133
921

THS

 

Heats

G

Illll111211111111111111111

3 01782 5013

 

LIBRARY
Michigan State
University

 

 

 

This is to certify that the

thesis entitled

GANG INVOLVEMENT AND ITS
CONNECTION WITH PHYSICAL ABUSE

presented by

Blythe A. Steele

has been accepted towards fulfillment
of the requirements for

M.S. degree in criminal JUStice

 

 

I
/ / Major professor

Date W997

0-7639 MS U is an Affirmative 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 RECAU£D with earlier due date if requested.

 

DATE DUE I DATE DUE DATE DUE

‘TEC 1 at
'1 2 01723?
AUG 3 0 £033

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

11” W14

 

 

 

 

GANG INVOLVEMENT AND ITS CONNECTION WITH PHYSICAL ABUSE
By

Blythe A. Steele

A THESIS

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

MASTER OF SCIENCE

School of Criminal Justice

1997

.ri >|‘

‘I 'frl" TO!

it’ll 11’

ABSTRACT
GANG INVOLVEMENT AND ITS CONNECTION WITH PHYSICAL ABUSE
By

Blythe A. Steele

The often dysfunctional homes of many gang members can be the center of
physical abuse by an adult. Unreleased anger, unsupported pain, and the lack of feeling
loved may force many of these youths to take vengeance by acting out within the family
of a gang. Robert Agnew’s revised strain theory offers support to the connection between
parental physical abuse and gang involvement. Agnew describes physical abuse as one of
many types of aversive situations that frustrate an adolescent, such that the adolescent
escapes in the form of anger-based delinquency. This study describes anger-based
delinquency as gang involvement. Data were used from the Cambridge Study in
Delinquent Development by David Farrington. A study from London, England was used
for the quality of its data and the fact that it was a longitudinal study spanning over 20
years. This present study was left with the suggestion that stronger measures of parental
physical abuse and data from the United States might produce different results indicating
an effect of parental physical abuse on gang involvement. To test the effect of parental
physical abuse on gang involvement, bivariate and multivariate analyses were performed.
Both crosstabulations and logistic regressions revealed no relationship between parental
physical abuse and gang involvement, with the exception of control variables such as a

peer rating of more or most troublesome.

TABLE OF CONTENTS

LIST OF TABLES .................................................................................. iv
INTRODUCTION ................................................................................... 1

GANG INVOLVEMENT ........................................................................... 2
AGNEW’S REVISED STRAIN THEORY ...................................................... 6

DEPENDENT VARIABLES ..................................................................... 12
INDEPENDENT VARIABLES .................................................................. l3
CONTROL VARIABLES ......................................................................... 14
ANALYTIC PROCEDURES ..................................................................... 16
FINDINGS ........................................................................................... 17
CONCLUSION ...................................................................................... 21
BIBLIOGRAPHY ................................................................................... 28

iii

LIST OF TABLES

Table 1: Codes and Frequencies for Dependent, Independent, and Control
Vanable524

Table 2: Crosstabulations of Indicator Variables and Gang
Involvement25

Table 3: Logistic Regression Results .............................................................. 26

Table 4: Logistic Regression Results With Close to Significant
Vanable327

iv

INTRODUCTION

Many youths involved in gangs come from a background that placed them in a
painful situation rendering them helpless because Of age and circumstance. The Often
dysfunctional homes of many gang members can be the center of physical abuse by an
adult (Spergel, 1990). Therefore, unreleased anger, unsupported pain, and the lack of
feeling loved has forced these adolescents to take vengeance by acting out within the
family of a gang.

Klein, a noted researcher of gangs and gang violence, defined the term gang to be
a denotable group of adolescents and young adults who recognize themselves with a
group name, and who have been involved in a number of illegal incidents that elicited a
negative response from law enforcement agencies and citizens (Klein, 1989).
Membership in a gang increases a youth’s probability of violent Offending (Sheldon,
Tracy, & Brown, 1997). First, membership in a gang increases overall criminal behavior.
Second, certain features of gangs may facilitate violence (Sheldon et al., 1997); Fagan
(1989) has found that gang members are far more likely than non-gang members to
engage in violent acts.

There are many sociological theories that attempt to explain crime and
delinquency, and specifically, the fascination with and the origins of adolescent gangs.

Likewise, the revised strain theory by Robert Agnew Offers an explanation as to why

2

adolescents join gangs. The original strain theory states that delinquency results when an
individual strikes out in frustration or turns to illegitimate means of achieving goals that
could not be achieved through legitimate means. The basis for the revised strain theory is
that adolescents are often placed in aversive or painful situations from which they cannot
legally escape. This frustrates the adolescent and can lead to illegal escape attempts or

anger-based delinquency (Agnew, 1985).

GANG INVOLVEMENT

Research has shown that gangs are a useful channel for the expression of hostility
and the strive for power (Yablonsky, 1969). Some gang members show a preference for
aggression based on their feelings of inferiority and their fear of being rejected or ignored
by others (Gerrard, 1964). Observation (Spergel, 1990) has shown that the leaders or
lieutenants in gangs are those who are the most psychologically disturbed and with the
lowest impulse control. A gang member is viewed as emotionally unstable and as having
difficulty making satisfactory interpersonal relationships. In turn, aggression becomes a
coping mechanism receiving constant reinforcement within the gang (Spergel, 1990). An
aversive family environment has been linked to gang membership: “Deficient home life
is Often cited as an explanation of the gang member’s disturbance and resort to gang
membership....[G]ang members come from stressful family situations” (Spergel, 1990, p.
231). However, Spergel (1990) claims that very little research has been done on the
relation between family variables and participation in gangs. What has been argued is

that the process of becoming a gang member occurs through an accumulation of parental

3

physical or emotional neglect, abuse frOm older street children, punitive educational
incidents, and poor role models (Vigil, 1988).

Sheldon, Tracy and Brown (1997) discuss several of the key family-related factors
that best distinguish delinquents. Of all the relevant factors, that of parental discipline
can be linked to gang involvement. Those parents who use consistently harsh and
physical discipline will produce the most habitual and violent offenders (Sheldon et al.,
1997)

Many mental health providers have seen signs of post traumatic stress syndrome
in children chronically exposed to violence in their homes and in their communities
(Prothrow-Stith, 1991). Given the connection between the violent environment and the
evidence of post traumatic stress, it is not surprising that these adolescents join a gang:

Studies have shown that gang members have had plenty of experience with
violence while growing up. Such youths have seen and have been victimized by
violence in their homes and in their communities. These experiences have shaped
the attitudes of these young people toward perpetuating violence. Gang youth
have accepted violence as the normal and appropriate way to resolve minor and
major disputes. These youth have come to believe that there is no nonviolent
method for dealing with daily disputes and other problems of life. (Sheldon et al.,
1997,p.89)

Many times, gangs will reinforce what their environment has taught them by
encouraging or praising another’s willingness to engage in violence:

If you do not care about yourself, then you are not going to care about what you
do to other people or their rights. This is the feeling ingrained in many gang
members, those from the inner city as well as suburbia. Some youngsters from
the inner city are hardened from a young age by the violence around
them....[T]hey become callous to others and do not really care if they kill
someone or die themselves. For those more middle-class youths...the violence
they experience in their young lives stems from parental abuse, indifference, and
neglect... (Wooden, 1995, p. 10)

4

Gang members are commonly perceived as poor, urban, and minority. Reality
presents any youth, regardless of social or economic status, as being susceptible to gang
recruitment if he or she displays certain characteristics. The characteristics might
include: [The adolescent does not receive positive, loving family attention. If they
receive any at all, it is usually violent, negative and without any rationale. The adolescent
is a victim of abuse or neglect. Youths that experience harsh punishment and/or
inconsistent discipline is a common factor; and the adolescent is usually anti-social.
aggressive, hostile, de-sensitized to violence, and sees violence as a reasonable way to
resolve conflicts (Schaumburg Police Department, n.d.).

No one can refute that sexual abuse within the home is an aversive stimuli for an
adolescent. Hence, studies have been conducted to determine the relationship between
sexual abuse and gang membership. Knox (1994) examined this relationship in an
anonymous survey conducted in five states studying confined juveniles during 1991.
Controlling for sex, it was discovered that male gang members and non-gang members
report about the same levels of sexual abuse. However, among female juvenile
delinquents a much different pattern emerges. Among females, sexual abuse history was
significantly differentiated by gang membership (Knox, 1994). Knox proposed several
hypotheses about sexual abuse and gang membership. Of those hypotheses, one in
particular determined that

Females who have been sexually abused are more likely to subsequently join a

gang out of a social-psychological breakdown. The logic being that the trauma,

personal outrage and anger that can be expected in the aftermath of the life of a

young child abuse victim may propel or predispose such a juvenile to join a gang
subculture. (Knox, 1994, p. 397)

5

A great many variations of families and family relationships can contribute to
youth delinquency and predispose a child to joining a gang. Defects in family
relationships, parental character, and early child-rearing practices are found to be the
origin of youth affiliation with delinquent groups, (i.e., gangs). Researcher Joan Moore
(1991) explained that many of these families from which gang youth derive, for example
the Mexican-origin population,

were not particularly happy, and in some cases, they were acutely unhappy,...

Much of the paternal violence -- towards the kids as well as towards the

mothers -- appears to have been associated with heavy drinking... More than

half of these gang members were clearly afraid of their fathers... [Twenty]

percent of the men and 45 percent of the women grew up with a heroin addict in
the home... And finally, 57 percent Of the men and 82 percent of the women saw
somebody in their home arrested when they were children... (Moore, 1991, pp.
89,92,100,101)

Gang experts insist that the only way to suppress gangs in a society is to restrict
their ability to attract new members. Reality presents new members to be children --
underage teens, pre-teens, adolescents, and even infants (Knox, 1994). Something is
leading these children to join gangs, and a large part of that “something” is violence and
abuse (Knox, 1994). A case in point is that of a child by the name of “Yummy.”
“...Yummy, a member of the Black Disciples, and a full eleven years Old, was executed
by his own gang for fear that he may testify against others in a gang-related killing”
(Knox, 1994, p. 360). By the age of eleven, “Yummy” had produced an extensive record

of arrests and convictions including robbery, arson, and murder. It was also discovered

that Yummy was a victim of child abuse and neglect (Knox, 1994).

6
AGNEW’S REVISED STRAIN THEORY

The revised strain theory by Robert Agnew provides one of the reasons why
adolescents may choose to join gangs. Strain theory was the original creation of Robert
Merton who borrowed Emile Durkheim’s concept of anomie. A definition of anomie
refers to the inconsistencies between societal conditions and opportunities for growth.
Merton’s theory states that there exists no unity between the culturally defined goals
(materialistic status achievement) and the legitimate ways and means to achieve those
goals. In other words, individuals who see themselves at a disadvantage relative to
legitimate economic activities are motivated to commit crime (Sheldon et al., 1997).

In 1960, Cloward and Ohlin expanded Merton’s strain theory. They argued that
poor self-concepts and frustration result from blocked opportunity aspirations; and that
delinquency stems from these frustrations, especially within a gang context. The main
emphasis of Cloward and Ohlin’s theory is that

The disparity between what lower class youth are led to want and what is actually

available to them is the source of a major problem of adjustment. Adolescents

who form delinquent subcultures, we suggest, have internalized an emphasis upon
conventional goals. Faced with limitations on legitimate avenues of access to
these goals, and unable to revise their aspirations downward, they experience
intense frustrations; the exploration Of nonconformist alternatives may be the

result. (Cloward & Ohlin, 1960, p. 86)

Each of this theory’s variations contains the common factor of a strain that
motivates adolescents to go towards a goal that has been blocked. Sociologists who
support strain theory suggest that adolescents turn to delinquency in order to achieve their

goals through illegitimate channels, or to vent their frustration (Agnew, 1984). It is

difficult or impossible to achieve most immediate goals through illegitimate channels.

7

The revised strain theory assigns greater importance to the second link between strain and
delinquency.

Robert Agnew has developed a revised strain theory that asserts adolescents are
forced by necessity to remain in certain environments (e.g., family or school). If those
environments are painful or repugnant to the point of wanting to get away, there is little
they can do legally to escape. Frustration from being unable to escape can lead to illegal.
anger-based delinquency (i.e., gangs) (Agnew, 1985). It is the negative relationships
with others within these environments that ignites the feelings of anger and fi'ustration.
Agnew and other supporters of strain theory argue that the motivation for deviance is
variable. For example, social control may create the possibility for delinquency, but it
may not take place unless the adolescent is strained or provoked in some way.

Criticism from researchers calling for revision or abandonment of the current
strain theory motivated the presentation of a new revised strain theory. Wavering support
of the idea that delinquency results from the blockage of goal-seeking behavior has led to
the idea that delinquency results from pain-avoidance behavior. Agnew (1985) states that
certain situations may be aversive to adolescents although they do not find achievement
of valued goals threatened. “Certain situations may be intrinsically aversive (e. g., the
infliction of physical pain, the deprivation of sensory stimuli); they may be conditioned
aversive stimuli (e. g., verbal insults); or the adolescent may simply be taught to
experience these situations as aversive” (Agnew, 1985, pp. 154-155). The adolescent
believes that exposure to the aversive stimuli is undeserved, which can lead to aggression.

Revised strain theory does not make conclusions about adolescent valued goals or how

8

aversive situations might interfere with the achievement of those goals. It simply
assumes the feeling of frustration from being unable to escape an aversive
situation (Agnew, 1985).

A significant feature of adolescence is lack of power, especially that of being
forced to remain in certain aversive situations (Agnew, 1985). The aversive situation
may be in the neighborhood in which they live, the group of peers with whom they
interact, or most Often, their family itself (siblings and/or parents). Whereas most adults
have legal avenues they can use, such as divorce, quitting one’s job, or moving to another
neighborhood, to escape aversive situations, adolescents lack the power to use those
avenues (Agnew, 1985). “When escape or removal of the aversive source is not possible,
the adolescent may become angry and strike out in rage at the source of aversion or a
related target” (Agnew, 1985, p. 156). Revised strain theory supplements theories such
as social control and differential association by describing how negative relationships
may lead to delinquency (Agnew, 1985). Differential association focuses on how
positive relationships with deviant peers contributes to the formation of negative
attributes in the adolescent. Social control theory, in contrast, focuses on the absence of
meaningful relationships, which convinces the adolescent he has nothing to lose by
engaging in delinquent activities. The revised strain theory supplements these two
theories by explaining that it is the negative relationships that ignite the feelings of
frustration. The frustration from being located in an aversive environment may manifest
itself in the form of anger-based delinquency. The aversive environment can come in
various forms. For example, literature has examined aversive stimuli such as “physical

pain, verbal threats and insults, unfavorable changes in positive reinforcement, unpleasant

9

odors, excessive noise, personal space violations, irritating cigarette smoke, high
temperatures, and disgusting scenes” (Agnew, 1989, p. 374).

Anger-based delinquency can also be operationalized in various ways. In
Agnew’s longitudinal study of his own revised strain theory, he measured anger-based
delinquency in an interpersonal aggression scale that included involvement in gang
fights. High scorers on this scale stated that they have gotten into serious fights, have
been in gang fights, have robbed others, or have hit their teacher and parents (Agnew,
1989). Agnew’s longitudinal study was different from this present study in that he
combined several stressful life events into what he called an Environmental Adversity
scale. He cautioned that one should not examine the impact of a single stressful life event
on the manifestation of anger-based delinquency, but rather one should combine stressful
life events into an overall measure of stress or aversion. In other words, the assumption
should be made that stress or strain is a fimction of the sum total of many events (Agnew,
1989). Although it is desirable to include many facets of aversive life circumstances
when testing the revised strain theory, this present study uses only one because of a lack
of quality measures of aversion within the data set used. It was determined that including
all possible measures of aversion in a scale would be difficult in that aversion is felt on an
individual basis. By selecting and analyzing only one of the more obvious aversive
stimuli (physical abuse), we are better able to examine its particular effect on gang

involvement.

10

DATA

To examine the relationship between gang involvement and physical abuse by a
parent, I used data from the Cambridge Study in Delinquent Development directed by .
David Farrington. The lack of data using subjects from the United States and containing
usable variables regarding gangs and parental abuse prompted my decision to choose the
Cambridge study.

This study by David Farrington can be used because the Cambridge researchers
defined a gang much like the previously noted definition by American gang researcher
Malcolm Klein (1989). In addition, there is a parallel between European white ethnic
gangs and American ethnic gangs. Most authors of gang research trace the evolution of
skinheads, which are one type of white ethnic gang, to the 19708 and the working class in
Great Britain. The basic values of the original skinheads were those of the British
working class (Covey, Menard, & F ranzese, 1992). The few differences between gangs
from Great Britain and gangs from America include: gangs in Great Britain have been
found to be more loosely organized and unstructured than American gangs; although
some gangs have a sense of territory, formal leadership may be absent, and some gangs
do not identify themselves as being gangs; and most importantly, gang membership
appears to have little effect on delinquency in London (Covey et al., 1992). With these
exceptions in mind, this present study continued to use the Cambridge study because of

the quality of its data.

11

The Cambridge researchers felt that no American longitudinal survey of crime
and delinquency was as extensive as the Cambridge Study in Delinquent Development.
The Cambridge study boasts of a unique combination of features that supported my
decision to use it for data analysis. 1) It is a prospective, longitudinal survey over 20
years. 2) The focus of interest is crime and delinquency. 3) Many variables were
measured before the boys were Officially convicted to avoid retrospective bias. 4)
Because of frequent personal contact with the group of boys and their parents, records
were supplemented with interviews, tests, and questionnaire data. 5) All degrees of
delinquency were present. 6) The officially delinquent minority became gradually
differentiated from their non—delinquent peers, avoiding the problem of selecting control
groups. 7) Official and self-report measures of delinquency were used. 8) There was a
very low attrition rate, at least up to age 21; and 9) The numerous variables from
different sources made it possible to test many hypotheses about delinquency.

The Cambridge Stuay in Delinquent Development began in 1962 with 411 males
between ages 8 and 9. AS of May 1987, the researchers were attempting to re-interview
as many of the original sample as possible. At the survey’s inception in 1962, the boys
were all living in a working class area of London, England. A majority of the boys were
taken from a sample of 8 to 9-year-Olds who were attending any Of six state primary
schools in the area of the research Office that chose to cooperate with the study. There
were 399 boys from those six schools, and 12 boys were from a local school for the
educationally subnormal. The additional boys from the educationally subnormal school
were chosen to make a more representative sample of the population of boys living in the

area. The population from which the sample was drawn consisted of mostly white

12

citizens of British origin. Based on their father’s occupations, 93.7% could be described
as working class. Only 12 boys were black and had at least one parent of West Indian

origin.

DEPENDENT VARIABLES

The hypothesis to be tested is whether gang involvement is affected by parental
physical abuse. A cross section of the Cambridge longitudinal study was used for the
purposes of this thesis This paper will use two years of data from the twenty year total.
These two years, when the subjects were between the ages of 18 and 19, reflect the time
during which gang activity is most likely to occur. Research has shown that street-wise
males aged 12 to 14 are the best target for recruitment into a gang (Covey et al., 1992).
For the next few years, these younger members are used for risky violent activities
because they are thought to be insulated from serious criminal penalties (Covey et al.,
1992). It must be stated that involvement in gang activities is not a foolproof indicator of
gang membership. However, most researchers agree that involvement in gang activities
reflects a desire to belong to the gang subculture (Covey et a1, 1992; Goldstein, 1991;
Taylor, 1990).

A dichotomous dependent variable was created to measure gang involvement.
Two sets of variables were combined to indicate membership in a gang and participation
in gang activities. Membership in a gang was defined as whether the boy said he went
around in a group of four or more mates that had a definite structure, area, pub, club,

name, badge, uniform, leader, or initiation requirement. Participation in gang activities

13

was described in various combinations of group fighting and vandalism. As seen in

Table 1, approximately 69% of the boys were involved in gangs in the last two years.

INDEPENDENT VARIABLES
As discussed previously in this paper, parental discipline has been linked

to gang involvement. Those parents who use consistently harsh and physical discipline
will produce the most habitual and violent gang Offenders (Sheldon et al., 1997). Each of
the independent variables selected reflects an aspect of parental discipline, with
categories indicating harsh or abusive discipline practices, along with disinterested or lax
discipline practices. Many of these variables are asked repeatedly during different stages
of the longitudinal study. The variables chosen were all determined prior to the boys
reaching the 18 to 19 year-old age bracket. This was done to explain parental discipline
prior to the boys’ involvement in any form of delinquency. Parental discipline practices
from the years prior to a child becoming a teenager most likely reflect that parent’s
perceived discipline ideals (Goldstein, 1991). It is when a young person enters the
teenage years, facing upheaval in physical, social and emotional development,
questioning adult standards and authority and weakening dependence on parents , that the
teenager becomes dependent on the involvement of a peer group (Goldstein, 1991). If
that peer group encourages the teenager to engage in delinquent activities, a parent’s
discipline practices are forced to change to reflect the child’s misbehavior.

To measure the concept of parental physical abuse, I chose several variables (see

Table 1). One measured both parents’ attitude toward their child by looking at whether

14

either parent was cruel to their child. Another measured whether either parent’s
discipline practices and rules were strict/harsh/rigid or, in contrast, disinterested/lax/
slack. The presence of physical neglect Of the boy was also measured. And finally, this
study measured whether the supervision by the parents was poor.

Looking at Table 1, it is noted that under 3/4 of the boys did not have either
parent treat them cruelly. Likewise, the parent’s discipline practices and rules had a fairly
even distribution among harsh/strict/rigid, disinterested/lax/slack, and neither harsh nor
lax. Over 80% of the boys reported no presence of physical neglect by their parents. The
variable indicating parental supervision showed 75% of the boys as not having poor
supervision. From this Table, it appears that the least skewed variable is parental
discipline practices and rules. It is possible that these results are distributed in the above
stated manner because all of the measures of parental discipline come from the boys’
perceptions of their parent’s practices. It is interesting to note, however, that given these

results, there are still a large number of boys who engaged in delinquent gang activities.

CONTROL VARIABLES

Many other factors besides family related factors can be linked to gang
involvement. Some of these factors were controlled for in this research in anticipation of
a challenge by different sociological theories. Of the many factors contributing to gang
origination, the Illinois State Police (1989) explain that the young men are rebelling
against their low social status. Many of these young gang members come from areas of
overcrowded, substandard housing, poor or nonexistent health care facilities, broken

homes, and few economic opportunities (Illinois State Police, 1989). Another

l5

explanation considers how social control and social learning approaches combine to
become the social development theory. Hawkins and Weis (1985) indicated that the
social influence of peers becomes salient during adolescence. If the social bond process
has been interrupted by, among other things, unconcerned parents with inconsistent
discipline, the youth is more likely to be influenced by peers to engage in delinquent
activities such as gangs (Hawkins & Weis, 1985).

The control variables selected for this present study consisted of ethnic origin,
type and care of family housing, family income, peers rated as like self, and peers rated as
troublesome. Although 87% of the boys in the sample were white and of British origin
according to Table 1, ethnic origin was still controlled for to possibly dispel the
perception of gang members as mostly minorities. The type and care of family housing
measured whether the boy’s living environment was tolerable or satisfactory. This
variable had the purpose Of determining how a tolerable versus an intolerable home might
effect gang involvement. Table 1 shows that 87% of the boys perceived their housing
conditions to be tolerable or satisfactory. Because most gang members have been
perceived as coming from low income families (Illinois State Police, 1989), this present
study measured whether the boy’s family income was adequate or comfortable. Over
75% of the sample perceived their family income to be adequate or comfortable.
Differential association was controlled for by examining how the boys rated their peers.
One variable measured whether the boy’s peers were like himself, and the other variable
measured whether the boy’s peers were more troublesome than himself. Table 1
indicates that both peer rating variables had an even distribution among their perspective

categories.

ANALYTIC PROCEDURES

The data was first analyzed through bivariate analysis in the form of
crosstabulations. Most of the indicator variables and the dependent variable originally
had a category of not known or unascertainable. All individuals who fell into that
category were taken out of the analyses and labeled as missing values. Crosstabulations
were calculated to demonstrate the relationship between each independent and control
variable and the dependent variable of gang involvement. Crosstabulation was the
chosen method of bivariate analysis because all of the variables are categorical. When
you want to look at the relationship between two variables that have a small number of
values or categories, a crosstabulation is the suggested procedure. A crosstabulation is a
table that contains counts and percentages of the number of times various combinations
Of values of two variables occur.

Multivariate analysis is the next step in the data analysis. It is used to look for
relationships between the independent and dependent variables while controlling for other
variables. In order to predict the absence or presence of an effect on gang involvement by
each predictor variable, logistic regression was used. Logistic regression is similar to a
linear regression model, but it is best used in models where the dependent variable is
dichotomous. The logistic regression directly estimates the likelihood of an event

occurring, as in here, gang involvement.

FINDINGS

After performing bivariate crosstabulations for each indicator variable, all but one
of the indicators did not have significant relationships with gang involvement. The
results can be interpreted to Show that the indicator variables of parental physical abuse,
along with the specific control variables Of housing condition, peer ratings, and family
income, are not related to gang involvement. The chi-square values calculated were far
less than the critical values of chi-square for the corresponding .05 alpha level and
degrees of freedom. When the calculated chi-square is less than the critical value, the
closen Significance level becomes greater than .05. Therefore, I must fail to reject the
null hypothesis that parental physical abuse indicators do not affect gang involvement.
Even when examining the strength of the relationships through various measures of
association, the results Showed very weak relationships.

The one variable that proved to be significant was a control variable representing
the theory of differential association. The significant variable was that of a peer rating by
the boy as more or most troublesome. This control variable had a calculated chi-square
of 5.30, which exceeds the critical value of chi-square at one degree of freedom and a .05
alpha level. The Significance level was .021. This indicates a 2.1% chance of making a
type I error by rejecting the null hypothesis. Thus, I can successfully reject the null
hypothesis that a peer rating as troublesome has no effect on gang involvement. This
means that if the boy rated his peers as more troublesome than himself, there is a Strong
possibility that the boy will be involved in a gang. However, peers rated as troublesome
is merely a control variable and not a primary indicator variable of parental physical

abuse.

18

Table 2 represents the crosstabulation for the indicator variables and the
dependent variable gang involvement. Several interesting Observations can be made from
reviewing the results listed in Table 2. For those boys involved in gangs, 73.4% had
parents who were not cruel to them, and only 68.3% had parents who were cruel to them.
It is also interesting to note that Of the boys involved in gangs, 75.9% had parents who
were not harsh or lax in their discipline practices (in other words “normal”). This
percentage is higher than both percentages for harsh or lax discipline and rules. Another
unique result was shown in the parental supervision variable. Of those boys involved in
gangs, 73.4% did not have poor supervision, whereas 71% did have poor parental
supervision. Family income produced yet another interesting result in the Table 2
crosstabulations. Of those boys who had gang involvement, 74% reported adequate or
comfortable family income. Only 70.8% reported that family income was not adequate
or comfortable. The remaining results were not surprising and could have been predicted
based on the previous review of gang literature. Even though none of the above measures
shown in the bivariate analysis has a significant relationship with gang involvement, it is
possible that in a multivariate context that includes the effects of other variables, may
uncover a significant relationship. As indicated in Table 2, the only variable that proved
to be significant was peers rated as more troublesome.

Given the results of the crosstabulations, a logistic regression was performed
using all of the indicator variables to examine the multivariate effects of the independent
variables, taken together, on the dependent variable. To determine how well a logistic
regression model fits the data, I used the Classification Table to compare predicted

outcomes to the observed outcomes.

19

Another way of assessing the goodness of fit of the model is to examine how
“likely” are the results, given the parameter estimates. SPSS assigns parameter codes to
the indicator variables. The probability Of the Observed results, given the parameter
estimates, is called the likelihood. Negative two times the log of the likelihood is used to
measure how well the estimated model fits the data. A good model is one that translates
to a small value for -2 Log Likelihood (-2LL). SPSS presents an initial Log Likelihood
function for a logistic regression model that contains only the constant. The -2LL for the
gang involvement regression had a value of 289.82. A second Log Likelihood is
presented that includes all of the indicator variables. The -2 LL for the gang involvement
regression at this point had a value of 276.55, that is a value smaller than the -2 LL for
just the constant. This indicates a slightly better model.

Yet another assessment of the goodness of fit of a model is the model chi-square.
The model chi-square is the difference between -2 LL for the model with just the constant
and -2 LL for the current model including the variables. It is a likelihood ratio test that
tests the null hypothesis that the coefficients for all of the terms in the current model,
except the constant, are 0 (F test of regression). The model chi-square for the gang
involvement regression had a value of 13.27 at 10 degrees of freedom, with a
Significance level of .21. Again, when the Significance level is greater than .05, the null
hypothesis cannot be rejected. What this means for the gang involvement regression, is
that the coefficients for each of the variables are much greater than 0. There appears to be
no effect of parental physical abuse on gang involvement.

Table 3 contains the results from the logistic regression equation for gang

involvement. The column labeled “B” represents the estimated coefficients (not

20

standardized) for each Of the variables in the regression. “SE.” represents the estimated
standard error; and the “Significance” column represents the Significance level. The
Significance level determines whether the estimated coefficient is significantly different
from 0. As with all levels of significance, if it is greater than .05, then one must fail to
reject the null hypothesis. A model should be created with all of the indicator variables
as well as a model with only the close to significant variables. The hypothesis test would
then be based on the change in the log likelihood.

Table 4 presents the logistic regression model that included only those variables
proving to be close to significant in the first regression. This model produced a slight
change in the log likelihood (-2LL). The variables of ethnic origin, parental discipline/
rules, peers rated like self, and peers rated as troublesome were the closest to the
Significance level of .05. They were placed in the new regression model to assess the
change in -2LL, as well as a change in the model chi-square.

It is particularly interesting to note that ethnic origin has now become a significant
indicator variable. According to Table 3, Non-British youth are less likely to be involved
in gangs than are British youth. This finding contradicts the results in the bivariate
analysis of ethnic origin and gang involvement (Table 2). In the bivariate crosstabulation
of ethnic origin and gang involvement, there was no relationship, but the Non-British
boys were slightly more likely to be involved in gangs than were the British boys.

The new regression calculated -2LL at 310.07. The model chi-square for the new
regression increased in value (13.94) and the degrees of freedom decreased (6). In turn,
the Significance level decreased to .03. This Significance level is less than .05, which

allows one to reject the null hypothesis that the logistic regression model is not a good

21

model for the data. The first regression model (Table 3) had a model chi-square
Significance level of .21. The new regression model (Table 4) proved to be more
significant at .03.

The standard error values decreased for each of the selected variables in the new
regression model (Table 4). The Significance levels for each Of the variables in Table 4
also changed by moving closer to .05. The original regression (Table 3) showed ethnic
origin to be the only significant variable at .04. In the new regression (Table 4), ethnic
origin still remains the only significant indicator at .02. Although the variable of peers
rated as troublesome proved to be significant in the bivariate analysis, once included with

other indicators in the multivariate analysis, it was no longer significant.

CONCLUSION

The results of the analyses indicate that there is no significant relationship
between parental physical abuse and gang involvement. It is interesting that ethnic origin
proved to be related to gang involvement, especially given its skewed distribution.
Recalling the earlier description of the data, 93.7% of the sample population was white.
Remember, however, that ethnicity in the Cambridge study was measured as British or
Non-British in origin. Non-British may not necessarily mean minority. Only 12 of the
411 boys were black. The remainder were white. The crosstabulations showed that of
the boys involved in gangs, a larger percentage was of Non-British origin. This can
easily mean that those boys were most likely white, but originated from possibly another
European country, or even the United States. Therefore, we need to look at the results of

these analyses very carefully.

22

Available data must be accepted cautiously and used conservatively. As in the
United States, accurate data on the number, nature, structure, and functioning of
delinquent gangs are hard to come by. Police, public service agencies, schools, media
representatives, and others regularly exposed to gang youths will often exaggerate or
minimize the numbers and illegal behaviors of gangs (Goldstein, 1991). These
individuals are serving political, financial, or other needs related to the management of
gang perceptions (Goldstein, 1991). Hagedom and Macon (1988) warn that often the
source we deem most reliable, that being gang members themselves, is subjective or
inaccurate.

Given the array of studies cited earlier, physical abuse has been found to be, by
definition, an aversive stimuli and adolescents may manifest aversion through anger-
based delinquency (Agnew, 1989). Gang activity continues to be increasingly more
violent (Knox, 1994). And violence can be referred to as a manifestation of anger-based
delinquency (Agnew, 1989).

Although this study did not find the specific relationship between gang
involvement and parental physical abuse to be significant, there is much reason to believe
that the findings presented here may be atypical. Perhaps a new study with data from the
United States, including stronger measures of parental physical abuse, would produce
different results. One of the limitations in this study was the reliance on the existing
measures for parental physical abuse and neglect supplied by the Cambridge study. If
new measures were created in the Untied States using survey interviews specifically
designed to measure parental abuse/neglect, we might find more significant results

linking parental abuse and gang involvement.

23

If a relationship does exist between physically abusive homes and a child’s
decision to join a gang or participate in gang activities, it was not shown in this study. It
should be noted that this study was limited in the measures available, and the results of
the anlayses could be caused by the skewness of the measures used. However, this study
can mark the beginning of a new direction in gang research. For years, researchers have
been describing gang members as oftentimes victims of child abuse. Yet, how many
studies have attempted to measure the relationship between child abuse and gang
involvement? The results Of this present study might serve as a motivation to find a
connection between physical abuse and gang involvement. If we ever expect to control
the gang problem in this country and throughout the world, we need to stop and/or limit

the source from which gangs draw their members.

24

Table l: CODES AND FREQUENCIES FOR DEPENDENT. INDEPENDENT AND

 

 

 

 

CONTROL VARIABLES
M Code Mlle—Ml
Dependent
Gang involvement in the 0 = no 104 (25.3%)
last two years 1 = yes 285 (69.3%)
2 = missing 22 (5.4%)
Independent
Either parent is cruel to the child 0 = no 299 (72.7%)
1 = yes 43 (10.5%)
2 = missing 69 (16.8%)
Either parent’s discipline and rules 0 = no (normal) 117 (28.5%)

is strict/harsh/rigid or disinterested/ l = harsh, etc. 121 (29.4%)
lax/slack. 2 = lax, etc. 105 (25.5%)
3 = missing 68 (16.5%)
Presence of physical neglect 0 = no 349 (84.9%)
of boy 1 = yes 49 (1 1.9%)
2 = missing 13 (3.2%)
Supervision by parents poor 0 = no 309 (75.2%)
1 = yes 74 (18%)
2 = missing 28 (6.8%)
Control
Housing conditions are tolerable 0 = no 41 (10%)
or satisfactory l = yes 358 (87.1%)
2 = missing 12 (2.9%)
Peers rated as more 0 = no 175 (42.6%)
like self 1 = yes 178 (43.3%)
2 = missing 58 (14.1%)
Peers rated as more 0 = no 178 (43.3%)
troublesome l = yes 175 (42.6%)
2 = missing 58 (14.1%)
Ethnic origin of boy 1 = British 357 (86.9%)
2 = Non-British 54 (13.1%)
Family income is adequate 0 = no 93 (22.6%)
or comfortable 1 = yes 318 (77.4%)

 

 

25

Table 2: CROSSTABULATIONS OF INDICATOR VARIABLES AND GANG

 

 

 

INVOLVEMENT

Variable Column % and N for yes on gang involvement
Independent
Either parent is cruel to child yes 68.3% (28)

no 73.4% (212)
Either parent’s discipline and no 75.9% (85)
rules is strict/harsh/rigid or harsh 69.2% (81)
disinterested/lax/slack lax 74.0% (74)
Presence of physical neglect yes 76.7% (33)

no 73.4% (245)
Supervision by parents is poor yes 71.0% (49)

no 73.4% (218)
Control
Housing conditions are tolerable yes 73.2% (248)
or satisfactory no 74.4% (29)
Peers rated as more or most ' yes 78.5% (135)
like self no 69.9% (1 l6)
Peers rated as more or most yes 79.8% (134)
troublesome no 68.8% (1 l7) *
Ethnic origin of boy British 71.8% (242)

Non-British 82.7% (43)
Family income is adequate or yes 74.0% (222)
comfortable no 70.8% (63)

* Indicates variable with Significance level less than .05

 

 

Table 3: LOGISTIC REGRESSION RESULTS

26

 

 

Variable

Either parent is cruel

Parent’s discipline/rules
strict/harsh/rigid
disint/lax/slack

Presence of physical neglect

Poor parental supervision

Adequate income

Tolerable housing condition

Peers rated like self

Peers rated as troublesome

Ethnic origin

Constant

B
21864
21447
n6570
a9027
.8584
a3038
.1334
.4787
.4421
-L0720

3 .0942

_S._E_.

.4678

.4778

.4729

.7735

.5580

.4089

.6364

.3037

.3102

.5228

.9820

* Indicates a variable with a Significance level less than .05

We
.69
.24
.76
.16
.24
.12
.46
.33
.12
.15

.04 *

.00

 

-2LL: 276.55

Model Chi-Square: 13.27 (10 d/f)

Significance: .21

 

27

Table 4: LOGISTIC REGRESSION RESULTS WITH CLOSE TO

 

 

 

SIGNIFICANT VARIABLES
Variable B §._E_. Significance
Parent’s discipline/mics .20
strict/harsh/rigid -.2574 .4354 .55
disint/lax/slack -.7048 .4349 .1 1
Poor parental supervision .6570 .4747 .17
Peer rated like self -.4626 .2804 .10
Peer rated as troublesome -.4676 .2868 .103
Ethnic origin -1.1815 .5091 .02 *
Constant 2.3670 .6080 .00

* Indicates a variable with 3 Significance level less than .05

 

-2LL: 310.07

Model Chi-Square: 13.94 (6 d/f) Significance: .03 *

 

BIBLIOGRAPHY

BIBLIOGRAPHY

Agnew, R. (1984). Goal achievement and delinquency. Sociology and Social
Research. 68(4), 435-451.

Agnew, R. (1985). A revised strain theory of delinquency. Social Forces. 64(1),
151-165.

Agnew, R. (1989). A longitudinal test of the revised strain theory. Journal of
Qualitative Criminology, 5(4), 373-387.

Agnew, R. (1993). Why do they do it? An examination of the intervening
mechanisms between “social control” variables and delinquency. Journjalof Research in
Crime and Delinquency. 30(3), 245-266.

Cloward, R. & Ohlin, L. (1960). Delinquency and Opportunity. New York: Free
Press.

Covey, H.C., Menard, S., & Franzese, RJ. (1992). Juvenile gangs. Springfield,
Illinois: Charles C. Thomas, Publisher.

Fagan, J .A. (1989). The social organization of drug use and drug dealing among
urban gangs. Criminology, 27, 633-667.

Gerrard, N.L. (1964). The core member of the gang. British Journal of
Criminology, 4. 361-371.

Goldstein, AP. (1991). Delinquent gangs: a psychological perspective.
Champaign, Illinois: Research Press.

Hagedom, J ., & Macon, P. (1988). People and folks. Chicago: Lake View.

Hawkins, J .D. & Weis, J .G. (1985). The social development model: an integrated
approach to delinquency prevention. Joumilof Primary Prevention, 6. 73-97.

Illinois State Police. (1989, January). Criminal intellgence bulletirL (42).
Springfield, IL: Author.

28

29

Klein, M. (1989). Street gang violence. In Wolfgang & Weiner, Violent crime,
violent criminals. Newbury Park, CA: Sage.

Knox, G.W. (1994). An introduction to gang; Bristol, IN: Wyndham Hall Press.

Loeber, R., & Stoutharner-Loeber, M. (1986). Family factors as correlates and
predictors of juvenile conduct problems and delinquency. In M. Tonry & N. Monis
(Eds.), Crime and justice. an annu_al_review of research: vol. 7. Chicago: The University
of Chicago Press.

Moore, J .W. (1991). Going down with the barrio. Philadelphia: Temple
University Press.

Prothrow-Stith, D. (1991). Deadly consequences. New York: Harper Collins.

Schaumburg Police Department, Schaumburg, Illinois. Gangs: fighting back in
our neighborhoods [Brochure]. Rizzo, Paul: Author.

Sheldon, R.G., Tracy, S.K., & Brown, W.B. (1997). Youth gangs in american
society. Detroit: Wadsworth Publishing Company.

Spergel, LA. (1990). Youth gangs: continuity and change. In Michael Tonry &
Norval Morris (Eds.), Crime and iustice. an annual review of research: vol. 12. Chicago:
The University of Chicago Press.

Spergel, LA. (1995). The youth gangproblem: a community approach. New
York: Oxford University Press.

Taylor, OS. (1990). Gang imperialism. In C.R. Huff (Ed.), Gangs in america.
California: Sage Publications, Inc., 103-115.

Vigil, J .D. (1988). Street socialization, locura behavior, and violence among
Chicano gang members. In J. Kraus & A. Morales (Eds.), Violence and homicide in

hispanic communities. Washington DC: National Institute of Mental Health.

Vold, G.B., & Bernard, T.J. (1986). Theoretical criminology (3rd ed.). New York:
Oxford University Press.

Wooden, W.S. (1995). Renegadegkids. suburban outla_vy_s_, Detroit: Wadsworth
Publishing Company.

Yablonsky, L. (1962). The violent gang. New York: Macmillan.

 

"I11111111111111111111111“