1984). An empirical test of Hirschi’s control
theory of delinquency. Deviant Behavior, 5. 11-22.

Tolone,W., & Tieman, C. (1990). Drugs, delinquency and "nerds": Are loners deviant?
Journal of Drug Education. 20. 153-162.

Trojanowicz, R. & Morash, M. (1992). Juvenile delinquency: Concepts and control.
Englewood Cliff, NJ: Prentice-Hall Inc.

Tygart, C. ( 1988). Strain theory and public school vandalism: Academic tracking, school
social status, and students’ academic achievement. Youth and Society, 20, 106-
118.

Warr, M. (1993). Parents, peers, and delinquency. Social Forces, 7; 247-264.

Windle, M. (1992). A longitudinal study of stress buffering for adolescent problem
behaviors. Developmental Psychology, 28, 522-530.

52

.ll :II I I

"Mal,

. . ., Bryant}.
r...IiLFII11. ”v.3... FEM“... 5|.

. 1:17.. .mmurwfilfl: as
Janina» r
6.51....

3.1.

.12.... J... .
HERE. out-3:."

. «Liam; mud. a... on...
$44.“. Irahnuwuuuir «.314...

1:.

a
I.
.A.

WH...

.r
.a