u“. .44
p... 4 Swan:

7 .9 AWN”

L

M v.
. s 2.!
A???

L».
«A;

z?!

u}?

.
in..-

a whit.

_ ..o .,
fitfizfiwflflwla
‘ . ‘ .H..1L.. .

Madam. 3!.

lift... ....
$.52 . . .. u
an n 5...:
g
. .

r“

‘ gift-
1,“.

 

L...
a £731 x:
.. I.

b
51.51:.”
. .3:

:Y. 2324;, , .3...:.-...

 

 

3 . LLIBRARY
55,36 W"? Michigan State
Untversnty

This is to certify that the
dissertation entitled

ASSESSING THE EFFICACY OF THE SOCIAL CONTROL
THEORY FOR EXPLAINING THE CIGARETTE SMOKING
AND ALCOHOL DRINKING BEHAVIOR OF AFRICAN
AMERICANS AND MEXICAN AMERICANS

presented by

Jessica Shovette Davis

has been accepted towards fulfillment
of the requirements for the

PhD. degree in School of Criminal Justice

flZ/é/y

Major P /oiéssors Sig a“

9/ /f/y

Date

MSU 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 RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/01 c:/ClFtC/DateDue.p65-p. 15

 

ASSESSING THE EFFICACY OF THE SOCIAL CONTROL THEORY FOR
EXPLAINING THE CIGARETTE SMOKING AND ALCOHOL DRINKING
BEHAVIOR OF AFRICAN AMERICANS AND MEXICAN AMERICANS
BY

Jessica Shovette Davis

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
DOCTORATE OF PHILOSOPHY

School of Criminal Justice

2003

ABSTRACT
ASSESSING THE EFFICACY OF THE SOCIAL CONTROL THEORY FOR
EXPLAINING THE CIGARETTE SMOKING AND ALCOHOL DRINKING
BEHAVIOR OF AFRIC N AMERICANS AND MEXICANS
BY

Jessica Shovette Davis

This research addressed the efficacy of the social
control theory in explaining the cigarette smoking and
alcohol drinking behavior of African American and Mexican
youth as well as it explains these behaviors for white
youth. This is a secondary data analysis that utilizes the
National Educational Longitudinal data (NELS:88).

Results from the data analyses indicate that the
drinking and smoking behavior of females was significant.
In addition, findings indicate the model was a better
explanation of the smoking and drinking behavior of whites
than African Americans and Mexicans. Also, the model was a
better explanation of the smoking and drinking behavior of
Mexicans than African Americans. Finally, the social
control model is a better explanation of the smoking
behavior of all adolescents than it is of their rinking
behavior.

Future research utilizing the social control variable

should (I) include original data collection to address the

limitations of having to compute variables from data that
has already been collected. This would insure the exact
specification of the social control indicators as well as
the control variables (i.e., race and class); and (2)
Include social indicators that are conceptualized taking
into account cultural differences. This would require the
researcher to understand what cultural differences might
exist when considering the conceptualization of attachment,

commitment, involvement, and belief.

 

Copyright by
JESSICA SHOVETTE DAVIS
2003

This is dedicated to my parents for their unwavering love
and support throughout this process and throughout my life.
They are the cornerstone of my foundation and my shelter in

time of storm, for this, I thank them.

ACKNOWLEDGEMENTS

I would like to thank my dissertation committee, my
family, friends, and mentors for their support throughout
this process. Specifically, I would like to thank my
committee for their guidance and instruction. I will take
with me the lessons learned and apply them throughout my
career. I would like to thank my family for their prayers
and words of encouragement when it seemed there was no end
in sight (I see the light at the end of the tunnel and I am
pretty sure it is not a train). I would like to thank my
friends for their love and support, as well. Your prayers
and words of encouragement have also inspired me to
complete what I started, what seems like a lifetime ago. I
would also like to thank Dr. Willie Tompkins for getting me
back on track when I was in a slump. I think God sent you
to me when I needed the right encouragement at the right
time. Your advice and constant “Jessica get it done” was
immeasurable and I will always be grateful. Finally, I
would like to thank Drs. Harvey L. McMurray, Sr. and George
P. Wilson for believing in me. I know I would nOt have
achieved this degree without your support and training as a
Master’s student at North Carolina Central University, and

for this I thank you.

TABLE OF CONTENTS

LIST OF TABLES .................................................................................................................................... X

CHAPTER I
INTRODUCTION/PROBLEM STATEMENT mmmmmmmmmmmmmmmmmmmmmmml
The Social Control Theory ... ... ”mmmmmmmml
Hispanics and the Social Control Theory ........................... 3
Smoking Cigarettes and Drinking Alcohol ........................... 5
Significance of the Problemmmmmmmmmmmmmmmmmmmmmmmmmmm6

CHAPTER 2
LITERATURE REVIEW ........................................................................................................................... 8
Theoretical Framework: Social Control Theory ........................... 8
Limitations of the Social Control Theory ..................... 12
The Social Control Theory and Race ....................................... 14
Traditional Conceptualization of Social Control
Elements ....................................................................................... 98
Cigarette Smoking and Alcohol Drinking: A Review of
the Literature .................................................................................................................. 33
Factors that Affect Cigarette Smoking and
Drinking Alcohol for Adolescentsmmmmmmmmmmmmmmm35
Cigarette Smoking and Alcohol Drinking:
Race/Ethnic Differences ..mm.. "mmm44
Significance of this Dissertation and Study ............ 47
Hypothesis ............................. .mu .mmmmmmmmmm49

 

 

CHAPTER 3
METHODOLOGY .mmm. n.mmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmm5O
Research Design, Data and Samplemmmmmmmmmmmmmmmmmmmm5O
Data Collection ................................................................................................ SI
Specification of Variables Under Study .......................................... S4
Dependent Variablesmmmmmmmmmmmmmmmmmmmmmmmmmmmm54
Independent Variables mmmmmmmmmmmmmmmmmmmmmmmmm54
The Control Measures ................................................................................................ 57
Race/ethnicity ................................................................................................... 57
Socio-Economic Status Indicators ............................................. 58
Data Cleanup ........................................................................................................................ 59
Missing Data ......................................................................................................... 59
Composite Measuresmmmmmmmmmmmmmmmmmmmmmmmmmmmmm6l
Data Analysis Strategy .......................................................................................... 64
Univariate Analysesmmmmmmmmmmmmmmmmmmmmmmmmmmmm64
Bivariate Analysesmmmmmmmmmmmmmmmmmmmmmmmmmmmmm64
Logistic Regression .................................................................................... 65

 

{-

Strengths and Limitations of the Study .......................................... 60

CHAPTER 4
FINDINGS AND DISCUSSION mmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmm68
Findings .................................................................................................................................... 68

Univariate Analysis .................................................................................... 68
Bivariate Analyses: The Social Control Indicators and
Gender ..mm.. .mmmmmmwmmmmmmmmmmmmmmmmmm74

Attachment and Cender .............................................................................. 74

nvolvement and Gender ........................................................................... 77

Commitment and Geno er mmmmmmmmmmmmmmmmmmmmmmmmm78
Belief and Gender ..mm”. .H M“.M8O
Bivariate Analyses: The Social Control Indicators and

Race/Ethnicity .mm.. .m. .mmmmmmmmmmm". 81
Attachment and Race/F. hniCIty ...................................................... 81
Involvement and Race/Ethnicity ................................................... 85

Commitment and Race/Ethnicitymmmmmmmmmmmmmmmmmm86

Belief and Race/Ethnicity .................................................................. 88
Bivariate Analyses with Smoking Cigarettes and the
Independent Variables ............................................................................................. 89

Smoking Cigarettes and the Control Variables ......... 89
Attachment and Cigarette Smokingmmmmmmmmmmmmmmm94
Involvement and Cigarette Smoking" .. .95

 

Commitment and Cigarette Smoking mmmmmmmmmmmmmm95

Belief and Cigarette Smoking ......................................................... 96
Bivariate Analyses with Drinking Alcohol and the
Independent Variables ..................................... .mmm 97

Drinking Alcohol and the Control Variables ............... 97

Attachment and Drinking Alcohol ............................................. 102

Commitment and Drinking Alcohol ............................................. 103

Belief and Drinking Alcohol . .m. "103
Gamma Measures of Association and Smoking

 

 

Cigarettes .mm.. nmmwmmmmmmmmmmmmmmmmmmmmm104
Attachment Indicators ........................................................................... 104
Commitment Indicators. ..mm. .mmmmmmmmmmmm106
Involvement and Smoking Cigarettes .................................... 109
Belief and Smoking Cigaret ttes ................................................... 112

Gamma Measures of Association and Drinking Alcohol 114
Attachment Indicators ........................................................................... 114
Commitment and Drinking Alcohol ............................................. 116
Involvement and Drin ing Alcohol mmmmmmmmmmmmm120
Belief and Drinking Alcohol. .mmmmmmmmmmmmmmmmIZ

Logistic Analyses and Cigarette Smoki ng mmmmmmmmmwm124
Entire Sample and Smoking Cigarettes ............................. 125
Whites and Smoking Cigarettes mmmmmmmmnmmmmmmml3l
African American 1outh and Smc<ing Cigarettes 137
Mexican American Youth and Smoking Cigarettes 140

vm

 

Logistic Analyses and Drinking Alcohol ..... 144

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Entire Sample and Drinking Alcohol 144
Whites and Drinking Alcohol 148
African American Youth and Drinking Alcohol ......... 153
Mexican American Youth and Drinking Alcohol ......... 156
Discussion ""159
CHAPTER 5
DISSERTATION SUMMARY 163
Purpose and Significance of this Research 163
Summary of the Literature Review 164
The Social Control Theory 164
Literature on Smoking and Drinking 167
The Methodology 168
The Data 168
Dependent Variables 168
Independent Variables ................................... 169
The Control Measures 171
The Findings ...... 171
Univariate Analysis ............................. 171
Bivariate Analyses 172
Social Control Indicators and Gender 172
Social Control Indicators and Race/Ethnicity ...... 173
Smoking Cigarettes, Control Variables, and Social
Control Indicators 174
Drinking Alcohol, Control Variables, and Social
Control Indicators _ 175
Gamma Measures of Association for Smoking
Cigarettes and the Social Control Indicators for
the Entire Sample, African Americans and
Mexicans .176
Gamma Measures of Association for Drinking
Alcohol and the Social Control Indicators for the
Entire Sample, African Americans and Mexicans 177
Logistic Regression Findings and Smoking
Cigarettes 178
Logistic Regression Findings and Drinking
Alcohol 179
Limitations of the Study "m. 181
Future Research -mm. 182
APPENDICES ...... 183
APPENDIX A 184
APPENDIX B 188
APPENDIX C 190
REFERENCES .......... nmmm. 198

 

 

Table

Table

Table

Table

Table

Table

Table

Table

Table

Table

10:

LIST OF TABLES

Frequencies of Dependent Variables, Gender,
Race/Ethnicity, Income, Parent’s Education, and
Social Bond Indicators (N=11,038) ............................................. 69

Results of the
Social Control

Bivariate Analysis of Gender and
Indicatorsmmmmmmmmmmmmmmmmmmmmmmm75
Bivariate Analysis of

and Social Control Indicators ............ 82

Results of the
Race/Ethnicity

Bivariate Analysis of the Dependent Variable
Smoking Cigarettes with Independent Variables ......... 9O

Bivariate Analysis of the Dependent Variable
Drinking Alcohol with all Independent

Variablesmmmmmmmmmmmmmmmmmmmm. .......................... 98

 

Gamma Measures of Smoking Cigarettes and the
Attachment Indicators for the Entire Sample,
Whites, African Americans, and Mexicans mmmmumm105
Gamma Measures of Smoking Cigarettes and the
Commitment Indicators for the Entire Sample,
Whites, African Americans, and Mexicans ........................ 107

Gamma Measures of Smoking Cigarettes and the
Involvement Indicators for the Entire Sample,
Whites, African Americans, and Mexicans mmmmmmmlll
Gamma Measures of Smoking Cigarettes and the
Belief Indicators for the Entire Sample, Whites,
African Americans, and Mexicans mmmmmmmmmmmmmmm113

Gamma Measures of Drinking Alcohol and the
Attachment Indicators for the Entire Sample,
Whites, African Americans, and Mexicans ..................... 115
Gamma Measures of Drinking Alcohol and the
Commitment Indicators for the Entire Sample,
Whites, African Americans, and Mexicans ..................... 118

Table

Table

Table

Table

Table

Table

Table

Table

Table

Table

Table

Table

Table

Table

12:

13:

14:

15:

16:

17:

18:

19:

20:

21:

22:

23:

24:

25:

Gamma Measures of Drinking Alcohol and the
Involvement Indicators for the Entire Sample,
Whites, African Americans, and Mexicans ..................... 121

Gamma Measures of Drinking Alcohol and the
Belief Indicators for the Entire Sample,
Whites, African Americans, and Mexicans mmmmmm123
Logistic Regression Results for the Entire Sample
with the Dependent Variable Smoking

Cigarettes 126

 

Logistic Regression Results for Whites and the

Dependent Variable Smoking Cigarettes 132

 

Logistic Regression Results for African Americans
and the Dependent Variable Smoking Cigarettesm138

Logistic Regression Results for Mexicans and the
Dependent Variable Smoking Cigarettes 141

 

for the Entire Sample
Drinking Alcohol ...... 145

Logistic Regression Results
with the Dependent Variable

for Whites and the
Alcohol

Logistic Regression Results

Dependent Variable Drinking 149

 

Logistic Regression Results for African Americans
and the Dependent Variable Drinking Alcohol mm154

Logistic Regression Results for Mexicans and the
Dependent Variable Drinking Alcohol 157

 

Comparison of Missing Respondents in Data Sample
to Population Parameters in the 1990 Census

 

 

 

Data 189
Correlations of Smoking Cigarettes and Social
Control Measures for Whites 192
Correlations of Drinking Alcohol and Social
Control Measures for Whites 193
Correlations of Smoking Cigarettes and Social
Control Measures for African Americans ........................ 194

xi

Table 26:

Table 27:

Table 26:

Correlations of Drinking Alcohol and Social
Control Measures for African Americans ........................ 195

Correlations of Smoking Cigarettes and Social
Control Measures for Mexican Americans ........................ 196

Correlations of Drinking Alcohol and Social
Control Measures for Mexican Americans ........................ 197

xii

CHAPTER I
INTRODUCTION/PROBLEM STATEMENT

There is a dearth of evidence that the social control
theory is a viable explanation of delinquent behavior for
youth of color. This paper will examine previous research
utilizing the social control theory and their findings on
race/ethnicity and test the applicability of the social
control theory to explain the cigarette smoking and alcohol
drinking behavior for African American and Mexican youth.
Specifically, this research examines (a) the current
research using the social control theory to explain
delinquency; (b) the use of the social control theory as it
pertains to African American and Mexican American youth;
and (c) the extent to which the social control theory
explains the cigarette smoking and the drinking of alcohol

among African American and Mexican American youth.

The Social Control Theory
The purpose and usefulness of any delinquency theory
is related to the extent to which it can explain
delinquency. Since the inception of Hirschi’s (1969) social
control theory there has been a vast amount of research
1

that has examined its ability to explain juvenile

delinquency. Hirschi’s (1969) control theory has been

F.)

tested and has received a wealth of empirical support
(Hepburn, 1977; Hindelang, 1973; Johnson, 1979; Krohn and

Massey, 1980; Wiatrowski, Griswold and Roberts, 1981).

The theory’s explanatory power has been described as
moderate to good depending on its use (Gardner and
Shoemaker, 1989). One of the main tenets that Hirschi
(1969) posited is the social control theories’
applicability to explain the behavior of youth across all
socio-economic categories. However, much of the research
that has utilized social control theory has tended to gloss
over the applicability of the model to youth of color.
Prior research has indicated the social control theory
loses its explanatory power when race is introduced into
the equation (Gardner and Shoemaker, 1989; Covington, 1988;
Liska and Reed, 1985; Weber, et. al., 1995; Cerkonvich and
Giordano, 1992). In addition, the amount of variance the
model explains in regression analyses is far less for youth
of color than for majority youth (Gardner and Shoemaker,
1989; Covington, 1988; Liska and Reed, 1985; Weber, et.
al., 1995; Cerkonvich and Giordano, 1992).

In addition to the problems with the efficacy of
social control theory to explain juvenile delinquency

relative to race or ethnicity, much of the research that

It)

does address race tends to focus mostly on the racial
dichotomy of African Americans and White Americans
(Covington, 1988; Liska and Reed, 1985; Cerkonvich and
Giordano, 1992; Gardner and Shoemaker, 1989). The
Black/White dichotomy does not address issues pertinent to
many racial/ethnic groups that find themselves within the
criminal justice system. Hence, there is a need to broaden
the research to include Other racial and ethnic groups

including immigrant populations.

Hispanics and the Social Control Theory

Research utilizing the social control theory has shown
there is a dearth of information assessing the utility of
the theory to explain the delinquent behavior of Hispanic
youth. There is very little published where any Hispanic
ethnicity is included in the analyses. The scant social
control research including Hispanics and more specifically
Mexican Americans is problematic, as Mexican Americans are
over represented in the criminal justice system as well as
African Americans. Data from the Office of Juvenile Justice

and Delinquency Prevention indicate that Hispanics commit

(I

delinquent and deviant acts in proportion to vheir numbers
in the general population (12.5%)(OJJDP, 1998). However,

arrest, adjudication, and detention data indicate they are

 

over represented in the criminal justice system (OJJD.,

1998).

The rapid growth of the Hispanic population in general
requires attention across all major institutions including
the criminal justice system. The Hispanic population in the
United States has increased dramatically since the 1990
census. The 1990 census indicates Hispanics represented
approximately 9% of the population (US Census, 1999). In
contrast, the 2000 census indicates Hispanics represent
12.5% of the total United States population (US Census,
2000). This increase makes Hispanics the largest minority
group in the United States out numbering African Americans.
The increase in the Hispanic population also affects their
presence in the criminal justice system (Note: there are
only recent data available on Hispanics in the criminal
justice system that allows comparison across all races).
Overall, the data illustrate Hispanics are over represented
at key points in the criminal justice system (i.e., arrest,
adjudication, and placement). Data suggests that Hispanic
males and Black males were more likely to be arrested than
white males (OJJDP, 1999; NLSY, 1997). In addition, a

greater proportion of Black and Hispanic males were

\0
SJ

\./
I

arrested more than white males (OJJDP, 1999; NLSY, 1 9

Hisqoanics represented 18% of all people in residential
placement in 1997 (OJJDP, 1999; NLSY, 1997). In addition,
Hispanics represented 15% of the overall population of
those who were admitted to adult prisons in 1997 (OJJDP,

1999).

It is pertinent to understand the extent to which the
social control theory explains the behavior of Hispanic
youth in order to effectively address the issue of over
representation of Mexican American youth overrepresentation

in the criminal justice system.

Smoking Cigarettes and Drinking Alcohol

The research on cigarette smoking and drinking alcohol
indicate that participation in these delinquent acts varied
by race. The data indicate African American males and
females were less likely to drink alcohol or smoke
cigarettes in the 30 days prior to being interviewed that
white or Hispanic youth.

The research on adolescent cigarette smoking and
drinking focuses mainly on two themes. First, parents and
peers have the most influence over adolescent cigarette
smoking and alcohol use than any other factors. Secondly,
is the differential smoking and drinking of youth in

reference to race/ethnicity.

SIGNIFICANCE OF THE PROBLEM

It is important that theories used as basis for policy
and programs address the specific and cultural differences
of youth of color. The social control theory is one such
theory that influences policy and programs for all youth;
however, current research lacks compelling evidence the
social control theory is a viable explanation of the
delinquent behavior for youth of color.

In addition, the social control research that does
address issues of race/ethnicity primarily focuses on the
Black/White race dichotomy and ignores the issues of
Mexican youth. Hawkins (1995) suggests much of the researc
in criminal justice tend to focus mostly on the racial
dichotomy of African Americans and white Americans. The
Black/White dichotomy does not address the specifics of
many racial/ethnic groups that find themselves in contact
with the criminal justice system. There needs to be a
broadening of the research to include other racial and
ethnic groups.

The current research will examine previous findings on
the social control theory and race, as well as, test the

applicability of the social control theory in explaining

ON

 

Cfi43arette smoking and alcohol drinking behavior of African
American and Mexican youth.
Does the social control theory explain the cigarette
smoking and alcohol drinking behavior of African American
and Mexican youth as well as it explains these behaviors

for white youth? This study aims to an wer this question

by:
a) examining the race findings from previous
research,
b) examining the traditional conceptualization and

operationalization of the social control theory
in the literature to understand how the
traditional treatment might affect research
findings, and

c) conducting analyses that include whites, African
Americans and Mexican Americans from a national

data set.

CHAPTER II
LITERATURE REVIEW
The literature review addresses three themes: (a) a
critique of the social control theory (b) the evidence of
the efficacy of social control theory to explain the
delinquency of youth of color; and (c) a review of the
literature on the cigarette smoking and alcohol drinking

behavior of youth.

THEORETICAL FRAMEWORK: SOCIAL CONTROL THEORY

Since the inception of Hirschi’s (1969) social control
theory there has been a plethora of research that has
examined its ability to explain juvenile delinquency
overall. Hirschi’s (1969) social control theory has been
tested and has received a wealth of empirical support
(Hepburn, 1977; Hindelang, 1973; Johnson, 1979; Krohn and
Massey, 1980; Wiatrowski, Griswold and Roberts, 1981). The
theory’s explanatory power has been described as moderate to
good depending on its use (Gardner and Shoemaker, 1989). One
of the tenets that Hirschi (1969) posited about social
control theory is its applicability across all socio-
economic categories. However, much of the research that has
utilized social control theory has tended to gloss over the
applicability of the model to youth of color. This, despite

the fact that the social control theory loses its

(1)

explanatory power or mixed results are obtained when race is
introduced into the equation (Gardner and Shoemaker, 1989;
Covington, 1988; Liska and Reed, 1985; Weber, et. al., 1995;
Cerkonvich and Giordano, 1992). In addition, the amount of
variance the model explains in regression analyses is far
less for adolescents of color than for majority youth
(Gardner and Shoemaker, 1989; Covington, 1988; Liska and
Reed, 1985; Weber, et. al., 1995; Cerkonvich and Giordano,
1992).

In addition to the problems with the efficacy of social
control theory to explain juvenile delinquency relative to
race or ethnicity, much of the research that does address
race tends to focus mostly on the racial dichotomy of
African Americans and white Americans (Covington, 1988;
Liska and Reed, 1985; Cerkonvich and Giordano, 1992; Gardner
and Shoemaker, 1989). The Black/White dichotomy does not
address issues pertinent to many racial/ethnic groups that
find themselves within the criminal justice system. There
needs to be a broadening of the research to include other
racial and ethnic groups including immigrant populations.

The underlying assumption of the social control theory
is that everyone has tendencies to commit criminal acts.
However, the desire to act upon those criminal tendencies
are suppressed by the bonds they form in society to

individuals (i.e., parents, teachers, peers) as well as to

socially accepted norms (i.e., completing high school,
attending to college, getting married)(Hirschi, 1969).
Elements of the social bond as explained by Hirschi
(1969) are attachment, commitment, involvement and belief.
Attachment is considered the most affective element of all
the social bonds and attachment to parents is the most
important bond within this element because parents are
viewed as the primary influence over behavior as well as
socializing youth to the acceptable norms in society. The
attachment element also consists of attachments to other
family members, friends, and the community at large (i.e.,
teachers, clergy, etc.). Commitment to conventional
attitudes or conformity is the second element in social
control theory. Hirschi (1969) specifically addressed
conventional attitudes such as going to college and
attaining a high status job upon graduating as indicative of
a juvenile committed to conventional attitudes. The element
of involvement refers to participation in conventional
activities such as school related (i.e., sports, band,
clubs, etc.) and other extracurricular activities.
Involvement assumes a person who is involved in
“conventional” activities will be too busy to engage in
illicit activities. The belief bond suggests an individual’s

belief in the rules of society as a common value system,

10

within the society, whose norms are being violated (Hirschi,
1969).

The bonds of the social control theory have a negative
effect on delinquency, the stronger the bonds the more
likely delinquency will decrease and vice versa. In
addition, Hirschi (1969) suggests the influence of the bonds
on delinquency is not equally distributed among the four
bonds. In fact, he suggests that attachment to parents is
seen as the most affective bond. If parental attachment is
strong then this lessens the influence of the other bonds on
delinquency. However, if the affects of parental attachment
or attachment in general is not strong, then one of the
other bonds would have to be strong in order to prevent or
reduce the likelihood of delinquent behavior. For example,
there are youth who grow up in environments where there is
not a strong parental influence as the result of the
parents’ absence due to drug use, imprisonment, abandonment,
etc. However, these youth may not grow up to be delinquents
because they have been influenced by other members of their
community. These youth may have formed an attachment to
clergy, teachers, coaches, etc. Moreover, the youth may be
involved in sports and other school-related activities and
realizes that his/her involvement in these activities is
important. In these examples, attachment to other community

members and/or involvement in conventional activities has

...—I
I...)

the affect that parental attachment should have had on the
behavior of the youth. Attachment to other community members
and/or involvement in conventional activities becomes a
stronger determinate for the youth when parental attachment

fails to be the most affective element.

Limitations of the Social Control Theory

Previous research indicates the social control is
limited in the types of delinquent behaviors it can explain
and for whom it can explain delinquent behaviors. Research
shows that the social control model predicts status or less
serious offenses better than more serious types of crimes
(Krohn and Massey, 1980; Agnew, 1985). This is important
because adolescents with ties to conventional others and
institutions may perceive that committing certain types of
delinquent acts will not affect their future whereas, other
forms of delinquency will have more serious consequences
(Friedman and Rosenbaum, 1988).

Age is another variable that limits the applicability
of the social control model. Age is perceived as the
strongest predictor of delinquent activity. Certain
components of the bonding process become less effective as
deterrents against delinquent activity as adolescents get
older. Older adolescents may begin to emphasize more career—

oriented goals (Friedman and Rosenbaum, 1988). Gender may

12

also limit the social control theory’s explanatory power.
Males and females encounter different experiences during
their growth due to unique socialization processes and role
expectations.

The social control theory may also be limited in
explaining delinquency depending on the initiation,
intensity, and cessation of delinquency (Akers, 1991).
Consequently, there may be differences in the factors that
determine a) whether an adolescent starts offending, b) the
rate at which an adolescent offends, and c) whether an
adolescent stops offending (Akers, 1991). Paternoster and
Triplett (1988) argue that social control theory is better
able to explain the initiation of offending than the
intensity of offending. That is, the social control theory
seems to explain why youth may begin to smoke but it does
not explain why one youth may smoke very little or a lot.
Research findings on these findings however are mixed
(Agnew, 1987; Agnew, 1991; Paternoster and Triplett, 1988;
Skinner, Massey, Krohn, and Lauer, 1985; Krohn, Skinner,
Massey, and Akers, 1985)

In short, previous research has offered support for the
social control model as an explanation for delinquent
behaviors. However, there is also evidence to support the
idea that the theory may need further specification of the

conditions under which the theory holds. One such condition

H
LA)

(in addition to the ones already mentioned) is the efficacy
of the social control theory to explain the delinquency of

youth of color.

The Social Control Theory and Race

Race, ethnicity, and crime are topics that are often
debated in the public/political arena (Junger and Marshall,
1997). Unresolved questions concerning the race/ethnicity
and crime link continue to capture the attention of
criminological studies in the United States (e.g., Tonry,
1995). The search for a theoretical framework that can
accommodate race/ethnicity as a variable continues (e.g.,
Georges—Abeyie, 1989; Hawkins, 1986). A more common approach
is to just include race/ethnicity as one of many variables
in a model that may be related to criminal behavior. Some
theorists argue that there should be a theory specifically
designed to explain the criminal behavior for people of
color (Harris and Meidlinger, 1995; Mann, 1993). This is
consistent with literature that stresses the importance of
understanding how specific ethnic groups’ sociocultural
situations influence their adolescents' delinquent
experiences (Rodriguez and Weisburd, 1991). The opposing
perspective, however, suggests that concepts derived from
mainstream criminological theories should be able to account

for delinquency among all people of color (e.g., Sampson and

Laub, 1993). Hirschi (1969) concluded in his findings on the
social control theory that the elements of the bond are
found in every section of society and therefore social
control theory can help explain crime across all socio—
economic lines including race/ethnicity. There is however,
evidence to suggest the social control theory does not
explain delinquency, comparably, in all sections of society,
particularly for youth of color (Gardner and Shoemaker,
1989).

Gardner and Shoemaker (1989) concluded that race
explains as much of the delinquency as does community
context. The study by Gardner and Shoemaker (1989) is a
multivariate analysis of social bonding theory of
delinquency, which includes an analysis of urban and rural
locations. The sample consisted of 733 eighth and twelfth
grade students from Virginia. Resource limitations
necessitated the use of a convenience sampling method in the
selection of schools and students. Subjects in the study
ranged in age from 13—20 and were evenly distributed in
categories on gender and race (White and Black). The sample
was divided into two categories of urban and rural. A total
of 277 youth were classified as urban and 456 were
considered from a rural area. Most of the urban sample
consisted of African Americans and the rural consisted

mostly of whites.

 

The results of the multivariate analysis indicate there
was very little support of the social bond model from the
urban sample as compared to the rural sample. The amount of
explained variance was much lower for the urban sample than
the rural sample where there were several significant
associations with delinquency and the amount of explained
variance was very high (Gardner and Shoemaker, 1989).

This study provides helpful information on the ability
of the social control theory to explain the delinquent
behaviors of African Americans and Whites. This study
however does have limitations. The sample drawn for the
study was not randomly selected and did not contain national
representation of subjects, as the subjects were all from
Virginia. In addition, Hispanics were not selected as a
separate ethnic group for the analyses.

The research conducted by Liska and Reed (1985)
examines the reciprocal effects of the social bond elements
on delinquency. Specifically, this study examines the
effects of the social bonds on delinquency and the effects
of delinquency on the social bonds. The data used for this
study was from the Youth in Transition Study, which is a
four-wave, multistage, national probability sample of boys
from 87 high schools. The researchers used the first and

second wave of the data to conduct their analyses. The

}_1
ON

 

sample was divided into four sub-samples of social class and
race.

Their findings suggest that the causal structure is not
contingent on class and pattern of delinquency, but on race.
The analysis suggests there is a reciprocal effect of the
social control variables. Specifically the findings suggest
that for whites parental attachment affects delinquency,
delinquency affects school attachment, and school attachment
affects parental attachment. The reciprocal effects were
different however for African Americans in the analysis.
Specifically, two compelling differences for African
Americans were, school attachment is affected by parental
attachment, whereas for whites parental attachment is
affected by school attachment. Moreover, for African
Americans, school attachment affects delinquency, whereas,
for Whites, delinquency affects school attachment. The
researchers concluded that the social bond model could not
ignore social categories such as race because differences
that were race-based were found.

The findings of Liska and Reed's (1985) suggest once
again that the use of the social control model to explain
delinquency does have different effects for different race

roups. The study was limited however in that not all of the
elements of the social control model were examined for their

reciprocal effects on delinquency. In addition, the

17

researchers limited their study of race effects to African
American males and White males excluding the race effects of
other race/ethnic groups.

Cerkonvich and Giordano (1992) researched how different
levels of attachment, commitment, involvement and belief in
school activities may vary across race. They hypothesized
that African American youth would be less attached to school
than white youth because being involved in school would be
seen among African American peers as going along with the
establishment or in essence “acting white” (Cerkonvich and
Giordano, 1992).

The researchers obtained a cross-section of youth
between 12 and 19 years of age in the Toledo, Ohio
metropolitan area. They achieved their objective of sampling
equal numbers of males, females, African Americans and
Whites by using a multistage, modified probability sampling
design in which geographically defined area segments were
selected with known probability. A total of 942 face-to—face
interviews were successfully completed. Delinquency was
measured by a modified version of Elliott and Ageton’s
(1980) self-report delinquency scale. Seven dimensions of
school attachment were used in the model. The seven
dimensions are: a) school attachment; b) attachment to

teachers; c) school commitment; d) perceived risk of arrest;

18

e) school involvement; f) parental communication; and, g)
perceived opportunity.

One-way analysis of variance was used to analyze the
data, as the researcher was particularly interested with in-
group effects. The results indicate there were no race-based
differences but the amount of explained variance was lower
for African—Americans than for Whites. The researchers
concluded that although the racial differences were not
found for school attachment and delinquency, the way school
attachment variables were conceptualize may have influenced
the outcomes of their research.

The researcher's conclusions are however contrary to
what they present in their article. Cernkovich and Giordana
conclude school bonding plays the same role in the
delinquency involvement of African Americans and whites.
However, an alternative explanation of the findings suggests
that African Americans are more strongly attached to school
than whites. Moreover, the model explained more variance for
African American males (R?=.154) than for white males
“8:.133). In addition, school attachment, school
commitment, and school involvement were all significant
predictors for delinquency for African American males;
whereas, only school commitment was a significant predictor
of delinquency for white males. These findings suggests that

the way in which the social control model is conceptualized

 

may influence the ability of the model to explain the
behavior of youth of color, in this case African American
youth. Limitations of this research are the fact that the
researcher’s focused only on school-related variables. In
addition, this study did not include a national sample, and
again the research focused on the Black/white dichotomy.

Research conducted by Weber, Miracle and Skehan (1995)
examined the predictive power of the social control theory
for African Americans, Hispanics and Whites. The research
specifically focused on the elements of social bond in the
context of family bonding among African American, Hispanic
and White adolescents. The sample consisted of sixth and
eighth graders who attended school in a single attendance
zone at three regional diverse urban sites representing the
American southwest, Midwest, and southeast. The researchers
administered the survey in a structured interview setting
where the interviewer read the questions out loud and the
youth circled answers on a separate sheet of paper (Weber,
et a1., 1995).

The delinquency measures were derived from self-report
data. The attachment bond was a measure of three components
1) affection for and sensitivity towards others; 2) intimacy
of communication (i.e., the extent to which a youth shares
his or her mental life with significant others); and, 3) the

respect, love, or affection the individual has for

significant others (Weber et a1., 1995). The commitment bond
was measured in terms of the strength of an emotional
response to an identity, wh ch should be reflected in actual
behavior. Involvement was measured in terms of the time
spent in conventional activities (Weber et a1., 1995).
Finally, belief was measured in terms of the extent to which
one feels obliged to conform to the rules of society, is
made possible by the absence of beliefs that prohibit
delinquency (Weber et a1., 1995).

The data analysis consisted of the researchers
investigating the reliability and validity of the social
bonding constructs. Next, they investigated the relationship
between the independent and dependent variables in the
analysis. Finally, the relationship between the social
bonding constructs and delinquency was analyzed and compared
across racial/ethnic groups. The reliability analysis
supported the use of the social bond constructs in racially
or ethnically diverse groups. The validity analysis provided
limited support for Hirschi’s (1969) theoretical proposition
that the elements of the social bonds are distinct
components. Other findings from this study indicate that
researchers should be cautious in the utilization of the
social bond elements in diverse populations. The reason
being, measures that may be valid indicators of the bonding

components for one group may be an indicator of another

I\)
...:

social bond component in another group (Weber, et a1.,
1995).

Weber et a1., (1995) found that family pride exerts
influence over delinquency for Whites and Hispanics.
Conversely, caretaker supervision exerts influence over
delinquency for Whites and African Americans. In addition,
female caretaker communication exerts influence over
delinquency for Hispanics (Weber et a1., 1995). Family
identity had an influence over delinquency for Whites and
Hispanics. The influence of the female caretaker varied
across all three groups, but female supervision appears to
have more influence over delinquency for Whites and African
Americans. However, female communication has an influence
over delinquency for Hispanics (Weber et a1., 1995).

The aforementioned research is important because the
researchers analyzed the independence of the control
components by using factor analysis. The loadings were the
same for the male attachment indicators across all racial
groups. However, loadings were different for the female
attachment indicators. The data indicated that female
supervision indicators loaded on a discrete factor for
Whites, however, other indicators were included for Blacks
and Hispanics. For African Americans the female supervision
factor also included the youths’ acknowledgement that one

should follow the rules of the family. For the Hispanics

A
,22
£-

female supervision was related to the number of school—week
afternoons and weekend—time spent with families (Weber et
a1., 1995).

Weber et a1., (1995) also found that the family
commitment variables loaded differently for each of the sub-
samples. This difference in loading could have been a
difference in culture. Family involvement also was different
for the racial groups. Time spent with family seem to have
relevance for African Americans and white. Time spent with
family is associated with female supervision for Hispanics.
The belief component was only relevant for Hispanics. The
indicators for belief did not load high on any factor for
the White sub-sample and for the African American sub-sample
it was associated with female supervision. The researchers
then created social bonding indices from indicators that
would be valid across all racial categories. The overall
results from the social bond elements indicate that age and
sex account for 28.4% of the variance in delinquency for
Whites, 22.2% for African Americans and 26.5% for Hispanics
(Weber et a1., 1995).

The researchers concluded that the reliability analysis
of the social bond constructs support the use of these
constructs and the social control theory to explain
delinquency in ethnically or racially diverse groups. The

reliabilities ranged from .68 to .95 (Weber et a1., 1995).

i\)
O)

However, the reliabilities were consistently higher for
Whites than for Blacks and Hispanics (Weber et a1., 1995).
In addition, the amount of explained variance for all three
groups was low.

Rodriguez and Weisburd (1991) tested social control
theory on a sample of Puerto Rican males. The researchers
hypothesized that the Puerto Rican family would have a
strong influence over delinquency. Therefore, peer
involvement was expected to be less important and the strong
influence of the family was expected to be at the expense of
other involvements (i.e., school, church, activities, etc.).
The data for this research were gathered from a longitudinal
survey of male Puerto Rican adolescents residing in South
Bronx, New York. The survey, which was administered as a
two-wave panel, examined a representative sample of the
South Bronx’s Puerto Rican male adolescent population, aged
12 to 19 years. The researchers compared their findings with
the National Youth Survey analyses.

The social bond measures included involvement with
delinquent peers, attitudes towards delinquency, family and
school normlessness, family and school involvement, and home
and school strain. Involvement with delinquent peers was
measured in terms of the amount of time spent with peers and
how many of the respondents’ friends had been involved in

delinquent acts during the past year (Rodriguez and

Weisburd, 1991). Attitudes toward delinquency were measured
in terms of how wrong six delinquent acts were (Rodriguez
and Weisburd, 1991). High scores reflect a conventional
orientation.

Family and school normlessness was measured in terms of
the extent the subject views his relationship to his family
and school as governed by conventional norms or as requiring
a transgression of these norms (Rodriguez and Weisburd,
1991). Family and school involvement were measured in terms
of the amount of time spent with the family and in academic
activities at school (Rodriguez and Weisburd, 1991). Home
and school strain were measured in terms of reported
discrepancy between aspirations and expectations in each of
five aspects of family and school life (Rodriguez and
Weisburd, 1991).

The results from this study indicate family involvement
was significant for the Puerto Rican youth but it was not
significant for the National Youth Survey. Also, peer
involvement was almost half as important in the Puerto Rican
sample as it was in the national sample. The finding of the
importance of family was predicted as well as the lack of
peer influence.

Rodriguez and Weisburd (1991) concluded that the social
bond model was applicable to Puerto Rican American males.

However, they found that school variables did not have a

direct effect on delinquency for Puerto Ricans as it did for
their comparison group. This research best illustrates how
the elements of the social bond model will vary in
significance among different racial and ethnic groups --
this was also demonstrated in the research conducted by
Weber et a1. (1995).

The research by Junger and Marshall (1997) tested the
social bond model on a population of juveniles in the
Netherlands. The study included individuals who were Turks,
Moroccans, Surinamese and Dutch. There were four hypotheses
tested: a) the social control theory predicts variations in
general delinquent involvement among varying ethnic male
youth; b) the social control theory predicts variations in a
variety of types of delinquency involvement and deviance
among varying ethnic male youth; c) Delinquent friends play
the same role in the causation of general delinquency among
varying ethnic male youth; and, d) The dimensions of the
social bonds are interrelated in the same manner among all
four ethnic groups (Junger and Marshall, 1997).

The data was obtained in a stratified random sample of
Moroccans, Turks, Surinamese, and Dutch boys. A total of
1,231 potential respondents were identified and 814
completed surveys. The instruments were conducted by a

professional survey agency and the respondents received

I\)
O\

about five dollars (American) to complete the survey (Junger
and Marshall, 1997).

Multiple regression analysis was conducted to test the
first three hypotheses; the fourth hypothesis was tested
using LISREL analysis. The amount of explained variance is
consistently higher for the Dutch population than for the
other ethnic/racial groups except in the case of the family
bond. The amount of explained variance for the family bond
is higher for Moroccans than for all other ethnic/racial
groups. However, the overall analysis of the combined social
control index indicates that social control helps explain
delinquent acts better for Dutch youth. Junger and Marshall
(1997) conclude that there was high compatibility of the
social bond elements and the entire model for each of the
ethnic/racial groups.

The study by Junger and Marshall (1997) is further
evidence that when the elements of the social control model
are operationalized in a manner that reflects the values of
the majority group (in this case the Dutch), the model works
best to explain the behavior of the majority and less for
all other groups who are represented in that society.

The literature presents evidence that the social
control theory works less effectively to explain crime for
people of color than the majority group. Researchers for the

most part concluded that social bond was a viable theory to

help explain delinquency in racially different populations
even though in all cases the amount of explained variance
was consistently higher for the majority population. In
addition, the research by Weber et a1., (1995) and Rodriguez
and Weisburd (1991) illustrate the point that although
bonds exist for all groups in society all groups bonds’ are
not the same. In the Weber et a1., (1995) research, the
results indicated that female supervision was associated
with different indicators for all racial groups. In
addition, the other indicators for attachment, commitment,
and involvement showed differences as well. The research by
Rodriguez and Weisburd (1991) also illustrated the
differences in the bond elements among Puerto Rican males.
Traditional Conceptualization of Social Control Elements

This section is a discussion on how the social control
model has been traditionally specified and the effects this
has on the explanatory power of the social control model for
people of color. This discussion is important because it
does not dismiss the use of the social control model for
people of color; however, it does highlight the fact that
the traditional way in which the social control model has
been conceptualized does not work for people of color. The
research presented, indicates the number of variables used
in a social control model will vary across research

depending on the level of specification a researcher chose.

 

The variables common in the research include attachment to
parents, peers and teachers, involvement in traditional
activities, commitment to traditional ideas, and beliefs.
Only two of these variables are used for the following
discussion. The two variables that will be highlighted for
discussion are attachment to parents and involvement in
traditional activities.

In the research by Gardner and Shoemaker (1989) there
were eleven variables that represented the social control
elements: 1) involvement in school; 2) involvement in
community; 3) attachment to parents; 4) attachment to peers;
5) attachment to teachers; 6) attachment to church; 7)
commitment to church; 8) commitment to school; 9)
conventionality to parents; 10) conventionality to peers;
and 11) beliefs. The following discussion will involve the
variables “attachment to parents" and “involvement in
school.”

Involvement in school was conceptualized as involvement
in school-related activities and the amount of time spent in
the school related activities (Gardner and Shoemaker, 1989).
Involvement in school related activities was conceptualized
as attendance at meetings involving the Scouts, 4—H, YMCA,
YWCA, or other youth groups and how often was the respondent
involved in religious service . The attachment to parents

was conceptualized in terms of the communication level of

 

the respondent and their parents (Gardner and Shoemaker,
1989). Meaning, do parents know where the respondent is when
s/he is not at home, whether the respondent feels
comfortable talking to their parents about goals, as well as
the respondent’s general thoughts and feelings about his/her
parents.

The findings indicate that attachment to parents was
significant for African Americans and not for Whites. Also,
the school involvement variable was significant for Whites
and not for African Americans.

Liska and Reed (1985) sought to understand the causal
path of delinquency using the social control model. The
researchers incorporated the elements of attachment to
parents and school for their analysis. Keeping with
traditional conceptualization of parental attachment, Liska
and Reed’s parental attachment variable emphasized parent
adolescent communication and affectivity. Parent adolescent
communication was conceptualized with questions such as “How
often do your parents listen to your side of the arguments?”
and affectivity was conceptualized with questions such as
“How often do your parents act as if they don’t care about
you anymore?"

School attachment was conceptualized as immediate
satisfaction with the school experience and the belief in

the long-term value of school (Liska and Reed, 1985). The

30

 

element of immediate satisfaction was conceptualized with
questions such as, “School is satisfying to me because it
gives me a sense of accomplishment.” The element of “belief
in the long-term value of school" was conceptualized with
questions such as, “School is boring to me.”

The results of structural equation modeling indicate
for Whites, three major causal effects underlie the
relationships between delinquency, school attachment, and
parental attachment. The findings indicate parental
attachment affects delinquency, which affects school
attachment, which in turn affects parental attachment. For
African Americans the causal structure is different. For
African Americans delinquency is affected by both school
attachment and parental attachment, in addition, parental
attachment affects school attachment, which can also affect
delinquency. The causal models for whites show a circular
one-way path to delinquency where as, the causal model for
African Americans does not. The African American causal
model of delinquency indicate that parental attachment
directly affects delinquency and it is also an intervening
variable that affects school attachment, which in turn,
affects delinquency for African Americans. In addition, the
amount of explained variance was consistently higher for
Whites than African Americans in each of the models tested

in this research.

 

The research of Cernkovich and Giordano (1992) assessed
what roles school bonding and race play in delinquency.
Parental communication and school involvement were two of
the social control variables the researchers used in their
research. Parental communication was a measure of the level
of parental interest in, and support for, school related
activities. Parental communication was conceptualized in
terms of questions that asked frequency of communication
between a parent and adolescent on school related problems.
School involvement was conceptualized in terms of the amount
of behavioral participation in various school activities. In
other words, how much time the adolescent spend
participating in a particular activity during the week.

The results of this research indicate that African
Americans are bonded to school in much the same way as
whites. The model explained more of the variance for Whites
at mixed schools (R? = .156) then African Americans at mixed
schools (R? = .094).

These three studies illustrate how the
conceptualization of the social control variables have
varied and also how they impact the way the social control
model can explain the delinquent behaviors of youth of
color.

Overall, the literature review examines the efficacy of

the social control model to explain delinquency in general

 

and more importantly and specifically for youth of color.
Although there is a plethora of research using the social
control model there is no definitive research that examines
the usefulness of the social control model for youth of
color using national data. In addition, there is no clear
understanding of the usefulness of the social control model
for other race/ethnic groups (e.g., Hispanics) to explain
their delinquent behavior. The following section outlines
the research on alcohol drinking and cigarette smoking

behaviors of youth.

Cigarette Smoking and Alcohol Drinking: A Review of the
Literature
Nationally, 82% of high school seniors indicate they

have drank alcohol at least once in their lifetime and
approximately 66% indicate they have tried cigarettes at
least once in their lifetime (U.S. Department of Health and
Human Services, 1997). National data also indicates about 1
million youth start smoking each year (Institute of
Medicine, 1994). Data from the Bureau of Labor Statistics in
1999 indicate that 55% of male youth and 47% of female youth
indicated they had smoked cigarettes within the 30 days
prior to be interviewed. Specifically, the data indicate

that 22% of white male youth, 14% of African American male

youth, and 19% of Hispanic male youth indicated they had

33

smoked within the 30 days prior to being interviewed.
Similarly, 23% of white female youth, 9% of African American
female youth and 15% of Hispanic female youth indicated they
had smoked cigarettes during the 30 days prior to being
interviewed.

Additionally, data from the Bureau of Labor Statistics
(1999) also indicate that 58% of male youth and 56% of
female youth indicated that they had drank alcohol during
the 30 days prior to be interviewed. Specifically, 23% of
white males, 13% of African American males, and 22% of
Hispanic males indicated they had drank during the 30 days
prior to being interviewed. Similarly, 23% of white females,
9% of African American females, and 15% of Hispanic females
indicated they had drank alcohol during the 30 days prior to
being interviewed.

The statistics presented illustrate the extent of the
problem of drinking alcohol and cigarette smoking for all
adolescents. The following section is a review of the
literature on adolescent cigarette smoking and alcohol
drinking. The literature focuses on the factors or
correlates (i.e., family and peer associations) that predict
which adolescents will turn to smoking cigarettes and
drinking alcohol. In addition, some of the research also
focuses on the racial/ethnic differences in those predictors

of smoking and drinking behavior of adolescents.

34

 

Factors that affect Cigarette Smoking and Drinking Alcohol
for Adolescents

The research of Foshee and Bauman (1992), examines the
smoking behavior of adolescents in the context of social
control and social learning theories. Specifically, Foshee
and Bauman’s (1992) research has to do with how parental and
peer behaviors and attitudes modify bond/behavior
relationships. Foshee and Bauman sought to expand the social
control theory by including elements of social learning
theory. They suggest that not only the bond elements should
be considered but that parental and peer behaviors and
attitudes should be considered when assessing the cigarette
smoking behaviors of youth.

Foshee and Bauman (1992) formulated two models.The
first model examines the interaction of weak bonds and a
social environment conducive to smoking. Therefore, youth
who are susceptible to initiation are the ones with weak
bonds to society and one with parental or peer smoking
models present in their environment. The second interaction
model looks at the interaction of the social control theory
and social learning theory and suggests that for adolescents
with parents who do not smoke, the likelihood of smoking
decreases as the social bonds increases.

The data for this study was gathered ticm ten

metropolitan statistical areas throughout the southeast

35

 

United States in two waves. The sample was attained from the
first wave by those respondents who indicated they had not
smoked at all prior to being interviewed and who also
indicated they lived with both parents/guardians.

The results indicate that for adolescents with parents
who smoke, smoking increases as attachment to that parent
increases. Similarly, adolescents with parents who do not
smoke, the adolescents’ smoking behavior decreases as
attachment to the parent increases. Overall, the results
indicate cigarette smoking might be a learned behavior that
is mediated by an adolescent’s attachment to his/her
parents.

The research by Ennett and Bauman (1993) examined how
peer group structures effects adolescent cigarette smoking.
Prior research indicates there is compelling evidence that
the peer group plays an important role in the initiation and
maintenance of smoking (Oetting and Beauvais, 1986;
Friedman, Lichtenstein, and Biglan, 1985; Bauman et a1.,
.1984; McAlister, Krosnick, and Milburn, 1984; Huba and
Bentler, 1980; Levitt and Edwards, 1970). The purpose of
Ennett and Bauman’s (1993) research was to determine whether
memberships in particular peer group structures i.e., clique
members, liaisons, and/or isolates vary the prevalence of

cigarette smoking.

LA)
0\

 

Ennett and Bauman (1993) recognize three peer group
structures that are the “regularized patterns of
interactions among adolescents in a social system” (Ennett
and Bauman, 1993, p.227). The three peer group structures
are: cliques, liaisons, and isolates. Cliques are clusters
of adolescents who spend more time with each other than with
other adolescents, and they also share the same attitudes
and beliefs (Ennett and Bauman, 1993). Liaisons interact
extensively with peers but not as members of cliques.
Liaisons tend to have friendships with members of different
cliques (Ennett and Bauman, 1993). Isolates have very little
or no interaction with peers (Ennett and Bauman, 1993).

The data were obtained from a county school system in
North Carolina. Friendship networks were defined based on an
adolescent identifying an individual as their friend and the
other friend also identifying that same individual as their
friend. The peer group structures (i.e., cliques, liaisons,
and isolate) were then identified through these
associations. The findings indicate that isolates tended to
smoke more often than clique members or liaisons. In
addition, although isolates have fewer friends overall,
proportionally more of their friends are smokers as compared

to the other two groups. The researchers offer four

(‘1‘

explanations for their findings. The firs explanation is

that social isolation may cause cigarette smoking. The

second explanation is that cigarette smoking causes social
isolation. Third, the relationship between social isolation
and cigarette smoking might be spurious meaning that both
social isolation and cigarette smoking might be caused by
the same factors. The final explanation for the findings is
that isolates are actually members of cliques made up of
adolescents outside the school that are dropouts or truants
who may encourage smoking. Overall, the findings of this
research add to the understanding of how peer group
associations affect the smoking behavior of adolescents.
Unger and Rohrbach (2002) examined how adolescents’
estimates of their peers' smoking prevalence can influence
their own smoking behavior. Specifically, Unger and
Rorhrbach (2002) wanted to understand the correlates of the
prevalence estimates for youth and why adolescents
overestimate their peers’ smoking prevalence. It is believed
that adolescents’ form their estimates of peer smoking
prevalence based on the adolescent’s own smoking behavior
and intentions, inferences based on close friends’ behavior,
inferences based on social comparison with similar others,
accurate observations of the true smoking prevalence in the
school, observations of smoking the media, perceived access
in cigarettes, cigarette offers (meaning when an adolescent

is offered a cigarette by a peer).

 

The data for this study were gathered from a
representative sample of 8Ch graders in California. The
results indicate that higher prevalence estimates or those
variables associated with overestimating the prevalence of
smoking among peers are age, female, non—Pacific Islander
ethnicity, low academic performance, school smoking
prevalence, best friends’ smoking, cigarette offers,
perceived access to cigarettes, and perceptions of smoking
in the media.

The results indicate adolescents’ observations of their
best friends’ and schoolmates’ smoking behavior may account
for a considerable amount of the variance in their peer
smoking prevalence estimates. Unger and Rohrbach (2002)
concluded that prevention and intervention programs should
demonstrate to adolescents that their perceptions of their
peers’ smoking prevalence tend to be inflated.

The research of Alexander, Piazza, Mekos, and Valente
(2001) focused on the importance of school environments for
understanding peer group influences on adolescent cigarette
smoking. Specifically, their research examined the
associations of popularity, having a best friend who smokes,
and the prevalence of smoking within the friendship network
on adolescents’ current smoking behavior (Alexander et a1.,

2002).

39

 

 

The data source for this study was the Longitudinal
Study of Adolescent Health (Add Health). The Add Health data
is a nationally representative sample of all public and
private high schools in the United States. Results indicated
that having best friends who smoked meant that respondents
were twice as likely to smoke as well. School smoking
prevalence was positively associated with being a current
smoker. For every 10% increase in school smoking prevalence,
there was a 73% increase in the likelihood of current
smoking (Alexander et a1., 2002). The results also indicated
that there was a significant risk of being a current smoker
for youth who were more popular and who had a school smoking
prevalence.

Overall, the results indicate popularity has a
differential risk for cigarette smoking. In schools with low
smoking prevalence, higher levels of popularity are
associated with reduce likelihood of smoking (Alexander et
a1., 2002). In contrast, popular students in schools with
high smoking rates are more likely to smoke cigarettes
(Alexander et a1., 2002). The findings from this study add
to the body of research that has found strong correlations
between adolescent and peer smoking.

The research of Ennett, Flewelling, Lindrooth, and
Norton (1997) examined school and neighborhood

characteristics associated with school rates of alcohol,

 

cigarette, and marijuana use. The data source was 36
elementary schools in a midwestern state. The measures for
this study were substance use (lifetime and current); school
characteristics (i.e., substance use norms in the school and
school climate); attitude toward substance use; perceived
acceptability of alcohol use; perceived acceptability of
cigarette use and marijuana use; availability of alcohol;
availability of cigarettes; school attachment; neighborhood
characteristics (climate and socioeconomic characteristics);
neighborhood attachment; neighborhood safety; and,
neighborhood drug activity.

The results indicate that rates of alcohol use are

higher in schools where there is a perception among students
there are higher levels of acceptability of use among peers,
where alcohol is more readily available, and where students
report lower attachment to school (Ennett et a1., 1997).
The results also indicate there are higher rates of alcohol
in schools located in areas where parents report higher
neighborhood attachment and safety and where neighborhood
population density and mobility are lower. These results
contradict the expectations of the research where
neighborhood cohesiveness was thought to be associated with
lower rates of alcohol use (Ennett et a1., 1997).

The findings for the lifetime cigarette use mirrored

the findings on lifetime alcohol use. Rates of lifetime

cigarette use are higher in schools in low-density
neighborhoods and where parents reported high neighborhood
attachment (Ennett et a1., 1997). The results of the current
cigarette use indicate that neighborhood drug activity,
attitude toward drug use, perceived acceptability of
cigarette use, and availability of cigarettes are
significantly associated with school rates of current
cigarette smoking (Ennett et a1., 1997). In addition, school
rates of current cigarette smoking are higher in
neighborhoods that parents described as supporting illegal
drug sales and availability.

The research of Scheer, Borden, and Donnermeyer (2000)
examined the relationship between family factors and
adolescent substance use (i.e., cigarette smoking, alcohol
drinking and marijuana use) in rural, suburban, and urban
settings. Specifically, the researchers wanted to understand
how families might differ in communication, sanctions, and
involvement according to their geographic location.

The source for the data was a stratified random sample
of 11"n grade students in public high schools in a
midwestern state. The measures used included a 34-point drug
involvement scale as well as measures for family
communication, family sanctions, family involvement in

school, and family care. The 34—point drug scale

 

specifically assessed the cigarette smoking, alcohol
drinking and marijuana use of the adolescents.

The results indicate there is no significant difference
between substance involvement and location (Scheer et a1.,
2000). The data also indicate for rural and suburban
adolescent substance abuse, that youth who reported that
their parents talked to them about the dangers of smoking
and drinking alcohol were less involved in substance use
(Scheer et a1., 2000). The family care and family
involvement in school was significant across all geographic
locations. Adolescents whose families were involved in
school events and activities were less likely to be involved
in substance use. In the multivariate analyses, family
involvement in schools, belief that their families cared
about them, and family talking about the dangers of
cigarettes predicted lower substance use in the regression
models.

Overall, the results indicate that location status does
not determine how family factors are related to adolescent
substance use. Specifically, there were no differences found
for location in relationship to substance use, in addition,
location did not significantly influence how the family

variables predicted substance use.

Cigarette Smoking and Alcohol Drinking: Race/Ethnic
Differences

The research of Flay, Hu, Siddiqui, Day, Hedeker,
Petraitis, Richardson, and Sussman (1994), examined the
differential influence of parental smoking and friends’
smoking on adolescent initiation and escalation of smoking.
Research indicates that the influence of parents and peers
over alcohol drinking and cigarette smoking varies depending
on the type of drug. Prior research indicates parents and
peers tend to influence adolescent use of alcohol equally
(Chassin et a1., 1986). However, peer influence tends to be
a better predictor of cigarette smoking (Alexander et a1.,
1983; Chassin et a1., 1986; McCaul et a1., 1982).

Flay et a1., (1994), propose a model of smoking
influence that examines the causal links of factors
associated with adolescent smoking behaviors such as
initiation and escalation (Flay et a1., 1994).

The results indicate that friends smoking had both
direct and indirect effects on smoking initiation. Parental
smoking had only an indirect effect on smoking initiation.
The results also indicate that neither friends’ smoking nor
parental smoking had direct effects on escalation. The
gender results indicate that the effects of parental smoking
are stronger for females than males. Ethnic group

comparisons showed differences in friends’ influences all

44

having to do with smoking initiation. There were no
significant findings on the influence of parents or peers on
smoking escalation (Flay et a1., 1994).

The research of Guthrie, Young, Boyd, and Kintner
(2001), examined the relationship between cigarette use and
daily life hassles among African American females. Previous
research indicates that although African American women
generally initiate smoking at a later age during
adolescence, African American women’s rates of smoking
increase with age. In addition, African American women have
high rates of recidivism for smoking than white women as
well as higher death rates for smoking-related diseases than
most other women.

The data source for the study was the Female Adolescent
Substance Experience Study (FASES). The results indicate
there were no significant differences found between the
group that indicated they had smoked in the past 30 days and
those who indicated they had not smoked in the past 30 days
on items such as age, primary means of support, social
class, religious preference, presence of mother or father in
home.

Further analyses indicate there was significant
difference in the number of daily hassles for girls who
indicated they had never smoked compared to those who

indicated they had ever smoked. Those who indicated they had

 

 

never smoked were significantly lower on daily hassles.
Finally analyses were conducted to determine whether there
was a significant difference in specific daily hassles and
whether a respondent indicated they had never smoked and
those who indicated they had ever smoked. Results indicate
there were significant differences on school/academic
hassles and family/economic hassles, which meant girls who
indicated they had ever smoked, were significantly higher on
those indicators. There were no significant differences in
social/peer and personal safety hassles between the two
groups.

Ellickson and Morton (1999) examined early risk factors
for initiation of hard drug use by 10Ch grade and compared
their results across race/ethnicity. Research indicates that
alcohol drinking and cigarette smoking are considered
gateway drugs. Cigarette smoking, alcohol drinking and
marijuana use are the most commonly used substances that
typically precede the use of hard drugs.

The data source for Ellickson and Morton’s (1999) study
were middle schools in California and Oregon. Results
indicate that major risk factors for initiation of hard drug
use include early marijuana and cigarette use, deviant
behavior, pro-drug attitudes and intentions, being offered
drugs, and poor parent-child communication. Findings

indicate white adolescents had the most risk factors,

followed by Hispanics, Asians, and African Americans. For
all groups except African Americans, early marijuana use was
the strongest predictor of hard drug use followed by early
smoking. Overall, the findings indicate that addressing
early initiation of cigarette and marijuana use might have
an impact on reducing the risk for hard drug use.

Overall findings of the cigarette smoking and drinking
alcohol behavior of adolescents indicate that adolescent
peers have a strong influence on the initiation and overall
use of cigarettes and alcohol by adolescents followed by
parental influence. The research also indicates that there
are differential influences on the smoking and drinking
behavior of adolescents when race/ethnic differences are

considered.

Significance of this Dissertation and Study

There is no definitive evidence on the usefulness of
the social control theory to explain the delinquent
behaviors for youth of color. All research using the social
control theory has been cautious in reporting findings where
race may have significantly contributed to the theory’s
(in)ability to explain the differences found. In addition,
the literature review illustrates the dearth of information
that addresses the relationship between the social bonds,

Hispanic populations and delinquency. Also, much of the

47

 

 

previous literature lacks an examination of the social bonds
using a national data set; instead majority of the research
has been done in specific regions of the country or
localized in a particular city.

Finding out the ways in which the social control theory
can or cannot explain the delinquent behavior of youth of
color, will help policy makers and program planners make
better policy and create better programs that effectively
address the needs of adolescents of color in the criminal
justice system.

Similarly, the literature on the cigarette smoking and
drinking behavior of adolescents illustrate (1) the
prevalence and incidence of the problem; (2) the impact that
families and peers have on the prevalence and incidence of
cigarette smoking and alcohol drinking of adolescents; and
(3) the differential effects of correlates of smoking and
drinking on youth of color. Therefore, the current study is

important in the following aspects:

1. It contributes to the current body of research by
clearly examining the social bond constructs in
relation to African American and Mexican Americans.

2.It makes the findings generalizeable to the
population as a result of the utilization of a
national data set.

3.It applies the principles of social bond theory to
the explanation of the smoking and drinking behavior
for whites, African Americans and Mexican youths in
the United States. The dichotomy of whites and

48

African Americans in the research literature has
always limited the generalizeability of results to
other groups.

Hypothesis

This study will test the following hypothesis,

comparing white, African American and Mexican youth:

Ho:

There is no relationship between attachment (to
parents, teachers, peers), commitment (to school and
furthering education or getting a job), involvement (in
sports related activities and other school-related
activities) and belief (in committing mild and serious
infractions in school)in regards to smoking cigarettes
and drinking alcohol. Moreover, these relationships

will be conditioned by gender, SES, and race.

49

 

CHAPTER III
METHODOLOGY
This section will outline the research strategy used
to answer the research hypothesis stated in Chapter II.
Included in this chapter is a description of the research
design, the sample population, the data collection method,

data analysis, and limitations.

Research Design, Data and Sample

This study uses secondary data from the National
Educational Longitudinal Study 1988-1994 (NELS:88). The
NELS:88 is the first nationally representative longitudinal
study of eighth grade students in public and private
schools. This study is sponsored by the National Center for
Education Statistics in conjunction with the Department of
Education (NCES). The NCES is mandated to “collect and
disseminate statistics and other data related to education
in the United States” and to “conduct and publish reports
on specific analyses of the meaning and significance of
such statistics” (Education Amendments of 1974—Public Law
93-380, Title V, Section 501, amending Part A of the

General Education Provisions Act).

50

The general aim of the NELS program is to
study the educational, vocational, and
personal development of students at various
grade levels, and the personal, familial,
social, institutional, and cultural factors
that may affect that development (NCES,
First Follow—up Final Technical Report,

October 1994).

These data were collected for the purpose of (1)
providing longitudinal data on the critical transitions
experienced by young people as they develop, attend school,
and start their careers; (2) informing decision makers,
practitioners, and parents about the changes in the
operation of the educational system over time, and the
effects of various elements of the system on the lives of

the individuals who pass through it (NCES, October 1994).

Data Collection

The NELS:88 uses a two-stage stratified probability
sampling design that selected a nationally representative
sample of 24,599 students from 1,052 randomly selected
schools. The first stage of the sampling method constituted

schools, which was the primary sampling unit. The target

51

sample for the schools was 1, 032. The original sampling
schema produced 1,032 schools but 30 of the schools were
ineligible for participation. Of the 1,002 eligible to
participate only 698 actually participated. An additional,
359 schools were then selected from a replacement pool
(NCES, October, 1994).

The purpose of the second stage of the data collection
was to obtain the student sample. The sampling schema
produced 26,432 students among participating sampled
schools, resulting in the participation of 24,599 eighth
grade students. On average, 23 student participants
represented each of the participating schools.

For the first follow—up another sampling strategy was
employed for the purpose of “freshening the sample”. The
sample was freshened in order to provide a representative
sample of students enrolled in the tenth grade during the
1989—90 school year. The first stage of sampling involved
the selection of 21,474 students who were in the 8th grade
for the Base Year data collection. For the first follow—up
these students were referred to as core students. The core
student sample was then augmented through freshening, which
added 1,043 eligible tenth graders who were not contained

in the base year sampling frame, either because they were

not in the country, or were not in the eighth grade in the
spring term of 1988 (NCES, October 1994).

All of the samples for NELS:88 contain an over-sample
of Latinos, Asian-Pacific Islanders, and Native American
students, which was sponsored by the 0.8. Department of
Education’s Office of Bilingual Education and Minority
Languages Affairs (OBEMLA). As a result of the over-
sampling, an additional 2,400 students were added to the
study.

Three follow—up surveys since the base year has been
conducted, the first in 1990, when most respondents were in
the tenth grade, the second in 1992, when the majority of
students were in twelfth grade, and the third in 1994. The
NELS:88 also interviewed respondents’ parents in 1988 and
in 1992, two of the respondents’ teachers, as well as a
school administrator. For the purpose of the present study,
the first follow—up is utilized when respondents were in
the tenth grade at the time that they were interviewed.
Research indicates the peak ages for delinquency are
between the ages of 14 and 18. The tenth grade sample is

when majority of the respondents were age 15.

53

Specification of Variables Under Study
Dependent variables

The dependent variables for this study are smoking
cigarettes and drinking alcohol.

Cigarette Smoking: This dependent variable is
conceptualized in terms of whether the respondent
indicated they had smoked cigarettes in the 30
days prior to being interviewed. This is a
dichotomous variable where one (1) is equal to
those respondents who indicated they had smoked
cigarettes during the 30 days prior to being
interviewed and where zero (0) is equal to
those respondents who indicated they had not
smoked cigarettes during the 30 days prior to
being interviewed. This dependent variable is
operationalized in Appendix A.

Ltdnking.Alcohol: This dependent variable is
conceptualized in terms of whether the respondent
indicated they had drank alcohol in the 30
days prior to being interviewed. This is a
dichotomous variable where one (1) is equal to
those respondents who indicated they had drank
alcohol during the 30 days prior to being
interviewed and where zero (0) is equal to those
respondents who indicated they had not
drank alcohol during the 30 days prior to being

(interviewed. This dependent variable is
operationalized in Appendix A.

Independent variables

The independent variables for this study are
attachment (parents, peer, and teachers), involvement (in
sports related activities and involvement in other school

related activities), commitment (to school, to marriage and

success, to furthering education or getting a job) and

54

 

 

belief (it is not okay to commit mild and serious
infractions in school). These variables are conceptualized
and operationalized according to traditional
conceptualizations and operationalizations in the
literature (see Hepburn, 1977; Hindelang, 1973; Johnson,
1979; Krohn and Massey, 1980; Wiatrowski, Griswold and
Roberts, 1981; Agnew, 1985 and 1991; Gardner and Shoemaker,
1989; Covington, 1988; Liska and Reed, 1985; Matseuda and
Heomer, 1987; Friedman and Rosenbaum, 1988; Rodriguez and
Weisburd, 1991; Weber, et. a1., 1995; Cerkonvich and
Giordano, 1992). All of the social bond indicators were
divided into three categories (low, medium, and high), as
indicated below in the “Composite Measures” portion of this

methods section.

.Attachment to Parents: Attachment to parents is
conceptualized as the respondent’s feelings about
his/her parents. The more positive responses
indicate higher attachment to parents. This
variable is operationalized in Appendix A.

Attachment to Peers: Attachment to peers is
conceptualized as how the respondent feels s/he
is thought of by his/her peers. The more positive
responses indicate higher peer attachment. This
variable is operationalized in Appendix A.

.Attachment to Teachers: Attachment to teachers is
conceptualized as how the teacher feels about the
respondent. The more positive responses indicate
higher teacher attachment. This variable is
operationalized in Appendix A.

OW
(11

 

Involvement in Sports-Related.Activities: This
variable is conceptualized in terms of the types
of sports the respondent participated while in
school. The more positive responses indicate
greater participation and therefore greater
involvement in the sports related activities.
This variable is operationalized in Appendix A.

Involvement in other School-Related.Activities: This
variable is conceptualized in terms of the
respondent’s level of participation in school
related activities such as band, orchestra,
plays, and school clubs. Therefore, greater
participation indicated greater involvement in
the other school-related activities. This
variable is operationalized in Appendix A.

Cummitment to School: Commitment to school is
conceptualized in terms of whether the respondent
comes to class prepared. Therefore, more positive
responses indicate higher levels of commitment to
school. This variable is operationalized in
Appendix A.

Cdmmitment to Further Education or Going to Trade
School: This variable is conceptualized in terms
of whether the respondent plans to take college
entrance exams or skill tests. Therefore, more
positive responses indicate higher levels of
commitment to further education. This variable is
operationalized in Appendix A.

Believing it is not okay to COmmit.Mild Infractions in

School: This variable is conceptualized in terms of
moral issues that involves whether the respondent
believes it is okay to skip school, cheat on
tests, be late for class, to copy someone’s
homework, etc. The more positive responses
indicate it is not okay to commit mild
infractions in school. This variable is
operationalized in Appendix A.

Believing it is not okay to COmmit Serious Infractions

in School: This variable is conceptualized in

terms of moral issues that involves whether the

respondent believes it is okay to get into

physical fights at school, belong to gangs, steal

belongings, and destroy school property. The more

positive responses indicate the respondent

believes it is not okay to commit serious

infractions in school. This variable is

operationalized in Appendix A.

The Control Mbasures
Race/Ethnicity
The analyses for this study will include analyses
where white youth are compared to African American youth
and Mexican youth. The Mexican youth variable includes only
third generation Mexican Americans. Third generation
Mexican youth were chosen so that the effects of
acculturation/assimilation would have taken affect by this
point. In addition, the cultural effects of recent Mexican
immigrants would not confound the analyses. Consequently,
any cultural effects that may be detected in the analyses
would not be confounded by recent immigrants in the sample.
Third generation Americans were selected by selecting

the parent of the youth who indicated their parents were
born in the United States. For youth who indicated their
parents were born in the United States, and who also

indicated they were born in the United States, these youth

were labeled as third generation Americans. It was assumed

57

that these youth had grandparents that were either born in
the United States or immigrated to the United States, in
which case, their grandparents were no less than first
generation Americans, which would make their grandchildren
no less than third generation Americans.

The race variable is a nominal level variable that
contains all groups (i.e., whites, African Americans, and
Mexican Americans).

Socio-Econcmic Status Indicators

There are two socio-economic status (SES) indicators
utilized in the data analyses. The first SES indicator used
was the SES Quartile. This variable was constructed by the
National Center for Education Statistics (NCES) by using
available parent data. The variables used to construct the
SES quartile were the father's education level, mother's
education level, father's occupation, mother's occupation,
and family income (BYP30, BYP31, FYP34B, BYP37B and BYP80).
Occupational data were recoded using the Duncan SEI scale.
Parent data were used to construct FlsES if at least one
component was not missing (NELS:88).

The second social economic indicator used in the data
analyses was parent’s education level. This variable is an
indicator of the highest level of education attained by

either of the parents of the respondent. It was

constructed using the parent questionnaire data (BYP30 and
BYP31). Base year student data (BYS34A & BYS34B) were used
for Base Year respondents whenever parent data were
missing. For Base Year non-respondents with missing parent
data and First follow—up freshened students, the New
Student Supplement questions F1N20A and F1N20B were used.
That is, the First follow-up composite starts with the base
year parent education variable. If the base year parent
education variable was missing or the case is a freshened
student, the follow up New Student Supplement data were
used. These measures were taken to reduce the amount of

potential missing data (NELS:88).

DATA CLEANUP

Afiseing'Data

There were two types of missing data for this data
set. The first type of missing data was the result of
matriculation problems from the first round data collection
to the second round data collection. There were 1,153
missing cases that did not matriculate from the first round
data collection to the second round data collection. The
1,153 cases were selected and demographic data (race and

sex) were used to compare the missing data from the group

59

that did not matriculate to the first follow—up. Table A in
the Appendix displays the data on missing respondents.

The missing data indicate that African Americans and
Mexican Americans are overrepresented in the missing
respondents from the base year sample to the first follow-
up sample. In addition, whites are overrepresented in the
not missing category. Also, males are slightly
overrepresented in the missing category whereas females are
underrepresented in the missing category. Although there
are discrepancies in the missing data as compared to the
first follow—up. The percentages, for the most part, are
representative of the census estimates for the year 1990
when this data were gathered (see Appendix B). This missing
data was dropped from the sample, as no other data was
available for these respondents beyond demographic data.

The second type of missing data were data missing
within the variables used for data analyses. Potential
biases in missing data were explored. For each variable, a
dummy variable was created that scored 1 if data were
missing, and zero if not. This dummy variable was then
correlated with race and sex as covariates to assess bias.
No significant correlations were detected, suggesting that
the data were missing at random rather than representing

systematic biases.

As for the range of missing data, the minimum number
of missing cases for a single variable was 73 (0.5%), and
the maximum number of missing data for a single variable
was 1182 (8.6%). This missing data was given the value of
the mean category so as not to lose any data in the

analyses.

Composite Measures

Composite measures were created for the social control
indicators. These indicators were created from a series of
steps, which included factor analyses, reliability analyses
and finally computing the measure. Factor analyses were
conducted in order to determine the way in which variables
grouped themselves statistically and to ensure the
indicators for each of the social control elements were
unidimensional. A factor analysis is a multivariate
analysis used to discover the patterns among the variations
in values of several variables. These patterns are
discovered through the computer generation of factors that
correlate highly with other variables and that are
independent of one another.

A confirmatory factor analyses was performed to create
the social bond measures, as many of the social bond

indicators were computed from scales and indices that were

61

used in the original data collection. Once a set of
indicators were determine to represent a component of
social control then the indicators were placed in a
reliability analysis in order to get a reliability score.
All reliabilities for the composite measures were at or
above 73% (the reliability scores for each composite
measure is located beside each composite measure in
Appendix A).

The composite measures were reduced from interval
level data to ordinal level data to indicate high, medium
and low attachments, commitments, involvements, and
beliefs. The categories of high, medium and low for all the
social bond measures were determined from the data. Once
each of the elements was computed the higher numbers
indicated more positive responses or higher attachment,
commitment, involvement, and belief. Lower numbers
indicated more negative responses or low attachment,
commitment, involvement, belief. Respondents who clustered
around the middle response range were placed in the medium
category. For example, if 8 items represented a computed
variable and each item contained 5 response categories from
0 to 1, once the variable was computed the range of the
variable was from 8 to 40. Individuals who fell into the 8

category were clearly individuals who had low attachment,

62

 

commitment, involvement, and belief. Similarly, individuals
who were in the 40 category were clearly individuals who
had high attachment, commitment, involvement, and belief.
After determining this, the next step was to get the middle
category and disperse the other respondents in a category
of low, medium, or high. This was done by taking the range
and dividing by three. In the above the example the range
is 32 (8 — 40 =32), therefore 32 divided by 3 gives a
category range of approximately 10. Meaning there should be
a range of 10 points included in each category from the
range of the original computed variable. Therefore,
utilizing the above example, the low category will include
individuals who fall in the range of 8-17; the medium
category will include the individuals who fall in the range
of 18-28, and the high category will include the
individuals who fall in the range of 29-40. This method was
checked with the data of previous researchers and was
believed to have face validity. This was surmised, because
a larger proportion of the respondents fell into the medium
categories for all of the computed social bond elements,
whereas the high and low categories had lower proportions
of respondents represented in these categories. The pattern
found in this study was consistent with prior research when

these categories were compared to the frequency of

63

respondents who represented these categories in prior

research.

Data Analysis Strategy
Uhivariate Analyses
Descriptive statistics were used to describe the
sample in terms of the frequency in which respondents
smoked cigarettes and drank alcohol, as well as, the
frequency of the respondents in the categories of the
independent and control variables. The result of this

analysis is summarized in Table 1.

Bivariate Analyses

The bivariate analyses were used to assess the
relationship between the social control indicators and the
dependent variables. In addition, bivariate statistics were
used to assess the relationship between the dependent
variables, social control variables, race/ethnicity,
gender, and income. There were two purposes for subgroup
comparisons, the first was to describe the data while also
including the element of comparison. The second purpose was
to describe the relationships among variables. The measure
of association that was used for the analysis was gamma, as

all the independent variables were ordinal level measures.

64

A Chi-square test was used in conjunction with the
measures of association to test whether the significance of
the relationship between two variables observed in the
sample could be inferred to the population. The Chi—square
is a test of significance that is based on the null
hypothesis that there is no relationship between two
variables in the total population. The measures of
association allows a researcher to test whether there is a
relationship between two variables in the sample and the
Chi-square test allows inferences from the observed
relationship in the sample to the population (Agresti and

Finlay, 1986; Babbie, 1995).

Logistic Regression

Logistic regression was used to analyze the effect of
the independent variables on the dependent variables
smoking cigarettes and drinking alcohol. A logistic
regression analysis allows the researcher to describe how
the proportion of responses in one of the two categories in
a dichotomous dependent variable depends on the independent
variable (Agresti and Finlay, 1986). The equation for

Logistic Regression is:

b%f¥;]=a+fl¥

Where,

z

1. The function bng——] is called the logistic
—z

transformation or logit.

2. As flincreases from 0 to 1, the logit increases
from —oo to co.

3.The probability 3 = H corresponds to a logit
of 0, and fl-values above 5 correspond to
positive logits. H-values below H correspond
to negative logits.

Several logistic regression analyses were conducted
because of the nature of the question under review. There
are two dependent variables and analyses were conducted for
each of the race groups for each of the independent
variables. Therefore there were 8 logistic regression
analyses conducted in total. The results of the analyses

are in Tables 14-21 and the discussion of these findings

are in the next chapter.

Strengths and Limitations of the Study

Using the NELS:88 data set has some limitations.
First, the analysis of secondary data always carries with
it inherent limitations. The data collected, at best
approximates the kind of data a researcher would prefer for

testing a hypothesis, frequently affecting the study

66

design, question wording and sequence, and details of the
interviews (Nachmias and Nachmias, 1994).

Second, there is a slight bias as a result of the
missing respondents as a result of some respondents not
matriculating from the base year data collection to the
first follow-up. As mentioned however, the sample that
remains does approximate the proportions for the categories
of race and gender from the 1990 census.

Third, there are measure limitations, for example, the
data did not contain a variable to clearly indicate the
class of the respondents. As a result, the variable used as
a proxy for class represents the education level of the
respondent’s parents as well as a SES quartile variable.

This section outlined the methods that will be
employed to test the efficacy of the social control model
to explain the smoking and drinking behaviors of youth of
color. This methodology was guided by the plethora of
research on the social control theory. This methodology
included a plan to situate race/ethnicity as key
independent variables in order to test the efficacy of the
social control model to explain the smoking and drinking

behavior of youth of color.

CHAPTER IV
FINDINGS AND DISCUSSION

FINDINGS

Univariate Analysis

The study sample comprised an estimated 11,038
respondents, who indicated they were in the tenth grade at
the time the first follow—up to the NELS:88 data was
collected. Table 1 represents the frequency distribution of
the dependent variables, the control variables, and the
social control indicators. The data indicate that less than
20% of the respondents indicated they had smoked cigarettes
30 days prior to being interviewed. However, more
respondents indicated they drank alcohol in the past 30
days (42.4%). The sample comprised of 51.4% females and
48.6% males. Indicating there are more females in the
sample than males. There are more whites in the sample
(78.4%) than African Americans (12.0%) and Mexican
Americans (9.6%) combined. Consequently, the smoking and
drinking behavior of white respondents drive the results
for the analyses conducted using the entire sample.

There are approximately the same percentages for the
income quartiles. Because the variable is a measurement of
quartiles, it is expected that equal proportions would be

in each of the quartiles. Parent’s highest level of

 

no.4
we.a
em.en
mm.ue
we.um
WS.S
we.em
wm.mm
wi.fim
we.em
wo.a
wo.wi
ne.as
Ii in
we.me
we.sn
we.me
wa.nm
flu ea
w

mow

5H9
more
emnv
Harm
mace

Ovem
Nemm
acSN
mmem

case

vaa
mmom

ween
Neva

wave
mace

meow
Name

>Oconvwwm

nocbo no .2.: ..a.;a
bcoep>esdm no m.uobmnz
oumzpoum vomeaou

ommaeou oEom

onpspmuo Hoorom nae:
Hoosom can: :mHCHw Do: pee

Arawzv v meeuumdo
m mewnung
N umLDGMDO

AZOJV fl meebumzo

:puwxuz
:cdflumsm :noeuu<

DBHLE

mmenzmm

mmenz

OZ
mm»

02
hwy

.An.u:ooc H madmw.

H0>0A coeumosfim ummfimex m.ucmumm

OEOUCH

hueOHG£uH\momm

HQDCOO

mflflmflHm¢> AOMBZOU

Honooem Scexawun

nmuumhmmwb uanOEm
mmAm<Hm4> Bzmazmmma

 

COHUMOSUE n.u:oumm

Ammo .Hauzc nuoumonocH econ Hmaoom 6cm
.OEOOCH .%UHoacnum\oomm .HQDGOO .anDMHHM> uCGUcOQOQ mo mafiocmsvmum .H Genny

 

wm.a
w~.flfi
wo.>m

wm.>o
wm.am
wc.N

wc.m
wm.flvv
wr.am

wm.mm
wa.m~
mo.m

wn.NN
wc.so
wn.oe

wa.ma
ww.mc
we.om

MV.N
wm.Nm
w>.v

«Na
mesa
Germ

NHVS
SSNm
5mm

can
cram
amen

room
were
m\m

33mm
Cams
once

Lava
mmsw
meam

New
memo
wsv

hocoavmwh

bcmEo>ao>cfl roe:
Dcqu>Ho>cL EzepoE

quuE®>HO>Cfl 30H

.D:mcnee:22oc F_$H__
DcmEDeEEOO Esepoi

u ENDS—u ACE—C»). 30. H

ocmieeEECO 23L:
ecmEDeEEOC Ezepvi
DcmEDHEEOo 33L

Dcvsuwzasfiu five:
ecmEDHEEo: Edflme
DcmEDHEEOO 33H

D:$::ramuem Azwer
D:®::7omnnm ::iepm:_

Dszzumem 30a

bch;omDon 50H:
ecoszcmbum Edepoi
ecuE:omDDn 30H

b:oE:cmubm roe:
Dcmszoouum Edwme
bcmszcmhem 30H

I III]! l iIil II

.lo.u:oov H deems

muwomm Hooeom Ge ucmfim>ao>cH

mm0003m DEM mmmHuumE Ou DCOEUHEEOO

Hoonon mnmwu ou oceom
no GOHHMOSUO OGHDCAUCOO ou ucmfiuafiEOO

fill

Hooson ou ucofiuHEEOO

mumgommu on ucmfifiomuud

mwoom ou ucmfinomuud

mucmumm on ucofifiomuud
mMOB‘UHQZH 920m Q<HOOm

 

wo.m

wv.o

wa.om

wm.c
wa.mm
we.va

mecca

emm

sv

memw

mamm

we
meow
Home

>Ocmfikuh

Hoocom :H mcoHDOQDLCH

museumm DHEEOO OD >nxo DO: me Delzdez

H0050m CH mcoebomCCCA muoHumm
OLEECO OD >nxo DmLZmEOm we DHIESHUvE

Loo:um CH mcoebumuwce Hoosom ca mcoeuomwmcH macaumm
nsoeuwm DHEEOO oh >pxo me DH liveufiafioo ou >mxo no: me DH ”Mmeamm

Hoozom CH m:OHLunHC:e

COLE OLEEOO ob >mxo Do: me Dfllroez

DAEEoo OD >nxo nmzzmsom he Delinepvi

3 c:

aCHZEOn Few nfizoanm:wuuitfl pi e:

aoozum :H mccmbomunce H0050» cw mcoeuomuuce

OLEEOO On >nxo m.nfid£o~vdfifi ufiafioo ou hmxo no: me UH ”wmeamm

Ecosv>eo>cfl nae:
bcoEm>eo>cfl EJHme
b :o:_o> H02: 33

amena>euom Hoonom Hmeuo CH DCOEO>HO>CH

education reveals that majority of the respondents’ parents
had some college education. In addition, 9.6% did not
finish high school, 21.8% completed high school, 14.2% were
college graduates, 8.4% indicated their parents’ had earned
a Master’s degree or equivalent, and 4.5% indicated their
parents’ had earned a Ph.D. or M.D.

The social bond indicators for attachment were
attachment to parents, peers, and teachers. The results
indicate 92.8% of the respondents had a medium attachment
to their parents, in addition, 65.6% had a medium
attachment to their peers and another 20.7% respondents had
a low attachment to their peers. Also, 67.0% of respondents
had a medium attachment to their teachers and 22.5% had a
high attachment to their teachers.

The social bond indicators for commitment were
commitment to school, commitment to continuing one’s
education or going to trade school, and commitment to
marriage and success. The results indicate a majority of
the respondents were high on their commitment to school
(83.5%). Over half of the respondents (51.3%) had a low
commitment for continuing their education or going to trade
school after they graduated from high school another 40.2%
had a medium commitment on continuing their education or

going to trade school after completing high school.

72

There are two social bond indicators for involvement,
they are involvement in school related sports or
involvement in other school related activities (i.e., band,
clubs, etc.). The results for involvement in school related
sports indicate that 87.6% of respondents had low
involvement in school related sports. In addition, 11.1% of
the respondents had medium involvement in school related
sports. Similarly, there were 85.1% of the respondents who
had medium involvement in other school related activities.
Another 14.7% had low involvement in other school related
activities.

The social bond indicators for belief are, believing
it is not okay to commit mild and serious infractions in
school. There were 76.5% of the respondents who were high
on belief. Meaning 76.5% of the respondents believed it was
not okay to commit mild infractions in school. Another
21.0% of the respondents were medium on belief, meaning,
these respondents believed it was somewhat okay to commit
mild infractions in school. Finally, an overwhelming
majority (96.6%) of the respondents felt it was not okay to

commit serious infractions in school.

BIVARIATE ANALYSES: THE SOCIAL CONTROL INDICATORS AND
GENDER

Attachment and Gender

The next sets of findings are the results of bivariate
analyses that compare the social control indicators on
gender (Table 2) and race (Table 3). The findings are very
similar when comparing males and females on the attachment
indicators of attachment to parents, peers, and teachers.
Comparing males and females on attachment to parents the
findings indicate 92.3% of males and 93.4% of females have
a medium attachment to their parents. In addition, the
findings indicate there are slightly more males (3.5%) who
have high attachment to their parents than females (1.4%).
In addition, there are more females who have low attachment
to their parents than males. This finding is significant
(12= 51.011; df=2; p<.01), which indicates the slight
differences with attachment to parents for males and
females can be inferred to the population with a 99%
accuracy.

The findings on attachment to peers indicate that
males (14.8%) have a higher attachment to their peers than
females (12.8%). More females (66.7%) have a medium
attachment to their peers than males (64.4%) and males and

females are nearly identical on their percentages on low

a ll ex..Mwm

No.m

‘Nmm.m

«llZEHm

E

w~.c
w~.vm
ww.me

wm.NN
wr.>m
wm.m

ww.NH
w>.©m
wa.cm

move
mmw

me
new
Nome

omafi
mwmm
NNn

Sac
_Smm
mace

we
momv
new

Swzcmzmxwum
nOHmEdh

wN.o
mm.ow
wn.mH

mo.VQ um unnowhwnmwm me mumsgmienbe
HQQ.VQ um ufiflOwacmwn mm mhmnbmiwsuie

HA see:
c w N v E: .L ~52
mom son

«new.omnzc nowuw>wu0< nouneomiaoocom uoauo cw unmEu>a0>2H

wm.~
wm.mH
wm.mw

wm.mm
mm.cw
wN.HH

wm.v~
wv.vo
ww.cm

noncoo

HS eel:
moo ESWUDZ
moan 23d

«NHntmflzy mmwufibwu0< DOuMHQMImuHOQW CH DCOEM>HO>CH

wood can:
mem E: 392%
men zen
«mm~.SHu2C mumcomme 0» unmezomSum
was set:
who, am E: NUDE
m3 ~OH 30H
«vmm.eauzcmummm ou uzmscomuum
moi eel:
nvvv Ezwfimz
mom son

«mwo.o~uzc mucoumm on unmanomuus

>9:m:vmuw
moan!

MHODMOHUCH Houucoo Hmfioom Dan HQDGOO Mo nenidmcd Oumflum>fim Mo nudznom .N wands

mcvnm um unmouenmwm me Quagmiwabx
Hoo.vnw um. uGMUfiMMBmMm aw OHMDHUMIHEO...‘

wo.am News en.em Loom emu:
em.n me w>.e mam szunmz
we.o a em.o me son

..amm.eeu «vam.omuzc Hoocom ca ncowuoaumcn msowuom “fiasco 0“ Smxo no: me uh . wmaamm
wa.oa meme we.ue moem cal:
wa.eu veer am.em SSNL Seine:
Wm.u Le wa.m mom 26S

..nia.apl «m~m.oauzc Hoonom cw muowuomuwcm new: “Aesop on Smxo “on me uh - wmaamm
we.ce maam WN.4S omem cal:
no.5N Lani we.mm Swen Salem:
wa.i SSH we.m use sou

..SSS.AS «mkm.oauzc unooozm use mmmwuumz 0» unosuessoo
we.e eam ws.a one zed:
we.me mean wm.em mash agents
wa.ae mepn wa.mm Siam 23S

< «m VN . NM «BNO \QNZ» HOOQUW GENRE OH UCHOU H0 EDMUNUSUH UCHDCMuCOU 0U “COS NCEnOU
wfl.aa else Se.e> aria can:
wa.a man wa.eu Lem .ssusiz
wo.m men w~.n nor sen

..o>.oSH «kmm.omuzc Hoonom on unmeueesoo

m Sucozgmuu w Secmzdmuk
E moan-50h MOHQX
HOUGOU

 

.Ao.u:ooc w menus

KO
All

...afif‘
‘ -
...--avuw

‘\I
(I)
(I)

era a:
..

vte‘d ~4-

n" -

Jud Q

-~‘- 'V‘

«vau-

.

A~-

.... ~
-_t._4
I...

_"
...,“
~-

- S.
..“v
‘5-

a a
" ...

v.4
~‘E‘

_‘ —

'~..

”'9,

“Ivl

5-‘

1.." ~

.A
‘S.
w,‘_‘

‘-

-u
~—

In“.

‘-

"

(s

\“rl

\I

-.-
-
A-
.—"t
‘4
u

*-

n"

..‘J R

A.
u
‘;
~"‘--
v..""
N“

v-

S- .

«.‘a—P

“‘

-

s".*

‘L 7‘s

. “e.

 

 

 

attachment (20.8% and 20.5%, respectively). These findings
are significant (12= 9.232; df=2; p<.05), which indicates
the differences observed with attachment to peers for males
and females can be inferred to the population with a 95%
accuracy.

The findings on attachment to teachers indicate that

males (11.2%) have a lower attachment to teachers than

females (9.8%). In addition, females (67.6%)have a higher
medium attachment to teachers than males (66.3%). Both

males and females have the similar percentages of high
attachment to teachers (22.55). The findings on teacher

attachment were not significant.

Involvement and Gender

The findings on involvement are again similar for
males and females, however, there are slightly more males
involved in school related sports and other school related
activities than females. The findings on involvement in
school related sports indicate that more females (89.5%)
had low involvement than males (85.5%). More males (13.2%)
had medium involvement in school related sports than
females (9.2%). Both males (1.3%)and females (1.2%) have
nearly the same rate of high involvement in school related

sports. The differences observed were significant (12:

38.741; df=2; p<.01), which indicates the differences
observed between males and females in regards to
involvement in school related sports can be inferred to the
population with a 99% accuracy.

The findings on involvement in other school related
activities indicate again that more females (15.8%) are
higher on low involvement in other school related
activities than males (13.5%). In addition, males
(86.3%)are higher on medium involvement in other school
related activities than females (84.1%). The percentages
for high involvement for males and females were almost
nonexistent (.1% and .2%, respectively). The differences
observed were significant (12: 12.232; df=2; p<.05), which
indicates the differences observed between males and
females with involvement in other school related activities

can be inferred to the population with a 95% accuracy.

Commitment and Gender

The findings for commitment to school, commitment to
continuing one’s education or going to trade school, and
commitment to marriage and success indicate females are
more committed to school and marriage and success than
males. In addition, there are mixed results for males and

females on continuing one’s education or going to trade

78

 

school. Specifically, the findings on commitment to school
indicate that more females (88.1%) have a high commitment
to school than males (78.6%). The findings are significant

(12: 180.70; df=2; p<.01), which indicates the differences

observed in the findings can be inferred to the population
with a 99% accuracy. The results for commitment to
continuing one’s education beyond high school or going to
trade school were mixed. There are more males (52.9%) who
are low on commitment to continuing their education or
going to trade school than females (49.8%), but, at the
same time, there are more males (9.6%) who are high on
their commitment to continuing their education or going to
trade school than females (7.6%). The findings indicate
that the males in the sample are least likely to continue
their education or go to trade school after graduation, on
the other hand more males indicated they are slightly more
likely than females to continue their education or go to
trade school after completing high school (see Table 2).
The differences observed were significant (12: 32.245;
df=2; p<.01), which indicates the differences observed
between males and females with commitment to continuing

one’s education or going to trade school after graduation

can be inferred to the population with a 99% accuracy.

79

 

The findings for commitment to marriage and success
indicate that more females are committed to marriage and
success than males. Specifically, the findings indicate
there are more females (70.7%) high on their commitment to
marriage and success than males (64.2%). There are more
males (32.4%) who are medium on their commitment to
marriage and success than females (27.5%). In addition,
there are more males (3.4%) who are low on commitment to
marriage and success than females (1.9%). The differences
observed with these findings were significant (12: 63.696;
df=2; p<.01), which indicates the differences observed
between males and females with commitment to marriage and
success can be inferred to the population with a 95%

accuracy.

Belief and Gender

Finally, the findings on whether the respondent
believed it was not okay to commit mild and serious
infractions in school indicate that a majority of the
respondents believed it was not okay to commit mild or
serious infractions in school with only a slight difference
between males and females. The first belief indicator on
whether it is not okay to commit mild infraCtions in school

indicate more females (80.9%) were high on this variable as

80

 

compared to males (71.8%). This indicates that more females
believed it was not okay to commit mild infractions in
school than males. More males (3.9%) were low on believing
it was okay to commit mild infractions in school than
females (1.3%). The differences observed were significant
(12: 159.115; df=2; p<.01), which indicates the differences
observed between males and females in regards to believing
it was not okay to commit mild infractions in school can be

inferred to the population with a 99% accuracy.

BIVARIATE ANALYSES: THE SOCIAL CONTROL INDICATORS AND
RACE/ETHNICITY

Attachment and Race/Ethnicity

Table 3 displays the findings from the bivariate
analyses with the social control indicators and
race/ethnicity. Overall, the findings are very similar
among the racial/ethnic groups for the social control
indicators.

The findings on attachment to parents, peers, and

teachers are similar for whites, African Americans and

o\0
v

Mexican Americans. There are more African Americans (5.4
who are high on parental attachment than whites (2.1%) or
Mexican Americans (1.7%). Whites and Mexican Americans have

higher percentages of medium attachment than African

(D
I-"-‘

w~.o a

w~.cm mom

wm.m mm
«4omo.mm

wo.~ we

wfl.oH Hm

wm.mm New
*«flo.mm

wH.om vmm

wm.mw own

wv.m m>
«<mem.mm

wm.c~ ma

wo.vo ham

wr.mm omm
«4wNO.©m

wm.a he

w~.mm firm

wm.m \v
TI©N©.N0

w Sn:~®:mxvek
g 53 was:

nuoumoevcH Houucou

wm.o
mo.mm
mH.SH

w>.m
wm.cH
wv.wm

op
[\

.Sm
.mo

3‘0
O

w>.ma
mm.mm
wv.om

wv.n
wm.mm
wa.m

m
0mm
mom

wN.o
mm.vm
mm.va

mQ.VQ.um DGHUHMHRURM me mhmnvnificbi
HQQ.VQ.ufl ucnowflwnmwn MN oumnvmiwani

«a
vcmo
mama

2S4:

Ez~3m2
20m

«ASN.6HuzC nowuw>wuo< nounaomisoocom umsuo cw unmeusaosce

om
ONH
emm

vfim
man
wee

Mme
vme
mmm

Socmzomue

wo.a
wN.HH
wn.>m

wv.HN
we.m©
wm.oH

wm.mH
wo.wo
wa.om

wH.N
wv.mm
wo.v

w

GMUHHOH< cmowflhw

Suaondrumxoomm
Huaoom Dan %DAOHGSUH\OUNM No awn>aflcfl Gunfium>fim mo uuafiaom .m QHQMB

om
mom
Home

col:
:5 2.52
33

«Nehtmflze nowuw>wu0< Daumemminuhomw Cw ucmfim>H0>cH

Ween 2S4:
mono Ezntuz
vmw 33m
«mm~.S~uzC neoconoe ou newshounue
nmee CS“:
Gown Ezepvz
Sena ion
«VMN \QHHZC “NOON OH “COEUMDUG
emu est:
mwme Ezwbbz
hem 30m
«mmotemuze nucmumm Ou unmacomuufi
Aocngmuw
Omega

A, A
AC

mc.Vnw an unmounmunflmwm min @Hmflvniecor.
HQQ.VQ um DENOfiNHnuH—d use @Hmnmumiecbl.

we.ea was em.ma Same em.om some SSH:
we.m mm em.o SH wm.m mew asunez
em.o N wH.o A no.0 as 26s
..Sme.um «vam.o~nzc eoocom an muofiuoauwdH unowuom season on Sane eon we he . umaemm
em es use we.me mean wm.me ewes :Su:
wo.am sum we.mi nah es.um Same Ears»:
we.e we em.L en we.m sew sou
t.msm.ae «mam.oflnzc Hoocom ca acowuounmna new: unease 0» Smxo add me ue . mmaamm
we NS mom w>.me SSS wa.mo mama :Sa:
we.«m Gym wS.Lm one wi.mm momw Ezuevz
em.m on we.m um SS.N Ham 204
.esm.ei «mum.oauze nuooosm new mmmwuumz on unmadesEOo
em.:e ea LS.ML men we.e ASS est:
wa.m~ New no.0? bee mm.ue eemm Essen:
em.os San em.me was wo.um emae 26d
..va.mm «Rum.muzt Hoocom opens ou Sawoo no cowumoauw Sewnceucoo ou u:¢8u45506
$2.3m mum wa.mm HSSH La.ma Skis cal:
em.pm spa $S.mn men wa.mu Seen szetmz
$9.4 we wa.m we AN.N emu see
..meo.am «kmm.o~uzc Hooaom on unusuassou
w hscmzwmum m Socngmuk w Aocngmuu
a .30 E32 53 eke—SN and ”FANG Ga .25
Spacecruu\oomm -

 

.lo_u:oocm manna

83

Americans (93.4%, 93.1%, and 88.9%, respectively). The
differences observed were significant (12: 52.626; df=2;

p<.01), which indicates the differences observed between
whites, African Americans, and Mexican Americans in terms
of parental attachment can be inferred to the population
with a 99% accuracy.

The findings on attachment to peers indicate that more
African Americans (15.7%) are high on attachment to peers
than whites (13.9%) or Mexican Americans (10.2%). The
findings also indicate that whites (66.0%) have a higher
rate of medium attachment to peers than African Americans
(63.8%) and Mexican Americans (64.0%). There are more
Mexican Americans who are low on attachment to peers than
African Americans (20.4%) and whites (20.1%). The

differences observed were significant (12: 26.026; df=2;

p<.01), which indicates the differences observed between
whites, African Americans, and Mexican Americans in terms
of peer attachment can be inferred to the population with a
99% accuracy.

The findings on attachment to teachers indicate that
more African Americans (27.7%) are high on attachment to
teachers than whites (21.4%) or Mexican Americans (26.1%).
Whites (67.7%), African Americans (63.0%), and Mexican

Americans (65.5%) are all high on medium teacher attachment

84

with only slight differences. Whites (10.9%) have the
highest percentages of low attachment to teachers than
African Americans (9.3%) or Mexican Americans (8.4%). The
differences observed were significant (12= 32.937; df=2;
p<.01), which indicates the differences observed between
whites, African Americans, and Mexican Americans in terms
of teacher attachment can be inferred to the population

with a 99% accuracy.

Involvement and Race/Ethnicity

The next findings are the results of race/ethnicity
and involvement in school related sports activities and
involvement in other school related activities other than
sports. Majority of the respondents were not involved in
any sports related activities. Mexican Americans (88.3%)
had the highest rate of low involvement in school related
sports activities than African Americans (86.4%) or whites
(87.7%). There were similar percentages of medium
involvement for all groups. However, African Americans
(2.7%) had higher percentages of high involvement in school
related sports activities than whites (1.0%) or Mexican
Americans (1.6%). The differences observed were significant

(z2= 26.01; df=2; p<.01), which indicates the differences

observed between whites, African Americans, and Mexican

85

Americans in terms of involvement in sports related
activities can be inferred to the population with a 99%
accuracy.

African Americans (17.1%) had the highest rate of low
involvement for other school related activities followed by
whites (14.9%) and Mexican Americans (9.8%), respectively.
Mexican Americans (90.1%) had the highest percentages of
medium involvement in other school related activities than
African Americans (82.6%) and whites (84.9%). The
differences observed were significant (12: 25.08; df=2;
p<.01), which indicates the differences observed between
whites, African Americans, and Mexican Americans in terms
of involvement in other school related activities can be

inferred to the population with a 99% accuracy.

Commitment and Race/Ethnicity

The indicators for the social bond of commitment are
commitment to school, commitment to continuing one’s
education or going to trade school, and commitment to
marriage and success. The results indicate that all groups
have a high commitment to school, but African Americans
(83.5%) and whites (83.9%) have higher commitment to school
than Mexican Americans (80.0%). The differences observed

were significant (z2= 29.073; df=2; p<.01), which indicates

86

the differences observed between whites, African Americans,
and Mexican Americans in terms of commitment to school can
be inferred to the population with a 99% accuracy.

The findings on commitment to continuing one’s
education or going to trade school indicates that majority
of the respondents had low or medium commitment to
continuing their education or going to trade school.
African Americans (13.6%) had the highest commitment to
continuing their education or going to trade school upon
graduation from high school than whites (7.7%) or Mexican
Americans (10.2%). The differences observed were
significant (2'2= 93.438; df=2; p<.01), which indicates the
differences observed between whites, African Americans, and
Mexican Americans in terms of commitment to continuing
one’s education or going to trade school upon graduation
can be inferred to the population with a 99% accuracy.

Majority of the respondents were either high or medium
on commitment to marriage and success. Whites (68.4%) had
the highest rating on high commitment to marriage and
success than African Americans (65.7%) and Mexican
Americans (62.9%). The differences observed were
significant (12: 16.374; df=2; p<.05), which indicates the
differences observed between whites, African Americans, and

Mexican Americans in terms of commitment to marriage and

87

success can be inferred to the population with a 95%

accuracy.

Belief and Race/Ethnicity

The indicators for the social bond belief are
believing it is not okay to commit mild infractions in
school and believing it is not okay to commit serious
infractions in school. The findings indicate that majority
of the respondents were high on believing it was not okay
to commit mild infractions in school. However, African
Americans (83.6%) had the highest percentage of respondents
who believed it was not okay to commit mild infractions in
school as compared to whites (75.3%) and Mexican Americans
(77.3%). The differences observed were significant (12=
48.565; df=2; p<.01), which indicates the differences
observed between whites, African Americans, and Mexican
Americans in terms of believing it is not okay to commit
mild infractions in school can be inferred to the
population with a 99% accuracy.

An overwhelming majority of the respondents indicated
it was not okay to commit serious infractions in school.
However, African Americans (99.2%)represented the highest
percentage of respondents who believed that it was not okay

to commit more serious infractions in school as compared to

88

 

whites (96.2%) and Mexican Americans (96.4%). The
differences observed were significant (12: 31.456; df=2;

p<.01), which indicates the differences observed between
whites, African Americans, and Mexican Americans in terms
of believing it is not okay to commit serious infractions
in school can be inferred to the population with a 99%

accuracy.

Bivariate Analyses with Smoking Cigarettes and the
Independent Variables

The bivariate analyses of the dependent variable
smoking cigarettes and the independent variables are
displayed in Table 4. The data indicates that most of the
variables are significant at p<=.001 or p<=.05. Attachment
to peers and involvement in other school related activities
were the only two measures that did show a significant

relationship with smoking cigarettes.

Smoking Cigarettes and Control variables

The first group of analyses compared smoking
cigarettes to the control variables (i.e., race, gender,
and $85). The data on cigarette smoking indicates that more
females (18.0%)than males (16.1%) smoke (12: 6.671; df=2;
p<.05). The data also indicate that more Whites (19.3%)

smoke than African Americans (4.8%) and Mexican Americans

89

 

 

«<Nmm.cm

44Ho~.ooa

*wa.w

E

wm.mm

wv.>w

wa.mm

wm.mm
w>.mh

$5.0m
wm.na
$5.0m

wnv.mc
wm.mm

mow

mwo

mrmfi
mmmm
emba

Ham
coma
nebw

cove
avav

xocmzvmuw
02

wa.VH
w®.NH
wm.VH
wm.ba

wm.om

wm.ma

wm.mH
mm.v
wm.m~

wc.mwa
w~.©H

mo

mmm
new
>mv

mra

ewe
on
New”

emm
mob

kocngmnm
no»

nouuouumdo mcfixofim

mo.vm an ufimowhflnmwn aw Qumnvmlfl£U1
Hco.vm.um unmowhwnmwm MM mumnvMIߣU**

motto

no 6.: Tod;

u2®~n>wzvm

no m.umbmw2

mbmzonuo mamANoL

momNNOQ meow

momzcnuo

~Oc£om so“:

Nocccm cow:

cmwzak no: UWQ
«mum.eauzc Hoboq cowumosnm m.ucmumm

:muwxmz
: to a.“ 95x : no flux <
maize
«Hov.oauzc momm

mmN 9:me
mmNmZ

«Hov.e~n2c umnnmo

 

moaanum>

ucovcmamvcH £ua3 mmuuoummfio mcdxofim OHQMHHM> HGOUGOQOQ Gnu MO uflnhaqu oumaum>fim .v manna

90

«inoH.am

«<0>@.NHN

«aMNH.mH

«4vwm.HN

E

wm.~m
wm.mm
wv.ow

wm.>w
wm.mm
w~.mm

wm.kfim
wm.mm
wm.Nw

wo.w>
wv.mw
wm.>>

wo.nw
wc.mw
we.~m
wv.aw

anew
mom
om

mob
ammo
Arma

Nova
wmvm
ovaa

comm
HONN
maam
meow

>0:m:@m~u
02

wa.mfi
wN.HH
wm.mfl

wm.mm
wm.©a
wo.~H

wm.ma
ww.ma
wm.ba

wv.mm
wo.©H
w>.mN

we.va
wo.wa
wo.mm
ww.wa

w

News
via
mm
«amw.mnzc

5mm
mmofl
sow

mom
mafia
mmm

moH
mama
Hm

amm
Hmv
ewe
owe

xocmzooum
no»

nouuouMOMU mcwxoem

 

 

mo.VQ.un unmofikwnmwn mH QHMDUMIHSU«
Hco.vm.um uQmUwMfinmwm nw oumnvmlwgoi«

 

204
Ezwme
can:

uOfiufl>HuU< UQuMHQNIMuHOQW Ce ucm8m>HO>QH

2m<~
Ezatmz
cal:
«Hhm\QH2\ numzommh Ou u=®E£OMuu<

30%
E: .2 Caz
can:
annexenuzemwmom cu unmSJUMuuw

2C4
Ezebmz
gun:

«omn.mu2\ mucmumm 0» ucmszomuuw

«zowzcv.m-oun:a
m mswunnzo
m mNNounzo
«264cm massunzo

«Hcv.0~n2» oawuumzc QSOUCH

.Av.ucoovv OHQMB

91

«ammw

«meO.

«aNNb.

«ammm.

n>m.m

.wmn

voH

mom

owa

mm.

wm

wm.
we.
wm.

wo.
wo.
wo.

ww.
we.
wfi.

Hm

.mo
we.

rm

wma
mama
Namw

we.
wm.
we.

mc.vm.un ucuowwwnmfim aw onmavm:w:0*
HQQ.VQ um “EGOfiMflnmwm MM mhmnvaH£U*t

we owe 30s
cm was Egonmz
NH mam cs4:

«vmm.oauzc Hooaom aw ucowuuuuuaH nawz.uwaaoo 0» sage no: ma “A . mmwamm

we
as
mm

.mb
.om
-.om

a>
H>

\J

Nm
N@
as

mma
mavm
ooow

Ovmm
mvmm
©H>

wm.
wm.
we.

wm.
WA.
ma.

«mmm\muzy Hecaom

wwa
omoH
Germ

Nmaa
ammo
Ha

>ocngmuk
02

We.
We.
we.

wm.
mm.
wm.

Hm mm 304
am mow Ezwbmz
so amen zoom
~w5m\omflzy mnmooam USN mUMHHHNE Cu acmEuHSEOU
ON wacfl keg
ma mom ESNUDE
mH on £0“:

OUNHE Ou ucfiou H0 COquUDUH UGwzcfluGOU Cu uCQEUHESOU

mm mm 20a
mm veg EzeCmZ
as mama can:
«me.o~u2\ Hoonom On ucmEuHSSOU
5H mew 20w
ea mmva Enfltmz
mm b nae:

«an\mfl2~ nowuwbwuofl heumflomnfloonom Monuo EH unmam>HO>GH

w

xocmsgmuw
mm»

nouuaunmflu andxoam

..u.u:oo.v manna

92

wn.mv om
wo.vm moa
wm.vm deem

«amqa.mmm

m xozmzvmnk

mc.vm um uQ~0wManowm an oumnvm-wco*
Hoo.vm an uanAMwnawm mw mumnmmuwgox.

am.mm mm 264
wo.om vow Ezetmz
wv.mH mmmfl cane

«HMM~0HH2~ Hoogom Cw MGOquNHMGH ”30flhom “WEEGU Cu EM¥O uo: mw “H I wafimm

m xocmzomuw
no»

mouuoummao acfixoam

..c.u:oo.v manna

93

(13.3%) (12: 160.171; df=2; p<.01). The percentage of
African Americans who smoke is much less than the other two
groups. In addition, respondents whose parents were high
school graduates had the highest percentages of smoking
(20.3%)than the other parental education levels followed by
respondent's parents who did not complete high school
(18.5%). The lowest percentages of smoking were among
respondents whose parents had a Master’s degree or
equivalent (12.6%)(12= 38.822; df=2; p<.01). The findings
on the income quartile indicate that respondents who were
part of the high SES quartile had low percentages of

smoking(14.4%) than the other SES quartiles (12: 38.822;

df=2; p<.01).

Attachment and Cigarette Smoking

The next group of analyses assessed the relationships
between the social control measures and smoking cigarettes.
Some of the results do not follow the expected pattern,
that is, youth who are highly bonded will refrain from
smoking cigarettes. The findings on attachment to parents
indicate that youth with high attachment (22.7%) have
nearly the same percentage of smoking to youth with low

attachment (23.4%)(12= 21.384; df=2; p<.01). The findings

on attachment to teachers indicate students who have a low
attachment to teachers have a significantly higher
percentage of smoking (32.2%) than students with high

(11.9%) or medium (16.5%)teacher attachment (12: 212.67;

df=2; p<.01).

Involvement and Cigarette Smoking

The findings on involvement and smoking cigarettes
indicate that respondents who were more involved in sports
related activities, meaning, they were high on involvement,
indicated a higher percentage of smoking cigarettes than

those who were medium or low on involvement (12: 31.11;

df=2; p<.01).

Commitment and Cigarette Smoking

The findings on commitment and smoking cigarettes
indicate that respondents who were low on commitment
displayed a higher percentage of smoking cigarettes than
the other respondents. Specifically, respondents who were
low (28.4%) or medium (28.4%) on commitment to school were
almost twice as likely to smoke than students who had high
(15.0%) school commitment (12: 189.89; df=2; p< .001).
Respondents who were low on commitment to continuing their

education or going to trade school had a higher percentage

95

of smoking (20.9%)than respondents who were medium (13.1%)
or high (13.1%) on commitment to continuing their education
or going to trade school (12= 102.722; df=2; p<.01). The
findings on the last commitment indicator indicate
respondents who were low on commitment to marriage and
success were over twice as likely to indicate they smoked
cigarettes (31.5%) than respondents who were high on

commitment to marriage and success (14.7%) (12: 104.66;

df=2; p<.01).

Belief and Cigarette Smoking

The final social control indicator, belief, showed
overwhelming support that respondents who were low on the
belief indicators had a high percentage of smoking than
respondents who were high on belief. Specifically,
respondents who believed it was okay to commit mild
infractions in school were four times more likely to smoke
than (48.7%) those who believed it was not okay to commit
mild infractions in school (12.4%)(12= 586.24; df=2;
p<.01). In regards to the second belief indicator, the
results indicate that respondents who were medium (66.0%)
or low (56.5%) on whether it was okay to commit serious
infractions in school had a high percentage of smoking than

respondents who were high on belief (15.4%), which

96

indicated they did not think it was okay to commit serious

infractions in school (,‘(22 593.15; df=2; p<.01).

Bivariate Analyses with Drinking Alcohol and the
Independent Variables
Drinking Alcohol and the Control variables

The data presented in Table 5 are the bivariate
analyses with drinking alcohol as the dependent variable.
The data indicate there are less of the independent
measures that are significant or as significant as compared
to the analyses where smoking cigarettes was the dependent
variable. This observation is the first inclination that
the social control model is better used to explain less
serious behaviors, as indicated in the literature review
(Akers, 1991). In this case, the social control model may
be a better model to explain the cigarette smoking behavior
of youth than it is to explain their drinking behavior.

The data in table 5 indicate that more youth overall
indicated they had drank alcohol 30 days prior to
interviewing than those who indicated they had smoked
cigarettes 30 days prior to being interviewed.
Specifically, the data indicate more males (45.3%) drank
alcohol in the 30 days prior to being interviewed than

females (40.2%)(z-= 24.698; df=2; p<.01). The data is also

97

«omN.oH

«Nem.ma

«immv.qmfl

«ammo.vm

E

wo.>m
wm.om
wm.wm
wm.oo

wm.mm
wm.ao
wo.>n
$9.9m
wm.mm
wo.mw

w>.om
wo.mm
ww.vn

wo.vm
wm.mm

meme
mom~
onH
memfi

amm
Ame
mom
vmmm
mmfia
mam

mew
mew
Cbav

Chem
momm

Xucmscmnu

02

Hogooaa unexcwua

mo.mV
mo.mv
w>.mv
wb.mm

wo.wv
w>.xm
wm.me
wa.mv
wv.vv
wo.mm

Mm.ave
wm.vN
wm.mv

wm.mv
wm.0e

w

mmofi

wooa

omoa
vwm

Gem
New
hme

mmom
coma

>U:mzvmum

no»

mc.vm.uu unmowmwQme aw OHMBoMIfl:U«
Hoo.vm.un “snowmwnown aw amazonuwaoii

«amazes m~flu~nzo
m meeunmzo
m mefluqmzo
«zones m~wn~mzo
«How.o~nzc oawuumno masoca

umcbo no .Q.Z.\.Q.:L
bcm~n>wzvm no m.nvomn2
memzbnum mvmaeob
momeeoo meow
memzbmub Noozom cue:
Hoocom cue: cmwcwt us: we:
«mvm\QHfl2» Habofl QOquUSUH m.ucmumm

:moexmz
CmVOHMmXSN. :muQ flamed.
when:
«Hov.e~nzc comm

mmemsmk
moan:

«Hov.oauzc umocmu

noHnMwnm>

uGGvCOQOUcH Ham Sud: HO£OOH4 Ucwxcaua mandflum> ucovcmmoo as» no afia>HMG< Gunfium>fim .m QHQMB

98

m>n.m
H>m.m
4<m~w.om~
4<n~©.ab
«ismv.wm

E

wm.wm
wm.>m
wm.mm

wo.>n
wm.mn
wv.mn

w©.mv
ww.mm
$0.5b

wa.mm
mm.@n
wr.wv

wv.mwv
wm.>fi
wo.mm

eve
Nmmv

bee
mmmm
Home

coma
mevm
Haw

com
msrv
mofi

Sh:cm:me~L
02

wN.mv
w>.Nv
wH.Hw

men
wmmm
Ha

no.vm an ufiuoflhwnmwu we OHMSvMIHSUi
Hoo.vm.ufl HGQOMMMRDwn aw @NMSUMIflcUii

zed
Ezwomz
new:

«omn.muzc nowuw>wuu<_nounHuMIHoonom umcuo cw unmsm>~o>ca

wv.mv
w>.ve
w©.©v

w~.om
wm.vv
wm.mm

w~.om
we.@v
wm.~n

ww.~n
mm.Ne
wv.>v

w

NmHm
mmv
we

30a
Ezwbmz
can:

«nnw.muze nowuwbwuod noumHQMInuwomw cw unmao>~o>cH

how
mmmm
Hem

CNN
omvm
mm

Sciemszam

no»

Hogooa< mcwxcwuo

304
5: H332
zen:

«~mm.muzc muonomme cu acoE:QMuuw

20g
Ezwbmz
zen:

«MMQ\QHH2\WHQQN On HGQESOMuufl

20a
Ezwomz
zen:
«omn.quze mucoumm 0» u:QSAUMuu<

.Ao.u:ooc m manna

99

«*mee.>m~

<4M>N.eee

<m>m.e

«mom.e

«iemm.mma

wN.
we.
we.

ma e
eN ee
we eemm

we.
mm.
wm.

em
me
He

MQ.VQ u! uanflanmfiu he Ohmnwulwabe
HQQ.VQ.um ufiMOHMwnmfln aw mumzvmlwaoai

am gee
mam Ezeemz
Hmsm cue:

«Hmm.o~uze Hoonom :w «cowuumnunm maowuom awaado ou knee “on we aw - mmwamm

we.
we.
we.

we.
we.

we.

we
we

wm.

we.
wm.
wm.

om Hm
mm HNA
em arse
«vmm.o~uzc
mm awe
mm ammo
we neem
.nm seem
.am ream
an awe

we.
we.

We
Hooaum

we.
we.

we

We.

we

mo
ee

.mm

was zoo
moms Ezeemz
mmqm gee:

cw mcoeuomumcm new: awaaoo ou ameo no: aw aw - umwamm

he
ee

.He

ee

.Oe
we.

He

mHH RafiN
mNmfi Ezebmz.
eeem cow:

«ehmtomflze unooosm ham mmmMHNNE Cu ucmfiquEoU

mass ace
asqa Ezetvz
vim can:

«mmmxmuze Hoonum ovmuh Ou DGMOQ Ho Gowumozvfl unwacquOU Ou ucmauflSEOU

en eaa
Ne men
ee mwbe

we.
we.
we.

xocozvmnk

02

en
be
em

one age
eeb EZNUmE
Harm can:

«Hmm.o~nzc Hoozom on acmauesaoo

AUcmzvmmw
no»

nwuuouamao ucwxofim

..n.ucooe m manna

100

indicative of the pattern observed in table 4 in regards to
race and smoking cigarettes. In table 5 more whites
indicated that they drank in the 30 days prior to being
interviewed as compared to African Americans (24.5%) and
Mexican Americans (40.3%). However, the data suggests that
African Americans drink less than all the groups, which is
the same pattern observed in table 4 with the smoking
behaviors. This pattern of behavior is similar to the
observations in national trend data on the behaviors of
youth by race. The data indicated that African Americans
tended to participate less often in smoking cigarettes,
drinking alcohol, and/or drug use. However, the data did
suggest African American youth tend to participate more
often in other risky behaviors such as having more sex than
any of the other race/ethnic groups as well as
participating in other more serious behavior more often
than other race/ethnic groups (i.e., property crimes)
(OJJDP, 2002).

Overall, percentages of drinking were very similar for
all groups when compared to respondents’ parents’ education
level in this sample. The highest percentages of drinking
were among respondents’ whose parents’ had Ph.D.’s or
M.D.’s (46.7%). This is contrary to the predicted pattern

of behavior. In fact the lowest rate of drinking (38.0%)

was among respondents’ who indicated their parents had not
completed high school. Again, when the highest SES quartile
and lowest SES quartile are compared, the respondents who
are in the highest SES quartile have a higher percentage of
drinking (43.0%) than respondents who are in the lowest SES

quartile (39.7%) (12: 18.342; df=2; p<.05).

Attachment and Drinking Alcohol

Focusing on the attachment indicators, the findings
indicate that youth who have low attachment to parents (had
a higher percentage of drinking alcohol (51.6%) than youth
who were medium (42.2%) or high (47.4%) on attachment to
parents (12: 71.02; df=2; p<.01). Respondents who were high
on attachment to peers had a higher percentage of alcohol
use (51.2%) than respondents who were low on peer
attachment (36.1%)(12= 71.015; df=2; p<.01). The findings
on attachment to teachers indicate that respondents who are
low on teacher attachment drink alcohol at a higher rate
than respondents who are medium (44.2%) or high (32.5%) on

attachment to teachers (12: 160.82; df=2; p<.01).

in:
O
f\)

Commitment and Drinking Alcohol

The findings on the commitment indicators illustrate
that high commitment was related to low percentages of
drinking. Specifically, the findings on commitment to
school indicate that low or medium commitment to school
corresponds to a high percentage of drinking(12= 153.53;
df=2; p<.01). Similarly, low commitment to continuing one’s
education or going to trade school corresponds to a high
percentage of drinking (44.1%) as compared to a high
percentage of commitment to continuing one’s education or
going to trade school (41.8%). The differences observed are
significant (12: 8.279; df=2; p<.05). Finally, in regards
to commitment, low percentages of commitment to marriage
and success corresponds to high percentages of alcohol
drinking (47.9%) as compared to high commitment to marriage
and success (47.9%) (12: 9.28; df=2; p<.05).
Belief and Drinking Alcohol

The final social control indicator of belief shows the
most dramatic differences between the levels of belief and
percentages of drinking alcohol. Specifically, findings
indicate that believing it is okay to commit mild
infractions in school (low belief) corresponds to
percentages of drinking alcohol (79.4%) that are more than

double the percentages of respondents who believe that it

103

is not okay to commit mild infractions in school (high
belief; 35.1%) (22: 695.27; df=2; p<.01). The final belief
indicator, which indicates whether a respondent believes it
is okay to commit serious infractions in school has similar
findings as the first belief indicator. Specifically, high
percentages of drinking alcohol (84.8%) correspond to low
belief meaning the respondent believes it is okay to commit
serious infractions in school. This is compared to low
percentages of alcohol drinking and high percentages of
belief meaning the respondent believes it is not okay to

commit serious infractions in school (12: 197.673; df=2;

p<.01).

Gamma Mbasures of Association and Smoking Cigarettes

Attachment Indicators

The next sets of bivariate analyses explore the
relationship of the dependent variables more closely with
the independent measures by race. These tables present the
first evidence that the social control model, as it has
been traditionally constructed, may be a model best suited
to explain the behavior of whites than for African
Americans or Mexican Americans (Note: the intercorrelations
among the social control variables are presented in

Appendix C). Table 6 presents the gamma measures for the

104

.¢Hoo. vo nan .mo. vm ”mao>oa unmuwwwcafim

 

 

 

 

 

 

 

 

 

Aoewu:e ieonuce isomeuce Aamm.au:e

i<em~.1 N23: «‘mmH.u «1OmH.I ...nAMfiCtDE 3e e:D:E.3h.:e.
immense leaoeuce imeomuce Ammo.oiuce 7:-1 - . .-((((--
smo.: Hmo. oHo.| moo.1 nuvmm an n:ms:unnt<
Amemice Ameoause Ammmence Aoee.muce ,@ -(()(J(s-(:(.-,.(((- .
mmo.z mee. emo.l DHO.I mingled a: ZlmllEUUIe.
(quumowocw. ( l - .

9.50%me ncmuofluwiix Empowhke. mmeflca mHQEmm $87.25 ucwfifiumuuia
auwowcgum\oomm -- -~:(:((

 

 

on» now

nCflUfiunoz U56

.mamoeuoafl GMUHH: .mouflé .OHmfimm @HHHCH
muouaUHUGH usagouuud any van nouuwuumfiu mcwxofim Ho noun—mum! .9536 .e 0.71;.

105

attachment indicators and cigarette smoking for the entire
sample, whites, African Americans, and Mexican Americans.
The findings indicate attachment to teachers show the only
significant association to smoking cigarettes for the
entire sample, whites and Mexican Americans. There was no
significant association between attachment to parents and
peers for any of the groups. In addition, none of the
attachment indicators were significant for African
Americans. The gamma associations show a moderate
association between attachment to teachers and smoking. The
findings indicate error is reduced when predicting the
smoking behavior of the respondents about 13% of the time
knowing their levels of attachment to teachers. In
addition, the relationship between attachment to teachers
and smoking cigarettes is negative which indicates that as
the percentages of smoking cigarettes increase, attachment

to teachers decreases.

Commitment Indicators

The relationship between commitment to school,
continuing one’s education, and commitment to marriage and
success and cigarette smoking is better than attachment and
smoking cigarettes as indicated in Table 7. The commitment

indicators are stronger predictor of the smoking behavior

106

.eaoo. vAw new *mo. vo "mam>oa unmoemncmwm

 

 

II (in ifn

 

 

 

 

 

 

 

 

 

7 romance Aaqeeuce “Hammacv imam.ofiv

«ieme.1 eeee.u i...:ea... «keea.| mmmoozm Ucm maneuumz 0o n:msowEEOL
immense imooHuce Amaeeuce Ammm.mucv Hoocom mnmmm on mango em-
mme.a eme.n *«Nee.x *xmme.u sewunosbm mewzzeu:00 ob mzmsewESOQ
Ammauce immosuce Amowmuce AHNm.oHucc a-
..emn.l aflo.1 ..emfi.- ..emfi.- Hoocom on nzmsnflssob

memoHHmE<
mcmoflxmz cmowum< mmuflgz meQEmm mueufim mHODflUHUGH HGOEHHEEOU
Spacecnum\oomm

 

 

 

or»

mfidOfixmz Ucm

.MGMOfiHOEG snowHM¢ .mouanz .oamfimm mueuam

Mom mHOuMUHUEH ucmfiuflfifiou may find nouuounmfiu mcfixofim no mounmmoz MEEMG .b OHQMH

107

SD‘ff vvhites than African Americans and Mexican Americans.
ESEilweacifically, commitment to school is significantly
Eissociated with smoking cigarettes for the entire sample
( }’==-.124; p<.01; n=10,321) and whites (y: -.136; p<.01;
r1==8,263) only. In fact, the findings indicate it is a
k3Gatter measure of association for whites overall. The
lflindings indicate that knowing the respondents level of
<3<3mmitment to school error is reduced when predicting
ESmoking behavior about 14% for whites and 12.4% for the
GBITtire sample. In addition, the relationship is negative,
nneaning that as percentages of smoking increases the level
(of commitment to school decreases. Commitment to continuing
cone’s education or going to trade school is significantly
associated to smoking cigarettes for the entire sample (7:—
.095; p<.01; n=9,533) and whites (y=-.102; p<.01; n=7,675).
Again, the findings indicate that commitment to continuing
one’s education or going to trade school is a better
predictor of respondents’ smoking behavior for whites than
the entire sample. Additionally, the findings indicate that
knowing the respondents’ commitment to continuing their
education or going to trade school error is reduced when
predicting their smoking behavior about 9.5% of the time
for the entire sample and 10.2% of the time for whites.

Finally, there is a negative relationship between

108

<342>Ieit;inuing one’s education or going to trade school and
Esrt\<>}<ing cigarettes, meaning, as the percentages of smoking
J—Ifi.<:::reases, commitment to continuing one’s education or

§3<>jgng to trade school decreases.

Commitment to marriage and success is significantly
Eissociated to smoking cigarettes for the entire sample (7:-
- 1.00; p<.01; n=10,376), whites (y=-.107; p<.01; n=8,29l),

ZXifrican Americans (7=—.075; p<.05; n=11,491) and Mexican

Z¥nnericans (y=—.094; p<.01; n=936). The findings indicate
t:hat commitment to marriage and success is a better

EDredictor of the smoking behavior for whites followed by

Dfiexican Americans, and then African Americans. In addition,

T:he findings indicate that knowing a respondents level of
<:ommitment to marriage and success helps to reduce the
error of predicting smoking behavior of respondents by
10.0% for the entire sample, 10.7% for whites, 7.5%

for

African Americans and 9.4% for Mexican Americans. Finally,
there is a negative relationship between commitment to

marriage and success. Meaning, that as commitment to

marriage and success decreases, the percentage of smoking

increases.
Involvement and Smoking Cigarettes

The two involvement indicators are involvement in

sports related activities and involvement in Other school

109

related activities (i.e., band, clubs, drama, etc.). The

findings from the gamma measures are presented in Table 8.
The results indicate that involvement in sports related
activities was significantly associated with smoking

cigarettes for the entire sample (y=-.044; p<.01; n=9,257),

whites (y=—.051; p<.01; n=7,440), African Americans (y:—

.084; p<.01; n=995) and Mexican Americans (y=-.070; p<.05;
n=8 22) . Involvement in other school related sports was not

significantly associated with smoking cigarettes for any
group. Involvement in sports related activities is a better
predictor of behavior for African Americans, followed by
Mexican Americans, and whites. There is a positive
aSSociation between smoking cigarettes and involvement in
sE>C>Jets related activities for African Americans. This
inclJLcates that as involvement in sports related activities
inc1*eases for African Americans so does their percentage of
SIT1C>1<ing. Conversely, there is a negative relationship
betvzeen smoking cigarettes and involvement in sports
related activities for whites and Mexican Americans. This

lrlClicates that as involvement in sports related activities

1nCrease for whites and Mexican Americans their percentage

0 5 smoking decreases.

F.)
in
C)

..Hoo. vm new «mo. vm "mam>ma unmowmwcmam

 

 

 

 

 

 

 

 

 

 

 

imeence Ameosuce ”memeuce Romemuce mmenl>enu< twinivz-
mNe.| eee.- eee. eee. Hoozom .4955 :e S:P:D>~o>:~
immense $3.1: SETS Swami: «”3333?(Wadflwm
«eee.: *«eee. «.Lmef i.eee.1 mmuogm. CH uzmsmxoeoE:
mcnoflumezx
mcmoexmz cmoeumm mob HES mHQEmm muaucm mHOUMUflVGH ucmfim>HO>GH
auwowcfifloomm --,.,_:1|tti,-i!-

 

 

 

ogu “om muoumofieeH ucoeosaoscH any new mouumuumao meexoem no monsmm

5352 2; 128223 5352 £3.22 633mm 2.3:.»

02 @3936 .m QHQME

.b

.1

11

 

Belief and Smoking Cigarettes
The two indicators for belief are (1) believing it is
FICDt: okay to commit mild infractions in school and (2)
k>€ajlieving it is not okay to commit serious infractions in
Escihool the gamma associations are presented in Table 9. The
Ireasults indicate there is a moderately strong, significant,
rnegative association between believing it is not okay to
Clommit mild infractions in school and smoking cigarettes
ifor the entire sample (7=—.238; p<.01; n=10,334), whites
(y=~»235; p<.01; n=8,265), African Americans (y=—.187;
ED<-Oli n=1,141) and Mexican Americans (y=-.256 p<.01;
n=928). Believing it is not okay to commit mild infractions
in school is a better predictor of Mexican Americans
behavior followed by whites then African Americans. The
negative relationship between smoking cigarettes and
believing it is not okay to commit mild infractions in
school indicate that the higher the respondent is on this
belief indicator their percentage of smoking decreases.

The results indicate there is a moderately strong,
significant, negative association between believing it is
not okay to commit serious infractions in school and
smoking cigarettes for the entire sample (7=-.224; p<.01;
n=10,331), whites (7=-.221; p<.01; n=8,258), African

Americans (y=-.199; p<.01; n=1,144) and Mexican Americans

112

.iaoo. vm new «mo. vm "mam>ma unmowmwcmwm

 

 

 

 

 

 

 

 

 

eoccon
Aemmuce Aeea.fiucv Aemw.mnce Aemmteauce Ce mCOeeomum:a mzoewmm SNEEoc
«xmmm.| *«mme.s «*emm.x «*eNN.I be xnxo no: we m~ uwvw~mm
.:!(i.I:wmomwwj
Aemmuoe Aaee.ance Aeem.encv Aemm.eauce CH mcoeeonawce UNWE nessoo
*«emm.u albea.r «immm.n ««mmm.| op >mxo no: we eH ”mue~vm
manoeum5<

memoexmz cmoflflw< mmbflzk mHQEmm muwocm mHOHMOHUGH HOHHmm

munoaarufloomm -I I..-

 

mcmoaxoz van inducenofid auowumd .mwufié .mdnmaflm

a muormofieca “gleam all as. moiiau.meu ma.ioam to ......»z .aa.w .. ..g.a
no . .

onenem can

113

 

( ;/==-.227 p<.01; n=929). Again, believing it is not okay to
czcermnit serious infractions in school is a better predictor
<>Lff Mexican Americans behavior followed by whites then
.Pxifrican Americans. The negative relationship between
=Srnoking cigarettes and believing it is not okay to commit
Searious infractions in school indicate that the higher the
Irespondent is on this belief indicator their percentage of

Esmoking decreases.

Gamma Measures of.Association and Drinking Alcohol

.Attachment Indicators

The next sets of bivariate analyses explore the
relationship of the independent measures and drinking
alcohol by race. The gamma measure of association is again
used to determine statistical significance, strength and
direction of the association. The overall results indicate
that the social bond measures appear to be better suited to
predict the smoking behavior of adolescents than it is to
predict their drinking behavior as evidenced by the number
of statistically significant associations that were
observed in the smoking data as compared to the drinking
data. Table 10 presents the gamma measures for the
attachment indicators and drinking alcohol for the entir

sample, whites, African Americans, and Mexican Americans.

114

115

*cHoo.flVQ ficm {mo.flvm umH0>¢H unmofiwficmfim

 

 

 

 

 

 

 

 

 

 

:5qu Swaui :Nmsni $351:
(I: ZNNTI ES.» l) .3me immef ..(nwWBune 3 2:35.13 5.
$31: 23.1: Sail: Salami )1 (14.11-31-
-.Iflkmigkilll ilzllSo. 1;on Emma. 22$ 3 212.5%; E
21:35 :31: SNNTE :3 .el: (111-111 (r ( : - -
ems. NS. 130.1 $3 I 3113mm.-.tadflehéCe.
m:noa9b5< mHODMUHUGH
nllmhfixmz fi :noficex mmewza m~Q=Em mhflntlm UCOEUMUUG
)1)(l1(1(|fi)\l!.l hufloflcnum\wumm (t (Ilwilkiiwiiilil(--)

 

 

mcmoflxoz Una imndofiumad :dUHHmd .mmufiSB .mHmEmm
MOHVCH unmEJUMuud Gnu flaw HOSOOH¢ mqfixnfiun Ho Mohammoz MEEMG .OH wanna

are S E as

’ITklea findings indicate attachment to parents, peers and
T:«E:achers show significant associations for the entire
Esaimple and whites. In addition, the findings indicate that
Eit1tachment to peers (7:.072; p<.05; n=804)and teachers (y:-

-i122; p<.01; n=780)were significantly associated to
Cirinking alcohol for Mexican Americans. Finally, attachment
tio teachers was the only significant attachment association
.for African Americans (y=-.075; p<.05; n=922). The findings
.Show that for whites attachment to teachers is the
strongest attachment indicator of the three indicators. In
addition, the findings indicate a negative weak to moderate
strength in association (7=—.136; p<.01; n=7,321) for
whites, meaning that knowing the level of teacher
attachment for white respondents reduces the error in
predicting drinking alcohol by 13.6%. Attachment to
teachers was a stronger measure of association for whites,

followed by Mexican Americans then African Americans.

Commitment and Drinking Alcohol

The association between commitment to school,
continuing one’s education or going to trade school, and
commitment to marriage and success and drinking alcohol
indicates commitment is a stronger predictor of the

drinking behavior of whites than African Americans and

116

EVIeaxican Americans as indicated in Table 11. Specifically,

C3<>mmitment to school is significantly associated with
ciirinking alcohol for the entire sample (y=-.112; p<.01;
r1==9,385), whites (y: -.128; p<.01; n=7,575) and African
.Punericans (y=—.111; p<.01; n=975). However, the findings
:indicate it is a better measure of association for whites
coverall. The findings indicate that knowing the respondents
.level of commitment to school, error is reduced when
goredicting drinking behavior about 13% for whites and 11%
Ifor the entire sample and 11% for African Americans. In
éaddition, the relationship is negative, meaning that as
goercentage of drinking increases, the level of commitment
1:0 school decreases. Commitment to continuing one’s
eaducation or going to trade school is significantly
associated to drinking alcohol for the entire sample (y:-
.026; p<.05; n=8,666) and whites (7=-.026; p<.05; n=7,030).
.Again, the findings indicate that commitment to continuing
one’s education or going to trade school is a better
predictor of respondents’ drinking behavior for whites than
the entire sample. Additionally, the findings indicate that
knowing the respondents’ commitment to continuing their
education or going to trade school, error is reduced when
predicting their drinking behavior about 3% of the time for

the entire sample and for whites. Finally, there is a

117

aeaoo.flvm van «mo.flvm umHm>0H QGMOHMflcmHm

 

 

 

 

 

 

 

 

 

 

 

 

 

3375 Semi: Sail: €351: ii---
Nae.( *Nhe.| ««.Nme.l *«Hme.l mmmoozm. Ucm metwuumi on i5:5.fi:.:ob
Aebence Aeemnce “emebuce Aeee.muce Noozom mbtua on eceou nnfl
OMO. VNO.I <©NO.I «0N0... COHQMUZEN UCflZCfluCOU Cu HZmuEquEOU
$5qu 33.1.5 $37.5 Smmaui -I. is
eee.: «.«HHH... :eNH... «*NHH.I Hoocom Om efllmiifiiziiecuc
mCQOeumE< i

memoexmz :moeum< memes: madame muHoSm muoumerGH usmfiuafifioo

auwowcfiamomm - El
mcmofixoz Ucm .mchHHoafl GdUeHM4 .umuHAE .mamfimm

mweucm How «HouMOHUCH ucoauefifioo on» flow Hosooa¢.mcexcfinn Ho mounmmoz M5800 .HH manna

 

118

negative relationship between continuing one’s education or
going to trade school and drinking alcohol, meaning, as the
percentage of drinking increases, commitment to continuing
one’s education or going to trade school decreases.

Commitment to marriage and success is significantly
associated to drinking alcohol for the entire sample (y:-
.031; p<.01; n=9,417), whites (y=-.032; p<.01; n=7,593),
and African Americans (y=—.072; p<.05; n=982). The findings
indicate that commitment to marriage and success is a
better predictor of the drinking behavior for African
Americans followed by whites. In addition, the findings
indicate that knowing a respondents level of commitment to
marriage and success helps to reduce the error of
predicting drinking behavior of respondents by 3% for the
entire sample and whites and 7% for African Americans.
Also, there is a negative relationship between commitment
to marriage and success. Meaning, that as commitment to
marriage and success decreases, the percentage of smoking
increases. Finally, there was no significant relationship
found between the commitment indicators for Mexican
Americans and drinking alcohol. This is contrary to the
findings on smoking cigarettes where commitment to marriage
was significantly associated to smoking for Mexican

Americans. This finding reiterates the idea the social

119

control model might be a better model to explain the
smoking behavior of all adolescents than it is to explain
their drinking behavior.
Involvement and Drinking Alcohol

The two involvement indicators are involvement in
sports related activities and involvement in other school
related activities (i.e., band, clubs, drama, etc.). The
findings from the gamma measures in Table 12 indicate that
involvement in sports related activities was significantly
associated with drinking alcohol for whites (y=.037; p<.01;
n=7,440) and Mexican Americans (7=-.100; p<.05; n=742)only.
Involvement in other school related sports was not
significantly associated with drinking alcohol for any of
the groups. Involvement in sports related activities is a
better predictor of behavior for Mexican Americans than for
whites. There is a positive association between drinking
alcohol and involvement in sports related activities for
whites. This indicates that as involvement in sports
related activities increases for whites so do their
percentage of drinking. Conversely, there is a negative
relationship between drinking alcohol and involvement in
sports related activities for Mexican Americans. This

indicates that as involvement in sports related activities

121

I. _ee.nv: 1:: i.me...1.v._ 3.9.3. ...:owzfirm

 

 

Ambeuce

 

 

 

 

 

 

 

 

Aofimucv Aemeeuce Ammm.mu:e mmani>nno< tmtnwmm:mam:ww-
hme. ewe.) eee.: flee... .5920 :w m:U~:.D>~C>:~
imeaucc Aammuce immmencc imme.wn:e Newmwwwaue.eunn~me
«eeH.| mee.( ......hme. bee. nenolm. :a ecvzru>~©>:~
m:moeumed. II- . -.-
mcmoexmea :MQMMMe. mmuwcz mNQEmm muwffi mHODMUHUGH UGO§>HO>GH
auaoaaruwxoomm (a--- -

 

 

mcmoaxoz Una .mcmOdHOEd GNOfiuwd .mmuws3 .oamfimm
ouaucm Mom mHODMOfiUGH ucosw>HO>CH on» man Honooad mcflxcfiun mo mmunmmoz MEEMG .NH manna

 

increase for Mexican Americans their percentage of smoking
decreases.
Belief and Drinking Alcohol

The two indicators for belief are (1) believing it is
not okay to commit mild infractions in school and (2)
believing it is not okay to commit serious infractions in
school. The results in Table 13 indicate there is a
moderately strong, significant, negative association
between believing it is not okay to commit mild infractions
in school and drinking alcohol for the entire sample (y:-
.270; p<.01; n=9,382), whites (y=-.277; p<.01; n=7570),
African Americans (y=-.152; p<.01; n=978) and Mexican
Americans (y=—.242 p<.01; n=834). Believing it is not okay
to commit mild infractions in school is a better predictor
of white respondents’ behavior followed by Mexican
Americans then African Americans. The negative relationship
between drinking alcohol and believing it is not okay to
commit mild infractions in school indicate that the higher
the respondent is on this belief indicator their percentage
of drinking decreases.

The results indicate there is a moderately strong,

significant, negative association between believing it is
not okay to commit serious infractions in school and

drinking alcohol for the entire sample (y=-.141; p<.01;

 

 

 

......ee.flv._ 6:: ..me.nva_ H.95— .=au..__=w_m

 

 

Noocom cw

 

«op:fsm

 

 

 

 

 

 

 

Ahmeuce Aemmnce Amemeuce Ambmtmuce mCOauomwmce m20e~vm CHEEo;
..mNH.I ..mo~.- ..oee.| *.Hea.- on ages no: me he ”umeavmi
Aemence Aebmncv Aebeeucv Amemtmncv Ca WCONCQMMMCe Bees massed
«imem.l .«mme.- ..eem.n ..oem.- an ammo no: we um Hemline
mcmoamem
mcmoeme :moeum< mmoeca mHQEmm muflmfim MHOHMOHUGH mafiamm
munoncnum\oomm .

 

munucm 0:»

mcmoflxoz van

.mcMOflHoad GMUHHM4 .mouag3 .mamamm

How muoumofiUGH mofiaom onu Una Hosooad mcfixcdua mo nouamdmz mfiamw .ma manna

123

 

n=9,379), whites (7=-.140; p<.01; n=7562), African
Americans (7=-.108; p<.01; n=980) and Mexican Americans
(y=-.125 p<.01; n=835). Again, believing it is not okay to
commit serious infractions in school is a better predictor
of whites respondents’ behavior followed by Mexican
Americans then African Americans. The negative relationship
between drinking alcohol and believing it is not okay to
commit serious infractions in school indicate that the
higher the respondent is on this belief indicator their

percentage of drinking decreases.

LOGISTIC ANALYSES AND CIGARETTE SMOKING

The multivariate analyses were conducted using
logistic regression, as both dependent variables are
dichotomous variables. The purpose of the multivariate
analyses is to understand the ability of the social control
model to explain the smoking and/or drinking behavior of
African American and Mexican youth as compared to white
youth. Tables 14-17 display the results of the logistic
analyses with cigarette smoking as the dependent variable
for the entire sample, white youth, Mexican youth and
African American youth. Tables 15-21 display the results of

the logistic analyses with drinking alcohol as the

dependent variable for the entire sample, white youth,

Mexican youth and African American youth.

Ho: There is no relationship between attachment (to
parents, teachers, peers), commitment (to school and
furthering education or getting a job), involvement
(in sports related activities and other school—related
activities) and belief (in committing mild and serious
infractions in school)in regards to smoking cigarettes
and drinking alcohol. Moreover, these relationships

will be conditioned by gender, SES, and race.

Entire Sample and Smoking Cigarettes

Table 14 displays the results of the logistic
regression analyses for the entire sample. The results
indicate that the model was significant (Chi-square =
796.490; df=14; p<.01; n=7,804). Majority of the variables
and social control indicators were also significant. The
variables that were not significant were the SES quartile,
attachment to parents and peers and involvement in other
school related activities. The relationships between the
social control variables were mostly negative; therefore,
the results indicate that a one unit increase in the social

control indicators that were significant, results in a

125

coco.
**eeee.
**cooo.
ehwm.
**__oo.
**Oooo.
*+ceoe.
**oeee.
**cooo.
NOON.
m_em.
*m_me.
moem.
**oooo.
**oooe.
Av-m

3.3.. «w.

waem.
«eke.
waver
_x_~.
mam“.
_wmk.
_m_a.
meek.
Nmeor
xkoir
egos.
mama.
cmkmr
case;
SEQ

N_em.
h_e_.
oveo.
came.
m_e_.
name.
Nmmo.
:ee.
N_ec.
ebme.
me__.
mmeo.
meee.
eheo.
eeee.

...m.m

wmeew
wmwmi-
e_e>r
exec.
Nemmr
eebmr
oehmr
mmmmr
eeomr
meee.
eme_.
meeflr
N_mer
meme.
emne.

33m inc—.0

:SeWV: 5:: «meWV: 3.95. ...aa=___w_w

 

eemeZ
......cov.ca\. ..I. u-§_._s.n-___.e

Efinzou
.0058 3 assumes 25:3 5:58 o. 53.: E: 2 : :23:
30:8 3 25:05.5: 2:: 25:59 2 x33 5: m. : $2.9:
moE>=o< 3.e_2-_oo__om 550 E E2:u>_c>:_

mo_:>:o< 3.32-3.25 5 25:52.3...

mmooozm ...:e nestle—2 o. 2.22:5ECU

30.3% 933... o. mEcO .5 22.335 w:_._2:.5.._ c. =§E::_:_o,.e
323m 2 :_u:_::_:.o..e

20:25... 2 EuEzoSE

was; 2 225322

2:95.. o. 22:—.322

5:323. $25.5;

2:333 mum

53:03

33..

"mm._m<_~_<>

U u:EoEm 3:325» 2.35252 2: 5?: 2955 out-...m 2: ...... 2:53— :o_aaa..w3_ 33%...— .e_ 2...;

126

decrease in the likelihood that the respondents would smoke
cigarettes by a multiplicative factor of the data displayed
in the Exp(B) column. Race is positive as indicated by the
odds ratio and significant in this model (odds ratio=.4756;
Exp(B)= 1.6090; p<.01). This finding indicates that for
each unit increase in race there is a positive increase in
smoking behavior. Specifically, this finding indicates that
controlling for all other variables, the model explains the
smoking behavior of whites more than it does for African
Americans or Mexican Americans.

The relationship between gender and smoking was a
positive significant relationship as indicated by the odds
ratio in this model (odds ratio=.4543; Exp(B)= 1.575;
p<.01). This indicates that for each unit increase in
gender there is a positive increase in smoking behavior.
Specifically, this finding indicates that females are more
likely to smoke than males, controlling for all other
variables.

Parent’s education had a significant negative effect
on smoking as indicated by the odds ratio (odds ratio=
-.lOO3; Exp(B)=.9046; p<.05). This indicates for each unit
increase in the parent’s education level there is a
decrease in smoking behavior. Specifically, this finding

indicates that the more education a respondent’s parent has

127

the less likely that individual is to smoke, controlling
for all other variables.

Attachment to teachers also shows a significant
effect. Attachment to teachers has a significant negative
effect on smoking as indicated by the odds ratio (odds
ratio: -.2906; Exp(B)=.7478; p<.01). This indicates for
each unit increase in the attachment to teachers there is a
decrease in smoking behavior. Specifically, this finding
indicates that the more a respondent is attached to a
teacher a respondent is less likely to smoke, controlling
for all other variables.

Commitment to school also shows a significant
relationship to smoking. Commitment to school has a
significant negative effect on smoking as indicated by the
odds ratio (odds ratio: -.3353; Exp(B)=.7lSl; p<.01). This
indicates for each unit increase in commitment to school
there is a decrease in smoking behavior. Specifically, this
finding indicates that the more a respondent is committed
to school, s/he is less likely to smoke, controlling for
all other variables.

Commitment to furthering education or going to trade
school shows a significant relationship to smoking.
Commitment to furthering education or going to trade school

has a significant negative effect on smoking as indicated

by the odds ratio (odds ratio: —.2769; Exp(B)=.7581;
p<.01). This indicates for each unit increase in the
Commitment to furthering education or going to trade school
there is a decrease in smoking behavior. Specifically, this
finding indicates that the more a respondent is committed
to furthering his/her education or going to trade school
s/he is less likely to smoke, controlling for the other
variables.

Commitment to marriage and success shows a significant
relationship to smoking. Commitment to marriage and success
has a significant negative effect on smoking as indicated
by the odds ratio (odds ratio: —.2760; Exp(B)=.7588;
p<.01). This indicates for each unit increase in the
commitment to marriage and success there is a decrease in
smoking behavior. Specifically, this finding indicates that
the more a respondent is committed to marriage and success
s/he is less likely to smoke, controlling for the other
variables.

Involvement in sports-related activities shows a
significant relationship to smoking. Involvement in sports-
related activities has a significant negative effect on
smoking as indicated by the odds ratio (odds ratio: -.2760;

Exp(B)=.7588; p<.01). This indicates for each unit increase

in involvement in sports-related activities there is a

129

 

 

decrease in smoking behavior. Specifically, this finding
indicates that the more a respondent is involved in sports-
related activities s/he is less likely to smoke,
controlling for all other variables.

Believing it is not okay to commit mild infractions in
school shows a significant relationship to smoking.
Believing it is not okay to commit mild infractions in
school has a significant negative effect on smoking as
indicated by the odds ratio (odds ratio= -.74l4;
Exp(B)=.4764; p<.01). This indicates for each unit increase
in believing it is not okay to commit mild infractions in
school there is a decrease in smoking behavior.
Specifically, this finding indicates that the more a
respondent believes it is not okay to commit mild
infractions in school the less likely s/he will smoke,
controlling for all other variables.

Believing it is not okay to commit serious infractions
in school shows a significant relationship to smoking.
Believing it is not okay to commit serious infractions in
school has a significant negative effect on smoking as
indicated by the odds ratio (odds ratio= -l.2738;
Exp(B)=.2798; p<.01). This indicates for each unit increase
in believing it is not okay to commit serious infractions

in school there is a decrease in smoking behavior.

130

 

Specifically, this finding indicates that the more a
respondent believes it is not okay to commit serious
infractions in school the less likely s/he will smoke,
controlling for all other variables.

In sum, females, low parental education, and
respondents who were low on the social control indicators
were more likely to smoke. Belief in committing mild
infractions in school had the most effect on the smoking
behavior of youth in the sample as indicated by the Wald
statistic (Wald = 134.0372), which was highest among the

social control indicators.

Whites and Smoking'Cigarettes

Table 15 displays the results of the logistic
regression analyses for whites. The results for the white
respondents indicate that the model was significant (Chi-
square = 630.604; df=13; p<.01; n=6,390). Gender and many
of the social control indicators showed significance for
the white respondents. Attachment to parents and peers and
involvement in other school related activities were the
social control indicators that were not significant. The
relationships between the social control variables and
smoking cigarettes were all negative; therefore, the

results indicate that a one unit increase in the social

 

oooc.
4*o0oo.
**oooo.
doom.
*@_oo.
**oooo.
**ooco.
**oooo.
**ooco.
mocm.
who~.
mono.
Omom.
**ooco.
Av—m

Boom.
Noam.
mmmci
vm_s.
amok.
wmms.
ems».
ewes.
oases
smm_o
sews.
cams.
_moeu
£15.

doom.
mem_.
mace.
mvoc.
awo_.
cmoc.
some.
ammo.
mmcc.
coca.
o_m_.
cove.
cave.
v_>o.

dud

mwxfim
mmmmg-
swap;
ammo.
mummr
mocmv
cmwmr
Nmmmv
mwwmr
coho.
bwc_.
~w>ov
mmoor
omhv.

35. 3:5

«...—acrhva _::.. ..m...HV: "29>“: 2.35:3m

 

candlz
ivccdmc I 9,522-3;

:zzazflu
_oo;om E 228295 95:3 3:53 3 5,23 5: fl : :23:
75:3. 3 mcotosfifi 2:: 5:58 o. bio 3: mm : ”bozo:
3:333.» BEBHAoozom 550 E 255232:

moE>co< co.a_u._-.£._oam 5 2.925232:

3325 cc: swabs—2 2 EuEzEEou

Bozom 25.51 0. mEoO ..o coteosnm w:_._2_::,.._ o. ...2:::_:_:U
_Aoiss_q.mu as. ...“u.__._______.ohno

$2.93.“. 3 :_u___._oa:<

23.. o. :5::_om:<

2:23.. 2 32:333.

.5335: ”2:95..

9.3.530 mzm

coczomV

"mm—J:<_~_<>

933.53 ...o—ogaaa 2: a... antennas mixes—m a...“ $555 ..8 31:3: :o_mmu..wu~_ 333:4 .m_ “...—a...

132

control indicators that were significant, results in a
decrease in the likelihood that the respondents would smoke
cigarettes by a multiplicative factor of the data displayed
in the Exp(B) column.

The relationship between gender and smoking was a
positive significant relationship as indicated by the odds
ratio (odds ratio=.4720; Exp(B)= 1.6031; p<.01) in this
model. This indicates that for each unit increase in gender
there is a positive increase in smoking behavior.
Specifically, this finding indicates that females are more
likely to smoke than males, controlling for other
variables.

Attachment to teachers show a significant relationship
to smoking. Attachment to teachers has a significant
negative effect on smoking as indicated by the odds ratio
(odds ratio= -.2673; Exp(B)=.7654; p<.01). This indicates
for each unit increase in the attachment to teachers there
is a decrease in smoking behavior. Specifically, this
finding indicates that the more a respondent is attached to
a teacher a respondent is less likely to smoke, controlling
for all other variables.

Commitment to school also shows a significant
relationship to smoking. Commitment to school has a

significant negative effect on smoking as indicated by the

odds ratio (odds ratio= —.3532; Exp(B)=.7024; p<.01). This
indicates for each unit increase in commitment to school
there is a decrease in smoking behavior. Specifically, this
finding indicates that the more a respondent is committed
to school a respondent is less likely to smoke, controlling
for all other variables.

Commitment to furthering one’s education or going to
trade school shows a significant relationship to smoking.
Commitment to furthering one’s education or going to trade
school has a significant negative effect on smoking as
indicated by the odds ratio (odds ratio= —.2826;
Exp(B)=.7538; p<.01). This indicates for each unit increase
in the commitment to furthering education or going to trade
school there is a decrease in smoking behavior.
Specifically, this finding indicates that the more a
respondent is committed to furthering his/her education or
going to trade school s/he is less likely to smoke,
controlling for all other variables.

Commitment to marriage and success shows a significant
relationship to smoking. Commitment to marriage and success
has a significant negative effect on smoking as indicated
by the odds ratio (odds ratio= —.2695; Exp(B)=.7638;
p<.01). This indicates for each unit increase in the

commitment to marriage and success there is a decrease in

 

A.-

 

smoking behavior. Specifically, this finding indicates that
the more a respondent is committed to marriage and success
s/he is less likely to smoke, controlling for all other
variables.

Involvement in sports-related activities shows a
significant relationship to smoking. Involvement in sports—
related activities has a significant negative effect on
smoking as indicated by the odds ratio (odds ratio= -.3378;
Exp(B)=.7134; p<.05). This indicates for each unit increase
in involvement in sports-related activities there is a
decrease in smoking behavior. Specifically, this finding
indicates that the more a respondent is involved in sports-
related activities s/he is less likely to smoke,
controlling for all other variables.

Believing it is not okay to commit mild infractions in
school shows a significant relationship to smoking.
Believing it is not okay to commit mild infractions in
school has a significant negative effect on smoking as
indicated by the odds ratio (odds ratio= -.6927;
Exp(B)=.5002; p<.01). This indicates for each unit increase
in believing it is not okay to commit mild infractions in
school there is a decrease in smoking behavior.
Specifically, this finding indicates that the more a

respondent believes it is not okay to commit mild

135

 

 

infraction in school the less likely s/he will smoke,

controlling for all Other variables.

Believing it is not okay to commit serious infractions
in school shows a significant relationship to smoking.
Believing it is not okay to commit serious infractions in
school has a significant negative effect on smoking as
indicated by the odds ratio (odds ratio= —1.2353;
Exp(B)=.2907; p<.01). This indicates for each unit increase
in believing it is not okay to commit serious infractions
in school there is a decrease in smoking behavior.
Specifically, this finding indicates that the more a
respondent believes it is not okay to commit serious
infractions in school the less likely s/he will smoke,
controlling for all other variables.

In sum, the social control model controlling for white
respondents indicate females and respondents who were low
on the social control indicators were more likely to smoke.
Belief in committing mild infractions in school had the
most effect on the smoking behavior of youth in the sample
as indicated by the Wald statistic (Wald = 102.912), which

was highest among the social control indicators.

1,...)
(A)
O\

 

 

 

African American Ybuth and Cigarette Smoking

Table 16 displays the results of the logistic
regression analyses for African Americans and cigarette
smoking. The results for the African American respondents
indicate the model was significant for African American
youth (Chi-square = 44.062; df=10; n=l,324; p<.01). This
significant result for the model is not unusual given the 1
sample size. However, there is a marked difference in the
significant findings of the social control indicators for
African American youth as compared to white youth. For
African American youth only three of the social control
indicators were significant for smoking cigarettes, as
compared to white youth, where all, but three, were
significant.

Involvement in other school related activities shows a
significant relationship to smoking. Involvement in other
school related activities has a significant negative effect
on smoking as indicated by the odds ratio (odds ratio= —
.9126; Exp(B)=.4015; p<.05). This indicates for each unit
increase in involvement in other school related activities
there is a decrease in smoking behavior. Specifically, this
finding indicates that the more a respondent is involved in
other school related activities s/he is less likely to

smoke, controlling for all other variables.

 

 

.omo.
.32...
.1525.
......mo.
:5...
Sue...
voxm.
83.
NNE.
2.2.
awe...
Max...
9%...
one.
...—m

 

~.s..
mma..
m.oe.
e.cs.
ease.
coex.
«New.
e..s.
case.
s.ss.
ewes.
memc_.
”Rem
5......

ommh.m
m3...
mmmm.
m2...
.mmm.
mwmm.
amen.
0......
mam.
mmvm.
mamm.
cmcm.
Saw.
3%.».

......m

mmvow
memm.-
mw.c.-
cN.or
.ccm.
wwvmr
mmc.r
seem.
cmoor
.oMm.
owmmr
oovmr
cove.
coon.

2.3. «.50

«...—.....V: ...... «m=.v.. 3.9.5. 2:35:37.

 

..mm M Z
3:...3. A 2.21%....

25.25...
59.8 E 95.53.... 25.5w ...:ECo o. .35 5: m. .. ”.23..
59.8 ... 5255...... 3.5 ...:Eoo o. .35 5: m. .. ”.23..
mo...>..o< co....u.-.02_om .223 ... 2.2.52.3...

mo...>..o< 5222-2..25 ... ...u...u>_o>:.

$32....” ...... amateE o. 255:5:59

.oo..om 0.52% o. $.50 .5 :o..mo...om. m.....u....:£ o. ...».....:_:5..
....o..om o. 25:37:53..

22.3.5... 3. ...2....o.:.<

23.. o. Eu::.o£.<

329...... o. ...o::.o..:<

23:325.. n....o......

9.2-520 mam
5.59.. .

"mm—J._<.=<>

o...........> 2.9.52.3. a... 2. “5.5.3.5 33.520 .....5 2.3.5:}. 53.....< 5.. 3.33. .....maotwa: 23......— 5. 9.2:...

l38

 

 

Believing it is not okay to commit mild infractions in
school shows a significant relationship to smoking.
Believing it is not okay to commit mild infractions in
school has a significant negative effect on smoking as
indicated by the odds ratio (odds ratio= -l.6183;
Exp(B)=.l982; p<.01). This indicates for each unit increase
in believing it is not okay to commit mild infractions in
school there is a decrease in smoking behavior.
Specifically, this finding indicates that the more a
respondent believes it is not okay to commit mild
infraction in school the less likely s/he will smoke,
controlling for all other variables.

Believing it is not okay to commit serious infractions
in school shows a significant relationship to smoking.
Believing it is not okay to commit serious infractions in
school has a significant negative effect on smoking as
indicated by the odds ratio (odds ratio= -l.6545;
Exp(B)=.l9lZ; p<.05). This indicates for each unit increase
in believing it is not okay to commit serious infractions
in school there is a decrease in smoking behavior.
Specifically, this finding indicates that the more a
respondent believes it is not okay to commit serious
infractions in school the less likely s/he will smoke,

controlling for all other variables.

 

 

 

In sum, the social control model controlling for

African American respondents indicate those who were low on
the involvement in other school related activities and low
on the belief indicators were more likely to smoke. Belief
in committing mild infractions in school had the most
effect on the smoking behavior of African American youth as
indicated by the Wald statistic (Wald = 20.931), which was

highest among the social control indicators.

Mexican American YOuth and Cigarette Smoking

Table 17 displays the results of the logistic
regression analyses for Mexican youth. The findings on
Mexican Americans indicate that the model was statistically
significant (Chi-square = 82.486; df=13; n=656; p<.01).
Again, this was expected, as the sample size for Mexican
youth is large (n=656). The findings indicate there were
five social control indicators that were statistically
significant (i.e., attachment to teacher, commitment to
school, involvement in sports related activities, Believing
it is not okay to commit mild infractions in school, and
Believing it is not okay to commit serious infractions in
school). Overall, the model explains the behavior of

Mexican youth better than African American youth. However,

t...)
41:
C)

oooo.

+vooo.

**Nooo.

vomw.
*mwmo.
Oboe.
x.mo.
*wmoo.
*whoo.
vam.
.omh.
oooo.
hmoc.
.v.m.

.rv.m

vmxm.
.N.v.

mmxo..

cmwm.
o.wo.
omwo.
.owm.
womm.
mm.x.
oonw.
.xmo.

owns.
veem.
5......

mono.

moxv.
.ovm.
mmvv.
oo.c.
mo.m.
woom.
Nxo..
mmmm.
m~.m.
momv.
..mh..
m.om.
ommm.

dfim

.ooNAW

comm.
ammo.

mvwm.
Nn.o.
mon.
.mmo.
voom.
ooN..
ovno.

onv~.-

cmo~.-

wono.
ow.m.

2.3. ......o

{i—OS. VA— 5:1. ...m=. V: "39>“: u=39m=__w_.f.

 

cm... 4 Z
1.8:»? I. 2.2.7.9....5

.zs;::..
.oocom ... 25.5.2.5. 25:3 .5555 o. .92: 5: m. : H.252
.oozom ... $522.5: 3.5 .558 o. >55 5.. m. .. 5.5m.
.5...>.5< 35.2-1545 5.23 ... 2525232..

.5...>.5< 35.25.25 ... 55:53.3...

mmooo..m .5... 53:53. 5. ..55...::5Q

.oo..om 55.2... o. 9.53 5 5252...: $5.555... 3. 25:55:59
52.5 o. ..5.....:........

”5.5.5... ... ..55..o..:<

$55.. 3. ..5.....o..:<

3.5:... 2 ..5.....on:<

:o...5.....n. n...5......

5.5.520 mum

5.55..

.m.....:<.~.<>

2.5.55» 25.2.2.2. a... an ”5.5.55.0 35.5wfiv ......» 5.5555. ...... 3.55: :5_mmo..wu~_ 5.5%.... H. 52......

141.

 

 

the model does not explain the behavior of Mexican youth as
well as it does for white youth.

Attachment to teachers shows a significant
relationship to smoking. Attachment to teachers has a
significant negative effect on smoking as indicated by the
odds ratio (odds ratio= -.6221; Exp(B)=.5368; p<.05). This
indicates for each unit increase in the attachment to
teachers there is a decrease in smoking behavior.
Specifically, this finding indicates the more a respondent
is attached to a teacher a respondent is less likely to
smoke, controlling for all other variables.

Commitment to school also shows a significant
relationship to smoking. Commitment to school has a
significant negative effect on smoking as indicated by the
odds ratio (odds ratio= -.5463; Exp(B)=.579l4; p<.05). This
indicates for each unit increase in commitment to school
there is a decrease in smoking behavior. Specifically, this
finding indicates the more a respondent is committed to
school s/he is less likely to smoke, controlling for all
other variables.

Involvement in sports-related activities shows a
significant relationship to smoking. Involvement in sports-
related activities has a significant negative effect on

smoking as indicated by the odds ratio (odds ratio= -

F4
J3
l\)

 

 

1.2636; Exp(B)=.2826; p<.05). This indicates for each unit
increase in involvement in sports—related activities there
is a decrease in smoking behavior. Specifically, this
finding indicates that the more a respondent is involved in
sports—related activities s/he is less likely to smoke,
controlling for all other variables.

Believing it is not okay to commit mild infractions in
school shows a significant relationship to smoking.
Believing it is not okay to commit mild infractions in
school has a significant negative effect on smoking as
indicated by the odds ratio (odds ratio= —.8866;
Exp(B)=.4121; p<.01). This indicates for each unit increase
in believing it is not okay to commit mild infractions in
school there is a decrease in smoking behavior.
Specifically, this finding indicates that the more a
respondent believes it is not okay to commit mild
infraction in school the less likely s/he will smoke,
controlling for all other variables.

Believing it is not okay to commit serious infractions
in school shows a significant relationship to smoking.
Believing it is not okay to commit serious infractions in
school has a significant negative effect on smoking as

-l.2470;

indicated by the odds ratio (odds ratio

Exp(B)=.2874; p<.05). This indicates for each unit increase

 

 

in believing it is not okay to commit serious infractions
in school there is a decrease in smoking behavior.
Specifically, this finding indicates that the more a
respondent believes it is not okay to commit serious
infractions in school the less likely s/he will smoke,
controlling for all other variables.

In sum, the social control model controlling for
Mexican respondents indicate respondents who were low on
the social control indicators were more likely to smoke.
Belief in committing mild infractions in school had the
most effect on the smoking behavior of youth in the sample
as indicated by the Wald statistic (Wald = 13.6298), which

was highest among the social control indicators.

LOGISTIC ANALYSES AND DRINKING ALCOHOL

Entire Sample and Drinking Alcohol

Table 18 displays the results of the logistic analyses
for the entire sample. The results indicate that the model
was significant (Chi-square = 796.490; df=l4; p<.01;
n=7,804). Many of the social control indicators were
significant. The relationships between the social control
variables were mostly negative; therefore, the results
indicate that a one unit increase in the social control

indicators that were significant, results in a decrease in

144

oooo.
*o.oo.
**oooo.
o.mm.
m.mo.
wvvw.
oo.m.
+ohoo.
**oooo.
+*oooo.
ovco.
oooo.
oom_.
moon.
...:m

mv.c.
comm.
osmo.
cmmo..
ammo.
.ooo.
.mmm.
vema.
.e.e.
vmmw.
coco.
ammo.
memo.
52.5

mmvm.
nom..
owmo.
oooo.
owoo.
wovo.
oovo.
womo.
onvo.
mmvo.
Nooo.
.mmo.
memo.
«.mo.

dfim

vmvm.u
ohxvr
omoo.-
.mcor
ommo.
No.ov
movov
omm.r
w.wmr
movm.
vmw.r
mvmor
«Nmor
mvvor

5.3. 253

«a.oo.v.. .2... «mod... 53>». 2:3...5wi

 

2.2.512
1o..m..oo n o...._....n-....o

2::£=xo
222.8 2. 50.83.... 5.3.3 2558 0. 2.3.0 .o: m. .. ooze:
39.8 2. 2.0.55.5 2.5 2553 o. >53 .3: n. .. ”.23:
33.3.9... 3.3.0..-.023m .220 ... 2.2.52.5...

3.238.... 3.22-3:am ... 25:53.5...

$800.5 1:: “flies—2 o. 2.0:...553U

.oo..um 22:? o. wioO .5 20.2353. $2.22.... 5 25:55:53
.......n.m. ... ...“..._..._.._.a.~o

m..u._oau... o. 2.25.32...

28.. c. 2.2:..2..:<

3.5:... 5 255.3326.

5.2.32.2. 32.21...

u=:g:~.m.m
:oo:uov

.m:J:<—~.<>

9.22.5.» Eco—.23: o... a... 5:32 95.5.3— 5...» «Es—am 955. o... .5. 2.33. 52.39.23. 33...... .2. 92......

145

 

the likelihood that the respondents would drink alcohol by
a multiplicative factor of the data displayed in the Exp(B)
column, for the negative relationships. None of the control
variables displayed any significant relationships to
drinking alcohol.

Attachment to peers shows a significant positive
relationship to drinking alcohol as indicated by the odds
ratio (odds ratio= .3465; Exp(B)= 1.4141; p<.01). This
indicates for each unit increase in the attachment to peers
there is an increase in drinking behavior. Specifically,
this finding indicates that the more a respondent is
attached to peers the more likely s/he is to drink,
controlling for all other variables.

Attachment to teachers shows a significant
relationship to drinking alcohol. Attachment to teachers
has a significant negative effect on drinking alcohol as
indicated by the odds ratio (odds ratio= -.2818;
Exp(B)=.7544; p<.01). This indicates for each unit increase
in the attachment to teachers there is a decrease in
drinking behavior. Specifically, this finding indicates
that the more a respondent is attached to a teacher a
respondent’s is less likely to drink alcohol, controlling

for all other variables.

 

 

Commitment to school also shows a significant
relationship to drinking. Commitment to school has a
significant negative effect on drinking as indicated by the
odds ratio (odds ratio= —.1589; Exp(B)=.8531; p<.05). This
indicates for each unit increase in commitment to school
there is a decrease in drinking behavior. Specifically, .2
this finding indicates that the more a respondent is
committed to school a respondent is less likely to drink
alcohol, controlling for all other variables.

Believing it is not okay to commit mild infractions in
school shows a significant relationship to drinking.
Believing it is not okay to commit mild infractions in
school has a significant negative effect on drinking as
indicated by the odds ratio (odds ratio= -1.00S;
Exp(B)=.3660; p<.01). This indicates for each unit increase
in believing it is not okay to commit mild infractions in
school there is a decrease in drinking alcohol.
Specifically, this finding indicates that the more a
respondent believes it is not okay to commit mild
infractions in school the less likely s/he will drink
alcohol, controlling for all other variables.

Believing it is not okay to commit serious infractions
in school shows a significant relationship to drinking

alcohol. Believing it is not okay to commit serious

infractions in school has a significant negative effect on
drinking alcohol as indicated by the odds ratio (odds
ratio= -1.2738; Exp(B)=.2798; p<.01). This indicates for
each unit increase in believing it is not okay to commit
serious infractions in school there is a decrease in
drinking alcohol. Specifically, this finding indicates that
the more a respondent believes it is not okay to commit
serious infractions in school the less likely s/he will
drink, Controlling for all other variables.

In sum, respondents who were low on the social control
indicators (that were significant) were more likely to
drink. The exception to this was the attachment to peers,
as high indicators of peer attachment were associated with
reSpondents who were more likely to drink. Believing it is
not okay to commit mild infractions in school had the most
effect on the drinking behavior of youth in the sample as
indicated by the Wald statistic (Wald = 294.233), which was

highest among the social control indicators.

Whites and Drinking'Alcohol

Table 19 displays the results of the logistic analyses
for whites and drinking alcohol. The results for the white
respondents indicate that the model was significant (Chi-

sguare = 608.13; df=13; p<.01; n=5,878). Many of the social

 

 

coco.
*on.c.
*+cooo.
hcm..
*vmmo.
mmhw.
amhm.
*.Nco.
**oooo.
**ococ.
*cmmo.
amom.
hNVm.
vmmo.
shy—m

some.
ocmm.
whom.
Nmo..
m.oo.
mono.
cwox.
mcvm.
cmom.
mmwn.
meo.
wvmo.
Nmoo.

5...»...

wwoc.
mom..
wvco.
mono.
muse.
mmmo.
xmvo.
.mco.
mmmo.
vows.
mm...
come.
move.
come.

dfim

comma
35?.
:8. . -
08. .-
82.
$59.
8.5..
mgmr
mmomr
Naov.
wcmmr
mover
mvmo.
£59..

5.5: «3.0

.....ScV: .2... ...maV: 3.95. ...:oc._.w.m

 

at} n z
:3: ...... u 2.5.51.5

555.5

59.8 5 5.5.53.5. 555m 2555. o. >9... 5: m. a. H.255
59.3. 5 5253.5 2.5 55.5.. o. .35 5.. m. 2 ”.252
33.356. 35.0..-.oo..om 5.20 :. 555352..

53.22;: 35.2-5..on ... 25:5>_o>:.

35.5.5 .5: $555.). 3. 255.2559

.oosom 5.59.5 o. $.50 5 25.525”... 95.5.2.5... o. 255.....5ou
55.5 o. 25:55:59

@5225... o. 255.552

$59. 5 255.552

3.5.5; 3 255.55%

5.5.3.5.»... 925......

2.23:0 vim
5.253

"mme<_~—<>

2:525» 25.25::— of m... .2153. 9.35.3— 5.3 5:55 .5. ...—:5: =o.a.no..uo~_ 5.53:1. .3 52:...

149

 

control indicators showed significance for the white
respondents in relationship to drinking alcohol. However,
none of the control variables showed significance. The
relationships between the social control variables were
mostly negative; therefore, the results indicate that a one
unit increase in the social control indicators that were
significant, results in a decrease in the likelihood that
the respondents would smoke cigarettes by a multiplicative
factor of the data displayed in the Exp(B) column, for the
negative relationships.

Attachment to parents shows a significant relationship
to drinking. Attachment to parents has a significant
negative effect on drinking alcohol as indicated by the
odds ratio (odds ratio= -.2568; Exp(B)=.7735; p<.05). This
indicates for each unit increase in the attachment to
parents there is a decrease in drinking alcohol.
Specifically, this finding indicates that the more a
respondent is attached to a parent a respondent is less
likely to drink alcohol, controlling for all other
variables.

Attachment to peers shows a significant positive
relationship to drinking alcohol as indicated by the odds
ratio (odds ratio= .4092; Exp(B)= 1.506; p<.01). This

indicates for each unit increase in the attachment to peers

 

 

 

there is an increase in drinking behavior. Specifically,
this finding indicates that the more a respondent is
attached to peers the more likely s/he is to drink,
controlling for all other variables.

Attachment to teachers shows a significant
relationship to drinking alcohol. Attachment to teachers
has a significant negative effect on drinking as indicated
by the odds ratio (odds ratio= —.2923; Exp(B)=.7465;
p<.01). This indicates for each unit increase in the
attachment to teachers there is a decrease in drinking
alcohol. Specifically, this finding indicates that the more
a respondent is attached to a teacher a respondent is less
likely to drink alcohol, controlling for the other
variables.

Commitment to school also shows a significant
relationship to drinking alcohol. Commitment to school has
a significant negative effect on drinking as indicated by
the odds ratio (odds ratio= -.2125; Exp(B)=.8086; p<.05).
This indicates for each unit increase in commitment to
school there is a decrease in drinking alcohol.
Specifically, this finding indicates that the more a
respondent is committed to school a respondent is less
likely to drink alcohol, controlling for the other

variables.

}_.|
(J1
F-‘

 

Involvement in sports-related activities shows a
significant relationship to drinking alcohol. Involvement
in sports—related activities has a significant positive
effect on drinking alcohol as indicated by the odds ratio
(odds ratio= .1503; Exp(B)=1.l622; p<.05). This indicates
for each unit increase in involvement in sports—related
activities there is an increase in drinking behavior.
Specifically, this finding indicates that the more a
respondent is involved in sports-related activities s/he is
more likely to drink alcohol, controlling for the other
variables.

Believing it is not okay to commit mild infractions in
school shows a significant relationship to drinking
alcohol. Believing it is not okay to commit mild
infractions in school has a significant negative effect on
drinking as indicated by the odds ratio (odds ratio=-
1.0311; Exp(B)=.3566; p<.01). This indicates for each unit
increase in believing it is not okay to commit mild
infractions in school there is a decrease in smoking
behavior. Specifically, this finding indicates that the
more a respondent believes it is not okay to commit mild
infractions in school the less likely s/he will drink,

controlling for all other variables.

152

 

 

 

 

, m_ _- _._____

Believing it is not okay to commit serious infractions
in school shows a significant relationship to drinking
alcohol. Believing it is not okay to commit serious
infractions in school has a significant negative effect on
drinking as indicated by the odds ratio (odds ratio= -
.4065; Exp(B)=.6660; p<.05). This indicates for each unit
increase in believing it is not okay to commit serious
infractions in school there is a decrease in drinking
behavior. Specifically, this finding indicates that the
more a respondent believes it is not okay to commit serious
infractions in school the less likely s/he will drink,
controlling for all other variables.

In sum, the social control model controlling for white
respondents indicate females and respondents who were low
on the social control indicators were more likely to smoke.
Belief in committing mild infractions in school had the
most effect on the smoking behavior of youth in the sample
as indicated by the Wald statistic (Wald = 253.23), which

was highest among the social control indicators.

African American Youth and Drinking Alcohol
Table 20 displays the results of the logistic
regression analyses for African Americans and drinking

alcohol. The results for the African American respondents

153

3:5.V: =5... ¢m=.v._ "£95. ....5=_..w_m

 

can .I Z

.3: . _ m u 2.2.31.3

Eco. mmmc.~ 32: 232.2;
cot. momm. m _ vw. SS..- .958 E mcocoacE 25:3 3:58 o. bio E: m. : “.53:
1.35:. 2.3. coom. comm; 15:8 E 20:03.2. 3:: 25:59 9 >33 5: fl : gaze:
com? Nwmx. cwmm. 332.. 323.3» co.m_o.-_oo;om 550 E ...»:5232;
Sam. mph. vcmm. Smmr muE>co< 3.53-3.25 E 22:023.;
mmmm. womm. Em _. cw: .- 3.325 9.... Emits—2 o. Eu::_:::.5
mm _ w. cmmo; .3 E. 3.3. _oo._um 2:29 2 3.60 E 52335 w:_..2:._:,._ 2 22:35.5;
2.2.. hm _ n. Rx _. mvmmr .223m 2 :_2:::::cc
xmmv. vowx. mmf. new..- £2.35... 2 525.332.
«59%. come; eee.. was. ”.59; o. 52:53?
v3.9 .23“. :3. xi d .- 3:21... 3 2.25.932
swam. Exam. 32. 3.2.. :ocnoztm 9:55..
vim. v22: QUE. ammo. 2:330 mum
mmmm. vmow. w _ 3. am. _ .- 51:03
.Em Evens ...; .53. 2:5 um.n._._=<=:..>

933:5» 32:53: a... as .2752 9.35.;— 53, m:3_.55< 53...?» ..8 3.33: :o_mma..wa~_ 33w...— .=N 33:...

154

 

 

indicate the model was significant for African American
youth (Chi-square=31.149; df=13; n=666; p<.05). Again, this
significant result for the model is not unusual given the
sample size. However, there is a marked difference in the
significant findings of the social control indicators for
African American youth as compared to white youth. For
African American youth there was only one social control
indicator that was significant for drinking alcohol, as
compared to white youth, where there were many more social
control indicators significant.

Believing it is not okay to commit mild infractions in
school shows a significant relationship to drinking alcohol
for African Americans. Believing it is not okay to commit
mild infractions in school has a significant negative
effect on drinking alcohol as indicated by the odds ratio
(odds ratio= -1.6183; Exp(B)=.l982; p<.01). This indicates
for each unit increase in believing it is not okay to
commit mild infractions in school there is a decrease in
smoking behavior. Specifically, this finding indicates that
the more a respondent believes it is not okay to commit
mild infractions in school the less likely s/he will drink
alcohol, controlling for all the other variables. Belief in
committing mild infractions in school had the most effect

on the drinking behavior of African American youth as

155

 

indicated by the Wald statistic (Wald = 12.46), which was

highest among the social control indicators.

beican American Youth and Drinking'Alcohol

Table 21 displays the results of the logistic
regression analyses for Mexican youth. The findings on
Mexican youth indicate that the model was statistically
significant (Chi-square = 56.537; df=13; n=592; p<.01).
Again, this was expected, as the sample size for Mexican
youth is large (n=592). The findings indicate there were
two social control indicators that were statistically
significant (i.e., involvement in sports related activities
and believing it is not okay to commit mild infractions in
school). Overall, the model explains the drinking behavior
of Mexican youth better than African American youth.
However, the model does not explain the drinking behavior
of Mexican youth as well as it does for white youth.

Involvement in sports—related activities shows a
significant relationship to drinking. Involvement in
sports—related activities has a significant negative effect
on drinking as indicated by the odds ratio (odds ratio= -
.5873; Exp(B)=.5558; p<.05). This indicates for each unit
increase in involvement in sports—related activities there

is a decrease in drinking behavior. Specifically, this

}__|
U1
0\

hvo_.
mcmm.
**ocoo.
coon.
*VNNO.
cowo.
0—wh.
vcw_.
0—5..
Nm_m.
wvvm.
vmmm.
meB.
Nwo_.
.U—m

mmme._

some. c_we.
mxcm. mm_m.
boom._ xo_m.
wmmm. mama.
mesa. mmm_.
ammo._ mem_.
emcms eee..
case. mmc_.
_ms_.i eee..
mmavr amvm.
.mm_r xw__.
ocvo._ wwm_.
omen. eee..
Eves. a;

GmoN
gov:
Nwoor
ZEN.
mummr
359.
N58.
emmm.
Ncmmr
m _ 2.
mmov.
mm. _.
cmvo.
wwmmr
2.3— 2.10

.2. :5. V.— 3.: «ms. V: "£95. 2.35231.

 

mam.1~4
Immmom n 95:73:;

23235

.oozom E 32895: 25:3 35:8 2 bio .o: .n_ : “.3292
Boson 3 20:25.: 2:: 55:09 o. 533 .3: fl : :23:
mo_:>_.o< 3.a_9._-_oo__om .950 5 2559232..

823:5. boa—2-2.01.0 E 3925332;

mmooozm can omits—2 c. :BEzEEoc

_oo;um aback o. wEoC co cosmozum $592.5... 2 :SEzEEoQ
15.—om o. Eu::_:_:_:.v

€2.23... 2 :_D:_._um:<

23.. o. Eu::_us:<

3:95; 2 E95533

:ozaozom 9:53:

2:520 mmm

.5350

"mm—‘:_<_~_<>

o_:a_..a> Eat—5:9: 2: an .2152. wig—EL: 5?: 32:25: ..8 3:53— ==_.nm9.w3_ 233.... ._~ 93:...

157

 

 

finding indicates that the more a respondent is involved in
sports-related activities s/he is less likely to drink
alcohol, controlling for all other variables.

Believing it is not okay to commit mild infractions in
school shows a significant relationship to drinking
alcohol. Believing it is not okay to commit mild
infractions in school has a significant negative effect on
drinking alcohol as indicated by the odds ratio (odds
ratio= -.8866; Exp(B)=.4121; p<.01). This indicates for
each unit increase in believing it is not okay to commit
mild infractions in school there is a decrease in drinking
alcohol. Specifically, this finding indicates that the more
a respondent believes it is not okay to commit mild
infractions in school the less likely s/he will drink
alcohol, controlling for all other variables.

In sum, the social control model controlling for
Mexican respondents indicate respondents who were low on
the social control indicators that were significant were
more likely to drink alcohol. Believing it is not okay to
commit mild infractions in school had the most effect on
the drinking behavior of youth in the sample as indicated
by the Wald statistic (Wald = 21.905), which was highest

among the social control indicators.

158

 

Discussion

This research seeks to know does the social control
theory explain the cigarette smoking and alcohol drinking
behavior of African American and Mexican youth as well as
it explains these behaviors for white youth? To answer this
question the research examined the race findings from
previous research, the traditional conceptualization and
operationalization of the social control theory to
understand how this may affect research findings, and
conducted analyses that included whites, African Americans
and Mexican Americans from a national data set.

The findings in this study illustrate the extent to which
the social control theory is useful to explain the
cigarette smoking and alcohol drinking behavior of white
youth, African American youth, and Mexican youth. The key
finding of this study is that the social control model is
limited in explaining the smoking and drinking behavior of
Mexican youth and African American youth, especially. This
finding is consistent with previous research conducted,
research where results would indicate a race difference in
the explanatory power of the social control model, but was
glossed over or not mentioned in the findings or discussion

sections by the researcher (Gardner and Shoemaker, 1989;

 

 

Covington, 1988; Liska and Reed, 1985; Weber, et. a1.,
1995; Cerkonvich and Giordano, 1992).

Other findings are that the social control model
explains the smoking behavior of all the youth better than
it explains the drinking behavior of the youth. This is
consistent with previous research (Agnew, 1985; Gardner and
Shoemaker, 1989; Akers, 1991) that suggests the social
control model explains less serious behavior better as
compared to more serious behavior.

Another finding is that the social control model fit
Mexican youth better than African American youth, which is
consistent with findings from previous research (Weber,
Miracle, and Skehan, 1995; (Rodriguez and Weisburd, 1991).
In addition, there were differential findings based on
race/ethnicity in the effectiveness of the social control
indicators. The social control variables that were
significant, differed, for the most part, for each group,
for each dependent variable. This finding is consistent
with Weber et a1.,(l995). Weber et al.’s research indicates
that there should be consideration of cultural differences
when constructing the social control indicators. The fact
that there is such a marked difference among the social
control variables and their effect on the dependent

variables by race/ethnicity, suggests that considering the

160

 

 

impact of race/ethnic differences when constructing the
social control indicators should be considered as an option
to maximize the effect of the social control model for all
race/ethnic groups.

Finally, the attachment to parents, attachment to peers,
and attachment to teacher variables were not significant in

many of the analyses. Attachment to teachers was

 

significant for white youth when smoking cigarettes was the
dependent variable. All of the attachment variables were
significant for white youth when drinking alcohol was
analyzed as the dependent variable. However, none of the
attachment variables were significant for African American
youth or Mexican youth for either dependent variable. This
finding is important because Hirschi (1969) suggests, that
the attachment variable, and parental attachment
specifically, is the most affective bond. However, in this
research there was no significant association between
parental and peer attachment and smoking cigarettes.

The literature on smoking and drinking suggests that
white youth smoke and drink more often than other youth
(i.e., African American and Hispanic youth) (Bureau of
Labor Statistics, 1999). The findings in this study confirm

previous research (Flay et a1., 1994; Ellickson and

 

Morton).

}-4
CW
[.4

Previous research suggests that parents and peers have
the most influence in the smoking and drinking behavior of
youth (Foshee and Bauman, 1992; Ennett et a1., 1997). The
literature suggests that peers probably have the most
influence in regards to the drinking and smoking behavior
of many adolescents (Ennet and Bauman, 1993; Getting and
Beauvais, 1986; Friedman, Lichtenstein, and Biglan, 1985;
Bauman et a1., 1984; McAlister, Krosnick, and Milburn,
1984; Huba and Bentler, 1980; Levitt and Edwards, 1970).
The results of the current research suggests that parents
and peers do not have the influence over the cigarette
smoking or alcohol drinking behavior of youth in general.
There were significant affects of attachment to parents and
peers for white youth but only in reference to their
drinking behavior. These results should be assessed with

caution do to the limitations of the current research.

162

 

 

Chapter V
DISSERTATION SUMMARY
Purpose and Significance of this Research

The purpose of this research was to answer the
question, does the social control theory explain the
cigarette smoking and alcohol drinking behavior of African
American and Mexican youth as well as it explains these
behaviors for white youth? The question was answered by
(1) examining the race findings from previous research, (2)
examining the traditional conceptualization and
operationalization of the social control theory in the
literature to understand how the traditional treatment
might affect research findings, and (3) conducting analyses
that included whites, African Americans and Mexicans from a
national data set.

The significance of this study is to understand the
conditions under which the social control theory is useful
in explaining delinquency. Specifically, this study is
important because (1) it contributes to the current body of
research by clearly examining the social bond constructs
relative to African American and Mexican youth; (2) it
makes the findings generalizeable to the population as a
result of the utilization of a national data set; and (3)

it applies the principles of social control theory to the

Fl
0\
(A)

 

 

explanation of the smoking and drinking behavior for
whites, African Americans and Mexican youth in the United
States. The dichotomy of whites and African Americans in
the research literature has limited the generalizeability
of results to other groups. Understanding the limitations
of the social control theory in regards to race/ethnicity,
the following hypothesis was tested comparing white,

African American and Mexican youth:

Ho: There is no relationship between attachment (to
parents, teachers, peers), commitment (to school and
furthering education or getting a job), involvement
(in sports related activities and other school-related
activities) and belief (in committing mild and serious
infractions in school) in regards to smoking
cigarettes and drinking alcohol. Moreover, these
relationships will be conditioned by gender, SES, and

race.

SUMMARY OF THE LITERATURE REVIEW
The Social Control Theory
The social control theory assumes that everyone has
the tendency to want to commit crimes but most individuals

are deterred from doing so because of the bonds we form to

 

individuals and/or social institutions in society.
Specifically, Hirschi (1969) suggests there are four bonds
that individuals form to significant others or to the
traditional norms in society that curb their natural
tendencies to want to commit crimes and they are:
attachment, commitment, involvement, and belief.

Attachment is the bond that is formed for individuals
(i.e., parents, peers, teachers, clergy, etc.) in society.
Commitment has to do with commitment to the traditional
ideas of success in society (i.e., continuing ones,
education after high school, finishing school, getting
married and being a success). Involvement has to do with
the individual’s involvement in traditional activities
(i.e., sports, clubs, band, drama, etc.). Belief has to do
with the belief in the morals and values in society.

The bonds of the social control theory have a negative
effect on delinquency, the stronger the bonds the more
likely delinquency will decrease and vice versa. In
addition, Hirschi (1969) suggests the influence of the
bonds on delinquency is not equally distributed among the
four bonds. In fact, he suggests that attachment to parents
is seen as the most affective bond. If parental attachment
is strong, this lessens the influence of the other bonds on

delinquency. However, if the affects of parental attachment

165

 

 

 

or attachment in general is not strong, then one of the
other bonds would have to be strong in order to prevent or
reduce the likelihood of delinquent behavior.

Previous literature suggests the social control theory
is a viable explanation for the delinquent and deviant
behavior of some youth. However, there is no clear evidence
to suggest that the social control theory is viable
explanation for the delinquent behavior of youth of color.

Much of the research that has utilized social control
theory has tended to gloss over the applicability of the
model to youth of color. Findings from previous literature
suggest the social control theory loses its explanatory
power or mixed results are obtained when race is introduced
into the equation (Gardner and Shoemaker, 1989; Covington,
1988; Liska and Reed, 1985; Weber, et. a1., 1995;
Cerkonvich and Giordano, 1992). Also, the amount of
variance the model explains in regression analyses is far
less for adolescents of color than for majority youth
(Gardner and Shoemaker, 1989; Covington, 1988; Liska and
Reed, 1985; Weber, et. a1., 1995; Cerkonvich and Giordano,
1992).

In addition to the problems with the efficacy of
scxzial control theory to explain juvenile delinquency

relative to race or ethnicity, much of the research that

}_ J
O\
OW

 

 

does address race tends to focus mostly on the racial
dichotomy of African Americans and white Americans
(Covington, 1988; Liska and Reed, 1985; Cerkonvich and
Giordano, 1992; Gardner and Shoemaker, 1989). The
Black/White dichotomy does not address issues pertinent to
many racial/ethnic groups that find themselves within the

criminal justice system.

Literature on Smoking and Drinking

The literature on smoking and drinking suggests that
white youth smoke and drink more often than other youth
(i.e., African American and Hispanic youth) (Bureau of
Labor Statistics, 1999). In addition, previous research
suggests that parents and peers have the most influence in
the smoking and drinking behavior of youth (Foshee and
Bauman, 1992; Ennett et a1., 1997). The literature suggests
that peers probably have the most influence in regards to
the drinking and smoking behavior of many adolescents
(Ennet and Bauman, 1993; Oetting and Beauvais, 1986;
Friedman, Lichtenstein, and Biglan, 1985; Bauman et a1.,
1984; McAlister, Krosnick, and Milburn, 1984; Huba and
Bentler, 1980; Levitt and Edwards, 1970). The research also

indicates that there are differential influences on the

167

 

 

smoking and drinking behavior of adolescents when
racial/ethnic differences are considered.

With knowledge of prior research this current project,
sought to contribute the current body of research by
clearly examining the social bond constructs relative to
the cigarette smoking and alcohol drinking behavior of

white, African American, and Mexican youth. In addition,

 

the current research sought to make the findings
generalizeable to the population by using a national data
set.
THE METHODOLOGY

The Data

This study utilized secondary data from the National
Educational Longitudinal Study 1988-1994 (NELS:88). The
NELS:88 was the first nationally representative
longitudinal study of eighth grade students in public and
private schools. The NELS:88 uses a two—stage stratified
probability sampling design that selected a nationally
representative sample of 24,599 students from 1,052

randomly selected schools.

Dependent Variables
The two dependent variables for this study are smoking

cigarettes and drinking alcohol. Both dependent variables

O\
(I)

 

 

were indications of whether the respondent had smoked
cigarettes or drank alcohol in the 30 days prior to being
interviewed. The variables were coded as one (1) indicating
the respondent had smoked or drank 30 days prior to being
interviewed or zero (0) where they indicated they had not

smoked or drank 30 days prior to being interviewed.

Independent Variables

The independent variables for this study are
attachment (parents, peer, and teachers), involvement (in
sports related activities and involvement in other school
related activities), commitment (to school, to marriage and
success, to furthering education or getting a job) and
belief (it is not okay to commit mild and serious
infractions in school). All of the social bond indicators
were computed then recoded into ordinal level measurements
with categories of high, medium, and low attachment,
commitment, involvement, and belief (see Appendix B for the
items that were used to create the social control
indicators).

Attachment to parents was conceptualized as the
respondent’s feelings about his/her parents. Attachment to
peers was conceptualized as how the respondent feels s/he

is thought of by his/her peers. -ttachment to teachers is

 

 

conceptualized as how the teacher feels about the
respondent. Involvement in sports-related activities is
conceptualized in terms of the types of sports the
respondent participated while in school. Involvement in
other school-related activities is conceptualized in terms
of the respondent’s level of participation in school
related activities such as band, orchestra, plays, and
school clubs.

Commitment to school is conceptualized in terms of
whether the respondent comes to class prepared. Commitment
to further education or going to trade school is
conceptualized in terms of whether the respondent plans to
take college entrance exams or skill tests. Believing it is
not okay to commit mild infractions in school is
conceptualized in terms of moral issues that involves
whether the respondent believes it is okay to skip school,
cheat on tests, be late for class, to copy someone’s
homework, etc. Believing it is not okay to commit serious
infractions in school is conceptualized in terms of moral
issues that involves whether the respondent believes it is
okay to get into physical fights at school, belong to

gangs, steal belongings, destroy school property.

The Control Measures

The control variables consisted of race/ethnicity,
gender, parental education level, and SES. The race
variable is a nominal level variable that contains all
groups (i.e., whites, African Americans, and Mexicans).
Parental education is an ordinal level measurement with
categories that range from “less than high school” to
“Ph.D.”. The SES quartile variable was constructed using
available parent data that included: father's education
level, mother's education level, father's occupation,

mother's occupation, and family income.

THE FINDINGS
Univariate Analysis
The data consisted of 11,038 respondents. The sample
was almost evenly divided between males (48.6%) and females
(51.4%). The race/ethnic categories in the sample were
comparable to the national population of whites (78.4%),
African Americans (12.0%) and Mexicans (9.6%) for the year
1990 (see Appendix A for comparison). Majority of the
respondent’s had parents who had completed some college
(41.5%). The findings indicate that many of the respondents

were medium on most of the social control indicators.

f..J
\J
i J

 

Bivariate Analyses
There were several bivariate analyses conducted
comparing the social control indicators to gender,
race/ethnicity, smoking cigarettes and drinking alcohol.
There were only small differences in the percentages on
gender and race/ethnicity as indicated in Tables 3 and 4.
However, the small differences were many times significant

as indicated in the tables.

Social control Indicators and Gender

In regards to the attachment indicators, males and
females were mostly medium on this bond. There was little
difference in the involvement indicators as well. Males and
females were mostly low on involvement in sports related
activities and mostly medium on involvement in other school
related activities. Males and females were mostly high on
their commitment to school and their commitment to marriage
and success. However, males and females were mostly low to
medium on their commitment to furthering education or going
to trade school. Finally, males and females were mostly
high on the belief indicators. Meaning, males and females
overwhelmingly believed it was not okay to commit mild or
serious infractions in school. The chi—square values were

significant for all the relationships except, attachment to

+4
\J
N

 

teachers. The significant relationships indicate the
associations observed in the sample can be inferred to the
population.
Social control Indicators and.Race/Ethnicity

The bivariate analysis with the social control
indicators and race/ethnicity showed very little difference
in the levels of the social control indicators and
race/ethnicity. Whites, African Americans and Mexicans were
mostly medium on the attachment indicators. However, the
respondents were mostly low on involvement in sports
related activities and mostly medium on their involvement
in other school related activities. All three race/ethnic
groups were overwhelmingly committed to school as well as
to marriage and success as indicated by the fact that most
were high on these commitment indicators. In contrast, the
respondents were low to medium on their commitment to
continuing their education or going to trade school.
Whites, African Americans and Mexicans were mostly high on
their belief that it is not okay to commit mild or serious
infractions in school. The chi-square values were
significant for all of the bivariate relationships, which
mean the differences observed in the sample can be inferred

to the population.

H
\J
(A)

Smoking Cigarettes, Control Variables, and Social Control
Indicators

The findings indicate that more females smoke than
males. The findings also indicate that more whites smoke
than African Americans and Mexicans. In fact African
Americans were three times less likely to smoke than
Mexicans and almost five times less likely to smoke than
whites. The data also suggests that there are marginal
differences in the respondents who smoked when parental
education is considered. Respondents whose parents had no
more than some college, were more likely to smoke than
respondents’ whose parents had at least a college degree.

The results of the social control indicators show very
little difference in those respondents who are high,
medium, or low on the social control indicators and their
smoking behavior. The exception to this were the belief
indicators, where respondents who were low on the belief
indicators were over four times more likely to smoke than
respondents who were high on the belief indicators. This
finding indicates that respondents who believed it was not
okay to commit mild infractions in school were less likely
to smoke than those who believed it was okay to commit mild

infractions in school.

174

 

Drinking.Alcohol, Control Variables, and Social Control
Indicators

The findings indicate that more females drink than
males. The findings also indicate that more whites drink
than African Americans and Mexicans. In fact African
Americans was almost two times less likely to drink than
Mexicans and whites. Also the data suggests that there are
marginal differences in the respondents who indicated they
drank in the 30 days prior to being interviewed when
parental education is considered. Respondents whose parents
had a Ph.D. or M.D. or equivalent showed the highest
percentage of drinking.

The results of the social control indicators show very
little difference in those respondents who were high,
medium, or low on the social indicators and their drinking
behavior. The exception to this were the belief indicators,
where respondents who were low on the belief indicators
were over two times more likely to drink than respondents
who were high on the belief indicators. This finding
indicates that respondents who believed it was not okay to
commit mild infractions in school were less likely to drink
than those who believed it was okay to commit mild

infractions in school.

in
\J
U1

 

Gamma Measures of Association for Smoking Cigarettes and
the Social Control Indicators for the Entire Sample,
African Americans and Mexicans

The results of the gamma measures of association is
the first evidence the social control model might be a
better explanation of the smoking and drinking behavior of
white youth than African American and Mexican youth. The
findings on the attachment variables indicate that
attachment to teachers was the only significant
association. This association showed a significant negative
relationship for all but African Americans. Commitment to
school, commitment to continuing one’s education or going
to trade school, and commitment to marriage and success
were significant for the entire sample and whites.
Commitment to school and marriage and success was
significant for Mexicans. However, only commitment to
marriage showed a significant relationship for African
Americans, in addition, this commitment was less
significant for African Americans (p<.05) than it was for
whites (p<.01) or Mexicans (p<.01). All of the significant
relationships were negative; meaning the higher the
respondent was on commitment the less likely s/he was to
smoke.

Involvement in sports related activities showed a

significant negative association among whites and Mexicans

 

and a significant positive relationship for African
Americans. The gamma value for this outcome was highest for
African Americans, which indicate knowing the level of
commitment of an African American respondent a better
prediction of their smoking behavior was more likely than
for whites or Mexicans.

The belief indicators showed the greatest association
among the race/ethnic groups and smoking behavior. The
findings indicate a negative, moderately strong,
association between the belief indicators and smoking
behavior. The negative relationship indicates that as the
belief that it is not okay to commit mild or serious
infractions in school goes up, smoking cigarettes is

reduced.

Gamma Measures of Association for Drinking.Alcohol and the
Social Control Indicators for the Entire Sample, African
Americans and Mexicans

The results of the gamma measures of association and
drinking alcohol indicates the social control model might
be a better explanation of the smoking behavior of youth
than it is to explain their drinking behavior. The findings
on the attachment variables indicate that attachment to

parents, peers and teachers showed a significant, negative

association for the entire sample and whites. In addition,

 

 

attachment to teachers showed a significant negative
association for African Americans and Mexicans. All of the
commitment indicators were significant for the entire
sample and whites. Commitment to school and marriage and
success showed a significant negative relationship for
African Americans. However, none of the commitment
indicators were significant for Mexicans.

Involvement in sports related activities showed a
significant negative association among whites and Mexicans.
The belief indicators showed the greatest association among
the race/ethnic groups and drinking behavior. The findings
indicate a negative, moderately strong, association between
the belief indicators and drinking behavior. The negative
relationship indicates that as the belief that it is not
okay to commit mild or serious infractions in school goes

up, drinking alcohol is reduced.

Logistic Regression Findings and Smoking Cigarettes

The logistic regression findings indicate that the
social control model explains the smoking and drinking
behavior of white youth better than it does for African
American and Mexican youth. Specifically, the regression
findings indicate there were significant findings for seven

(7) of the ten social control indicators controlling for

178

all other variables for whites as compared to three (3) for
African Americans and six (6) for Mexicans. In addition,
the social control indicators tended to be more significant
for whites than African Americans and Mexicans (see p-
values in Tables 15-17). Also, the amount of explained
variance was much less for African Americans and Mexicans

as compared to whites.

 

Logistic Regression Findings and Drinking Alcohol

The logistic regression findings indicate that the
social control model explains the drinking behavior of
white youth better than it does for African American and
Mexican youth. Specifically, the regression findings
indicate there were significant findings for seven (7) of
the ten social control indicators controlling for all other
variables for whites as compared to one (1) for African
Americans and two (2) for Mexicans. In addition, the social
control indicators tended to be more significant for whites
than African Americans and Mexicans (see Tables 18-21).
Also, the amount of explained variance was much less for
African Americans and Mexicans as compared to whites in the
model with alcohol drinking in the dependent variable.

Finally, as compared to the model with cigarette

smoking as the dependent variable, the social control

179

 

V

theory explains the smoking behavior of all adolescents

better than it does their drinking behavior (compare the

results of Tables 15-17 and Tables 19-21).

There were several important findings in this study.

>

‘7

Findings indicate that females have higher
percentages of smoking and drinking than males. In
addition, the drinking and smoking behavior of
females was significant in the logistic regression
models. These findings are not consistent with
national data that suggests that females smoke and
drink at a lower rate than males.

Overall, the model was a better explanation of the
smoking and drinking behavior of whites than African
Americans and Mexicans.

The model was a better explanation of the smoking
and drinking behavior of Mexicans than African
Americans.

The social control model is a better explanation of
the smoking behavior of all adolescents than it is
of their drinking behavior. This is consistent with
prior research that suggests the social control
model explains the less serious behaviors better

than it does more serious behaviors (Akers, 1991).

180

 

 

Limitations of Study

There were some limitations in this study. First, the
analysis of secondary data always carries with it inherent
limitations. The data collected, at best approximates the
kind of data a researcher would prefer for testing a
hypothesis, frequently affecting the study design, question
wording and sequence, and details of the interviews
(Nachmias and Nachmias, 1994).

Second, there is a slight bias as a result of missing
respondents data due to some respondents not matriculating
from the base year data collection to the first follow—up.
However, the sample that remains does approximate the
proportions for the categories of race and gender from the
1990 census (see Appendix A).

Third, there are measure limitations, for example, the
data did not contain a variable to clearly indicate the
class of the respondents. As a result, the variable used as
a proxy for class represents the education level of the
respondent’s parents as well as a SES quartile variable.

Fourth, there were limitations determining the
Hispanic ethnic group. Originally, all Hispanic ethnicities
were to be used in order to compare findings among the
ethnic groups. However, the Cuban and Puerto Rican groups

were too small to use in the analyses. In addition, there

181

.‘L‘ __

 

was an “other” category that represented all the other
Hispanic ethnicities in the United States. This was
problematic, as it was not clear what ethnicities these
were. As a result this “other” Hispanic ethnicity was
removed from the analyses. In the end, the Mexican
ethnicity was the only one utilized in the data analysis,
as they represented the largest of the Hispanic ethnicities

and there was no ambiguity surrounding these respondents.

Future Research

Future research should address the limitations of this
study as well as began to build on the findings of this
study to address the issues of the limitation of the social
control theory as it is traditionally conceptualized to
explain the smoking and drinking behavior of youth of

color. Specifically future research should,

> Include original data collection to address the
limitations of having to compute variables from data
that has already been collected. This would insure the
exact specification of the social control indicators
as well as the control variables (i.e., race and
class).

2 Include social indicators that are conceptualized
taking into account cultural differences. This would
require the researcher to understand the specific
cultural differences might exist when considering the
conceptualization of attachment, commitment,
involvement, and belief.

 

 

 

 

APPENDICES

 

 

APPENDIX.A

ASSESSING THE EFFICACY OF THE SOCIAL CONTROL THEORY FOR
EXPLAINING THE CIGARETTE SMOKING AND ALCOHOL DRINKING
BEHAVIOR OF AFRICAN AMERICANS AND MEXICAN AMERICANS

 

 

The Dependent Variables

How many cigarettes does the respondent smoke per day (This
variable was recoded where l= respondents who indicated
they smoked less than I cigarette a day or more and 0 where
the respondents indicated they “don’t smoke at all”)

I don't smoke at all

Less than 1 cigarette a day

l-5 cigarettes per day

About % pack per day

More than % pack but less than 2 packs per day
2 packs per day or more

In the last 30 days, number of times respondent drank
alcohol (This variable was recoded where 1: respondents who
indicated they drank on l-2 occasions or more and 0 where
the respondents indicated they drank on 0 occasions in 30
days prior to being interviewed)

0 occasions

1-2 occasions
3-19 occasions
20+ occasions

The Social Control Indicators
Attachment to Parents (Alpha = .8931)

How often r does things with mother/father
Important living close to parents
Important getting away from parents

R’s parents treat r fairly

Does not like his parents very much

R gets along well with parents

Parents disappointed with what r does
Parents understand him/her

Involvement in Religious Activities (Alpha = .7365)

Important to participate in religious activities
R thinks he is a religious person

Involvement in Sports Related Activities (Alpha = .9774)

Played baseball/softball at school

185

 

 

Played basketball at school

Played football at school

Played soccer at school
Participated on swim team at school
Played other team sport at school
Played an individual sport
Participated in cheerleading
Participated on drill-team

Involvement in Other School Related Activities (Alpha =
.9750)

Participated in school band/orchestra
Participated in school play or musical
Participated in student government
Participated in academic honor society
Participated in school yearbook, newspaper
Participated in school service clubs
Participated in school academic clubs
Participated in school hobby clubs
Participated in school fta, fha, ffa

Commitment to School (Alpha = .8175)
Often go to class without pencil/paper
Often go to class without books

Often go to class without homework done

Commitment to College or Trade School by Taking Entrance
Exams (Alpha = .8958)

Does r plan to take the Pre-SAT test

R plans to take college board SAT test

R plans to take the ACT test

R plans to take the ASVAB

R plans to take the PACT

Commitment to Marriage and Success (Alpha = .61)

Important being successful in line of work
Important finding the right person to marry

Peer Attachment (Alpha = .7258)

Students think of r as being popular
Students think r is socially active

186

Students think of r as being important

Teacher Attachment (Alpha

.7348)

Teachers expect r to succeed in school

Teaching is good at school

Teachers are interested in student

When r works hard teachers praise his/her efforts

Belief it

(Alpha

t's
It’s
It’s
It's
It’s
It's
It’s

okay
okay
okay
okay
okay
okay
okay

Belief it

(Alpha

Feel
Feel
Feel
Feel

its
its
its
its

is

Okay to Commit Mild Infractions in School

.8313)

to
to
to
to
to
to
to

is

be late for school

cut a couple of classes

skip school a whole day
cheat on tests

copy someone else’s homework
talk back to teachers
disobey school rules

Okay to Commit Serious Infractions in School

.8355)

okay to
okay to
okay to
okay to

get into physical fights
steal belongings from school
destroy school property
bring weapons to school

 

APPENDIX B

ASSESSING THE EFFICACY OF THE SOCIAL CONTROL THEORY FOR
EXPLAINING THE CIGARETTE SMOKING AND ALCOHOL DRINKING
BEHAVIOR OF AFRICAN AMERICANS AND MEXICAN AMERICANS

pm
CO
(I)

Table 22. Comparison of Missing Respondents in Data Sample to
Population Parameters in the 1990 Census Data

 

 

 

 

 

 

 

 

 

Variable Missing (%) Not Missing (%) 1990 Census (%)
Base Year First Follow-up
Respondents
Race
African-American 14 11.4 12.1
White 63 74.6 75.6
Mexican American 23 14 9.0
Gender
Males 49.9 47.5 48.7
Females 50.1 52.5 51.3

 

 

 

 

“Q
I
.h‘

 

APPENDIX C

190

 

APPENDIX C

.ASSESSING THE EFFICACY OF THE SOCIAL CONTROL THEORY FOR
EXPLAINING THE CIGARETTE SMOKING AND ALCOHOL DRINKING
BEHAVIOR OF AFRICAN AMERICANS AND MEXICAN AMERICANS

...—85.. ...... .33.. 5:». 235.55

 

32% E 3283.:

 

 

 

 

 

 

 

 

 

 

 

 

o; team. So: So. ..iofi. .308. :02. ......X: - So. .83. .....EN. 523m EEEoU 8 >85 E E 32.3 .:
Begum
o._ 2o: doc. .15.. .....moo. icon. imam - 2o. 1&8. 3.28. E mcogaaE 2:: E550 8 >85 E E u>o:om d—
o._ m 5. woo. one. Go: «owe. 355.- So: doc. 8:333. coca—omioocum 850 E EoEo>_o>E .o
o; coo. 25cc. moo. 130.. 1.03.- NS. 3:8; 8253.. Eager 5on E 35.523:— .w
o._ ......oQ. .3009. six— _.- .53; .- ......omc. tabofif mmooosm use WwIflE—Z 9 “Cognac—ECU .N.
.0080
or 33c. 2.2: .- .12 .- 2o. 1&2; SEE 8:83pm mEtontzm 8 EoEEEEoU .e
or .... _ m _ .- 368.- .3qu 102.- Begum 9 552550 .m
o._ §*0N_ . .....OBOC f..mm—. whosomoh 8 acoE£omt< .v
o._ 1.80.. o5. 28d 8 E08585. .m
o._ for 3:23 8 EoEEomz< .N
3 .282 waxed ._
_ _ 2 o w h o w v m N _

 

 

 

 

 

 

 

 

 

 

 

 

 

853? com 35302 3:50 Eoom Ea 3:856 wEonm mo moons—8.80 .mN 033.

192

.. Saul. ...... .33.. 5.5. 2.3535

 

 

 

 

 

 

 

 

 

 

 

 

 

Begum E 2288.:—
o. 8.. NS.- ..5 :8. :08. :8... ....e... 8... :2... :3..- mzo.sm..eeooo.§ow:.2.2.8...
.Sfim
.... 2...- 8.. ...: :3... :3... :Q? as. ....8. IE; 58252........E..aeooo;s.om...o>o..om....
.... ms. 3... ON... 08. .02. ...-.....- so.- 8? 8:35... ..o.~.3.-.8..om§..o 5852.25.
o. 8... .38. .38. :3...- ...e..... S... :2... 8:38... 83.3355 5.55225...
.... .32. :8... 2......- :$...- :08. :NS; mmoszm usowagzozse..ee°os
.830
.... :3... :8... :3..- 2... $8.- 5...... 5:833. misstaozséieoo...
.... ... m . .- .38.- ...-NS. :2..- .228 2 525.558 .m
.... .32. .38.- :8. 22.23 2 202.83. ...
.... :2..- :.o..- 28. o. .8553... .m
.... .18.. 253.. 2 .5583 .N
o. 6.82%.}...5 ..
.. o. o w n o w v m N .

 

 

 

 

 

 

 

 

 

 

 

 

 

8.23 8.. 85982 .8280 Eoom new .382. mac—5.5 mo macaw—oboo .vm £an

193

...S.uv.. ...... .35.. 3.2:. ...-3:55

 

 

 

 

 

 

 

 

 

 

 

 

 

 

.825 E 8288...:
o.. 1.3.. boo. ovor :mo. mmo. 3o. omor woo. omor :3..- 32.3 2.880 8 >85 m. E o>o=om ...
Begum E
o._ ...wwo. 2o: :3. :wno. imm— iw. . .- moor go. 1.5..- 80:03:. 2:: 28.80 2 >85 m. E 05:3 .o.
3:.>:o<
o.. So. ..ooo. .58. 2o; ..ooor ....mo. .- wmo. Duo: B.a_om-_oo..om .050 E 82529.... .0
o._ .8: Nmo mmo; nmo. tow . .- ..ono. .....vwo. 8:3:3. 3.33. 38% E 88529:. .w
o.. 18o. ......So. mvor ...ooor voor ..mnor 9883 Us. $852 8 808.8800 .5
.885
o.. 1.36. :8. .- :3..- 8o. omor 8%.... 8:835 8.55.5... 8 8088.85 .0
o.. mmor omo. to; So; 89.0w 2 .EcEEEoU .m
o. tom .. So; do. 223$ 8 .5858... .v
o.. ..mcor .mo: 88.. o. .5858... .m
o.— moo. 855$ 0. 808583. .N
9. 820820 mEonm .—
2 o. o w h o m w m m _

 

 

 

 

 

 

 

 

 

 

 

 

88:08.... Swot-«<- 8m 8833). .2800 38m .28 mozoawmo wEonw mo 828.280 .mm 033.

194

...wa. ...... .35.. "...... 3355.5

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Begum E 80:00:...
0.. 2.3.. 80. $0.- .38.. 20. 0.0. 08.- TMB 08.. :8..- mzoEm ......80 01.3.0 m. .. 2.2.00 ...
.33
0.. .08. 20.- :5. :08. :mm. 1.....- 80.- .30. :Nm .- ... 22.22.... 2.... ......80 808.0 m. 0. ”>23 .0.
0.. 5.0. 38. as. 0.0.- .80.- :8..- .8. .8.- 8...>.6< 83.3.-.Sfim .050 ... .5823... .0
0.. .3- NS. NS.- Ro. 1%.. .08. 8..- 80.2.3. .853. 3.5% ... 25.522... ...
0.. :80. :30. $0.- tomfi .00.- LE..- W885 ...... 08...“: o. 22.0.5500 ..
.0800
0.. :30. :00..- :8. .- 0.... VS.- .oaca 808:3 082...... 2 225.5500 .0
0.. $0.. 08 :0.- ..L . ..- .023 o. 2.25.5050 .m
0.. :02. .8.- ..WS. 32.8.; o. .5558... ...
0.. .30.. .00.- 38.. o. .5582... .m
0.. NS. 3.8.0.. 8 808208;. .N
.... .282 80.5.0 ..
.. 0. o w n 0 w v m N _

 

 

 

 

 

 

 

 

 

 

 

 

 

8.8.805. 50...?- 80 88802 8.800 300m 05. 880?. mExERH .0 80:30.80 0N 030-.-

195

2.5.5.. ...: .35.. 3.2;. Ease...

 

 

 

 

 

 

 

 

 

 

 

 

 

.00..om ... 9.0.8.2....
.... imam. umwo. vs.- 2.0. .. me... 102. 2.2. - xv... mm... :3... 30.8w 2.5.50 0. >85 m. z 96:5 ...
.00..om E
0.. 0.0. no... ...-o... .30... {.mmm. 3N.~.- 1.00... .8. ...».nm: 20.62.... “...... ..EEoU 0. ......O m. .. o>o=om ....
80.36...
o. 0.0.- .08. .mo. 0....- Noor E...- .03.- onor 33.3.-.8zum .050 0. .coEo>.0>... .0
0.. 3.... In»... one. own..- one . m8.- ovo; 8:353. 0323. mtomm .... ...oEo>.0>... .w
0.. :03. 3.8.. :8..- 9m...- cmc. 1.3... $805 0.... “Eat-as. 0. .55....500 .5
.0980
0.. 8.3.. ....NN. - :00...- N-vo. 03.. .20.... 00:830.. mutants“. 0. 2.25:5..00 .0
.... 2.3...- ......mm..- ow... .Imm..- .00..om 0. 20.5.8800 .m
o.— ..ivg. ......wc. .. ......NM.. 95:08... 0. EoEzomz< .v
0.. 08.- ...... $8.. 0. ....oE..o§< .m
.... v8.- 3.5.9. 0. EoEۤ< .m
.... 830.890 m:...0...w ..
.. o. o w b c w v m N .

 

 

 

 

 

 

 

 

 

 

 

 

 

£80.58... 58063. .0. 3.093.). .8300 Eoom ...... mot-«....mG $6.005 ..0 9.0.8.880 ...m 0......-

196

...—Suva ...: .2?!— 525 ...-95.55

 

330m E 2283*:—

 

 

 

 

 

 

 

 

 

 

 

 

o._ :93. $3. vmo; :2 _. mvo. :62. 1.2—; «3. N8. 33...- Serum EEEoU 8 >85 a E o>ozom .:
Gogow E
o._ 20. 8o. 33¢. 3.8. 8.3.0.. Ina- 258. Rs. 2.63.. mcozomtE 2E. EEEoU 2 >85 E E 96:5 .2
moE>zo<
o._ 2o: 38. So. 2o; moor NS: m8; *ooo. coa_om-_oo:ow 550 E “5:523:— .a
3 Re: :25. 23. one: 3.28.- omo. ...oeo; 833:3. Essa 3.5mm E :55..ng .w
o._ .13.. 3.0.2. :3. .. tome: So. 8o: $085 new cunts—z 9 EucEEEoU ..\.
“8.80
o._ 1mg. .32 - 2&2 .- owe. _No.- BEEN 8:833. wESESE 8 EoEEEEoU .o
o._ 3.88.. coo {03. So; Begum 2 EoEEEEoo .m
o; €32. :2: .. ...LS. 30589 8 “SE—383 .v
o._ 08.- 33.- 38m 2 “55582 .m
o; woo. 3:85 8 E08582 .m
3 .28? micro ._
__ 2 ¢ x N. c w v m N E

 

 

 

 

 

 

 

 

 

 

 

 

 

éotog 5&on 8m $5302 .9280 Boom 98 3:82 mExctD mo mcosfluboo .mm 2an

197

REFERENCES

Agnew, R. (1985). Social control theory and
delinquency: A longitudinal test. Criminology, 23, 47—61.

Agnew, R. (1991). The interactive effects of peer
variables on delinquency. Criminology, 29(1), 47-72.

Agnew, R. (1991). A longitudinal of social control
theory and delinquency. JOurnal of Research in Crime and
Delinquency, 28, 126-156.

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

Akers, R. L. (1991). Self-control as a general theory
of crime. JOurnal of Quantitative Criminology, 7(2), 201-
211.

Alexander, C., Piazza, M., Mekos, D., Valente, T.
(2001). Peers, schools and adolescent cigarette smoking.
Journal of Adolescent Health, 29, 22—30.

Arneklev, B. J., Grasmick, H. G., Tittle, C. R., &
Bursik, R. J. (1993). Low self-control and imprudent
behavior. Journal of Quantitative Criminology, 9(3), 225-
247.

Brownfield, D., & Sorenson, A. M. (1991). Religion and
drug use among adolescents; A social support
conceptualization and interpretation. Deviant Behavior, 12,
259-276.

Brownfield, D., & Sorenson, A. M. (1993). Self—control
and juvenile delinquency: Theoretical issues and an
empirical assessment of selected elements of a general
theory of crime. Deviant Behavior: An Interdisciplinary
Journal, 14, 243-264. .

Burton, V. S., Cullen, F. T., Evans, T. D., Alarid, L.
F., & Dunaway, R. G. (1998). Gender, self-control, and
crime. JOurnal of Research in Crime and Delinquency, 35(2),
123-147.

198

 

Cernkovich, S. A., & Giordano, P. C. (1987). Family
relationships and delinquency. Criminology, 25, 295-321.

Cochran, J. K., Wood, P. B., Sellers, C. S.,
Wilkerson, W., & Chamlin, M. B. (1998). Academic dishonesty
and low self-control: An empirical test of a general theory
of crime. Deviant Behavior: An Interdisciplinary JOurnal,
19, 227-255.

Conger, R. D. (1976). Social control amd social
'learning models of delinquent behavior. Criminology, 14(1),
17-40.

Covington, J. (1988). Crime and heroin the effects of
race and gender. J0urnal of Black Studies, 18, 486-506.

Deng, X., & Zhang, L. (1998). Correlates of self-
control: An empirical test of self-control theory. JOurnal
of Crime and Justice, 21(2), 89—110.

Ellickson, P.L., and Morton, S.C. (1999), Identifying
adolescents at risk for hard drug use: Racial/ethnic
variations. JOurnal of Adolescent Health, 25, 382-395.

Ennett, S.T., Flewelling, R.L., Lindrooth, R.C., and
Norton, E.C. (1997), School and neighborhood
characteristics associated with school rates of alcohol,
cigarette, and marijuana use. JOurnal of Health and Social
Behavior, 38, 55-71.

Ennett, S.T. and Bauman, K.E. (1993), Peer group
structure and adolescent cigarette smoking: A social
network analysis. Journal of Health and Social Behavior,
34, (3), 226—236.

Evans, T. D., Cullen, F. T., Burton, V. S., Dunaway,
R. G., & Benson, M. L. (1997). The social consequences of
self-control: Testing the general theory of crime.
Criminology, 35(3), 475-504.

Finley, N. J., & Grasmick, H. G. (1985). Gender roles
and social control. Sociological Spectrum, 5(4), 317—330.

199

Flay, B.R., Hu, F.B., Siddiqui, 0. Day, I.E., Hedeker,
D., Petraitis, J., Richardson, J., and Sussman, S. (1194),
Differential influence of parental smoking and friends'
smoking on adolescent initiation and escalation of smoking.
JOurnal of Health and Social Behavior, 35, 248-265.

Forde, D. R., & Kennedy, L. W. (1997). Risky
lifestyles, routine activities, and the general theory of
crime. Justice Quarterly, 14(2), 265-294.

Foshee, V. and Bauman, Karl, E. B. (1992), Parental
and peer characteristics as modifiers of the bond-behavior

relationship: An elaboration of control theory. J0urnal of
Health and Social Behavior, 33, 66—76.

Friedman, J., & Rosenbaum, D. P. (1988). Social
control theory: The salience of components by age, gender,
and type of crime. JOurnal of Quantitative Criminology, 4,
363-381.

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

Gibbs, J. J., & Giever, D. (1995). Self—control and
its manifestations among university students: An empirical
test of Gottfredson and Hirschi's general theory. Justice
Quarterly, 12(2), 231-255.

Gibbs, J. J., Giever, D., & Martin, J. S. (1998).
Parental management and self-control: An empirical test of
Gottfredson and Hirschi's general theory. JOurnal of
Research in Crime and Delinquency, 35(1), 40-70.

Gottfredson, M. R., & Hirschi, T. (1990). A General
Theory of Crime. Stanford: Stanford University Press.

Grasmick, H. G., Tittle, C. R., Bursik, R. J., &
Arneklev, B. J. (1993). Testing the core empirical
implications of Gottfredson and Hirschi's general theory of
crime. JOurnal of Research in Crime and Delinquency, 30(1),
5-29.

Guthrie, B.J., Young, A.M., Boyd, C.J., Kintner, E.K.
(2001), Dealing with daily hassles: Smoking and African
American adolescent girls. JOurnal of Adolescent Health,
29, 109—115.

200

 

Hawkins, D. F. (1994). Ethnicity: The forgotten
dimension of American social control. In G. S. Bridges & M.
A. Myers (Eds.), Inequality, Crime, & Social Control (pp.
99—116). Boulder: Westview Press.

Hepburn, J. (1977). Testing alternative models of
delinquency causation. JOurnal of Criminal Law and
Criminology, 67, 450-460.

Hindelang, M.J.(l973). Causes of delinquency: A
partial replication and extension. Social Problems, 20,
471-487.

Hirschi, T. (1969). Causes of Delinquency. Berkeley:
University of California Press.

Hirschi, T., & Gottfredson, M. (1993). Commentary:
Testing the general theory of crime. JCurnal of Research in
Crime and Delinquency, 30(1), 47-54.

Institute of Medicine (1994). Growing Up Tobacco Free:
Preventing Nicotine Addiction in Children and YOuth.
Washington: National Academy Press.

Johnson, K. A. (1984). The applicability of social
control theory in understanding adolescent alcohol use.
Sociological Spectrum, 4, 275-294.

Junger, M., & Marshall, I. H. (1997). The interethnic
generalizability of social control theory: An empirical
test. JOurnal of Research in Crime and Delinquency, 34(1),
79-112.

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

Kandel, D., & Davies, M. (1991). Friendship networks,
intimacy, and illicit drug use in young adulthood; A
comparison of two competing theories. Criminology, 29(3),
441-469.

Keane, C., Maxim, P. S., & Teevan, J. J. (1993).
Drinking and driving, self-control, and gender: Testing a
general theory of crime. JOurnal of Research in Crime and
Delinquency, 30(1), 30-46.

201

 

Krohn, M.D. and Massey, J.L. (1980). Social control
and delinquent behavior: An examination of the elements of
the social bond. Sociological Quarterly, 21, 529—544.

Krohn, M.D., Skinner, W.F., Massey, J.L., & Akers,
R.L. (1985). Social learning theory and adolescent
cigarette smoking: A longitudinal test. Social Problems,
32, 455—471.

Linden, R. (1978). Myths of middle—class delinquency:
A test of the generalizability of social control theory.
Youth and Society, 9(4), 407-432.

Liska, A. E., & Reed, M. D. (1985). Ties to
conventional institutions and delinquency: Estimating

reciprocal effects. American Sociological Review, 50, 547—
560.

Longshore, D. (1998). Self—control and criminal
opportunity; A prospective of the general theory of crime.
Social Problems, 45(1), 102—113.

Longshore, D., & Turner, S. (1998). Self-control and
criminal opportunity. Criminal Justice and Behavior, 25(1),
81-98.

Lyerly, R., R., & Skipper, J. K., Jr.. (1981).
Differential rates of rural-urban delinquency. Criminology,
19, 385-399.

Marcos, A. C., & Bahr, S. (1988). Control theory and
adolescent drug use. Youth and Society, 19, 395-425.

Matseuda, R. L., & Heimer, K. (1987). Race, family
structure, and delinquency: A test of differential
association and social control theories. American
Sociological Review, 52, 826—840.

Miller, S. L., & Burack, C. (1993). A critique of
Gottfredson and Hirschi's general theory of crime:
Selective (in) attention to gender and power positions.
Women and Criminal Justice, 4(2), 115-134.

Paternoster, R., & Brame, R. (1998). The structural
similarity of processes generating criminal and analogous
behaviors. Criminology, 36(3), 633-669.

202

 

Paternoster, R.& Triplett, R. (1988). Disaggregating
self-reported delinquency and its implications for theory.
Criminology, 26, 591-625.

Piquero, A. R., & Rosay, A. B. (1998). The reliability
and validity of Grasmick et al.‘s self-control scale; A
comment on Longshore et al. Criminology, 36(1), 157-173.

Pope, C.E. and Feyerherm, W. (1995). Minorities and
the Juvenile Justice System. Office of Juvenile Justice and
Delinquency Prevention: Washington, DC.

Office of Juvenile Justice and Delinquency Prevention
(1999). Minorities in the Juvenile Justice System. U.S.
Department of Justice: Washington, DC.

Rodriguez, 0., & Weisburd, D. (1991). The integrated
social control model and ethnicity. Criminal Justice and
Behavior, 18, 464-479.

Scheer, S.D., and Borden, L.M., Donnermeyer, J.F.
(2000), The relationship between family factors and
adolescent substance use in rural, suburban, and urban
settings. JOurnal of Child and Family Studies, 9, (1), 105-
115.

Skinner, W.F., Massey, J.L., Krohn, M.D.,& Lauer, R.M.
(1985). Social influences and constraints on the initiation
and cessation of adolescent tobacco use. JOurnal of
Behavioral Medicine, 8, 353—376.

Sorenson, A. M., & Brownfield, D. (1995, January).
Adolescent drug use and a general theory of crime; An
analysis of a theoretical integration. Canadian JOurnal of
Criminology, pp. 19-37.

Unger, J.B. and Rohrbach, L.A. (2001), Why do
adolescents overestimate their peers’ smoking prevalence?
Correlates of prevalence estimates among California 8th
grade students. JOurnal of Youth and Adolescence, 31, (2),
147—153.

203

U.S. Department of Health and Human Services (1997).
Youth Risk Behavior Surveillance. Atlanta: U.S. Department
of Health and Human Services, Public Health Service,
Centers for Disease Control and Prevention, National Center
for Chronic Disease Prevention and Health Promotion. Office
of Adolescent and School Health.

Warren, C. W., Kann, L., Small, M.L., Santelli, J.S.,
Collins, J.L., Kolbe, L.J., (1997), Age of initiating
selected health-risk behaviors among high school students
in the United States. JOUrnal of Adolescent Health, 21,
225-231.

Weber, L.R., Miracle, A., and Skehan, T. (1995).
Family bonding and delinquency: Racial and ethnic
influences among U.S. Youth. HUman Organizations, 54, 363- 1
372. ‘

Wiatrowski, M. D., Griswold, D. B., & Roberts, M. K.
(1981). Social control theory and delinquency. American
Sociological Review, 46, 525-541.

Winfree, L. T., Backstrom, T. V., & Mays, G. L.
(1994). Social learning theory, self-reported delinquency,
and youth gangs. Youth and Society, 26(2), 147-177.

Winfree, T. L. J., Backstrom, T. V., & Mays, G. L.
(1994). Social learning theory, self-reported delinquency,
and youth gangs a new twist on a general theory of crime
and delinquency. Youth and Society, 26(2), 141-177.

Witchcoff Knight, K., & Tripodi, T. (1996). Societal
bonding and delinquency: An empirical test of Hirschi's
theory of control. JOurnal of Offender Rehabilitation, 23,
117—129.

204