. v 13F:

9.

h.

t.
5.1

r I .
.Jn..am«.j.z. -.
yflpaflufl... LR

 

 

 

uttct . . . $59,...A, . z u .4
. K 3»: I... .. I .. .7 3:31.. ‘ i . . RI}?!
0 - ‘ . . .u‘ T. .. , . ‘ . ,v.

 

THESIS

.10 0‘

LIBRARIES
MICHIGAN STATE UNIVERS
ITY
EAST LANSING, MICH 48824-1048

This is to certify that the
dissertation entitled

EFFECTS OF DIFFERENTIATION ON COLLEGE STUDENT
DRINKING

presented by

CHRISTOPHER R. LATTY

has been accepted towards fulfillment
of the requirements for the

Doctoral degree in Family and Child Ecolgg!

 

 

MTWF/Ld.

Major Professor‘s Signatfire

’7'22— 05’

 

 

Date

MSU Is an Affirmative Action/Equal Opportunity Institution

“—

PLACE IN REI'URN 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

 

AUG 0 6 2007

 

APR 0 2 290;

 

:5Z409

 

 

 

 

 

 

 

 

 

 

 

 

2/05 mlClFlCIDeteDuehdd-pds

 

EFFECTS OF DIFFERENTIATION 0N COLLEGE STUDENT DRINKING
By

Christopher R. Latty

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
DOCTOR OF PHILOSOPHY
Department of Family and Child Ecology

2005

ABSTRACT

EFFECTS OF DIFFERENTIATION ON COLLEGE STUDENT DRINKING
By

Christopher R. Latty

This study examined if Bowen’s theory of differentiation was related to general
alcohol use (GAU) and binge drinking among college students. It was hypothesized that
students with lower levels of differentiation would be more likely to have greater
amounts of general alcohol consumption and a higher prevalence of binge drinking. It
was also predicted that students with lower levels of differentiation would have greater
amounts of general alcohol use and binge drinking to the extent that they had greater:
perceptions of other Michigan State University (MSU) student’s alcohol use, perceptions
of their best friend at MSU’s use, expectations of alcohol use as a tension reducer and
social lubricant, and were a child of an alcoholic (COA).

A total of 447 participants (246 females, 201 males) between the ages of 18 and
22 were analyzed for this study. Due to a high level of skewness in the continuous
variables, GAU and binge drinking were each transformed into four categories. The GAU
categories were Abstainers, Low Drinkers, Moderate Drinkers, and Higher Drinkers; the
binge drinking categories were Abstainers, Nonbinging Drinkers, Occasional Binge
Drinkers, and Frequent Binge Drinkers.

Overall, the hypotheses received little support from the logistic regression
analyses. The hypothesis of differentiation and GAU was only supported in the Drinkers
versus Abstainers comparison for males. Following the hypothesized relationship, as

levels of differentiation increased in males the likelihood they drank alcohol decreased.

The differentiation hypothesis was not supported by any of the binge drinking
comparisons.

Again, the majority of the moderator hypotheses were not supported by the
model. Additionally, when significant interactions were found it was interpreted that
differentiation acted as a moderating variable. The only significant interaction in the
GAU analyses was with differentiation and perception of MSU student alcohol use. This
interaction was significant for the High versus Moderate Drinker comparison (males) and
the High versus Low Drinker comparison (females). In relation to binge drinking the only
significant interactions were for the female analyses. There was a significant interaction
between perceptions of MSU student use with Frequent Binge Drinkers versus
Nonbinging Drinkers. For these interactions it was interpreted that higher levels of
dilferentiation served as a protective factor against the disparate risk factors analyzed, as
students with lower levels of differentiation were more vulnerable to being in the higher
drinking category in relation to the risk factor.

The other significant interaction was with COA status in the Frequent versus
Occasional Binge Drinking comparison. Paradoxically, high levels of differentiation
appeared to be a risk factor with COAs, as females were nearly two-and-a-half times
more likely to be Frequent versus Occasional Binge Drinkers when they had higher levels
of differentiation. Explanations for the respective interactions, clinical implications and

recommendations for future research are provided.

COFyright by
Christopher R. Latty
2005

This work is dedicated to Sarah and Olivia.

Your patience, love, and support allowed me to complete a long and wonderful journey.

ACKNOWLEDGMENTS

I am deeply indebted to a number of individuals that allowed me to achieve this
degree. My wife, Sarah, has shown an invaluable level of patience, support, and
compassion throughout my academic career. You are an incredible person and I thank
you for all that you did to help me achieve this goal. I want to thank my daughter for
bringing such joyous light into a journey that was marked by many dark moments. My
parents, Dale and Pat, have provided a lifetime of support and unconditional love that
afforded me an opportunity to explore a road of academic curiosities which led me this
degree. I would also like to thank my brother, Shawn, and fi'iends Matt, Whitney, Jim,
Scott, Jessica, and Henry for providing respite and much needed fun and balance
throughout my graduate training. I also need to thank my colleagues and mentors
Kathleen, Lori, Christie, Tianna, Al, Jason, Brandon, Dr. Lee, Dr. Soderman, and Dr.
Bristor who supported, challenged, and inspired me to be a better teacher, researcher, and
clinician. I am also indebted to NIAAA and the Department of Family and Child Ecology
whose fimding was an invaluable resource that allowed me to invest the time and energy
needed to complete this dissertation.

Finally, I need to thank my committee for all of their time, patience, and energy.
Dr. Meece and Dr. Luster provided many insights along the way regarding data analysis
and editing. Dr. Fitzgerald, my dissertation co-chair, supported, challenged, and inspired
me to stretch and develop my skills as a researcher. Your sponsorship of my NIAAA
fellowship and mentoring were a fantastic capstone to my graduate training. I cannot

thank you enough for all that you did for me. Finally, I would like to extend my thanks to

Dr. Carolan, my major professor, dissertation co-chair, and wearer of so many hats in my
academic development. You were an unbelievably supportive constant throughout a time
of much transition. I will always cherish your encouragement, support, guidance, and

sense of humor. Thank you.

vii

TABLE OF CONTENTS

LIST OF TABLES ......................................................................................................... xi
LIST OF FIGURES ....................................................................................................... xvii
CHAPTER ONE
Introduction .................................................................................................................... 1
Statement of the Problem ............................................................................................... 3
General Alcohol Use ................................................................................................ 3
Binge Drinking ......................................................................................................... 5
Importance of the Problem ............................................................................................. 5
Theoretical Framework ..... 5
Human Ecological Theory ....................................................................................... 5
Bowen Theory .......................................................................................................... 7
Conceptual Models .................................................................................................. 12
General Alcohol Use Model .............................................................................. 14
Binge Drinking Model ....................................................................................... 15
Research Questions .................................................................................................. 16
General Alcohol Use .......................................................................................... 16
Binge Drinking ................................................................................................... 16
CHAPTER TWO
Introduction .................................................................................................................... 17
Literature Review ........................................................................................................... 17
College Student Drinking ........................................... . ............................................. 17
Binge Drinking ......................................................................................................... 18
Ecological Influences on College Student Drinking ................................................ 22
Individual Factors .............................................................................................. 22
Space and Location Influences .......................................................................... 23
Family of Origin Influences ............................................................................... 23
Family of Origin Influences through a Bowenian Lens ..................................... 25
Role of Anxiety and Stress on Drinking ............................................................ 27
Perceptions of Others’ Alcohol Use .................................................................. 32
Summary .................................................................................................................. 34
CHAPTER THREE
Introduction .................................................................................................................... 35
Methods .......................................................................................................................... 35
Conceptual and Operational Definitions .................................................................. 35
Research Objectives ................................................................................................. 39
Hypotheses ............................................................................................................... 41
Research Design ....................................................................................................... 43
Measures ........................................................................................................................ 44

viii

Differentiation .......................................................................................................... 44

Parental History of Alcohol Use. ............................................................................. 45
General Alcohol Use ................................................................................................ 46
Binge Drinking ......................................................................................................... 46
Expectancies of Alcohol as a Tension Reducer and Social Lubricant ..................... 47
Perception of the Average Michigan State Student’s Alcohol Use ......................... 47
Perception of Their Best Friend’s Alcohol Use ....................................................... 47
Demographic Information ........................................................................................ 48
Sample ............................................................................................................................ 48
Data Analysis ................................................................................................................. 50
CHAPTER FOUR
Introduction .................................................................................................................... 53
Analyses of General Alcohol Use .................................................................................. 53
Differentiation and GAU ......................................................................................... 54
Four-F actors of the DSI-R and GAU ....................................................................... 56
Moderator Analyses ................................................................................................. 57
Comprehensive Models ........................................................................................... 61
Analyses of Binge Drinking ........................................................................................... 67
Differentiation and Binge Drinking ......................................................................... 68
Four-Factors of the DSI—R and Binge Drinking ...................................................... 68
Moderator Analyses ................................................................................................. 70
Comprehensive Models ........................................................................................... 72
CHAPTER FIVE
Introduction .................................................................................................................... 83
Discussion of Results ..................................................................................................... 83
Differentiation and Alcohol Consumption. .............................................................. 84
F our-F actors of the DSI-R and Alcohol Consumption ............................................ 85
Moderator Hypotheses ............................................................................................. 87
Limitations ..................................................................................................................... 93
Recommendations for Future Research ......................................................................... 94
Clinical Implications ...................................................................................................... 97
Summary ........................................................................................................................ 98
REF EEN CES ................................................................................................................. 100
APPENDICES
Appendix A: Differentiation of Self Inventory — Revised (DSI-R) .............................. ll 1
Appendix B: Analysis of the DSI-R with College Students ......................................... 116
Appendix C: Children of Alcoholics Screening Test (CAST) ...................................... 129
Appendix D: Children of Alcoholics Screening Test—Mother (M-CAST) ................... 131
Appendix E: Children of Alcoholics Screening Test-Father (F-CAST) ..................... 134
Appendix F: Demographic Information and Alcohol Use ........................................... 136
Appendix G: Expectancies of Alcohol Use .................................................................. 140
Appendix H: Consent Form .......................................................................................... 144

Appendix I: Debriefing Form ...................................................................................... 145
Appendix J: Transformation of Continuous Alcohol Measures into Categorical

Data ............................................................................ 147
Appendix K: Individual Variable Descriptions ............................................................. 154
Appendix L: Analyses of General Alcohol Use ........................................................... 165
Appendix M: Analyses of Binge Drinking .................................................................... 220

Table 1.
Table 2.

Table 3.
Table 4.

Table 5.
Table 6.
Table 7.

Table 8.

Table 9.

Table 10.

Table 11.
Table 12.
Table 13.
Table 14.
Table 15.
Table 16.
Table 17.

Table 18.

LIST OF TABLES

Logistic Regression Predicting Male Students that High versus

Moderate Drinkers - Perceptions of MSU Student Use .............................. 58
Logistic Regression Predicting Female Students that High versus Low

Drinkers - Perceptions of MSU Student Use .............................................. 59
Comprehensive Models for Male GAU Categories ..................................... 64
Comprehensive Models for Female GAU Categories ................................. 65

Logistic Regression Predicting Male Students thatare Frequent Bingers
versus Occasional Bingers - Four Factors ................................................... 69

Logistic Regression Predicting Female Students that are Frequent

Bingers versus Nonbingers - Perceptions of MSU Student Use ................. 71
Logistic Regression Predicting Female Students that Are Frequent

Bingers versus Occasional Bingers - COA Status ...................................... 72
Comprehensive Models for Male Binge Drinking Categories .................... 78
Comprehensive Models for Female Binge Drinking Categories ................. 79
Summary of the Significant Individual Predictors Specifically Related

to Differentiation for GAU .......................................................................... 82
Means and Distribution for the Total DSI-R and Subscales ........................ 117
Correlations for the Total DSI-R and Subscales .......................................... 1 l7
Correlation Matrix for the DSI-R Item Clusters .......................................... 121

Means and Distribution for the Revised Total DSI-R and Subscales .......... 123

Correlations for the Revised Total DSI-R and Subscales ............................ 123
Correlation Matrix'for the Revised DSI-R Item Clusters ............................ 124
Frequency of Binge Drinking Categories by Gender .................................. 152
Frequency of GAU Categories by Gender ................................................... 153

Table 19.
Table 20.
Table 21.
Table 22.
Table 23.
Table 24.
Table 25.
Table 26.
Table 27.
Table 28.
Table 29.
Table 30.

Table 31.

Table 32.

Table 33.

Table 34.

Table 35.

Table 36.

Table 37.

Table 38.

Frequency of Race by Gender ..................................................................... 155
Frequency of Parental Alcoholism by Gender ............................................. 156
Frequency of Parental Alcoholism and Binge Drinking by Gender ............ 157
Frequency of Parental Alcoholism and GAU by Gender ............................ 157
Frequency of COA Status and Binge Drinking by Gender ......................... 158
Frequency of COA Status and GAU by Gender .......................................... 158
Frequency of Independent Variables by Gender ......................................... 160
Correlations for Independent Variables - Male Subjects ............................ 162
Correlations for Independent Variables — Female Subjects ......................... 163
VIP for Independent Variables .................................................................... 164
Logistic Regression Predicting Students that Drink - Differentiation ........ 167
Logistic Regression Predicting Students that Drink - Four Factors ............ 167
Logistic Regression Predicting Students that Drink — Perceptions of

MSU Student Use ........................................................................................ 169
Logistic Regression Predicting Students that Drink — Perceptions of

Best Friend’s Use ......................................................................................... 172
Logistic Regression Predicting Students that Drink - COA Status ............. 174

Logistic Regression Predicting Students that Drink — Expectations of
Tension Reduction ....................................................................................... 176

Logistic Regression Predicting Students that Drink - Expectations of
Social Lubrication ........................................................................................ 179

Logistic Regression Predicting Students that Drink — Comprehensive
Model ........................................................................................................... 181

Logistic Regression Predicting Students that are Moderate versus Low
Drinkers — Differentiation ............................................................................ 182

Logistic Regression Predicting Students that are Moderate versus Low
Drinkers—Four Factors ............................................................................... 182

xii

Table 39.

Table 40.

Table 41.

Table 42.

Table 43.

Table 44.

Table 45.

Table 46.

Table 47.

Table 48.

Table 49.

Table 50.

Table 51.

Table 52.

Table 53.

Logistic Regression Predicting Students that are Moderate versus Low

Drinkers - Perceptions of MSU Student Use .............................................. 184
Logistic Regression Predicting Students that are Moderate versus Low
Drinkers — Perceptions of Best Friend’s Use ............................................... 186
Logistic Regression Predicting Students that are Moderate versus Low
Drinkers — COA Status ................................................................................ 187
Logistic Regression Predicting Students that are Moderate versus Low
Drinkers - Expectations of Tension Reduction ........................................... 189

Logistic Regression Predicting Students that are Moderate versus Low
Drinkers- Expectations of Social Lubrication ............................................ 193

Logistic Regression Predicting Students that are Moderate versus Low
Drinkers - Comprehensive Model ............................................................... 194

Logistic Regression Predicting Students that are High versus Low
Drinkers — Differentiation ............................................................................ 195

Logistic Regression Predicting Students that High versus Low Drinkers
- Four Factors .............................................................................................. 195

Logistic Regression Predicting Students that High versus Low Drinkers
- Perceptions of MSU Student Use ............................................................. 197

Logistic Regression Predicting Students that High versus Low Drinkers
- Perceptions of Best Friend’s Use .............................................................. 199

Logistic Regression Predicting Students that are High versus Low
Drinkers — COA Status ................................................................................ 201

Logistic Regression Predicting Students that High versus Low
Drinkers - Expectations of Tension Reduction ........................................... 203

Logistic Regression Predicting Students that High versus Low Drinkers
- Expectations of Social Lubrication ........................................................... 205

Logistic Regression Predicting Students that High versus Low Drinkers
- Comprehensive Model .............................................................................. 207

Logistic Regression Predicting Students that High versus Moderate
Drinkers — Differentiation ............................................................................ 209

xiii

Table 54.

Table 55.

Table 56.

Table 57.

Table 58.

Table 59.

Table 60.

Table 61.

Table 62.

Table 63.

Table 64.

Table 65.

Table 66.

Table 67.

Table 68.

Logistic Regression Predicting Students that High versus Moderate

Drinkers — Four Factors ............................................................................... 209
Logistic Regression Predicting Students that High versus Moderate

Drinkers - Perceptions of MSU Student Use .............................................. 212
Logistic Regression Predicting Students that are High versus Moderate
Drinkers — Perceptions of Best Friend’s Use ............................................... 214
Logistic Regression Predicting Students that are High versus Moderate
Drinkers - COA Status ................................................................................ 215
Logistic Regression Predicting Students that are High versus Moderate
Drinkers - Expectations of Tension Reduction ........................................... 216
Logistic Regression Predicting Students that are High versus Moderate
Drinkers - Expectations of Social Lubrication ............................................ 218
Logistic Regression Predicting Students that are High versus Moderate
Drinkers — Comprehensive Model ............................................................... 219
Logistic Regression Predicting Students that are Occasional Bingers

versus Nonbingers — Differentiation ............................................................ 222
Logistic Regression Predicting Students that are Bingers versus

Nonbingers — Four Factors ........................................................................... 222
Logistic Regression Predicting Students that are Bingers versus

Nonbingers - Perceptions of MSU Student Use .......................................... 223
Logistic Regression Predicting Students that are Bingers versus

Nonbingers — Perceptions of Best Friend’s Use .......................................... 225
Logistic Regression Predicting Students that are Bingers versus

Nonbingers -— COA Status ............................................................................ 227
Logistic Regression Predicting Students that are Bingers versus

Nonbingers - Expectations of Tension Reduction ....................................... 229
Logistic Regression Predicting Students that are Bingers versus

Nonbingers - Expectations of Social Lubrication ....................................... 233

Logistic Regression Predicting Students that are Bingers versus
Nonbingers - Comprehensive Model .......................................................... 234

xiv

Table 69.
Table 70.
Table 71.
Table 72.
Table 73.
Table 74.
Table 75.
Table 76.
Table 77.
Table 78.
Table 79.
Table 80.
Table 81.
Table 82.

Table 83.

Logistic Regression Predicting Students that are Occasional Bingers
versus Nonbingers — Differentiation ............................................................ 236

Logistic Regression Predicting Students that are Occasional Bingers
versus Nonbingers — Four Factors ............................................................... 236

Logistic Regression Predicting Students that are Occasional Bingers
versus Nonbingers - Perceptions of MSU Student Use ............................... 237

Logistic Regression Predicting Students that are Occasional Bingers
versus Nonbingers — Perceptions of Best Friend’s Use ............................... 238

Logistic Regression Predicting Students that are Occasional Bingers
versus Nonbingers - COA Status ................................................................ 239

Logistic Regression Predicting Students that are Occasional Bingers
versus Nonbingers - Expectations of Tension Reduction ........................... 242

Logistic Regression Predicting Students that are Occasional Bingers
versus Nonbingers — Expectations of Social Lubrication ............................ 243

Logistic Regression Predicting Students that are Occasional Bingers
versus Nonbingers - Comprehensive Model ............................................... 245

Logistic Regression Predicting Students that are Frequent Bingers
versus Nonbingers - Differentiation ............................................................ 246

Logistic Regression Predicting Students that are Frequent Bingers
versus Nonbingers — Four Factors ............................................................... 246

Logistic Regression Predicting Students that are Frequent Bingers
versus Nonbingers - Perceptions of MSU Student Use ............................... 248

Logistic Regression Predicting Students that are Frequent Bingers
versus Nonbingers - Perceptions of Best Friend’s Use ............................... 251

Logistic Regression Predicting Students that are Frequent Bingers
versus Nonbingers - COA Status ................................................................ 252

Logistic Regression Predicting Students that are Frequent Bingers
versus Nonbingers — Expectations of Tension Reduction ........................... 255

Logistic Regression Predicting Students that are Frequent Bingers
versus Nonbingers — Expectations of Social Lubrication ............................ 258

XV

Table 84.

Table 85.

Table 86.

Table 87.

Table 88.

Table 89.

Table 90.

Table 91.

Table 92.

Logistic Regression Predicting Students that are Frequent Bingers
versus Nonbingers — Comprehensive Model ............................................... 259

Logistic Regression Predicting Students that are Frequent Bingers
versus Occasional Bingers — Differentiation ............................................... 261

Logistic Regression Predicting Students that are Frequent Bingers
versus Occasional Bingers - Four Factors ................................................... 261

Logistic Regression Predicting Students that are Frequent Bingers
versus Occasional Bingers — Perceptions of MSU Student Use .................. 263

Logistic Regression Predicting Students that Are Frequent Bingers
versus Occasional Bingers — Perceptions of Best Friend’s Use .................. 265

Logistic Regression Predicting Students that Are Frequent Bingers
versus Occasional Bingers - COA Status .................................................... 268

Logistic Regression Predicting Students that Are Frequent Bingers
versus Occasional Bingers - Expectations of Tension Reduction ............... 269

Logistic Regression Predicting Students that Are Frequent Bingers
versus Occasional Bingers - Expectations of Social Lubrication ............... 271

Logistic Regression Predicting Students that are Frequent Bingers
versus Occasional Bingers - Comprehensive Model ................................... 273

xvi

Figure 1.

Figure 2.

Figure 3.

Figure 4.

Figure 5.

Figure 6.

Figure 7.

Figure 8.

Figure 9.

Figure 10.

Figure 11.

Figure 12.

Figure 13.

Figure 14.

Figure 15.

LIST OF FIGURES

Conceptual Model for College Students’ General Alcohol Use ................ 14

Conceptual Model for College Students’ Binge Drinking ......................... 15

The differentiation and perception of MSU student alcohol use
interaction for the comparison of male High versus Moderate Drinkers... 60

The differentiation and perception of MSU student alcohol use
interaction for female High versus Low Drinkers ...................................... 62

The differentiation and perception of best Friend at MSU’s alcohol use
interaction for female High versus Low Drinkers ...................................... 66

The differentiation and perception of MSU student alcohol use
interaction for female Frequent Binge Drinkers versus Nonbinging
Drinkers ...................................................................................................... 73

The differentiation and COA status interaction for female Frequent
versus Occasional Binge Drinkers ............................................................. 74

The differentiation and expectations of tension reduction interaction for

male Frequent versus Occasional Binge Drinkers ..................................... 77
The differentiation and expectations of tension reduction interaction for
female Binge Drinkers versus Nonbinging Drinkers ................................. 80
The differentiation and expectations of tension reduction interaction for
female Occasional Binge Drinkers versus Nonbinging Drinkers .............. 81
Four-F actor Analysis of the DSI-R with Males and Females .................... 125
Differentiation as a Higher Order Latent Factor with Males and

Females ....................................................................................................... 126
F our-F actor Analysis of the Revised DSI-R with Males and Females ...... 127

Differentiation as a Higher Order Latent Factor with Males and
Females in the Revised DSI-R ................................................................... 128

General Alcohol Use (Quantity X Frequency) of Male Drinkers for the
Past 30 Days .............................................................................................. 148

Figure 16. General Alcohol Use (Quantity X Frequency) of Female Drinkers for

the Past 30 Days ......................................................................................... 149
Figure 17. Frequency of Male Binge Drinking Over the Past 2 Weeks ...................... 150
Figure 18. Frequency of Female Binge Drinking Over the Past 2 Weeks ................... 151

xviii

Chapter One
Introduction

Alcohol consumption has a high rate of prevalence among college students.
Estimates indicate that around 81 percent of college students will drink alcohol during the
current year, and approximately 66 percent of these students will have consumed alcohol
in the past month (Johnston, O'Malley, Bachman, & Schulenberg, 2004). Additionally,
many of these students will consume alcohol to the point of intoxication. According to
Woltcrsdorf (1997), anywhere from 26 to 48% of college students drink to intoxication
during any given month. Moreover, he asserts that nearly 4% of college students drink
daily. Furthermore, it is estimated that within the past 2 weeks over 40% of college
students will have engaged in binge drinking (Wechsler, Lee, Kuo, et al., 2002).
Estimates indicate around 31% of college students have a DSM-IV diagnosis of abuse,
and 6% have a diagnosis of dependence (Knight et al., 2002). This high prevalence of
abuse puts students at greater risk for academic problems, sexual abuse, risky sexual
behaviors, physical injury, assault, violence, and death.

As college students are in a period of developmental transition into young
adulthood, it is important to understand the continued relationship college students have
with their families of origin. This study sought to understand how the relationship with
one’s parents affects the decision-makingfprocesses of alcohol consumption. Central
constructs to this study were Bowen’s (1966) concepts of differentiation and triangling.
Differentiation exists on inter- and intrapersonal levels. Intrapersonally, differentiation
refers to a person’s ability to remain clear-minded and objective in the face of anxiety

rather tlmn becoming emotionally reactive. It is also an interpersonal dynamic reflecting

an individual’s ability to maintain a sense-of-self while remaining emotionally connected
in relationships. Individuals with low levels of differentiation struggle to form identities
with convictions and are more likely to adopt the beliefs and values of those for whose
approval and closeness they yearn. Moreover, they are dependent on others for validation
and ofien lack a firm foundation of self belief and approval. Therefore, when there is
some form of relational anxiety (e.g., a criticism) individuals with lower levels of
differentiation are not able to tolerate this stress and become emotionally reactive. When
emotional disequilibrium occurs in a dyad, poorly differentiated individuals are likely to
triangle in other people or things to diffuse this anxiety. For example, a couple on the
cusp of fighting about themselves may channel this energy into a conflict about their
children. Moreover, individuals may triangle things such as work or alcohol as a means to
difl‘use the relational tension building in the relationship, resulting in a couple having
conflict over the object rather than over their relationship. Patterns of differentiation are
acquired in families and influence how anxiety is tolerated and whether triangulation will
serve as a dysfunctional means to alleviate the anxiety. Accordingly, the more
differentiated individuals are, the more likely they are to tolerate anxiety. Greater levels
of differentiation result in greater potential to objectively respond to anxiety, decreasing
the likeliness that an individual will need to triangle someone or something to regulate
anxiety. The central hjpothesis of this study is that there is a relationship between

dtflcrentiation and alcohol consumption in college students.

It is crucial to understand the relationship between differentiation and alcohol
consumption for this population for two reasons. First, college students experience a

unique push-pull with their families of origin as they are likely geographically separated

from their family for the first time. At this time, college students are at a developmental
point of transition from adolescence to young adulthood and may not be able to tolerate
the relational anxiety with their parents while at school. Secondly, college students may
not be able to tolerate the anxiety of interpersonal relationships with peers and may
triangle alcohol consumption, particularly via abuse and binge drinking, to manage their
anxiety. As millions of students attend universities every year, the seriousness of this
issue points to the necessity to understand ecological influences on their decisions to
consume alcohol. This is especially true for those that acquire drinking as a means to
cope with the new stressors of higher education. Attitudes and behavior towards alcohol,
adopted while at college, may set a stage for a lifetime of dysfunctional usage and
addiction.
Statement of the Problem

This study investigated the relationship between levels of differentiation and
alcohol use for college students. Additionally, moderating effects on the relationship
between differentiation and alcohol use were investigated for the following variables:
status of being a child of an alcoholic (COA), perceptions of the average Michigan State
University (MSU) student’s alcohol use, perceptions of best fiiend at MSU’s use, and the
expectations of alcohol use as a tension reducer and social lubricant.
General Alcohol Use

Specifically, this model will investigate the following:

1. the relationship between differentiation, as measured by the Differentiation of

Self Inventory—Revised (DSI-R, Skowron & Schmitt, 2003), and general

alcohol consumption, as measured by quantity and frequency of alcohol use
over the past 30 days

. the relationship between the four factors of the DSI—R (Emotional Reactivity, 1
Position, Emotional Cutoff, and Fusion with Others) with general alcohol use
. the moderating effects of COA status, as measured by parent-specific versions
of the Children of Alcoholics Screening Test (Jones, 1983), on the

relationship between differentiation and general alcohol use

. the moderating effects of perceptions of the average MSU student’s alcohol
use, as measured by students’ perceptions of the quantity and frequency of
alcohol use over the past 30 days for the average MSU student, on the

relationship between differentiation and general alcohol use

. the moderating effects of perceptions of their best friend at MSU’s alcohol
use, as measmed by students’ perceptions of the quantity and frequency of
alcohol use over the past 30 days for their best fiiend at MSU, on the

relationship between differentiation and general alcohol consumption

. the moderating effects of expectations of alcohol serving as a tension reducer,
as measured by the Tension reduction subscale (Kushner, Sher, Wood, &
Wood, 1994), on the relationship between differentiation and general alcohol

consumption

. the moderating effects of expectations of alcohol serving as a social lubricant.
as measured by the Social Lubrication subscale (Kushner et al., 1994), on the

relationship between differentiation and general alcohol consumption

Binge Drinking

The outcome of binge drinking will also be investigated using the same seven
points of investigation as mentioned for general alcohol use.

Importance of the Problem

This project is innovative in that it incorporates the systemic concept of
differentiation with one’s family of origin to individual decision-making processes,
which has not been previously applied to the understanding of alcohol consumption. If
the hypothesized relationship between levels of differentiation and alcohol consumption
exists, not only will it prove to be a unique contribution to the alcohol consumption
literature, but will also yield a greater understanding of how to further understand the
decision-making processes of college students. Additionally, mental health professionals
working with adolescents and adolescent substance use will gain empirical evidence for
concepts comprising Bowen Theory.

Theoretical Framework

Human Ecological Theory

Human ecological theory and a systemic perspective consider how an individual,
a family, and their total environment are intimately intertwined with one another
(Andrews, Bubolz, & Paolucci, 1930; Bristor, 1995; Bronfenbrenner, 1977, 1999;
Bubolz, Eicher, & Sontag, 1979; Bubolz & Sontag, 1993; Griffore & Phenice, 2001; von
Bertalanfl‘y, 1980). The family ecosystem is comprised of a collection of interdependent

people interacting, sharing resources, goals, and space (Andrews et al., 1980; Bubolz &

Sontag, 1993; Griffore & Phenice, 2001). As children within the family ecosystem
mature, they are greatly influenced by familial subsystems (e.g., parents, siblings) and the
respective rules, roles, boundaries, hierarchy, patterns of interaction, perceptions, and
expectations of the greater family system (Andrews et al., 1980; Bristor, 1995; Bubolz &
Sontag, 1993; Griffore & Phenice, 2001). Just as an individual’s development is
interconnected with other members of her family ecosystem, they individual shares a
connection with her surrounding environment and contexts (Andrews et al., 1980;
Bronfenbrenner, 1977, 1999; Bubolz et al., 1979; Bubolz & Sontag, 1993; Lerner, 1991,
1992, 1993; Lerner & Galambos, 1998; Lerner & Lerner, 1987, 1989; von Bertalanffy,
1980). The development and wellbeing of individuals is greatly influenced by these
unique contextual relationships.

Although the field of alcohol research has become more systemically oriented, it
only became commonplace within the past decade, both for the general population and
specific to college students (Baer, 2002; Zucker, Fitzgerald, & Moses, 1995) (see
Schulenberg & Maggs, 2002, for an ecological review of empirical studies for adolescent
and young adults). This absence of ecological context is evident due to an emphasis on a
minimal number of variables, absence of multivariate or interactive models, and a, lack of
consideration of multiple developmental pathways to differing types of alcoholisms
(Fitzgerald, Zucker, Puttler, Caplan, & Mun, 2000; Zucker, 1994; Zucker et al., 1995;
Zucker, Reider, Ellis, & Fitzgerald, 1997). Consequently, while it is important to
understand the role of individual variables on college students’ drinking, Zucker,
Fitzgerald, and Moses (1995) suggest that risk for alcohol use be considered as a dynamic

process consisting of varying degrees of risk for usage throughout the lifespan.

Bowen Theory

Around the time von Bertalanffy published his general systems theory in 1968,
Murray Bowen was honing his systems theory of emotional systems in 1966 and 1967
(published as Anonymous in 1972). Bowen’s theory is grounded in four interconnected
concepts: differentiation, triangles, emotional system, and the multigenerational
transmission process.

Drflerentiation. The driving force of Bowen’s theory is the concept of
difl‘erentiation. As an intra-psychic construct, differentiation refers to the relationship
between being able to utilize one’s cortex (thinking clearheadedly and objectively) over
the limbic system (emotionally reactive) in the presence of anxiety. Accordingly,
reactivity is inversely related to objectivity (Friedman, 1991). By maintaining a
differentiated position, individuals can objectively think about stress and respond to it
with a variety of clearheaded options, rather than merely being emotionally reactive.

In the context of relationships, differentiation refers to how individuals and dyads
maintain a sense of self while remaining connected to others. The ability to maintain a
sense of self (basic-self) is described as having a firm notion of personal beliefs and
values that will not be surrendered in the context of a relationship. Conversely, an
undifferentiated individual makes decisions based on what feels like the right thing to do
at the time rather than on a reasoned principle. The undifferentiated person has an .
inability to form and develop his own convictions and will adopt those of significant
others or popular ideologies. Unlike a differentiated person, who is able to generate
unique “I believe” statements, an undifferentiated person will evidence laws or religion to

define his dogma (Anonymous, 1972; Becvar & Becvar, 1993; Bowen, 1966). In the

struggle between a desire for connectedness and separateness, an undifferentiated person
fails to recognize where he ends and others begin, whereas a differentiated person is able
to maintain independent thoughts and feelings within relationships (Wetchler & Piercy,
1996). At the crux of the struggle to remain differentiated is the ability to maintain a
connection with others while at the same time being able to maintain individuality.
Titelman (2003) describes this balance as “the ability to act for oneself without being
selfish and the ability to act for others without being selfless” (p. 20).

Bowen conceptualized differentiation as existing on a continuum with the
outcome of complete differentiation as not realistically obtainable (Anonymous, 1972;
Bowen, 1966). Individuals in the upper half of the continuum possess an increased ability
to differentiate between objective reality and feelings and they primarily can filnction in
ways that remain true to their sense of self—although some decisions are based on
feelings in order to not risk disapproval fiom significant others. Finally, these individuals
are much less reactive to praise and criticism as they have a much more firm identity,
whereas people with lower levels of differentiation are highly dependent on others for
external validation.

Theoretically, people tend to couple with others who share similar levels of
differentiation. When these levels are low, the two individuals become fused with one
another and little diflerentiation of self remains for the individuals. Individuals with low
levels of differentiation are dependent on others for approval and validation for a sense of
self. Thus, they are more susceptible to have dysfunctional relationships and. symptoms to
the extent that they experience disrupted emotional equilibrium and individual comfort

levels through criticism, conflict, or relational anxiety from others (Bowen, 1966).

Individuals are at greater risk for developing depressed, somatic, or alcohol related
symptoms to the extent that they have lower levels of differentiation; conversely they can
tolerate more intense anxiety as their differentiation levels increase (Friedman, 1991;
Skowron & F riedlander, 1998). In a well differentiated relationship, each member of the
couple “is more of an autonomous self: there is less emotional fusion in close
relationships, less energy is needed to maintain self in the fusions, more energy is
available for goal-directed activity, and more satisfaction is derived hour the directed
activity” (Anonymous, 1972, p. 119). As anxiety or tension exceeds the comfort level of
a relational dyad, it is more likely that another person or thing (e.g., alcohol or work) will
be triangled into the relationship as a means of shifting the tension away from the dyad
(Bowen, 1966; Friedman, 1991; McGoldrick & Carter, 2001; Titelman, 1998).

Triangles. According to Bowen (1966, Anonymous, 1972), dyads within
emotional systems are naturally unstable and will form a triangle, with another person
(people) or object(s), when under an unmanageable amount of stress. The intensity of the
triangle relationship is dependent on the level of differentiation as well as the importance
of the relationship that is experiencing the anxiety (Anonymous, 1972). In the context of
an extended family or a work environment, the system is comprised of multiple
interlocking triangles. The push-and-pull forces of the desire for togetherness and
individuality in a relationship create varying levels of stress and anxiety for dyads. As
mentioned, when this anxiety or tension exceeds the comfort level of a relational dyad, it
is more likely that a triangle will be brought into the relationship as a means of shifting
the tension away from the dyad (Bowen, 1966; Friedman, 1991; McGoldrick & Carter,

2001; Titelman, 1998). Some of the ways in which a dyad could triangle in another

person or object are: changing the conversation by talking about a third party, one or both
individuals may bring another person in to take sides, one person continually venting her
frustrations through a third person rather than directly with the other, one person shifting
her energy to work or alcohol rather than toward the conflict with the other. No matter
what the method of triangling is, the ultimate goal is to reduce the amount of tension, at
the expense of dealing with it directly, as the ability to remain objective and nonreactive
is lost. Although occasional triangling is not necessarily a dysfunctional behavior, it
becomes problematic when it is utilized as a constant diversion that prevents individuals
and couples from directly dealing with their problems.

The emotional system This concept, originally termed undifferentiated ego mass
(Bowen, 1966), refers to the overall level of differentiation, or fusion, in a system (e.g.,
the nuclear family emotional system) (Anonymous, 1972). As this concept is interrelated
with difl‘erentiation, the extent of emotional fusion between spouses is a product of the
level of difl‘erentiation within the individuals. Bowen (1966, Anonymous, 1972)
identified three ways symptoms fi'om undifferentiated parental couples can be expressed
in a nuclear family: marital conflict, projection to a child(ren), or dysflurction in a
parent/spouse. Families vary greatly regarding which of the areas receive the dysfunction
and in the degree to which it is shared across the areas. For example, if one member of
the parental dyad merges his identity into the couple’s identity-losing his sense of self-
then he would be likely to experience some form of dysfirnction (e.g., developing
depression, somatic illness, or alcoholism), which may in turn prevent dysfimction from

occurring in marital conflict or fiom being projected onto the children.

10

Multigenerational transmission The concepts of differentiation and triangles
operate in concert within families through multigenerational transmission. As parents
triangle children in order to diffuse anxiety within the marital relationship, children may
be inducted into multigenerational transmission of firsed relationships. As children
become more mature, they are consequently faced with the task of becoming more
differentiated from their family in order to become higher functioning individuals
(Titelman, 1998). Children who are fi'equently triangled at a young age into their parents’
relationship may develop lower levels of differentiation than their parents due to their
stunted development as triangled individuals. If this child later marries, it will
theoretically be to someone sharing a similar level of differentiation (Anonymous, 1972;
Bowen, 1966) and he will have a child that may have an even lower level of
differentiation. However, children are not automatically destined to be at lower levels of
differentiation than their parents, other factors may disrupt this inevitably. In addition,
greater levels of differentiation can be obtained via a lifelong process.

Research with Bowen theory. Although this can be an important lens through
- which to view symptom formation, few studies have explored this postulated relationship
in college students by examining the degree to which differentiation influences disparate
types of psychopathology, including alcohol use. Protinsky and Gilkey (1996) found a
relationship between females’ sense of individuation and self-esteem, personal health,
and grade point average. Bartle-Haring, Rosen, and Stith (2002) found that possessing
lower levels of differentiation with one’s mother was related to increased psychological
symptoms and increased reports of stressful life events. Research has also found

consequences of triangulation within families to be related to intimacy and academic

ll

difficulties in young adults, and substance abuse in adolescents (Larson & Wilson, 1998).
Studies evidence that individuals with higher levels of differentiation have been found to
be more flexible, better able to cope with stress, and have less chronic anxiety (Larson &
Wilson, 1998; Skowron & F riedlander, 1998). Additionally, Elieson and Rubin (2001)
found that college students with depression had lower levels of differentiation than did
traditional student groups. These reports indicate a need to further examine how
individuation in college students influences personal health, particularly how well-being
is compromised through decisions to triangle alcohol as a means to cope with anxiety.
Conceptual Models

The model for this study is grounded in the concepts of Bowen theory via two
essential constructs relevant to this study—differentiation and triangulationDifferentiation
influences how an individual tolerates anxiety, whereas triangling is a dysfunctional
means to alleviate the anxiety. According to this theory, the more differentiated
individuals are, the more likely they are to tolerate anxiety, resulting in greater potential
for an objective response to anxiety. More differentiated individuals will be less likely to
exhibit a need to triangle someone or something to regulate anxiety. Therefore, the
general model of this study is that individuals with higher levels of differentiation will be
less likely to triangle alcohol as a means to diffuse their anxiety. The relationship
between differentiation and alcohol use is also expected to be moderated by child of an
alcoholic (COA) status, perceptions of the average MSU student’s alcohol consumption,
and the perception of best friend at MSU’s alcohol consumption, expectations of alcohol

as a tension reducer, and expectations of alcohol as a social lubricant.

12

It is expected that children of alcoholics will be more likely to triangle alcohol as
they will possess a history of having parents model alcohol use as a triangle to tolerate
anxiety. The moderating effect of being a COA is further based upon the premise of
multigenerational transmission. As Bowen theory suggests, children have similar levels
of differentiation as their parents; it is assumed that individuals with low levels of
differentiation have parents who also have low levels of differentiation. It is expected that
alcoholic parents with low levels of differentiation chose alcohol as a triangle, therefore
providing a model of managing stress-COA students who are away from home may feel
comfortable triangling alcohol in relation to parent modeling and due to alcohol’s role as
an integral part of the college culture.

It is expected that less-differentiated students will generally consume greater
amounts of alcohol to the extent that they perceive greater amounts of student use
presenting as normative means by which to diffuse anxiety. Finally, it is expected that
students possessing lower levels of differentiation will have greater amounts of general
alcohol consumption to the extent that they have higher expectations that alcohol will
serve as a social lubricant or a tension reducer as a means to help tolerate relational

anxiety (see Figure 1, p. 14, and Figure 2, p. 15).

I3

 

Parental Alcohol Use

 

 

 

 

 

L

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Level of I genial]
Differentiation I co 0.
Consumptlon
Perceptions of the Average MSU Student’s Use
Level of L 216mm
Differentiation a co 0.
Consumptlon
Perceptions of their Best Friend at MSU’s Use
Level of I 23mg
Differentiation T on 0.
Consumptlon
Expectations of Alcohol as a Tension Reducer
Level of J filmefi
Differentiation ' co 0 .
Consumptlon
Expectations of Alcohol as a Social Lubricant
Level of l finial
Differentiation ' co 0.
Consumptlon

 

Figure 1. Conceptual model for college students’ general alcohol use.

14

 

 

 

 

 

 

 

Parental Alcohol Use

 

 

 

 

 

l _

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Level of Binge
Differentiation Drinking
Perceptions of the Average MSU Student’s Use
Level of Binge
Differentiation 4 a Drinking
Perceptions of their Best Friend at MSU’s Use
Level of I _ Binge
Differentiation ' Dn'nking
Expectations of Alcohol as a Tension Reducer
Level of L Binge
Differentiation 4 Drinking
Expectations of Alcohol as a Social Lubricant
Level Of 1 Binge
Differentiation Drinking

 

 

 

Figure 2. Conceptual model for college students’ binge drinking

15

 

 

 

 

 

Research Questions

The overarching research questions for this study are based on two models. The
review of literature in Chapter Two will support and guide the relationships for the
models to be tested.
General Alcohol Use for College Students

Is there a significant relationship between differentiation and general alcohol
consumption for college students? Are there moderating effects on the relationship
between differentiation and general alcohol consumption for the following variables:
COA status, perceptions of the average MSU student’s alcohol consumption, perceptions
of best friend at MSU’s alcohol consumption, expectations that alcohol will serve as a
tension reducer or a social lubricant?
Binge Drinking for College Students

Is there a significant relationship between differentiation and binge drinking for
college students? Are there moderating effects on the relationship between differentiation
and binge drinking for the following variables: COA status, perceptions of the average
MSU student’s alcohol consumption, perceptions of best fiiend at MSU’s alcohol
consumption, expectations that alcohol will serve as a tension reducer or a social

lubricant?

16

Chapter Two

Chapter Two presents a review of literature regarding general trends for college
student alcohol consumption and provides a description of the concept and consequences
of binge drinking. The role of individual and ecological factors will be considered along
with the role of anxiety and stress. Additionally, evidence will be provided for the
inclusion of the moderating variables of COA status, perceptions of others’ alcohol use,
and expectations of alcohol use for this study’s models.

Literature Review

College Student Drinking

Adolescence is a time when many individuals are at risk for developing
problematic drinking behaviors. This is particularly true for adolescents who are
attending college. Not only are many of these students away from parental supervision
for the first time, but also there is an increased availability of alcohol to be found at
college compared to high school, even for those who are legally underage (Dreer, Ronan,
Ronan, Dush, & Elliott, 2004; Wechsler, Lee, Nelson, & Kuo, 2002; Zucker et al., 1997).
At most colleges and universities, there is a normative component in the student culture
involving alcohol consumption for reasons to socialize, relax, celebrate, and for pleasure
(Demers et al., 2002; Dreer et al., 2004; Schulenberg & Maggs, 2002; Wechsler,
Davenport, Dowdall, Moeykens, & Castillo, 1994; Wechsler, Isaac, Grodstein, & Sellers,
1994; Zucker et al., 1997). Many students will normatively drink alcohol and will never
experience consequences of drinking beyond a hangover. Unfortunately, some students’
alcohol consumption may evolve into dangerous levels of consumption, result in various

negative consequences, or begin a pattern for a life-long battle with dependence.

17

Beyond the potential long-term consequences of addiction or continued abuse,
students are at risk for short-term effects of driving while intoxicated, experiencing a
blackout, sufl‘ering an injury, poor academic performance, skipping class, dropping out,
getting into fights, doing something they regretted, getting arrested, participating in
unsafe sex, experiencing/initiating unwanted sexual activity, attempting/committing
suicide, and unintentional death (Bishop, Lacour, Nutt, Yamada, & Lee, 2004; Cell,
Shott, &- Morris, 1999; Kaplowitz & Campo, 2004; Larirner, Lydum, Anderson, &
Turner, 1999; Presley, Meilman, & Lyerla, 1994; Smith, Wells, & Abdul-Salaam, 1997;
Waite-O'Brien, 1992; Woltcrsdorf, 1997). Unfortunately, there appears to be a negative
cycle between drinking and consequences as the more you drink the more likely you are
to experience a consequence, and not knowing about the health risks of drinking is
related to greater consumption (Jones, Harel, & Levinson, 1992). F urtherrnore, although
everyone who drinks is susceptible to consequences, men are reported to experience them
more oflen (Benton et al., 2004; Billingham, Wilson, & Gross, 1999).

Binge Drinking

Binge drinking has historically been defined as five or more drinks in a row at one
sitting (Presley et al., 1994; Wechsler, Isaac, et al., 1994). However, after comparing
odds ratios in the consequences of binge drinking with the traditional five drink standard,
Wechsler, Dowdall, Davenport, and Rim (1995) assert a more equitable standard of
four or more drinks be considered as binge drinking for females. This would account for
differences in body mass and metabolism, and avoid the underreporting of binge drinking
women. The standard for female binge drinking at four drinks in a row per setting has

since been adopted in multiple studies (Benton et al., 2004; Wechsler, Davenport, et al.,

1994; Wechsler, Dowdall, Davenport, & Castillo, 1995; Wechsler, Dowdall, Maenner,
Gledhill-Hoyt, & Lee, 1998; Wechsler, Lee, Kuo, & Lee, 2000; Wechsler, Lee, Kuo, et
al., 2002; Wechsler, Lee, Nelson, et al., 2002).

Binge drinking is prevalent in nearly half of all college students. A series of
national studies, covering 140 four-year colleges, have shown about 44% of college
students binge drink-a stable percentage spanning 1993 (44%, Wechsler, Davenport, et
al., 1994), 1997 (42.7%, Wechsler et al., 1998), 1999 (44%, Wechsler et al., 2000), and
2001 (44.4%, Wechsler, Lee, Kuo, et al., 2002). Although a similar rate of 42% was
found in a national study by Presley (1994), this figure likely reflects an underestimated
total of female binge drinkers as the study used the 5-drink standard for both genders.

Although the overall percentage of college students that binge drink remains static
over time, there are individual shifts in binging. Studying the changes of drinking
behaviors during the transition of first- to second-year attendance at a Massachusetts
college, Wechsler, Isaac, et al. (1994) found that 78% of males and 63% of females who
consumed at binge levels during their first year continued to do so in their second year.
Additionally, 37% of men and 19% of women who were nonbinge drinkers began to
drink at binge levels dining their second year of attendance. The authors report that the
overwhelming majority of the first-year binge drinkers were also binge drinkers in high
school (62% of the females and 77% of the males). Although this may suggest a
developmental trend for all adolescents, it appears to be specifically disturbing for
college students. There is trend for college bound students to binge drink at lesser rates
than non-college bound peers until they enter college, at which point they will binge

drink at greater rates than their peers not attending college (Schulenberg et al., 2001).

19

This trend for increased prevalence of binge drinking with the transition into college was
illustrated by Wechsler, Davenport, et al. (1994) as 10% of high school binge drinkers
were no longer binging in college, while 22% of non-bingeing high school students
started in college.

Some of the documented consequences associated with binge drinking include:
fighting, damaging property, experiencing blackouts, driving while intoxicated, being a
passenger in a car with an intoxicated driver, doing something they regretted, missing
class, acquiring a sexually transmitted infection, experiencing a sexual assault, having
unplanned or unprotected sex, lower grade point average, and trouble with police (Benton
et al., 2004; Clapp & McDonnell, 2000; Presley et al., 1994; Smith et al., 1997;
Wechsler, Davenport, et al., 1994; Wechsler, Dowdall, Davenport, & Castillo, 1995;
Wechsler et al., 1998; Wechsler et al., 2000; Wechsler, Lee, Kuo, et al., 2002). It is
crucial to understand the prevalence of binge drinking among college students as Smith,
Wells, et al. caution that there is a high risk of experiencing the consequences associated
with binge drinking even if a person only has an isolated episode.

Similar to alcohol consumption in general, infrequent binge drinking for college
students has been described as a normative part of the university culture (Bladt, 2002).
Nevertheless, some individuals may be at greater risk for this behavior than others. The
2001 estimates of Wechsler, Lee, Kuo, et al. (2002) indicate that the 40.9% of college
females and 48.6% of college males binge drank at some level during the year and these
gender differences do not appear to differ across races (O'Malley & Johnston, 2002).
Underage students were reported to drink less frequently but were more likely to binge

when they drank (57.8% of underage men to 41.9% of-age; 53% underage women to

20

37.2% of-age) (Wechsler, Lee, Nelson, et al., 2002). Additionally, regardless of age,
students living in fi'aternity or sorority houses were more likely to binge drink than those
living in any other type of residence (Wechsler et al., 1998; Wechsler, Lee, Nelson, et al.,
2002). Caucasian students have been shown to be the most likely to binge drink,
followed by Hispanic students, with African American students binging at much lower
levels (O'Malley & Johnston, 2002). Frequent heavy binge drinkers have 19 times greater
odds of meeting the DSM-IV criteria for dependence (Knight et al., 2002) than those who
drink at lesser levels.

Although it has never been studied specifically through a Bowenian lens, it can be
inferred that people with lower levels of differentiation may be at greater risk for binging.
Bladt (2002) examined mental health in first-semester college students that were
classified as being abstainers, drinkers without binge behaviors, infrequent binge drinkers
(l or 2 times in a 2-week period), and fiequent binge drinkers (3 or more times in a 2-
week period). Frequently binging female students had significantly lower self-esteem
than females in the other categories. Additionally, female frequent binge drinkers had a
significantly greater feeling of insecure attachment than females who abstained fi'om
drinking and male frequent binge drinkers. The author describes individuals with high
levels of insecure attachment as having a strong and conflicted desire for closeness,
consequently leaving them vulnerable to become enmeshed and concerned with
abandonment. This study’s results are in concert with Bowen’s theory as individuals who
are emotionally fused with significant others—being afiaid of criticism and
abandonment, without being able to validate their own sense-of-self—would be unable to

form secure attachments with others as they constantly react to others’ compliments and

21

criticisms. Consequently, these people would then be more likely to triangle someone or
something, such as alcohol, as a means to tolerate inter- and intrapersonal anxiety.
Ecological Influences on College Student Drinking

Individual factors. Overall, college students appear to have a higher prevalence of
drinking than their same age peers who are not attending college (O'Malley & Johnston,
2002). The most reported individual factor for the amount of alcohol consumed by
college students is being male (Allgood-Merten, Lewinsohn, & Hops, 1990; Benton et
al., 2004; Camatta & Nagoshi, 1995; Cooney & Nonnarnaker, 1992; Cotton, 1979;
Demers et al., 2002; Dreer et al., 2004; Gisske & Adams, 1988; Grant et al., 1994;
Harford, Wechsler, & Seibring, 2002; Johnson & Pandina, 2000; Jones et al., 1992; Lo,
1995; O'Malley & Johnston, 2002; Perkins & Berkowitz, 1986; Pullen, 1994; Sax, 1997;
Sher, Walitzer, Wood, & Brent, 1991; Woltcrsdorf, 1997; Zucker et al., 1995). However,
Demers et al. (2002) concluded that “the relationship between the frequency of drinking
and the consumption per occasion is stronger for female than for male students” (p. 421).
Moreover, recent changes in some methodologies have demonstrated that the gender gap
may be smaller than typically reported when considering body weight and alcohol
metabolism. Furthermore, as previously mentioned, after an analysis of the odds ratios of
negative consequences with binge drinking, Wechsler, Dowdall, Davenport, and Rimm
(1995) suggested changing the binge drinking criteria in women from 5 drinks (the same
standard as men) to 4 drinks in a row per sitting for a more equitable standard. The
change makes it more likely that binge drinking women will be accurately identified
rather than be underreported in studies. Lange and Voas (2001) studied 18— to 33-year-old

pedestrians returning from a night of drinking in Tijuana, Mexico. The authors found that _

22

men and women had similar levels of mean blood alcohol concentration. Although
females may be consuming less quantities of alcohol than males, they nevertheless may
be drinking to similar levels of intoxication (Korcuska & Thombs, 2003) or possibly
higher levels of intoxication and drinking problems (Fitzgerald, Zucker, Mun, Puttler, &
Wong, 2002; Waite-O'Brien, 1992).

Space and location influences. Additionally, the ecological consideration of
location of dwelling place and where one drinks alcohol has an effect on consumption.
College students who live away from their parents or are campus residents are more
likely to consume alcohol (Cooney & Nonnamaker, 1992; Demers et al., 2002; Harford et
al., 2002; Jones et al., 1992). Additionally, Harford et al. found that compared to women,
men are more likely to attend dorm parties and off-campus parties and are less likely to
attend off-campus bars. Interestingly, although a lower proportion of students attended
fraternity/sorority parties than attended off-campus parties and bars, higher proportions of
students attending fraternity/sorority parties chose to drink. Similarly, there is an
incredible prevalence of binging for those who live in fratemity/sorority houses (83.4%
in 1993 decreasing to 75.4% in 2001, see Wechsler, Lee, Kuo, et al., 2002).

Family of origin influences. Fifteen percent of the nation’s children under 17 are
estimated to live in a household with at least one adult with an abuse or dependence
diagnosis over the past year (Grant, 2000; Zucker & Wong, in press). Studies indicate
that approximately one-fifth (Rodney, 1995) to one-third (Landers & Hollingdale, 1988)
of college students have at least one alcoholic parent.

Although most children of alcoholics (COAs) will be resilient to the risks of being

raised in an alcoholic family (O'Sullivan, 1992), many will experience a variety of

23

consequences. Ellis, Zucker, and Fitzgerald (1997) presented the importance of
considering the roles of alcohol-specific and non-alcohol-specific risk factors on the lives
of COAs. Research on non-alcohol-specific risk factors for COAs has revealed evidence
of a long list of issues that include: learning disabilities, poorer intellectual fimctioning,
conduct disorders, eating disorders, antisocial behavior, truancy, stress-related illnesses,
anxiety, depression, increased rates of psychopathology, suicide attempts, inability to
trust, denial of or inability to express feelings, an excessive sense of responsibility, poor
communication, difliculty developing peer relationships, greater involvement in other
drug use, experiencing sexual abuse, experiencing emotional neglect, experiencing family
violence and conflict, having less cohesive families, greater levels of social isolation,
having to undertake a greater number of parenting responsibilities, having greater
parental marital instability, having parents with impaired cognitive abilities and comorbid
psychopathology, and marrying someone who is an alcoholic (Black, Bucky, & Wilde-
Padilla, 1986; Ellis et al., 1997; Garbarino & Strange, 1993; Hewes & Janikowski, 1998;
Jones & Kinnick, 1995; Lawson, 1992; Marlatt ct al., 1998; O'Sullivan, 1992; Poon,
Ellis, Fitzgerald, & Zucker, 2000; Rodney, 1995; Sher, Gershuny, Peterson, & Raskin,
1997; Yeatman, Bogart, Geer, & Sirridge, 1994; Zucker et al., 2000).

A strong alcohol-specific consequence of being a COA also exists, as there is a
relationship of COAs having drinking patterns closely related to their parents and greater
alcohol usage than non-COAs (Ellis et al., 1997; Gisske & Adams, 1988; Gotham, Sher,
& Wood, 2003; Jones & Kinnick, 1995; Yeatman et al., 1994), though this relationship
was not supported by Engs (1990). This relationship also appears to transcend ethnicities

as the transmission has also been illustrated in a sample of Afiican American students

24

(Rodney, 1995). Although the prevalence rates have some variability across studies, in a
seminal review of 39 studies of COAs, Cotton (1979) found that alcoholics are more
likely to have an alcoholic relative than nonalcoholics and that about 30% of any sample
of alcoholics will have at least one alcoholic parent.

A number of studies focus on the genetic heritability of alcoholism, however,
some of the psychosocial variance related to multigenerational transmission of
alcoholism in families has been described to be a learned behavior through modeling
parental behaviors and adopting respective expectations of alcohol and its use (Ellis et al.,
1997).

Family of origin influences through a Bowenian lens. Reports demonstrate that
COAs have less communication with their parents, have witnessed more family
arguments (Sher et al., 1991), and have less social support and less healthy family
environments (Rodney, 1995). From a Bowenian fiarnework, this lack of support and
inability to tolerate anxiety and manage conflict results in a need to triangle an outside
entity in order to diffuse interpersonal conflicts. Perhaps the triangled item of choice in
many alcoholic families is alcohol itself, which then is modeled for future generations to
observe as an appropriate coping mechanism. In the case of alcoholic families, the ability
to maintain a difl‘erentiated sense-of-self would seem to be stymied by the typically
inadequate responsive caregiving, retarded development of affect regulation, and poor
self-esteem maintenance found in many COAs (Bladt, 2002). Furthermore, Clair and
Genest (1987) found that COAs tended to use more emotion-focused, rather than
problem-focused, strategies in coping with their problems. As poorly differentiated

individuals lack the ability to maintain an objective presence in the face of anxiety, they

25

become more emotionally reactive and may have a need to triangle a person/object to
alleviate their anxiety. COAs may be more apt to triangle alcohol as they having
witnessed and adopted their parents’ coping strategies and attitudes of alcohol
consumption in their own lives (Archambault, 1992; Howard, 1992).

Ellis et al. (1997) surmise that increased conflict in alcoholic families may be the
result of the combined poorer problem-solving skills and communication that are
prevalent in so many of these families. This corresponds with Bowen’s theory that a lack
of a difl‘erentiated position in families will result in poorer conflict management and an
inability to remain objective in the face of anxiety or stress. Similarly, Garbarino and
Strange (1993) found that adult COAs had less family expressiveness and high degrees of
family conflict. The inability to maintain a differentiated sense of self is illustrated in
their finding that COAs have difficulty expressing or identifying feelings.

Karwacki and Bradley (1996) examined correlations between drinking
motivation, coping responses, goal expectations, and family-of-ori gin drinking problems
with college students' alcohol use. Their findings lend support to social learning theory
and conclude that family models make significant contributions to excessive alcohol use
in college students. Social learning theory is consistent with a Bowenian lens as children
may observe their parents, with low levels of differentiation, triangle alcohol in order to
shift the existing tension/anxiety away from the marital dyad. In turn, these children
would then choose alcohol as a triangle to aid in their own anxiety tolerance.

It is prudent to also consider how the concept of differentiation can be a useful
lens through which to view familial influence on triangling alcohol for college students

who may not necessarily be COAs. Haemmerlie, Steen, and Benedicto (1994)

26

investigated the role of conflictual independence (CI), adjustment to college, and
consumption of alcohol by college students. The construct of conflictual independence
was measm'ed fiem the Psychological Separation Inventory, which purports to delineate
having a relationship with one’s parents that is free fi'om excessive guilt, anxiety, anger,
and resentment-—all key Bowen concepts. A significant main effect was found for CI-
Mother, as subjects with high drinking had lower CI-Mother scores than did subjects with
low drinking scores. There were no significant findings for CI-Father or CI-Total.
Additionally, when compared to how well students were adjusting to college, a
significant positive relationship between Cl-Total was found with the overall scores of
Student Adaptation to College Questionnaire. A significant correlation was also found
between CI-Mother and CI-Father and the personal-emotional scores on the Student
Adaptation to College Questionnaire. The authors concluded that lower C I fiom both
parents was linked to the use of alcohol as a means to relieve emotional pressure. This
relationship supports the concept of differentiation, as students with lower levels of
conflictual independence were more likely to triangle alcohol as a way to tolerate their
emotional anxiety.

Role of anxiety and stress on drinking. As previously mentioned, the consumption
of alcohol in the student population does not remain static throughout their academic
career. Marlatt et al. (1998) report that expectancies for alcohol use may mediate peer and
normative influences at this time. Expectancies of alcohol have been reported to have a
mediating relationship in how COAs internalize parental drinking with their decisions
about drinking (Reese, Chassin, & Molina, 1994). However, Ellis et a1. (1997) state that

little is known about the formation of positive or negative expectancies with the severity

27

of parental alcoholism. Nonetheless, multiple studies have looked at the role of alcohol
consumption related to anxiety.

Essential to the framework of Bowen Theory is how an individual or dyad, acts in
the face of anxiety. It is especially prudent to consider the role of anxiety tolerance for
college students as Jones et al. (1992) state that perceived stress is greater for college
students compared to their same age peers who are not enrolled in college. Bartlett (2002)
reported on the adjustment of first year college students and showed 44.3% of students
felt overwhelmed by the demands college had on their time (compared to 31.6% at the
beginning of the year), 16% reported depressive symptoms (compared to 8.2%), and
44.9% rated their emotional health as being above average (compared to 52.4%). These
reports indicate a trend for maladjustment for a large portion of students. Students with
low levels of problem-solving abilities may struggle with managing the unique stressors
of college including handling situations when consuming alcohol, such as knowing when
to stop drinking, drinking and driving, and binge drinking (Dreer et al., 2004). Williams
and Kleinfelter (1989) found that students possessing greater confidence in their
problem-solving skills reported less use of alcohol to escape responsibilities and to cope
with negative emotions. This finding, which highlights an inverse relationship between
increased confidence and decreased alcohol use, supports Bowen’s theory of triangling in
order to alleviate emotional overload.

McCormack (1996) asked students to indicate if it was acceptable for a student to
use alcohol for the following situations: academic pressure, financial problems, family
problems, peer pressure, enhance sexual pleasure, at a party, on a date, to relax, and when

under stress in general. After comparing this to a 1990 survey with a similar list, the

28

author found an increased percentage (23% to 36%) of students advocated drinking when
under stress. Additionally, Lecci, MacLean, and Croteau (2002) found that drinking to
cope occurred to help reduce the level of perceived distress toward attaining life goals.
Other studies have found small effects (Camatta & Nagoshi, 1995) or no significant
relationship (Jones et al., 1992) for alcohol consumption and stress, concluding that
college students may be drinking more for socialization than for the alleviation of stress.

Specific to the role of triangling alcohol as a means to tolerate anxiety are studies
that focus on tension reduction theory (TRT). The premise of TRT is that alcohol is used
to reduce stressful states. Kushner et al. (1994) tested the prediction that this theory
would be evidenced with a strong correlation between anxiety and alcohol use as they
related to alcohol outcome expectancies. Their sample showed females scored
significantly higher in interpersonal sensitivity and general anxiety, while males were
significantly higher on the quantity/frequency of alcohol consumption and heavy drinking
composite. Men and women did not differ in terms of their perceived levels of alcohol
tension reduction expectancies. Additionally, males with higher tension reduction
expectancies had stronger associations between level of anxiety and alcohol
consumption.

Similarly, Pullen (1994) sought to find the relationships between alcohol abuse
and psychological/demographic variables, including anxiety. The author found that
students with high state anxiety abused alcohol more than students with low state anxiety.
Similarly, those with high trait anxiety, compared to low trait anxiety, also abused
alcohol more significantly. Students who reported abusing alcohol were more likely to be

from alcoholic families. Finally, the model that explained 59% of the variance in alcohol

29

abuse contained the following predictive variables: family abuse of alcohol, depression,
low levels of self-esteem, state anxiety, assertiveness, and having a lower grade point
average.

In a longitudinal study of Canadian students, Sadava and Pak (1993) investigated
the conditions under which stress can lead to substance abuse. Their findings showed that
alcohol use was related to stress, external locus of control, coping functions of drinking,
perceived support or sanctions for alcohol use, and the absence of social support.

Stewart, Karp, Pihl, and Peterson (1997) conducted two studies that examined
students’ use of substances in response to anxiety sensitivity, which was defined as
perceiving symptoms of anxiety as signs of catastrophic consequence. The first study
focused on substance use in general (non-specific to alcohol) and found female substance
users had a positive correlation between anxiety sensitivity scores and the use of
alcohol/drugs for reasons related to anxiety, while males also had non-significant positive
correlations. The second study focused on specific substances and found a significant
positive correlation among female alcohol users, regarding anxiety sensitivity scores and
drinking primarily for coping reasons. A similar pattern was found for men, but it did not
reach statistical significance.

In studying the various expectations students had towards alcohol consumption,
Sher, Walitzer, Wood, and Brent (1991) found that men had greater expectations for
activity enhancement, performance enhancement, and social lubrication with their
drinking than did women. Additionally, COAs reported stronger tension reduction,
performance enhancement, and social lubrication expectations than did non-COAs.

Moreover, students were found to drink more frequently if they had a greater motivation

30

to drink in order to increase comfort with others and for those who supported drinking in
social situations (Williams & Kleinfelter, 1989).

Although these studies showed a significant relationship, Noel and Cohen (1997)
did not find a significant relationship between stress and alcohol consumption. The
authors assessed how college students reacted to a specific time of stress by measuring
self-reported substance use during the week before exams and during a typical week in
the semester. Students reported drinking a lesser average of Standard Drinks per day
during the week before exams (1.06) compared to a typical week (1.48). This decrease
contradicts tension reduction theory by illustrating a negative correlation between
drinking and anxiety. The authors speculate that studying, which requires sobriety, serves
as an effective technique for managing the anxiety related to exam performance.

The increasingly large population of college students and their families will
undoubtedly face new stressors and anxieties throughout their academic careers. The
literature implies individuals with high levels of differentiation are more successful at
coping with college adjustment stressors. Johnson and Pandina (2000) found that students
who used avoidant strategies in the face of stress had greater levels of alcohol
consumption than those with active coping styles.

It is crucial to discover how the differentiation process interacts with levels of
anxiety and how alcohol is triangled as a dysfunctional coping strategy to assist college
students and their families with this life period. Specifically, Goluke, Landeen, and
Meadows (1983) detail the processes through which coping to stressors via alcohol can
eventually snowball into alcoholism. Although periodic use of alcohol to alleviate stress

may not be a risk factor for developing alcoholism, the gateway to addiction may be more

31

prevalent for students with low levels of differentiation, as they may become more reliant
on using alcohol as they have a restricted ability to react objectively to stressors.
Perceptions of others ’ alcohol use. Multiple studies have illustrated that the
perceptions students have towards the amount of alcohol their peers drink is related to the
amount that they themselves drink (Clapp & McDonnell, 2000; Demers et al., 2002;
Korcuska & Thombs, 2003; Perkins & Berkowitz, 1986; Perkins & Wechsler, 1996).
Perkins and Wechsler found that if a student believes normative student consumption of
alcohol is in heavy amounts, they are more likely to become involved in alcohol abuse.
Yet if a student personally feels very strongly about abstinence or restrained drinking,
then perception of other students’ drinking has very little effect. This dynamic relates to
Bowen’s notion of having a stable sense-of-self, as students choose behaviors based on
prevailing group norms and dogma, rather than from their own convictions.
Unfortunately, many students overestimate the amount of alcohol their friends
drink (Korcuska & Thombs, 2003; Perkins & Berkowitz, 1986; Perkins & Wechsler,
1996). Additionally, there seems to be a tendency to overestimate the amount of alcohol
students consume the more removed the comparison group is from the student. For
example, Korcuska and Thombs illustrated that students perceive their same-sex peers to
consume more alcohol than they do themselves and that the average same-sex student
consumes more than their same-sex peers. An immediate consequence of maintaining this
misperception is the greater tendency to consume more alcohol relative to the perceived
norm. For example, students that binge drink have been show to report higher perceptions
of fellow student alcohol consumption (Dreer et al., 2004). This highlights the notion that

even perceptions of one’s environment have a strong ecological influence on individuals.

32

Although there does not appear to be gender-based differences in levels of
perceived alcohol consumption (Demers et al., 2002), the effect of having a higher
perception related to greater alcohol consumption has been reported to be stronger in men
(Korcuska & Thombs, 2003). There is also evidence that drinking behaviors of males and
females may be quite similar as there is a stronger association between drinking
behaviors to perceived same-sex peer norms than differences in consumption based
solely on gender (Fromme & Ruela, 1994).

Fromme and Ruela (1994) studied the perceptions and actual alcohol use of
undergraduate students along with their parents and fiiends. The authors found that the
quantity of alcohol students consumed per drinking occasion was positively correlated to
the actual quantity mothers’ consumed per drinking occasion, the students’ perceptions of
the mothers’ alcohol consumption, the students’ perceptions of fathers’ alcohol
consumption, but not to the actual quantity of alcohol per drinking occasion. Positive
correlations were also found between students’ consumption and their friends’ actual
quantity of alcohol consumed per drinking occasion, frequency of days consuming
alcohol per week, and the total amount of alcohol consumed per week. Similar to other
research, positive correlations were also found for students’ actual use and their
perceptions of their fiiends’ use for actual quantity of alcohol consumed per drinking
occasion, frequency of days consuming alcohol per week, and the total amount of alcohol
consumed per week. These findings indicate both direct (actual use) and indirect
(perceptions) effects for student alcohol consumption. This suggests that important
individuals are modeling alcohol use, and that students’ perceptions regarding alcohol

consumption of important others is influential in their own use of alcohol. In terms of

33

Bowen Theory, the perceptions college students have about the normative use of alcohol
by family and fiiends may make it more likely that they choose alcohol as a means to
diffuse anxiety (a triangle) than some other method (i.e., other illicit substances).
Summary

Although no studies to date have directly examined the relationship between
Bowen’s theory of differentiation and alcohol use, inferences related to some of the
fimdamental principles have been demonstrated ill the literature. Evidence has been
presented to support the proposed model of study. Additionally, the literature also
suggests that effects of COA status (modeling triangle behaviors via alcohol
consumption) and perceptions of others’ use (the normative acceptance to triangle
alcohol as opposed to other substances) and expectations of alcohol reducing anxiety
(triangles) will influence the relationship between differentiation and alcohol

consumption and will also provide significant paths in the proposed model.

34

Chapter Three
Chapter Three provides conceptual and operational definitions of the variables
studied, along with the research objectives and hypotheses that described the predicted
relationship of these variables. Additionally, the sampling procedure, data collection, and
the measurements used will be described.
Methods
Conceptual and Operational Definitions
Dependent variables.
0 General alcohol consumption
Congptualization: The amount of alcoholic drinks generally consumed by
the student.
gmbnalizafion: This was acquired through a self-report measure. The
total amount of general alcohol consumption was the product of the
number of times the student drank in the past 30 days and the average
amount of alcohol consumed per occasion. One drink was equal to a
12-02 beer, a 4-oz glass of wine, a 12-oz wine cooler, or a 125-02 shot
of liquor, either straight or in a mixed drink.
0 Binge drinking
Conc_eptualization: Binge drinking is a behavior in which males have 5 or
more alcoholic drinks per occasion and females have 4 or more.
gmtionalization: This was acquired through self-report. The frequency
of binge drinking consisted of the number of times over the past 2

weeks that the student consumed 5 or more (if male, otherwise 4 or

35

more) drinks per occasion. One drink was equal to a 12-02 beer, a 4-02
glass of wine, a 12-oz wine cooler, or a 125-02 shot of liquor, either
straight or in a mixed drink.

Independent variables.

0 Differentiation

Conc_eptualization: An individual's emotional reactivity, ability to take an
“1” position, their degree of emotional cutoff from their family-of-
origin, and their degree of fusion with others in the face of
anxiety/stress.

Qm_ra_tionalization: Measured by the Differentiation of Self Inventory-
Revised. The DSI-R is a 46 item, six-point Likert scale with increasing
scores reflecting higher levels of differentiation.

0 Gender
M'onalization: Measured by self-report.
0 Age
M'onalization: Measured by self-report. Participation was restricted to
those who were between the ages of 18-22.
a Ethnicity/Race
M'onalization: Measured by self-report.
a BMI

Comtualization: An index used in the medical community for assessing

if an individual is within normal or obese ranges will be used as an

estimate for the body’s ability to metabolize alcohol.

36

Omtionalization: Height and weight was measured by self-report. The
value of BMI was derived by dividing weight in pounds by height in
inches squared and multiplying that product by 703 [BMI =
(lb/in2)703] (Centers for Disease Control and Prevention, 2004).

Moderator variables.
0 Parental alcoholism

Comtualization: The student’s feelings or perceptions that his mother or
father was an alcoholic.

ngg tionalization: Measured by an investigator created gender-specific
version of the Children of Alcoholics Screening Test. Both the mother
W-CAST) and father (F -CAST') versions consisted of 29 yes/no
questions. The presence of 6 positive responses indicated the presence
of an alcoholic parent.

0 Perceptions of normative student drinking

Comtualization: The amount of alcohol believed to be consumed by the
average student at MSU.

M'onalization: This was acquired through self-report. The total
amount of general alcohol perceived to be consumed was the product
of what the participant believed to be the number of times the average
student drank in the past month and the average amount of alcohol the
average student consumed per occasion. One drink was equal to a 12-
oz beer, a 4-02 glass of wine, a 12-oz wine cooler, or a 125-02 shot of

liquor, either straight or in a mixed drink.

37

o Perceptions of best friend’s use

Congptualization: The amount of alcohol believed to be consumed by
their best friend attending MSU.

Qperationalization: This was acquired through self-report. The total
amount of general alcohol perceived to be consumed was the product
of what the participant believed to be the number of times her best
friend at MSU drank in the past month and the average amormt of
alcohol consumed per occasion. One drink was equal to a 12-oz beer, a
4-02 glass of wine, a 12-oz wine cooler, or a 125-02 shot of liquor,
either straight or in a mixed drink.

0 Expectations of alcohol as a tension reducer

Comtualization: The student’s belief that consuming alcohol will
relieve tension and anxiety.

gmbnalization: The Tension reduction subscale was from Kushner et
al. (1998). Higher scores on the 9 item, 5-point Likert scale, indicated
greater expectations of tension reduction.

0 Expectations of alcohol as a social lubricant
Comafiizafion: The student’s belief that consuming alcohol will aid in
social interactions.
M'onalization: The social lubricant subscale is from Kushner et al.
(1998). Higher scores on the 8 item, S-point Likert scale, indicated

greater expectations of social lubrication.

38

Control variables.
0 Traditional student

Conggptualization: Participating students will be within the traditional age
range of starting college after high school and graduating within five
years.

M'onalization: As previously mentioned, participation was restricted
to those who were between the ages of 18-22. Screened for by listing
this as a restriction on the online version and verbally when students
were recruited in classrooms for paper-and-pencil versions.

0 Student is a US. citizen

Cor_rc_eptualization: The student must be a US. citizen whose family lives
in the US.

Qaglionalization: Screened for by listing this as a restriction on the
online version and verbally when students were recruited in
classrooms for paper-and-pencil versions.

Research Objectives

The overall objective of this research was to investigate the relationship between

levels of differentiation and alcohol use for undergraduate students. In order to reach this

objective it was necessary to understand the interactions between differentiation and

moderating variables. The following research objectives began with the goal of gaining

knowledge of the relationship between differentiation and alcohol use and then proceeded

with understanding the interaction relationship between differentiation and the potential

moderating variables.

39

Diflerentiation and General Alcohol Use and Binge Drinking
1. To identify the relationship with differentiation and general alcohol use and

binge drinking.

Four Factors of the DSI-R (Emotional Reactivity, “I ” Position, Emotional C utofif And
Fusion with Others) with General Alcohol Use and Binge Drinking
2. To identify the relationship of the four factors of the DSI-R with general

alcohol and binge drinking.

Moderating Eflects on the Relationship between Diflerentiation and General Alcohol Use
and Binge Drinking for College Students
3. To identify the moderating effects of COA status on the relationship between

differentiation and general alcohol use and binge drinking.

4. To identify the moderating effects of perceptions of the average MSU
student’s alcohol use on the relationship between differentiation and general

alcohol use and binge drinking.

5. To identify the moderating effects of perceptions of best friend at MSU’s
alcohol use on the relationship between differentiation and general alcohol use

and binge drinking.

6. To identify the moderating effects of expectations of alcohol serving as a
tension reducer on the relationship between differentiation and general alcohol

use and binge drinking.

7. To identify the moderating effects of expectations of alcohol serving as a
social lubricant on the relationship between differentiation and general alcohol

use and binge drinking.

Hypotheses
The following hypotheses of alcohol consumption were predicted based upon the
Bowenian concepts of differentiation and triangling and the literature related to COA
status, perceptions of alcohol use, and expectations of alcohol use as a tension reducer or
a social lubricant.
Difl'erentiation and General Alcohol Use and Binge Drinking
Hal: It was expected that college students with lower levels of differentiation
would be more likely to have greater amounts of general alcohol and greater
amounts of binge drinking consumption when controlling for BMI and age.
Four Factors of the DSI-R (Emotional Reactivity, “I " Position, Emotional C utofif and
Fusion with Others) with General Alcohol Use and Binge Drinking
Ha2: It was expected that college students with lower levels of the four factors of
the DSI-R would be more likely to have greater amounts of general alcohol
consumption and greater levels of binge drinking when controlling for BMI
and age. Further it was predicted that the interpersonal subscales (emotional
cutoff and fusion with others) would provide the greatest amount of
explained variance in alcohol use for males as it was thought they would
have a greater tendency to drink for social reasons. Additionally, it was
predicted that the intrapersonal subscales (emotional reactivity and “1”

position) would provide the greatest amount of explained variance in

41

alcohol use for females as it was thought they would have a greater tendency
to drink due to problem-solving and for coping reasons.
Moderating Eflects on the Relationship between Dfirentiation and General Alcohol Use
and Binge Drinking

Ha3: It was predicted that having an alcoholic parent would have a moderating

effect on the relationship between differentiation and general alcohol use
and binge drinking. Students that possessed lower levels of differentiation
would have greater amounts of alcohol consumption to the extent that they
had a greater number of alcoholic parents. It was thought COAs would have
a history of parents modeling alcohol use as a triangle to tolerate anxiety.

Ha4: Perceptions of the average MSU student’s alcohol use would have a
moderating effect on the relationship between differentiation and general
alcohol use and binge drinking. Students who had lower levels of
differentiation would have greater amounts of alcohol consumption to the
extent that they had greater amounts of perceived student use, presenting as
a normative means by which to diffuse anxiety.

Ha5: Perceptions of student’s best friend at MSU’s alcohol use would have a
moderating effect on the relationship between differentiation and general
alcohol use and binge drinking. Students who had lower levels of
differentiation would have greater amounts of alcohol consumption to the
extent that they had greater amounts of perceived best fiiend use,

presenting as a normative means to diffuse anxiety.

42

Ha6: Expectations of alcohol serving as a tension reducer would have a
moderating effect on the relationship between differentiation and general
alcohol use and binge drinking. Students who had lower levels of
differentiation would have greater amounts of alcohol consumption to the
extent that they had higher expectations that alcohol would alleviate
anxiety.

Ha7: Expectations of alcohol serving as a social lubricant would have a
moderating efl‘ect on the relationship between differentiation and general
alcohol use and binge drinking. Students who had lower levels of
differentiation would have greater amounts of alcohol consumption to the
extent that they had higher expectations that alcohol would serve as a social
lubricant as a means to help tolerate relational anxiety.

Decision Rule: A chance probability of .05 or less (p < .05) was required to reject

the null hypotheses.
Research Design

Individual undergraduate college students were the unit of analysis for this cross-

sectional, survey research. The dependent variables studied were college student alcohol
consumption (averaged Quantity X Frequency over a 30 day period) and frequency of
binge drinking (over a 2 week period). A minimum of 500 students were to be sampled as
this number would exceed Cohen’s (1992) criteria for a medium effect size with an alpha

value of .05 if there were a need to analyze males and females separately.

43

Measures

Dfirentiation

The Differentiation of Self Inventory-Revised (DSI-R) measured the degree of
differentiation for college students (see Appendix A). The DSI-R was chosen as it is one
of the few self-report scales that broadly cover the construct of differentiation. The
original DSI scale (Skowron & Friedlander, 1998) was a 43 item, six-point Likert
instrument consisting of four subscales: Emotional Reactivity (ER), 1 Position (1P),
Emotional Cutoff (EC), and Fusion with Others (F0). Estimates of internal consistency
reliability using Cronbach’s alpha yielded high values for the D81 full scale (a = .88) and
each ofthe subscales (ER or = .84, IP or = .83, BC or = .82, F0 or = .74). The DSI was
revised to improve the psychometrics of the Fusion with Others (F O) subscale (Skowron
& Schmitt, 2003). The revised FO scale changed from 9 items to 12, resulting in a 46
item scale. The internal consistency reliabilities improved for both the F O subscale (a =
.86) and the DSI—R full-scale (a = .92). Each item in the DSI-R presents a 6-point Likert
scale ranging from “not at all true of me” to “very true of me.” After reversing the
responses on select questions, scores on the subscales and total DSI-R are obtained by
summing the respective raw scores and dividing by the total number of items per scale.
For example, the DSI-R full-scale is obtained by summing all of the values and dividing
by 46, with scores reflect a range of 1 (low differentiation) to 6 (higher differentiation).

The D81 and DSI—R were originally tested with a sample that was 25-years old
and older. After analyzing the four-factor structure with the college students in this study,
three of the items (#11 fiom the “1” Position subscale, #32 from the Emotional Cutoff

subscale, and #38 hen) the Emotional Reactivity subscale), were omitted in order to

maintain a significant fora-factor structure for both males and females when analyzed
separately (see Appendix B for detailed analysis of the DSI-R with college students). The
full-scale and subscale Cronbach’s alpha values remained similar to the earlier studies
(ER 0. = .86; IP or = .77; BC or = .80; F0 or = .69; DSI-R or = .90), with the exception of
the FO subscale which was .17 lower than the Skowron & Schmitt (2003) sample.
Parental History of A lcohol Use

A minimally altered version of the Children of Alcoholics Screening Test (CAST)
(see Appendix C for the original CAST) was used to assess parental alcoholism. The
CAST is a 30—item instrument that asks about perceptions regarding parental alcohol
consumption. The total number of questions to which subjects reply “yes,” yields a total
score that can range from 0 to 30. Scores of six or more indicate the subject is likely the
child of an alcoholic (Jones, 1983; Kelly & Myers, 1996; Yeatman et al., 1994). The
CAST has demonstrated high alpha coefficients with studies of randomly selected
adolescents (0.95 - males, 0.97 - females, and 0.96 - combined gender, Dinning & Berk,
1989) as well as for adolescents who were selected from intact alcoholic families (0.90
and 0.88, Clair & Genest, 1992). Charland and Cdté (1992) found an extremely high test-
retest reliability (k = .83) and concurrent validity with the Structured Clinical Interview
for the DSM—III-R (SCID) and high CAST scores (k = .78). Additionally, Charland and
Cété’s study with 376 college students illustrated that the CAST yielded a false-negative
rate of 9.3% and a false-positive rate of 1.2%. In a study of adolescent offspring of
alcoholic fathers, Clair and Genest (2002) found test-retest reliability coefficient of .88

after a period of8 weeks.

45

As this study is interested in examining the effects of the number of alcoholic
parents, the CAST was rewarded to create a specific version for the mother (M-CAST,
see Appendix D) and the father (F-CAST, see Appendix E). The revised inventory
replaced the CAST’s terminology of “a parent” used throughout the inventory to
“mother” and “father” respectively. There are two questions on the CAST related to
specific parent use (“Did you ever think your father was an alcoholic?”/“Did you ever
think your mother was an alcoholic?"). The question for the father was omitted on the M-
CAST and the question for the mother was omitted on the F -CAST. The creation of
separate inventories afforded a more specific analysis for perceptions in which the
mother, the father, or both parents had a history with alcoholism. The total number of
alcoholic parents (0, 1, or 2) was indicated by the presence of six positive responses on
the M- and F -CASTs.

General Alcohol Use

Subjects were asked about their alcohol use over the past month in terms of
fiequency and quantity (see Appendix F, questions 18 and 19). Subjects reported how
many times they drank in the past 30 days and the average amount of drinks they
consumed per occasion. One drink was defined as a 12-02 beer, a 4-oz glass of wine, a
12-oz wine cooler, or a 125-02 shot of liquor, either straight or in a mixed drink. A total
average of drinks-per—month was calculated by multiplying the number of days by the
average number of drinks.

Binge Drinking
Male subjects were asked on how many occasions over the past 2 weeks they

consumed 5 or more beverages, while females were asked on how many occasions they

46

consumed 4 or more beverages (see Appendix F, questions 20 and 21). These numbers
reflected criteria for binge drinking according to Wechsler, Lee, Nelson, et al. (1991).
Expectancies of Alcohol as a Tension Reducer and Social Lubricant

Kushner et al.’s (1994) measure of alcohol outcome expectancies was used to
evaluate motivation for drinking (see Appendix G). Sher et al. created the scale after
being dissatisfied with other expectation measures that solely used dichotomous criteria
by including questions fi'om various measures into 12 a priori domains and running a
factor analysis to streamline their scale. The current scale consists of 35 questions using a
5 -point scale that load on the following four factors: Tension Reduction (9 items;
questions 1-9), Social Lubrication (8 items; questions 10-17), Performance Enhancement
(9 items), and Activity Enhancement (9 items), where higher sums of scores reflect
greater prevalence of respective expectations. The subscales have a common variance
ranging from .54 to .70 (Kushner et al., 1994). For the purposes of this study, the Tension
Reduction (a = .89) and Social Lubrication (a = .88) subscales were used in the analyses.
Perception of the Average Michigan State Student 's Alcohol Use

Subjects were asked how much alcohol over the past month in terms of fiequency
and quantity they believed the average Michigan State student consumed (see Appendix
F, questions 24 and 25). A total average of drinks-per-month was calculated by
multiplying the number of days by the average number of drinks.
Perception of Their Best Friend at Michigan State 's Alcohol Use

Subjects were asked about how much alcohol over the past month in terms of

fiequency and quantity they believed their best fiiend at Michigan State consumed (see

47

Appendix E, questions 28 and 29). A total average of drinkscper-month was calculated by
multiplying the number of days by the average number of drinks.
Demographic Information

In addition to the measurements, subjects completed a demographic sheet
recording their gender, age, race, grade level, body mass index, and living location (see
Appendix F).

Sample

It was originally proposed that 500 subjects would be recruited fi'om the
Department of Psychology’s Human Subjects P001 to voluntarily complete the battery
online for extra credit in their respective courses. Due to a high number of studies taking
place during the semester, a recruiting restriction was set by the department and
permission was given to recruit 162 subjects. In order to meet Cohen’s (1992) criteria for
a medium effect size, undergraduate students were then recruited from Family and Child
Ecology (F CE) undergraduate courses to anonymously complete paper-and-pencil
versions of the battery. Students were instructed that they could only complete the
measme once, so that if they had completed it from another class they were told not to
complete another one. As a recruiting incentive, six of the paper-and-pencil subjects were
randomly selected to win one of six $50 prizes. After recruiting subjects from the Human
Subjects Pool and F CE courses, there was a shortage of males so additional subjects were
recruited fi'om two courses in the Department of Computer Science. At the same time as
the Computer Science students were being recruited, the restrictions imposed by the
Human Subjects Pool were lified and additional subjects were recruited online. All

subjects read and agreed with the consent form prior to participation (see Appendix H).

48

Additionally, debriefing information regarding the nature of the study and instructions for
contacting the researcher and the University Committee on Research Involving Human
Subjects was provided (see Appendix 1). Responses to all measures were anonymous.

The proposal to University Committee on Research Involving Human Subjects
(UCRIHS) stated that efforts would be made to keep the total number of subjects
recruited to be around the range of 500 students. The data from the Computer Science
courses were never entered into the data set as they were the last to be recruited and their
paper-and-pencil forms were returned after the total number of subjects was extending
beyond the target range. Although the student responses were not entered into the data
set, their names were entered into the $50 prize lottery. All data collected from the PCB
courses were entered into the data set. Since it was a goal to recruit an approximately
equal number of males and females, the second wave of females recruited from the
Human Subjects Pool were omitted from the data set as the maximum number of female
subject had already been reached.

A total of 447 subjects (246 females, 20] males) were analyzed for this study
after omitting 39 subjects for missing data, 23 for having impossible or incongruent
information (e. g., what they reported their gender to be differed from the gender-specific
question related to binge drinking, or 17 days of binge drinking over the past 2 weeks),
and 13 for being outside of the l8-22-year old age range, The final sample exceeded
Cohen’s (1992) criteria for a medium effect size. A total of 256 individuals (102 females,
154 males) completed the online version of the battery via the Human Subjects Pool and

191 students (144 females, 47 males) from six FCE courses completed the paper-and-

pencil version of the battery.

49

Data Analysis

Once data entry into SPSS 13.0 was completed, the data were cleaned making
sure all values were within the appropriate ranges, omitting cases for missing data, and
examining outliers prior to running analyses.

Examination of the frequency distribution of general alcohol use (GAU) and
binge drinking outcome variables indicated the data were both positively skewed and
were subsequently transformed into categorical variables (see Appendix J). Males and
females were tested for differences in alcohol consumption with the respective categories.
The chi-square test of association indicated that males and females were different in
binge drinking behavior and therefore were analyzed separately for the remaining
analyses in order to avoid interpreting possible three-way interaction effects (see
Appendix K).

All of the variables were centered around their means prior to running the
regressions in order to minimize multicollinearity. Following the variance inflation factor
protocol (von Eye & Schuster, 1998), diagnosis for multicollinearity was implemented
prior to all regression analyses by regressing all of the predictors in the respective models
onto each other (see Appendix K). There were no issues of multicollinearity discovered
between the variables.

The hypotheses were tested through a series bf logistic regression analyses. In
' order for the difl‘erentiation and four factor models to be considered significant, the
overall model had to be significant (p < .05, for example see Equation 1) and the
respective variables had to significantly alter the odds of drinking classification (Wald

statistic was significant at p < .05), regardless of the magnitude of the odds ratio. As a

50

guide for the magnitude of effect size, Hopkins (2002) reports that an odds ratio of 1.50 is

considered small, 3.5 is moderate and 9.0 is large.
GAU = b0 + erifferentiation + szMI + b3Age (1)

When testing the moderating hypotheses the overall models and individual
interaction terms had to be significant. Additionally, in order to preserve the most
parsimonious model, the Interaction Model (for example see Equation 3) had to show a
significant block improvement over the Additive Model (for example see Equation 2) or
the Basic Model (if the Additive Model did not show a significant improvement over the

Basic Model) to justify adding additional variables.
GAU = b0 + erifferentiation + szMI + b3Age + b4COA (2)

GAU = bo + leifferentiation+ szMI + bgAge + b4COA + b5(Differentiation X

COA) (3)

Logistic regression does not possess an equivalent to the OLS R2. Although, pseudo R2
estimates are available, they do not describe the proportion of variance explained by the
predictors nor do they compare directly to other R2 measures. The Cox and Snell R2
(RICO, a 5...") and Nagelkerke’s R2 (RZNWC) are provided. The RIC“ a 5,... attempts to

create a value similar to the OLS R2 but the maximum range is typically below 1, while

51

the Rzngem“, manipulates the R20,x a 3.3.. so that the value ranges from 0 to 1. Both R2
are provided for the significant analyses.

Chapter Four presents the results of the logistic regressions as they relate to the
respective hypotheses. Complete analyses, including non-significant results, can be found

in Appendices L and M.

52

Chapter Four

This chapter presents the results of the logistic regression analyses for general
alcohol use (GAU) and binge drinking as they relate to the specific hypotheses. Only
results specifically related to the hypotheses are presented. For a more detailed
presentation of the logistic regressions, including the main effects found in the Additive
Models, see Appendix L for GAU analyses and Appendix M for binge drinking analyses.

Analyses of General Alcohol Use

As previously described, the continuous General Alcohol Use (GAU) measure (30
day Quantity X Frequency) was converted into a categorical variable consisting of 4
levels: Abstainers, Low Drinkers, Moderate Drinkers, and High Drinkers (see Appendix
J). As the four categories reflect increasing degrees of alcohol consumption, the nature of
this study’s originally stated hypotheses for a continuous dependent variable (e. g.,
students with lower levels of differentiation will be more likely to have greater amounts
of general alcohol consumption) was maintained with the categorical analyses (e.g.,
students with lower levels of differentiation will be more likely to have a greater
relationship with High Drinkers than Low Drinkers when controlling for BMI and age). .

As logistic regression is based on a dichotomous dependent variable, the four
GAU categories were contrasted in ways such that four tmique outcome dichotomies
were created:

1. Drinkers (Low, Moderate, High) versus Abstainers

2. Moderate Drinkers versus Low Drinkers

3. High Drinkers versus Low Drinkers

4. High Drinkers versus Moderate Drinkers

53

In each instance, the variable with the greater amount of alcohol consumption was coded
as a “1” in SPSS 13.0 and the lesser category was coded as a “0.” For example, in the
Drinkers versus Abstainers analyses, all persons in the Low, Moderate, and High
Drinking categories were coded as a “1” and Abstainers were coded as a “0.”

The analyses began by testing the hypotheses of differentiation and the four-factor
model on levels of alcohol consumption for each of the four dichotomies. The regression
equation that consists of differentiation, age, and BMI is referred to as the Basic Model
(see Equation 1, p. 51). The second step was to simply add the moderating variable of
interest to the regression equation, creating the Additive Model (see Equation 2, p. 51).
Thirdly, the Interaction Model was tested by adding the interaction term to the regression
equation (see Equation 3, p. 51). The moderating hypotheses were then tested with each
of the dichotomies with the criteria that the overall models and individual interaction
terms had to be significant. Additionally, in order to preserve the most parsimonious
model, the Interaction Model had to show a significant block improvement over the
Additive Model (simply controlling for the added variable) or the Basic Model (if the
Additive Model did not show a significant improvement over the Basic Model) to justify
adding additional variables. Finally, a Comprehensive Model was created by adding all of
the variables into one regression equation which was compared to the Basic Model for a
significant block improvement.

Dlfl'erentiation and GA U

It was hypothesized that college students with lower levels of differentiation

would be more likely to have greater amounts of general alcohol consumption when

controlling for BMI and age.

Males. There were differing levels of support for the hypotheses across the
drinking category comparisons. In the Drinkers versus Abstainers category the Basic
Model was significant (38(3) = 18.00, p < .001, 12%,, g 3,... = .09, mm“. = .20).
Differentiation had a significant negative coefficient indicating that the odds of drinking
were multiplied by .190 (95% CI = .068, .530) with a one unit difference in
differentiation -- an 81% (l - .19) decrease. As differentiation levels increased the
likelihood these male students would drink decreased, which was predicted in the
hypothesis. Additionally, age was significant as students were more likely to drink as
they got older.

The hypothesis was not supported in the comparisons of Moderate versus Low
Drinkers (38(3) = 2.92, p = .405), High versus Low Drinkers (36(3) = 2.21, p = .530), and
High versus Moderate Drinkers (38(3) = 5.35 p = .119) as the Basic Model failed to reach
a level of significance.

Females. The differentiation hypothesis was not supported in any of the
comparisons for females. All of the models failed to reach a level of significance:
Drinkers versus Abstainers (£0) = 6.91, p = .08), Moderate versus Low Drinkers (12(3)
= 2.62, p = .45), High versus Low Drinkers (36(3) = 2.06, p = .56), and High versus
Moderate Drinkers (38(3) = 4.10, p = .25). However, although the overall model was not
significant in relation to the population, differentiation was a significant predictor in
changing the odds of being a Drinker versus Abstainer for the sample by .510 (95% CI =

.268, .970) for every one unit increase in differentiation.

55

F our-F actors of the DSI-R with GA U

It was hypothesized that college students with lower levels of the four factors of
the DSI-R (Emotional Reactivity, “1” Position, Emotional Cutoff, and Fusion with
Others) would be more likely to have greater amounts of general alcohol consumption
when controlling for BMI and age. It was also predicted that the interpersonal subscales
(emotional cutoff and fusion with others) would provide the greatest amount of explained
variance in general alcohol use for males as there would be a greater tendency to drink
for social reasons. Additionally, it was predicted that the intrapersonal subscales
(emotional reactivity and “1” position) would provide the greatest amount of explained
variance in general alcohol use for females as there would be a greater tendency to drink
due to problem solving and coping reasons.

Males. There were difl'ering levels of support for the four-factor hypotheses
across the drinking category comparisons. The F our-F actor Model was supported in the
High versus Moderate Drinker comparison (38(6) = 16.35, p = .012, Ric,“ M= .13,
RZNM = .17). Although the interpersonal subscales (emotional cutoff and fusion with
others) were predicted to provide the greatest contribution for male drinking; the
intrapersonal subscales (emotional reactivity and “1” position) emerged as the significant
individual predictors. A one unit increase in Emotional Reactivity (being less emotionally
reactive) increased the likelihood of being a High Drinker (OR = 1.928, 95% CI = .999,
3.720), whereas increasing the “1” Position value by one unit decreased the likelihood of
being a High versus a Moderate Drinker (OR = .374, 95% CI = .198, .706). For males it

appears that maintaining a sense-of-self (“1” Position) made it less likely that one would

56

escalate into the High Drinking category, while remaining less emotionally reactive to
stress increased the likelihood of being a High versus a Moderate Drinker.

The overall models were not significant for the Moderate versus Low Drinkers or
the High versus Low Drinkers comparisons. Although the Drinkers versus Abstainers
comparison was significant as an overall model, none of the four factors significantly
changed the odds of group membership.

Females. The four-factor hypothesis was not supported in any of the comparisons.
Each of the models failed to reach a level of significance: Drinkers versus Abstainers
(38(6) = 9.21, p = .162), Moderate versus Low Drinkers (38(6) = 2.97, p = .813), High
versus Low Drinkers (£0) = 3.41, p = .757), and High versus Moderate Drinkers (x2(6)
= 4.49, p = .610).

Moderator Hypotheses

It was hypothesized that students with lower levels of differentiation would have
greater amounts of general alcohol use to the extent that they had greater perceptions of
others’ use (the average MSU student’s use and best friend at MSU’s use), greater
expectations of alcohol as a tension reducer and social lubricant, and if they were a child
of an alcoholic (COA). The male or female analyses failed to provide support for the
perceptions of best fiiend at MSU’s use, COA status, tension reduction, or social
lubricant interaction hypotheses as the interaction term failed to reach a level of
significance across all of the comparison groups. The perception of MSU student use was
partially supported by both the males and females.

Males. The High versus Moderate Drinkers comparison did provide support for

the interaction hypothesis as the Interaction Model was significant (xz(5) = 13.21, p =

57

.022, RICO, g 5.,“ = .10, Rznwgk, = .14) and showed a block improvement over the Basic
Model (38(2) = 7.36, p = .025), as the Additive Model failed to reach significance. The
interaction of differentiation and perception of MSU student use was a significant
individual predictor with individuals being .978 (95% CI = .957, .999) times as likely to
be a High Drinker- versus a Moderate Drinker (see Table 1 and Figure 3, p. 60, for
illustration of the interaction effect). Subjects with a standard deviation difference (20.25)
of the interaction were .63 times as likely to be a High Drinker versus a Moderate
Drinker.

The overall Interaction Model failed to reach a level of significance for the
Moderate to Low Drinkers model. Although the overall models were significant for the
Drinkers versus Abstainers and the High versus Low Drinkers comparisons, neither had a

significant block improvement over the Additive Model nor had a significant interaction

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

term.
Table 1
Logistic Regression Predicting Male Students that are High versus Moderate Drinkers
- Perceptions of MSU Student Use
Interaction Model b SE Wald OR 95% CI
DSI—R -0.23 0.40 0.34 0.798 0.375 1.699
BMI 0.11 0.07 2.28 1.116 0.968 1.288
Age 0.12 0.16 0.56 1.124 0.827 1.528
MSU Use 0.01 0.01 1.43 1.006 0.996 1.017
DSI-R X MSU Use -0.02 0.01 4.23“ 0.978 0.957 0.999
* p < .05, two-tailed. ** p < .01, two-tailed.

 

Females. The High versus Low Drinkers comparison was significant as an overall
model (38(5) =' 18.01, p = .003, RICOXgM= .11, Rimm = .15) and had a significant

block improvement over the Additive Model (38(1) = 4.59, p = .032). The interaction of

58

differentiation with perception of MSU student use was significant with an odds ratio of
.983 (95% CI = .967, .999) for every unit increase in the interaction term (see Table 2 and
see Figure 4, p. 62, for illustration of the interaction effect). Subjects with a standard
deviation difference (21.57) of the interaction were .65 times as likely to be High versus
Low Drinkers.

The overall Interaction Model failed to reach a level of significance for the
Drinkers versus Abstainers and Moderate versus Low Drinkers comparisons. Although
the Interaction Model was significant as an overall model in the High versus Moderate
Drinkers comparison, the interaction term failed to reach a level of significance and the

model did not demonstrate a significant block improvement over the Basic Model.

 

Table 2

Logistic Regression Predicting Female Students that are High versus Low Drinkers —
Perceptions of MSU Student Use

 

 

 

 

 

 

 

 

 

 

 

 

 

[ b 1 SE [ Wald [ 0R 1 95% CI
Interaction Model
DSI-R -0.03 0.30 0.01 0.968 0.539 1.741
BMI -005 0.05 0.89 0.951 0.858 1.055
Age 0.05 0.14 0.14 1.056 0.797 1.398
MSU Use 0.02 0.01 9.85” 1.015 1.006 1.025
DSI-R x MSU Use 0012 0.01 434* 0.983 0.967 0.999

 

 

" p < .05, two-tailed. *" p < .01, two-tailed.

 

59

 

 

 

 

 

,_____,

.‘ ...,- - - MSU Use—High ’
i — + - MSU Use Average
* —-— MSU UsefiLow

 

 

 

 

Predicted Probability of Drinking Category

 

 

Low Average High
Differentiation

Figure 3. The differentiation and perception of MSU student alcohol use interaction for
the comparison of male High versus Moderate Drinkers. Note: Predicted probability

scores > .50 indicate a greater probability for being a High Drinker and < .50 indicate a
greater probability for being a Moderate Drinker. Values are when controlling for BMI

and age.

The Comprehensive Model

All of the variables of interest were entered simultaneously into a logistic
regression in order to identify significant contributions of individual variables while
controlling the influence of the remaining variables. All of the Comprehensive Models
showed a significant improvement over the Basic Model.
Males. Each of the comprehensive models was significant: Drinkers versus Abstainers
(38(13) = 66.18, p < .001, 8%.... 5.... = .28, 192.4%“... = .66), Moderate versus Low
Drinkers (38(13) = 34.99, p = .001, 8%.... 5,... = .24, 18......“ = .33), High versus Low
Drinkers (38(13) = 86.95, p < .001, 8%.... 3.... = .51, mm... = .68), High versus
Moderate Drinkers (38(13) = 51.44, p < .001, 18c... 5.... = .34, 82...“... = .46) (see
Table 3, p. 64). Interestingly, there were no individual predictors in this model that
significantly changed the odds of being a Drinker versus Abstainer. Perceptions of best
fiiend use (OR = 1.020) and expectations of tension reduction (OR = 1.109) significantly
improve the odds of being classified as a Moderate Drinker versus a Low Drinker.
Perception of best fiiend usage improved the likelihood of being classified as a high
drinker (OR = 1.037) while being a COA actually decreased the likelihood of being a
High Drinker versus Low Drinker (OR = .188). Perception of best fiiend use was the only
significant individual variable in the comparison of High versus Moderate Drinkers (OR

= 1.026).

61

 

0.8 —. - .- __.__n _ Lb-

 

 

 

 

 

 

 

 

or
'5 e
«g 0.7 —»——~—~ Jae-iff- ——-e e “n .2 w
3 0.6 Lani. __,____m_
a ~~..,
e e 0'5 ‘F‘T' -‘=*<-_-’—‘E::::r-r— '7‘ won-MSU Use High 3
‘= o . .
§ 5’ 0-4 ‘F -*——~ —~---~ ~ i-f-MSU Use Average:
‘2: 8 0 3 _ ___ g—I——MSU Use Low
11. '/ — — ~
3 02 — ____- 4 ~ — —_ .4 _
ii 0.1 e——-- —~~ ~ «—
IL Ol-—————e._.___.-__,

Low Average High

Differentiation

Figure 4. The differentiation and perception of MSU student alcohol use interaction for
female High versus Low Drinkers. Note: Predicted probability scores > .50 indicate a
greater probability for being a High Drinker and < .50 indicate a greater probability for

being a Low Drinker. Values are when controlling for BMI and age.

62

Females. The Comprehensive Model was significant for each of the female
comparisons: Drinkers versus Abstainers (38(13) = 90.70, p < .001, RIC... .. 5.... = .31,
alum = .57), Moderate versus Low Drinkers (38(13) = 45.72, p < .001, Ric... 5,... =
.28, 82......” = .38), High versus Low Drinkers (38(13) = 88.84, p < .001, 8%., g 5,... =
.45, 1?sz = .60), High versus Moderate Drinkers (38(13) = 23.41, p = .037, 82.... ..
5...... = .16, Rznmm = .21) (see Table 4, p. 65). Only the tension reduction variable
showed an individual level of significance in changing the odds of being a Drinker versus
Abstainer (OR = 1.237). Age (OR = 1.461), perceptions of best friend use (OR = 1.028),
and expectations of tension reduction (OR = 1.118) were significant in changing the
likelihood of being classified as a Moderate versus Low Drinker. Perceptions of MSU
student use (OR = 1.016), perceptions of best friend use (OR = 1.048), and the interaction
between differentiation and the perception of best fiiend use (OR = .970, see Figure 5, p.
66) and tension reduction (OR = 1.135) were significant for the High versus Low Drinker
comparison. No individual predictors in this model significantly changed the odds of

being a High Drinker versus a Moderate Drinker.

63

 

Hugo w

Ooaearoammfin Kean; men Kan Q>C nefiomonam

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

01388 8 38038 Unis—w Em... Una—83 8 Em: 01383 8
>cm§=na 8 has Dan—83 rose 01383 £9338 Una—88
gag OW $35 0% gig OW $35 ON
02% rue Noam o3 rune o.oo o.ooo o.oo oxfio
wZ: o.oo o.oo; one o.ooo 93 926 NS :3
>mo Pom n; 3 o.oo o.omu o.oo fowo P: o.ooo
ch enaooeaosm 9: o.ooo NB _.o_~ Ne: roww o.oo o.ooo
OMEN X ch eonoooaoam o.on o.ooo obo o.ooo o.om o.oom o.oo o.ooo
mom. manna o.: o.ooo flaws... o.ouo No.8... _._ rowu 5.3... a. o.oNo
Um?” X mom" 113a o.oo o.ooa o.~o o.ooo rum Bo; ohm o.oow
00> o.oo o.ooo o.ow o.ooo #5... 95¢ who o.ooo
UmEN X 00> o.oo N30 95 ream o.om PNE row was
H038: x2330: o.oo rmoo 93... :8 Now Co.» o.oo rowu
Um?” X H0390: ”mucosa: o.oo _.~ou on: 93¢ 98 o.ooo o.oo : G
moor: Economic: N3 Zoe o.oo o.ooo o.oo o.omm o.oo o.ooo
Um?” X meow: reggae: o.oo _.oom o.oo row— 98 fowo ohm obs:

 

 

 

 

 

 

 

 

 

 

a. .e A .om. "gram—on. a... a A .or gem—=09.

 

 

 

A020 A

0033030380 7888 8.. 330.0 O>C nefiomoag

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

9.8.83 8 78838 9.8.88 2mm: 9.8.88 8 Em: 9.8.88 8
2008803 8 ~82 Unified. hos. 01383 32.088 01383

$08 0w $88 0% (<08 0% £05 OX

Om?” _.oo o. A oo 98 o.ouo o. _ o ..~u~ o. _ A :3

95 o.oA o.oo. o.uo o.omo o._A o.ooo 93 9vo

>m0 ouw Zmo A. A we _.Ao_ o. A N room o.oo o.oow

ch $80383 o.oo o.ooo o.oo rooo who... :vo row o.ooo

Um?” X ch $80388 o.oo room :A o.ooo Nam o.omw foo o.omA
mg» m. 103 Poo _.o_ m o.oo: #on No.3: _.oAm Now _.ooo

Um?” X w0m. 3.88 93 _.ooo o.A~ o.ooo Aha... o.ooo obo room

00> om» o.ooo o.oo o.ooo o. _o o.og ohm o.ooo

Um?” X 00> o.oo :3 PAN room PAN #on o.oo o.ooo

H038: ”0828: 93...... Emma ohms... _._ 3 Saws... :3 Now Foam

09% X H038: ”0898: PS _.oo.~ “Co o.ooo o.oo o.omu foo rom—
moomfl 581038: o.oo :AN o.oo fou— o.~o o.oNA o.o_ row—

Umzw X 808. C8188: o.oo :om row :3 PS rouse o.oo o.vo

 

 

 

 

 

 

 

 

 

 

_.. e A .3. gods—0a. are A .3. 30.5.0?

 

 

65

 

 

 

 

2'
g 0 9 ,E__ __.___\_;: --r ‘ if: i _

é.‘; _ ___,:_ __,
E 0.8 ‘T— \\ . .
'5 0.7 4%»: r ._ .___.\_ mm * - " ‘ r r + a- My _---- ____ ___.
z. a, 0.6 4“” ,__._ _,__ _ _ .-__\‘___ __.____n__ . W ' - - -Q- . - BF U88 High -
E O a ---- 1‘ '
'8 3 05 ll_____,,_ WW—W—— _ 4 *——~ _ —+-BF Use Average'
g 8 0.4 LN: A A __ -_ _________.._ v -- -~ ~ ‘—I—BF Use LOW
n. #5. _-_.,__ _- _ _ --
i
g T a .

Low Average High

Differentiation

Figure 5. The differentiation and perception of best friend at MSU’s alcohol use
interaction for female High versus Low Drinkers. Note: Predicted probability scores > .50
indicate a greater probability for being a High Drinker and < .50 indicate a greater
probability for being a Low Drinker. Values are when controlling for all other variables

in the Comprehensive Model.

Analyses of Binge Drinking
As was done with the General Alcohol Use measure, the continuous Binge
Drinking variable (number of days in the past 14 days the subject drank at binge levels - -
5 drinks per occasion for males, 4 drinks per occasion for females) was converted into a
categorical variable consisting of 4 levels of use: Abstainers (those that did not drink),
Nonbinging Drinkers (subjects that consumed alcohol, but not at binge levels),
Occassional Binge Drinkers (l or 2 times in the past 2 weeks), and Frequent Binge
Drinkers (3 or more times in the past 2 weeks) (see Appendix J). As the four categories
reflect increasing degrees of alcohol consumption, the nature of this study’s hypotheses
which were stated for a continuous dependent variable (e. g., students’ lower levels of
differentiation will be more likely to have greater amounts of general alcohol
consumption) was maintained with the categorical analyses (e.g., students’ lower levels
of differentiation will be more likely to have a greater relationship with High Drinkers
than Low Drinkers when controlling for BMI and age).
The four General Alcohol Use categories were contrasted in ways such that four

unique outcome dichotomies were created:

1. Binge Drinkers (Occasional & Frequent Bingers) versus Nonbinging Drinkers

2. Occasional Binge Drinkers versus Nonbinging Drinkers

3. Frequent Binge Drinkers versus Nonbinging Binge Drinkers

4. Frequent Binge Drinkers versus Occasional Binge Drinkers
The identical coding procedure and process of analysis from the General Alcohol Use

tests were implemented with the Binge Drinking analyses.

67

Differentiation and Binge Drinking

It was hypothesized that college students with lower levels of differentiation
would be more likely to have greater amounts of binge drinking when controlling for
BMI and age.

Males. The differentiation hypothesis was not supported in any of the
comparisons. All of the Basic Models failed to reach a level of significance: Binge
Drinker versus Nonbinging Drinkers (£6) = .75, p = .861), Occasional Binge Drinkers
versus Nonbinging Drinkers (x20) = .50, p = .929), Frequent Binge Drinkers versus
Nonbinging Drinkers (12(3) = 2.48, p = .479), Frequent Binge Drinkers versus
Occasional Binge Drinkers (76(3) = 6.83, p = .078).

Females. The differentiation hypothesis was not supported in any of the
comparisons. All of the Basic Models failed to reach a level of significance: Binge
Drinker versus Nonbinging Drinkers (36(3) = 3.38, p = .337), Occasional Binge Drinkers
versus Nonbinging Drinkers (12(3) = 1.40, p = .705), Frequent Binge Drinkers versus
Nonbinging Drinkers (x20) = 5.81, p = .121), Frequent Binge Drinkers versus
Occasional Binge Drinkers (38(3) = 4.59, p = .205).

F our—F actors of the DSI-R with Binge Drinking

It was hypothesized that college students with lower levels of the four factors of
theDSI-R (Emotional Reactivity, “1” Position, Emotional Cutoff, and Fusion with
Others) would be more likely to have greater amounts of binge drinking when controlling
for BMI and age. Further it was predicted that the interpersonal subscales (emotional
cutoff and fusion with others) would provide the greatest amount of explained variance in

binge drinking for males as there would be a greater tendency to drink for social reasons.

68

Additionally, it was predicted that the intrapersonal subscales (emotional reactivity and
“1” position) would provide the greatest amount of explained variance in binge drinking
for females as there would be a greater tendency to drink due to problem solving and
coping reasons.

Males. The four-factor hypothesis was partially supported in the binge drinking
comparisons for males. The F our-F actor Model was supported in the Frequent versus
Occasional Binge Drinkers comparison (38(6) = 15.82, p = .015, R20», & 5...." = .10,
Rszlkm = .13). Interestingly, it was predicted the interpersonal subscales (emotional
cutoff and fusion with others) would provide the greatest contribution for male drinking;
however, the “1” Position subscale was the only significant predictor of binge drinking
categories, while the Emotional Cut-off subscale approached significance (p = .056).
Students who had a one unit higher “1” Position score were .488 (95% CI = .268, .889)
times more likely to be Frequent Binge Drinkers versus Occasional Binge Drinkers (see
Table 5). This finding indicates that as students become more competent in the
development of an “1” position, they are less likely to be Frequent Binge drinkers in

comparison to Occasional Binge Drinkers.

 

Table 5

Logistic Regression Predicting Male Students that are Frequent Bingers versus
Occasional Bingers — Four Factors

 

 

 

 

 

 

 

b SE Wald OR 95% CI
Emotional Reactivity 0.42 0.31 1.91 1.524 0.839 2.769
“I” Position -0.72 0.31 5.50“ 0.488 0.268 0.889
Emotional Cutoff -0.50 0.26 3.67 0.606 0.363 1.012
Fusion with Others -0. 14 0.41 0.12 0.868 0.389 1.937
BMI -0.05 0.05 0.96 0.952 0.862 1 .05 1
Age 0.20 0.15 1.91 1.222 0.920 1.623

 

 

 

 

 

 

 

 

* p < .05, two-tailed. ** p < .01, two-tailed.

 

69

 

The overall model was not significant for the Binge Drinker versus Nonbinging
Drinkers (x2(6) = 1.327, p = .970), Occasional Binge Drinkers versus Nonbinging
Drinkers (x2(6) = 2.77, p = .838), Frequent Binge Drinkers versus Nonbinging Drinkers
comparisons (12(6) = 3.78, p = .706).

Females. The four-factor hypothesis was not supported in any of the comparisons.
All of the Four-F actor Models failed to reach a level of significance: Binge Drinker
versus Nonbinging Drinkers (x2(6) = 4.99, p = .546), Occasional Binge Drinker versus
Nonbinging Drinkers (12(6) = 4.58, p = .599), Frequent Binge Drinkers versus
Nonbinging Drinkers (38(6) = 7.39, p = .286), and Frequent Binge Drinkers versus
Occasional Binge Drinkers (36(6) = 9.05, p = .171).

Moderator Hypotheses

It was hypothesized that the students with lower levels of differentiation would
have greater amounts of binge drinking to the extent that they had greater perceptions of
others use (the average MSU student use and best friend at MSU’s use), greater
expectations of alcohol as a tension reducer and social lubricant, and they were COAs.
The male or female analyses failed to provide support for the perceptions of best friend at
MSU’s use, tension reduction, or social lubricant interaction hypotheses as the interaction
term failed to reach a level of significance across all of the comparison groups.

Although the perception of MSU student use interaction hypothesis was not
supported by males, it was partially supported by the females. The Frequent Binge
Drinkers versus Nonbinging Drinkers comparison had a significant Interaction Model
06(5) = 14.01, p = .016, RZCM 5,... = .10, 82%“, = .13) and had a significant block

improvement over the Additive Model (380) = 6.16, p = .013). The differentiation X

70

perception of MSU student use interaction term was the only significant individual
predictor (see Table 6 and Figure 6, p. 73 for interaction effect). Students with a one unit
higher interaction value were .975 (95% CI = .955, .996) times more likely to be a
Frequent Binge Drinker than a Nonbinging Drinker. Students who had a difference in the
interaction term equivalent to one standard deviation (21.57) were .58 times more likely

to be a Frequent Binge Drinker than a Nonbinging Drinker.

 

Table 6

Logistic Regression Predicting Female Students that are Frequent Bingers
versus Nonbingers —- Perceptions of MSU Student Use

 

 

 

 

 

 

 

 

 

 

 

Interaction Model b SE Wald OR 95% CI
DSI-R -0.24 0.34 0.48 0.790 0.403 1 .546
BMI -O. 12 0.07 3.38 0.884 0.776 1.008
Age 0.09 0.15 0.35 1.096 0.810 1.482
MSU Use 0.01 0.01 2.03 1.008 0.997 1.019
DSI-R X MSU Use -0.03 0.01 5.48‘I 0.975 0.955 0.996

 

 

 

 

* p < .05, two-tailed. " p < .01, two-tailed.

 

Similarly, the COA status interaction was not supported for males, but was
partially supported for females. The overall Interaction Model for Frequent versus
Occasional Binge Drinkers was significant (12(5) = 11.19, p = .048, Rza,“ 5n,“ = .07,
R3natueumke = .10). The only significant predictor was the interaction between
differentiation and COA status. Students with a one unit higher interaction value were
6.773 (95% CI = 1.209, 37.949) times more likely to be a Frequent Binge Drinker than an

Occasional Binge Drinker (see Table 7, p. 72, and Figure 7, p. 74, for interaction effect).

71

 

Table 7

Logistic Regression Predicting Female Students that are Frequent Bingers versus

Occasional Bingers — COA Status

 

 

 

 

 

 

 

[ b 1 SE 1 Wald OR 95%01
Interaction Model
DSI-R -0.27 0.34 0.61 0.767 0.395 1.489
BMI -0.12 0.06 3.75 0.887 0.786 1.001
Age -0.08 0.15 0.28 0.926 0.697 1.231
COA 0.35 0.46 0.58 1.413 0.579 3.447
DSI-R x COA 1.91 0.88 473* 6.773 1.209 37.949

 

 

 

 

 

 

 

 

* p < .05, two-tailed. *"‘ p < .01, two—tailed.

 

The Comprehensive Model

All of the variables of interest were entered simultaneously into a logistic
regression in order to identify significant contributions of individual variables while
controlling the influence of the remaining variables.

Males. The Occasional Binge Drinkers versus Nonbinging Drinkers failed to
reach a level of significance when all of the predictor variables were entered into one
model (.603) = 19.27, p = .115). The overall models were significant for the Binge
Drinkers versus Nonbinging Drinkers (36(13) = 41.1 1, p < .001, 8%,, r 3,... = .20,
RZNM“. = .34), Frequent Binge Drinkers versus Nonbinging Drinkers (3803) = 64.24,
p < .001, R20... 3. s..." = .40, Rsz = .60), and Frequent Binge Drinkers versus
Occasional Binge Drinkers (.603) = 46.1 1. p < .001, RIC... 5,... = .26, 82m...“ = .35)
(see Table 8, p. 78). Perceptions of best friend use (OR = 1.019) was the only individual

variable to significantly improve the odds of being classified as a Binge Drinker versus a

72

 

 

 

 

 

 

 

 

 

 

 

 

 

0.8
° 9
g 0.7 —W—‘—~»W—— ———— ____ we —
E
045 E 0.6 —#
§§05 ° — iu-ou-MSUUseHigh .
E: 0.4 / “ 1-+-MSUUseAverage;
2 .5 03 -kww I—u—MSU Use Low
“"8 - 1
t _ __- rrrrrr
ég 0.2 F-
‘g 0.1 a
n.
O 1 r —~r
Low Awrage High

Differentiation

Figure 6. The differentiation and perception of MSU student alcohol use interaction for
female Frequent Binge Drinkers versus Nonbinging Drinkers. Note: Predicted probability
scores > .50 indicate a greater probability for being a Frequent Drinker and < .50 indicate
a greater probability for being a Nonbinging Drinker. Values are when controlling for

BMI and age.

73

 

9.0.0
mum
l
I
l
i
I

if,
I
'1
il
('0
1
Ii
1.
‘l

 

 

 

 

A ' + Non-COAS‘

p
00
i
i
l
i
i
i
i
l

 

 

 

pp
AN

 

Predicted Probability of Binge
Drinking Category
0 O
A or

ii

 

O
a

Low Average High

Differentiation
Figure 7. The differentiation and COA status interaction for female Frequent versus
Occasional Binge Drinkers. Note: Predicted probability scores > .50 indicate a greater

probability for being a Frequent Drinker and < .50 indicate a greater probability for being

an Occasional Binge Drinker. Values are when controlling for BMI and age.

74

Nonbinging Drinker. Age (OR = 2.431) and perceptions of best fiiend use (OR = 1.040)
were the only individual variables to significantly improve the odds of being classified as
a Frequent Binge Drinker versus a Nonbinging Drinker. Age (OR = 2.43), perceptions of
best friend use (OR = 1.018), and the differentiation X expectations of tension reduction
(OR = .826, see Figure 8, p. 77, for interaction) were the only individual variables to
significantly improve the odds of being classified as a Frequent Binge Drinker versus an
Occasional Binge Drinker.

Females. The Binge Drinkers versus Nonbinging Drinkers (3603) = 85.81, p < .001,
R20... 8 3...... = .33, Rznmm = .47) , Occasional Binge Drinkers versus Nonbinging
Drinkers (1203) = 45.05, p < .001, R20... 8; 5...... = .27, 182.....3...r....1,.3 = .36), Frequent Binge
Drinkers versus Nonbinging Drinkers (x203) = 100.54, p < .001, R20,” 3..." = .52,
1?sz = .69), and Frequent Binge Drinkers versus Occasional Binge Drinkers (3803)
= 39.46, p < .001, RIC,“ s..." = .24, Rszmag. = .32) comparisons were each significant
as overall models (see Table 9, p. 79). Perceptions of best fiiend use (OR = 1.03 7),
expectations of tension reduction (OR = 1.166), and the differentiation X expectations of
tension reduction interaction (OR = .862, see Figure 9, p. 80, for interaction) were the
only individual variables to significantly improve the odds of being classified as a Binge
Drinker versus a Nonbinging Drinker. Perceptions of best friend use (OR = 1.030),
expectations of tension reduction (OR = 1.147), and the differentiation X expectations of
tension reduction (OR = .861, see Figure 10, p. 81, for interaction) significantly improved
the odds of being classified as an Occasional Binge Drinker versus a Nonbinging
Drinker. Perceptions of best fi'iend use (OR = 1.050) and expectations of tension

reduction (OR = 1.251) were the only individual variables to significantly improve the

75

odds of being classified as a Frequent Binge Drinker versus a Nonbinging Drinker.
Perceptions of best friend use (OR = 1.019) and expectations of tension reduction (OR =
1.081) were the only individual variables to significantly improve the odds of being
classified as a Moderate Drinker versus a Low Drinker.

Overall, few hypotheses received support from the disparate analyses. As an
individual predictor, differentiation only changed the odds of group membership in the
male Drinkers versus Abstainers comparison (see Table 10, p. 82). The differentiation X
MSU perceptions interaction for the female high versus low drinker comparison indicated
that differentiation moderated the main effect of MSU perceptions, which was contrary to
the proposed hypothesis (see Table 10). Additionally, although there were significant
interactions for male High versus Moderate drinkers, female frequent Binge Drinkers
versus Nonbinging Drinkers, and female frequent versus Occasional Binge Drinkers, in
each instance there was no significant main effect in either the differentiation variable or
the other respective interaction variable (see Table 10). Chapter Five provides a more

thorough discussion of the key findings.

76

 

 

 

 

 

 

 

 

 

 

 

 

 

 

8 0.9 i t W _.___,_L_
‘5 08
‘ m e —.:‘—.‘—“'-rr—r—’———"-—~~— —~~—
'5 8 0.7 “ ~~~~ “5:4 Wm
’33 0.6 W ‘%r —W =men-TR High
g ‘2 0,5 //W.W/—____°Q___ i-1--TRA\eragei
g; as W—W W—W ___, _._ ”.5 L...
3 .g 0.3 WWWWW W WW
3°02 ___._..w------
E 0.1 WWWW W __
n.
0 “*‘f-"“"—‘ C“. T7
Low Average High
Differentiation

Figure 8. The differentiation and expectations of tension reduction interaction for male
Frequent versus Occasional Binge Drinkers. Note: TR = Tension reduction. Predicted
probability scores > .50 indicate a greater probability for being a Frequent Drinker and <
.50 indicate a greater probability for being an Occasional Binge Drinker. Values are

when controlling for all other variables in the Comprehensive Model.

77

 

H020 w

008308380 788; 84 78.0 AwA:m0 01:55 00.08100

 

AwA:m0 61:88

88:0 2::AOA:mA:m

33:03 AwA:m0
9.888 58:0

33:03 wA:m0
©1288 88:0

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

. 78:38 8 000883: E: 0
:38; germ; 0 9588 m
€95 OAw €08 OAw $85 CAN
DE.” A .Nm N. A uA 08m A .03 A .3 chum
szAA 92 A .23 obA o.ooo A .Nm obwo
1.9mm 9.3 A .NoA Sum: NAu A Aka... A .AVN
736 380982.. Abe AbA A 93 Abcm 0.”: Abcw
UwArAw X ASmC $080388 93 Abom 9cm Abou who 0.3m
w0mA 3.88 9qu AbAo A Aha... AbAo Auuc: AbA w
DmAAw X w0mAm10RA obm Abou AQA AbNA VAN Ab;
00> Nuo 9qu NuA obAm and 93:
DE.” X 00> obo Abwm Ab.» Abuo PS Amum
H088: Away—28: A .3 A boa PA: A be» Now A .ouo
UmAAw X H038: ”0898: 0.3 Abao 98 930 A80... 93:
mooAuA E81088: Abe Aqu PS A.A AA Pun ob;
UmAAw X 808A 5810030: 93 o.ooo PaA A.AcN Nam A.Amw

 

 

 

 

 

 

 

 

.._ A bu. £98.88. 3.. h A .3. 38-88:.

 

 

78

 

 

 

H020 0

0033030880 78108 80 00:88 A0800 013080 0.3000100

 

 

 

 

9:00 01188

<088 2088080

00088:». A0800
01388 <088

300003 A0800
01188 <088

080:2: A0800
01188 <088

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

. 2088080 2088080 00088:»: A0800
95.8 9882 gas; gas;

€08 0A~ €08 0A~ €08 0A~ €08 0%
00A-” 0.00 0.000 0.00 0.000 0.0m 0.000 A .3 A .30
0?: A .00 0.000 0.00 0.000 0.3 0.0 A 0 0.00 0.0.8
>00 A .0 A A .000 A .3 A bum 0.00 0.0a 0.2 0.00m
:00 0080088 0.00 A000 0.00 A.00m 0.00 0.00.0 0.3 0.000

UmAAw X 300 0080088 0.00 0.00m 0.00 0.02 0.00 0.000 080 0.000
A000: 010:1 3.3.1. A .000 0.00: A .000 A9001. A .000 A931. A .0A 0

00A-” X A000: 010:1 :0 0.000 0.3 0.08 0.00 0.000 92 A000

00> 0.00 0.000 A00 0.00. 0.00 0.03 0.00 0.08

00:» X 00> 0.00 0.000 0.3 0.03 A.N0 0.000 0.00 0.00m

H088: ”0808: A0801. :8 3.003.. :3 A000: A.NmA “TAN... A.00A

09-x X H088: ”01:08: 0.00... 0.000 A.A A... 0.03 :0 0.000 0.00 0.08
000:: 081008: 0. A m 0.000 0.3 0.30 0.00 0.00.0 A . A 0 A .000
003w X 000101081008: 0.00 A000 0.00 A08 0.00 A100 0.3 A000

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

.. m A .09 35-8181. 3.. .u A .0A. 23-81.00.

 

 

79

 

 

 

 

 

 

 

 

 

 

 

 

Predicted Probability of Binge
Drinking Category

 

Low Average High
Differentiation

Figure 9. The differentiation and expectations of tension reduction interaction for female
Binge Drinkers versus Nonbinging Drinkers. Note: TR = Tension reduction. Predicted
probability scores > .50 indicate a greater probability for being a Binge Drinker and < .50
indicate a greater probability for being a Nonbinging Drinker. Values are when

controlling for all other variables in the Comprehensive Model.

80

 

 

 

 

 

 

 

 

 

0
on
c.
E
'5 0.6 ._-________,*_n-_,____‘ _ f ' Z
3 <3, 0'5 / +TRAveragef
2g 0.4 ~—--———1_._TRLow ‘
l C 03 ./ _ ,-._m-__, ,_ ‘ L. , _.H_H_,_._,__- _
E3 0.2 ——-—«--—i+ - Wu- mm.___ _ r
E 0.1%.”- _ m - u_ _H--_.-a ,_.__

0 . _

Low Average High
Dlflorentiatlon

Figure 10. The differentiation and expectations of tension reduction interaction for female
Occasional Binge Drinkers versus Nonbinging Drinkers. Note: TR = Tension reduction.
Predicted probability scores > .50 indicate a greater probability for being an Occasional
Binge Drinker and < .50 indicate a greater probability for being a Nonbinging Drinker.

Values are when controlling for all other variables in the Comprehensive Model.

 

1.01.0 _o

9::an 01. ":0 mmmamga 1:11:15. $011082; $000502? ”0:201 8 9320:1010: 1.2 021

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

_Sw—om moan—om
D1282; 2.01220 21m: <22; Em: 01:18.; 7101220 Em: <22; Em:
<22; <22; res. r0«< <22; <22; <22; r0«< r05 <22;
202052; 01:18.; 01:18; 301230 2:23:02; U118; 01:18.; 7101220
01::2; 010:2;
ww;10 71012 cum—aw "0:: 0:5 - -
511115: 583010: 2010—;
Um?” X _Smc $200013; . n .
Um?» x 00>
Um?” x w0;:..10:1
Um?” x ._.0:;10: 131510:
_‘ 02-x x m0010_r=:10010:
w. 0008103. 30000:" 33:03 . 00820:». 3.09.03 3.09.03
_:m2; 9: 2; E: 2; <22; w_:m2; mm: 2; E: 2; <22;
<22; m m . <22; m m .
20::Em1:m <2.2;. <2.2;. 003.30. 20::1:m1:m <2.2;. <2.2;. 00320:».
U132; 20:05m5m 20:0_:m_:m w.:m2; 01:.»2; 20:0_:m_:m 20:0_:m_:m w_:m2;
95:2; 95:2; 95:2; 01:18.;
mm;10 71010. 02-1223 0:. V
51111:». 588010: Z0101;
Um?» x ch 1.2000103 - u
0;; x 00> + .
Um?” x $210131
Um?” X 10:10: ”011010:
Um?” x M0012 r5183:

 

 

 

+ ;1m:1100:1< 520801 ":0 011; 0m :01:m 1: ”:0 :1m:2 11:11:m 030m0Q B h A .o; 3.5-8103
- - 2211003; 1020201 ":0 011; 0». :01:m 1: ":0 :1m:2 11:11:m 020m0Q 2 u A .2 33.01.03
- ;1m:1100:1v. 1020801 ":0 011; 0». Sim 1: ":0 :1m:2 125—”Em Samoa. 2 n A .8 35.21.08
0 n >=:0=m: 50 1320010: $01020 «<8 amino»? 520 «<20 :0 301: 014.00; 01.0192 <210:_0.
208” 4:0 1:131:01 2011082; 20 32x01 0:; «<:0: ”:0 $20 71011 «<0; ;1m:110~:~ m; 0.: 0<0B= 3010. 2 <<:0: ":0 1303010: 3012 1:021:01 m 200:
13 S<030:3<2~m:1: ":0 011110: 01. 01111031 5.10202

 

82

Chapter Five

This chapter provides a discussion of the results and an interpretation of the
interaction effects, as well as a discussion on the limitations and implications of this
study.

Discussion of Results

This study sought to understand the role of differentiation on college student
alcohol consumption in order to address a lack of empirically validated studies regarding
the validity of the tenets of Bowen theory. Additionally, this project’s incorporation of
differentiation with one’s family of origin as a predictor in understanding individual
decision-making processes had not been previously applied to the understanding of
alcohol consumption. A series of logistic regression analyses were conducted to
investigate the relationship of differentiation with general alcohol use (GAU) and binge
drinking. Additionally, five variables that have been linked to college student alcohol
consumption (perceptions of peer use, perception of best friend use, COA status,
expectations of alcohol as a tension reducer or as a social lubricant) were examined for
potential moderating effects on the relationship between differentiation and alcohol
consumption.

Overall the sample was comprised of a high number of students that drank alcohol
and drank in high quantities. In a review of five national longitudinal studies, O’Malley
and Johnston (2002) reported that in the 19903 around 85% of college students drank in
the past year, approximately 70% drank in the past 30 days, and 40% had binge drinking
experiences over the past two weeks. In this study, 89% of the subjects reported drinking

in the past 30 days, which was not only higher than the longitudinal ranges for 30 day

83

use, but exceeded the levels of annual use. Additionally, the percentage of students that
drank at binge levels (67% for the total sample, over 76% for males, and over 59% for
females) exceeded the longitudinal studies’ ranges as well. Particularly striking was the
fact that nearly half of the males (46.8%) drank at binge levels 3 or more times in the past
two weeks. Comparatively, Hembroff and Kothari (2004) conducted a health assessment
study of students at Michigan State University. A total of 88.8% of the 1329 students in
the survey (six did not answer the question) reported consuming alcohol in the past 30
days. This study again points to the overwhelming use of alcohol by students attending
Michigan State University. Additionally, students were asked how many times they had
consumed five or more drinks at a sitting in the past two weeks. Forty seven percent of
the respondents reported drinking at that level at least once in the previous two weeks.
Although this figure is much closer to the 44% prevalence rates of national studies, it
may still be an underestimate as females were screened with the five-drink criteria rather
than the gender adjusted four-drink rate. However, the discrepancy between the two
Michigan State populations may highlight that the students in this study were higher
drinkers or more prone to report higher drinking than in previous studies.
Dijfirentiation and Alcohol Consumption

The predicted relationship between levels of differentiation and alcohol
consumption received very little support. The only GAU comparison that had
differentiation as a significant predictor of level of alcohol consumption was the male
Abstainers versus Drinkers. The odds ratio supported the predicted effect of
differentiation that students with lower levels of differentiation would be more likely to

consume greater amounts of alcohol. This same relationship was found in the female

Abstainer versus Drinker comparison, although the model was not significant. The
findings indicate that, by itself as a risk/protective factor, differentiation only contributes
in the decision to drink (or not) for males, and does not contribute to the likelihood that a
student will be in one particular level of drinking category versus another.

The differentiation hypothesis was not supported by any of the binge drinking
comparisons. The results of the binge drinking analyses echo what was found in the GAU
analyses, in that by itself differentiation contributes little as a risk or protective factor for
how much alcohol students consume per drinking occasion.

Four Factors of the DSI-R and Alcohol Consumption

A closer examination of differentiation with alcohol consumption was
implemented using the DSI-R’s subscales as separate predictor variables for each of the
comparisons. Again, this model was very poorly supported by the analyses. The only
. significant comparisons for the F our-F actor Model were in the two comparisons of
extreme versus moderate drinking in males: High versus Moderate Drinkers and Frequent
versus Occasional Binge Drinkers. Interestingly, taken as a full-scale, the DSI-R did not
significantly predict either drinking category membership, yet the “1” Position subscales
did predict both GAU and binge drinking categories, while Emotional Reactivity
predicted GAU membership. Males with a greater sense of self, “1” Position, were less
likely to be High Drinkers than those with a less developed sense of self. Similarly, males
with a greater “I” position were less likely to be Frequent Binge Drinkers than Occasional
Binge Drinkers. This finding indicates that individuals who have a more highly
developed set of personal beliefs, values, and convictions are more likely to maintain a

sense of moderation when they are drinking. These individuals with a more highly

85

developed sense of self drank five or more alcoholic beverages per drinking occasion
once or twice in the previous two-week period. It can be interpreted from the GAU and
binge drinking results that regardless of one’s level of “1” Position, one is no more likely
to be a drinker or abstainer; however, when one does drink, he is more likely to do so in
moderation in terms of monthly use (Moderate versus High Drinkers) and amount
consumed per occasion (Occasional versus Frequent Binge Drinking). This suggests that
maintaining a high sense of self could serve as a protective factor for consuming alcohol
in more normative terms versus in extreme amounts.

Surprisingly and counter intuitively, males who were more emotionally reactive
were more likely to be Moderate versus High Drinkers. According to Bowen Theory,
individuals who are more emotionally reactive are more likely to triangle something or
someone in order to tolerate anxiety rather than deal with it directly. According to the
results of this study, males who are more emotionally reactive are no more likely to be
drinkers than not, but rather are more likely to stay within a parameter of moderate
consumption than heavy usage. As the drinking categories were based on a Quantity X
Frequency measure, it may still be true that students who were more emotionally reactive
turned to alcohol more frequently but chose to drink less per drinking occasion. There
appears to be a relationship supporting this notion as individuals who were more
emotionally reactive were more likely to be Occasional versus Frequent Binge Drinkers;
however, the odds of membership were not statistically significant in the model due to a
large range in the odds ratio confidence interval (students were anywhere from .84 to
2.77 times likely to be a Frequent Binge Drinker). This suggests that emotionally reactive

students may indeed drink more often, but at lesser amounts per drinking occasion.

86

Moderator Hypotheses

Although there was very little support for the moderator hypotheses and alcohol
consumption, some interesting interactions did occur. The only significant Interaction
Models in the GAU analyses for males and females was for the interaction of
differentiation with perception of MSU student alcohol use. For males, the interaction
was significant for the High versus Moderate Drinker Comparison and for females it was
in the High versus Low Drinker comparison.

Taken by itself, as perceptions of MSU student consumption increased it was
more likely that females were categorized as being High versus Low Drinkers. However,
as the differentiation and MSU perception interaction term increased (high levels of
each), it became less likely that the females were categorized as High Drinkers (Figure 4,
p. 62). The effect of high or low perceptions of MSU student use only appears to be
significant for females with low or average levels of differentiation. This indicates that as
differentiation increases, there appears to be a protective factor against having high
perceptions of use. Interestingly, as females with lower levels of perception achieve a
higher level of differentiation, the likelihood they will be Low Drinkers diminishes, yet
they retain an overall probability of remaining Low Drinkers. This suggests that as
females become more highly difl‘erentiated they are more likely to drink at levels
independent of what they feel is normative use. The nature of this interaction was
different from what was hypothesized in that differentiation appears to moderate the
relationship of perceptions of MSU student use.

Similarly for the males, when perceptions of MSU student consumption were

considered by themselves, it was more likely that students were High versus Moderate

87

Drinkers as their perceptions increased. This effect only appeared to be significant for
males with lower levels of differentiation, although it remained minimally true with
average levels of differentiation as they were 1.16 times more likely to be High Drinkers
(Figure 3, p. 60). When males with high perceptions transitioned from average to higher
levels of differentiation they actually become more likely to be Moderate Drinkers,
indicating increased differentiation serves as a protective factor against high levels of
alcohol consumption. Additionally, there was a trend for students with lower perceptions
to be more likely to be High Drinkers as their levels of differentiation improved fiom
average to high; although, this was a very minimal effect (they were only 1.05 times
more likely when having a one standard deviation higher level of perception).
Nevertheless, there is a unique relationship for males in that they are more vulnerable to
be more extreme, rather than moderate, drinkers if they have higher perceptions of
others’ use and low levels of differentiation. Moreover, as students obtain higher levels of
differentiation there is a much lowered effect of their perceived level of student alcohol
use. As was the case with females, the interaction was in the reverse order hypothesized
in that differentiation appears to moderate the relationship of perceptions of MSU student
use. These findings indicate that students with low levels (and moderate for females) of
differentiation are more vulnerable to their conceived notions of how much alcohol the
average student at MSU consumes. It is impossible to discern if the students create a
reality of the MSU culture that fits their own established drinking behaviors or if their
behavior matches what they feel is normative consumption. Nevertheless, having higher
levels of differentiation appears to be a protective factor as there is a very minimal effect

of perceptions in terms of students’ own drinking behaviors.

While there were significant interactions for both males and females in terms of
monthly alcohol use, there were only significant interactions for females in relation to
binge drinking. The interaction of differentiation and perception of MSU student use had
the same relationship for female Frequent Binge Drinkers versus Nonbinging Drinkers as
it did for High versus Low Drinking females. Females with high perceptions were more
likely to be Frequent Binge Drinkers when they had low levels of differentiation and
actually reached a point where they were more likely to be Nonbinging Drinkers when
they had high levels of differentiation (Figure 6, p. 73). Females with low levels of
perception became slightly more likely to be Frequent Bingers when they had high levels
of differentiation, although this effect was minimal in terms of predicting drinking
categories. Again, this supports the notion that students with higher levels of
difl‘erentiation drink independently of what they perceive to be normative use.
Interestingly, the interaction did not fit the hypothesized relationship between the
variables as neither was significant as an individual variable (according to the Wald
statistic) and one variable did not significantly moderate the other.

The other female binge drinking analysis that showed a significant interaction
with differentiation was COA status in the Frequent versus Occasional Binge Drinking
comparison. Similar to the previously described interaction, neither variable was
significant by itselfin terms of predicting binge drinking categories. The non-COA
females had a greater likelihood to be in the Occasional Binge Drinking category
regardless of level of differentiation (Figure 7, p. 74). Female COAs were two-and-a-half
times more likely to be Frequent versus Occasional Binge Drinkers when they had higher

levels of differentiation. In terms of the theory of differentiation and triangling alcohol it

89

would be more congruent if COAs with higher levels of differentiation would binge drink
less frequently as they had an ability to adopt different coping strategies than their
alcoholic parents had. However, this hypothesized relationship purports that students are
binge drinking in response to stress based on levels of differentiation rather than drinking
for social reasons. Perhaps these students have normalized levels of acceptable use
modeled by their parents and are not drinking as a means to alleviate anxiety. As there
was no COA interaction for monthly use, another explanation for why COAs with higher
differentiation may be more likely to be more extreme bingers is due to a perceived
freedom fi'om their families-of-origin and are more apt to drink at higher levels (as
modeled by an alcoholic parent) as they have afforded themselves greater fieedom to
consume more per drinking occasion.

As previously described, there has traditionally been a shortage of ecological
studies regarding college student alcohol consumption. In order to test the effects of
multiple variables and the context of varying degrees of differentiation all of the logistic
regression equations were analyzed utilizing all of the variables of interest. As would be
expected with the number of variables included in the analysis, all of the comparisons
were significant with the exception of the male Occasional Binge Drinkers versus
Nonbinging Drinkers. The most commonly occurring variable across the models was that
students with higher levels of perception regarding their best fiiend at MSU’s alcohol
consumption were more likely to be in the higher drinking category for the following:
Moderate to Low Drinkers, High to Low Drinkers, Higher to Moderate Drinkers (males
only), and in all of the binge drinking comparisons. When controlling for all of the other

variables, differentiation failed to be a significant predictor of drinking category when

considered independently. However, there was an interaction effect for perceptions of
best fiiend usage (High versus Low Drinkers - females) and for tension reduction
(Frequent versus Occasional BingeDrinkers - males; Binge Drinkers and Nonbinging
Drinkers - females; Occasional versus Nonbinging Drinkers - females). Interestingly, the
previously described interactions of perceptions of MSU student use and COA status
were no longer significant when controlling for the other variables in the Comprehensive
Model.

There was an interesting interaction between differentiation and perceptions of
best fiiend at MSU’s use as females had extreme probabilities of being a High or Low
Drinker if they had low levels of differentiation (see Figure 5, p. 66). Females with high
or average perceptions are almost always a High Drinker and those with low perceptions
are almost always Low drinkers. As differentiation levels increase, students with average
perceptions are only slightly more likely to be High Drinkers, while females with high
perceptions are likely to remain High Drinkers and those with low perceptions are likely
to remain Low Drinkers. The main effect of having highor low perceptions is minimally
altered based on degree of differentiation, while increasing levels of differentiation
appears to serve as a protective factor for females with moderated levels of perception as
they are almost as likely to be Low Drinkers as High Drinkers. As students with low
levels of difl'erentiation are extremely likely to either be a High or Low Drinker based on
their perceptions of what their best friend drinks, it appears that these students have an
inability to drink at differing levels than their immediate peer group. The probability of
being in a particular drinking category became slightly less certain as levels of

differentiation increased, regardless of level of perception.

9l

Females with high and average expectations of alcohol as a tension reducer were
always more likely to be a Binge Drinker versus Nonbinging Drinker (Figure 9, p. 80) or
an Occasional Binge Drinker versus a Nonbinging Drinker (Figure 10, p. 81), regardless
of their level of differentiation. This makes sense in a tension reduction fiamework as
individuals would be more likely to drink at higher levels when they believed alcohol
would help alleviate the anxiety. Although this study did not directly measure current
levels of anxiety, the ability to tolerate anxiety was measured via differentiation. As
students with high or average expectations exhibited increased abilities to tolerate anxiety
increased (higher levels of differentiation) they were slightly less likely to be Binge
Drinkers versus Nonbinging Drinkers (although minimal) or Occasional Binge Drinkers
versus Nonbinging Drinkers, indicating a lesser prevalence of relying on binge drinking
as a means cope with anxiety. Contrary to the tenets of differentiation, females with low
tension reduction expectations in both comparison groups had an increased prevalence of
binge drinking as their levels of differentiation increased; however, they consistently
remained less likely than those with moderate to high expectations. One explanation of
this relationship is that these students are merely drinking for the sake of drinking, rather
than as a reaction to anxiety tolerance. As these students have low expectations that
alcohol actually reduces tension, it is unlikely that they would rely on it is a means to
cope with anxiety. Therefore, the increased prevalence of binge drinking based on
increased levels of differentiation may be more attributable to a greater sense of fieedom
that they have afforded themselves to drink more beverages when they chose to drink.

The identical trend of tension reduction just described for females was applicable

to the males. Gender differences demonstrated that in more highly differentiated males,

92

those with low tension reduction expectations have a greater likelihood to be a Frequent
versus Occasional Binge Drinker, while males with high expectations are almost as likely
to be in either particular category (Figure 8, p. 77). For females, the relationship of
differentiation and high expectations of tension reduction were based on the comparisons
at the lower end of the spectrum of binge drinking behaviors. For males this interaction
was significant only for the more problematic and moderate categorical comparison.
These gender differences indicate that differentiation as a potential protective factor may
work differently for males and females. For females, increased differentiation may serve
as a protector from escalating into binge drinking behavior as a coping mechanism. For
males, greater difl'erentiation may protect them from a gateway to frequent incidents of
heavy drinking as a means of coping.
Limitations

One limitation of the study is related to self-reporting of alcohol use. As students
knew the purpose of the study was to understand how variables were related to alcohol
use, there may have been a tendency to misreport their level of alcohol consumption.
However, there were no difi'erences between categories of GAU or binge drinking for
those that took the survey online or the paper-and-pencil version for either males or
females. This finding suggests that the likelihood to over-report might be marginal as
those who did not have contact with the researcher reported similar levels of use as those
who did have contact.

Another limitation for this study is the limited generalizability of the results to
other populations. First, the sample was predominately Caucasian, which restricts the

understanding of how the relationship differentiation has with drinking applies to a more

93

diverse population. Additionally, this sample appeared to be comprised of more drinkers
than have been found in other studies. As mentioned, 89% of the subjects reported
drinking in the past 30 days, which was not only higher than the longitudinal ranges for
30 day use, but exceeded the levels of annual use. Additionally, the percentage of
students that drank at binge levels (67% for the total sample, over 76% for males, and
over 59% for females) exceeded the longitudinal studies’ ranges as well. This may
suggest that the sample may not be representative of other university students across the
nation. This may also indicate that trends for use are much higher than the past orthat
MSU has a unique culture of high alcohol consumption. Additionally, the effects of
difl'erentiation may have been masked with such a high drinking sample.
Recommendations for Future Research

Although there was not overwhelming evidence supporting the role differentiation
plays in college student drinking, the relationships that were significant may warrant
future research to better understand this dynamic. It appears that differentiation by itself
does not play a substantial role in the varying levels of alcohol consumption and binge
drinking behaviors. Findings indicated that differentiation had significant interactions
with perceptions of peer use, best fiiend use, COA status, and expectations of tension
reduction. These interactions indicate that differentiation does indeed have some
implications for college student drinking. Future studies should further explore the
interaction of gender differences with differentiation as a risk/protective factor for
alcohol use as well as for other behavioral coping strategies. .

Future work with differentiation and college students should also incorporate

longitudinal designs. As mentioned in the review of literature, the college culture is

94

unique in that a majority of students who drink at higher levels will drink at lesser
amounts after graduation. However, for others, drinking behaviors in college may be a
gateway to a lifetime of misuse or abuse. It would seem that differentiation would be one
stable factor that would contribute to maintaining drinking behaviors beyond college. A
person’s level of differentiation is unlikely to change without concerted effort on the part
of the individual. Therefore, if an individual has a low level of differentiation in college
she is likely to have a low level after college. It would be intriguing to examine the
contexts in which levels of differentiation would continue to be risk or protective factors
for students while they are in college as well as after graduation.

Overall there was very little effect of COA status across all of the GAU and binge
drinking analyses. This finding suggests that COAs attending college may have a greater
sense of resiliency than non-college attending COAs. Another reason for a lack of a
significant efl‘ect of COA status may be due to just screening for parental alcoholism
rather than the contextual variation of parental alcoholism. Perhaps differentiation acts
differently as a risk/protective factor under different circumstances. For example, a male
COA with low levels of differentiation might be at greater risk if he had an antisocial
alcoholic father with a 15 year history of alcoholism than he would be if he had a high
functioning alcoholic mother with a recent onset of alcoholism. Future research should
include typology of parental alcoholism, gender of the alcoholic parent, a gender of
alcoholic parent-gender of COA interaction, the comorbidity of other mental health
illness in the alcoholic parent, the presence of abuse either witnessed by or perpetrated on
the COA, approximate number of years the parent has been drinking at alcoholic levels,

and if the parent is in remission or is an active alcoholic. Higher levels of differentiation

95

may serve as a protective factor with these dynamics just as low levels of differentiation
may be an additional risk factor.

Not only are more comprehensive studies with types of alcoholics warranted,
future research should incorporate a more ethnically diverse sample to enhance
understanding of the relationship between differentiation and alcohol consumption, as
well as the concept of differentiation in general. The DSI-R was normed on a
predominately Caucasian sample and the few psychometric changes implemented in this
study were with a predominately Caucasian sample. Future research with differentiation
needs to examine if different ethnic groups are influenced in the same way by the various
theoretical tenets of differentiation. Different ethnic groups may react quite differently
from one another in their struggle for togetherness and individuality, anxiety tolerance,
and triangling behaviors. Although Bowen states that it is theoretically universal to all
cultures, there needs to be empirical validation of this assumption.

Perhaps the most surprising discovery in the Comprehensive Model was the
change in significance of individual variables. When controlling for all of the other
variables the interactions between differentiation and perceptions of MSU student use
and COA status were no longer significant. Additionally, the interactions between
differentiation and perception of best friend use and expectations of tension reduction
emerged as being significant contributors to the prediction of drinking categories.
Although evidence has been provided for differentiation serving as a risk/protective
factor, further research needs to analyze contexts in which differentiation remains a

significant factor.

Clinical Implications

Although the results minimally supported for the differentiation hypotheses,
certain clinical implications can be understood. Many universities offer or require alcohol
counseling for students who misuse alcohol while they are residents of the university
dormitories. Although there is a time limited nature to these counseling sessions,
implementing a family-of-origin component might prove to be beneficial. If the scope of
addressing the relationships between students and their families-of-origin is outside of the
time parameters or the clinical training of the counselor, referral to a campus-based
family or psychology clinic or a community-based family therapist might best serve the
short- and long-term needs of students. As indicated by the various interactions in the
analyses, levels of differentiation have a unique effect on risk factors for college
drinking. If students become more aware of how their relationships with their families-of-
origin are continuing to effect how they respond to anxiety tolerance while they are at
school, they have a greater likelihood to respond in more objective than reactive ways.

Family therapists and psychologists trained in family systems can be excellent
resources for maintaining the overall well-being of university students. The
psychotherapy literature refers to dynamic change process as being comprised of first-
and second order change. First order change refers to the alleviation of symptomatology.
In the case of alcohol misuse, this would be evidenced in a change of drinking behaviors.
Second order change deals with addressing the dynamics involved in individuals’ lives
that illicit symptomatic behavior. A treatment modality that emphasizes second order
change might be particularly beneficial for college students based on the unique

combination of stressors that they face: adjusting to living away fiom their family-of-

97

origin for the first time, a disparate number of demands in order to obtain academic
success, balancing financial struggle of limited resources and paying tuition, living with
peers, and becoming acclimated to the university culture. Students will react to these
pressures in a myriad of ways. It is crucial that college counselors be aware of deeper
motivations for problematic behaviors, such as alcohol abuse. If the first order goal of
reduction in alcohol abuse is met, some students may simply develop a new form of
psychopathology as a result of family relationships and coping styles learned in family
contexts. As previously mentioned, a parental dyad can diffuse the relational anxiety by
triangling their child. Although this may result in long-term consequences for the child, it
brings short-term stability to the family. A consequence of parents triangling the child
might be a deflated sense of self. Although being away at college may be a cathartic
healing experience for some, it may be an overwhelming period for others. By gaining an
understanding of students’ contextual dynamics, therapists, psychologists, and counselors
will be able to help them adjust to university life.
Summary

The significant main effects of perception of the average MSU student’s use and
expectations of alcohol as a tension reducer indicate that universities have the power to
take a proactive stance in providing education regarding misperceptions that students
may have about alcohol. One mechanism for providing reliable information about alcohol
and how much is typically consumed is through various first year orientation programs.
Additionally, information can be disseminated to the residence halls at various times
throughout the year through student affairs programming or by providing information

through campus newspapers as a way to educate both on- and off-campus students. The

98

focus of any of these educational programs should be on clarifying misperceptions
regarding peer use as well as providing suggestions of coping strategies outside of
alcohol use. As alcohol consumption is a normative part of the college culture, the
university has an obligation to provide appropriate and factual information to its students
regarding the use and misuse of alcohol. It is clear from the high numbers of students

consuming alcohol that this problem could plague university life.

REFERENCES

Allgood-Merten, B., Lewinsohn, P. M., & Hops, H. (1990). Sex differences and
adolescent depression. Journal of A bnormal Psychology, 99, 55-63.

Andrews, M. P., Bubolz, M. M., & Paolucci, B. (1980). An ecological approach to study
of the family. Marriage & Family Review, 3, 2949.

Anonymous. (1972). Toward the difl‘erentiation of a self in one's own family. In J. L.

F ramo (Ed.), Family interaction: A dialogue between family researchers and
family therapists (pp. 111-173). New York: Springer Publishing Company, Inc.

Archambault, D. L. (1992). Adolescence: A physiological, cultural, and psychological no
man's land. In G. W. Lawson & A. W. Lawson (Eds), Adolescent substance
abuse: Etiology, treatment, and prevention (pp. 11-28). Gaithersburg, MA: Aspen
Publishers.

Baer, J. S. (2002). Student factors: Understanding individual variation in college
drinking. Journal of Studies on Alcohol, Supplement No. 14, 40—53.

Bartle-Haring, S., Rosen, K. H., & Stith, S. M. (2002). Emotional reactivity and
psychological distress. Journal of Adolescent Research, I 7, 568-585.

Bartlett, T. (2002). Freshmen pay, mentally and physically, as they adjust to life in
college: New survey provides more data on what happens during that crucial first
year. The Chronicle of Higher Education, February 1, A35.

Becvar, D. S., & Becvar, R. J. (1993). Family therapy: A systemic integration (2nd ed.).
Boston: Allyn and Bacon.

Benton, S. L., Schmidt, J. L., Newton, F. B., Shin, K., Benton, S. A., & Newton, D. W.
(2004). College student protective strategies and drinking consequences. Journal
of Studies on Alcohol, 65, 115-121.

Billingham, R. E., Wilson, W., & Gross, W. C. (1999). Parental divorce and
consequences of drinking among college students. College Student Journal, 33,
322-327.

Bishop, J. B., Lacour, M. A. M., Nutt, N. J., Yamada, V. A., & Lee, J. Y. (2004).

Reviewing a decade of change in the student culture. Journal of College Student
Psychotherapy, 18(3), 3-30.

mo

Black, C., Bucky, S. F., & Wilde-Padilla, S. (1986). The interpersonal and emotional
consequences of being an adult child of an alcoholic. The International Journal of
the Addictions, 21, 213-231.

Bladt, C. W. (2002). Towards a psychodynamic understanding of binge drinking
behavior in first-semester college freshmen. Journal of College Student
Development, 43, 775-791.

Bowen, M. (1966). The use of family theory in clinical practice. Comprehensive
Psychiatry, 7, 345-3 74.

Bristor, M. W. (1995). Individuals and family systems in their environments (2nd ed.).
Dubuque, IA: Kendall/Hunt.

Bronfenbrenner, U. (1977). Toward an experimental ecology of human development.
American Psychologist, 32, 513-531.

Bronfenbrenner, U. (1999). Environments in developmental perspective: Theoretical and
operational models. In S. L. Friedman & T. D. Wachs (Eds), Measuring
environment across the life span: Emerging methods and concepts (pp. 3-28).
Washington, DC: American Psychological Association.

Bubolz, M. M., Eicher, J. B., & Sontag, M. S. (1979). The human ecosystem: A model.
Journal of Home Economics, 71, 28-31.

Bubolz, M. M., & Sontag, M. S. (1993). Human ecology theory. In P. G. Boss, W. J.
Doherty, R. LaRossa, W. R. Schumm & S. K. Steinmetz (Eds), Sourcebook of
family theories and methods: A contextual approach (pp. 419-450). New York:
Plenum Press.

Byrne, B. M. (2001). Structural equation modeling with Amos: Basic concepts,
applications, and programming. Mahwah, NJ: Lawrence Erlbaum Associates.

Camatta, C. D., & Nagoshi, C. T. (1995). Stress, depression, irrational beliefs, and
alcohol use and problems in a college student sample. Alcoholism: Clinical and
Experimental Research, 19, 142-146.

Centers for Disease Control and Prevention. (2004). BMI - body mass index: BMI for
adults: Body mass index formula for adults. Retrieved December 20, 2004, from
http://www.ch.gov/nccdphp/dnpa/bmi/bmi-adult-fonnula.htrn .

Clair, D., & Genest, M. (1987). Variables associated with the adjustment of offspring of
alcoholic fathers. Journal of Studies on Alcohol, 48, 345-355.

IOI

Clair, D. J ., & Genest, M. (1992). The children of alcoholics screening test: Reliability
and relationship to family environment, adjustment, and alcohol-related stressors

of adolescent offspring of alcoholics. Journal of Clinical Psychology, 48, 414-
420.

Clapp, J. D., & McDonnell, A. L. (2000). The relationship of perceptions of alcohol
promotion and peer drinking norms to alcohol problems reported by college
students. Journal of College Student Development, 41, 19-26.

Cohen, J. (1992). A power primer. Psychological Bulletin, 112, 155-159.

Coll, K. M., Shott, D., & Morris, J. (1999). An assessment of drinking patterns and
drinking problems among community college students: Implications for
programming. Journal of College Student Development, 40, 98-100.

Cooney, R. 8., & Nonnamaker, J. B. (1992). Alcohol behavior: Commuting versus
resident students. Journal of College Student Development, 33, 395-402.

Cotton, N. S. (1979). The familial incidence of alcoholism: A review. Journal of Studies
on Alcohol, 40(1), 89-116.

Demers, A., Kairouz, S., Adlaf, E. M., Gliksman, L., Newton-Taylor, B., & Marchand,
A. (2002). Multilevel analysis of situational drinking among Canadian
undergraduates. Social Science & Medicine, 55, 415-424.

Dinning, W. D., & Berk, L. A. (1989). The children of alcoholics screening test:
Relationship to sex, family environment, and social adj ustrnent in adolescents.
Journal of Clinical Psychology, 45, 335-339.

Dreer, L. E., Ronan, G. F., Ronan, D. W., Dush, D. M., & Elliott, T. R. (2004). Binge
drinking and college students: An investigation of social problem-solving
abilities. Journal of College Student Development, 45, 303-315.

Elieson, M. V., & Rubin, L. J. (2001). Differentiation of self and major depressive

disorders: A test of Bowen theory among clinical, traditional, and intemet groups.
Family Therapy, 28, 125-142.

Ellis, D. A., Zucker, R. A., & Fitzgerald, H. E. (1997). The role of family influences in
development and risk. Alcohol Health and Research World 21, 218-226.

Engs, R. C. (1990). Fanrily background of alcohol abuse and its relationship to alcohol

consumption among college students: An unexpected finding. Journal of Studies
on Alcohol, 51, 542-547.

102

Fitzgerald, H. E., Zucker, R. A., Mun, E.-Y., Puttler, L. 1., & Wong, M. M. (2002).
Origins of addictive behavior: Structuring pathways to alcoholism during infancy
and early childhood. In H. E. Fitzgerald, K. H. Karraker & T. Luster (Eds), Infant

development: Ecological perspectives (pp. 221-252). New York:
RoutledgeFalmer.

Fitzgerald, H. E., Zucker, R. A., Puttler, L. 1., Caplan, H. M., & Mun, E.-Y. (2000).
Alcohol abuse/dependence in women and girls: Aetiology, course and subtype
variations. Alcoscope International, Review of Alcoholism Management, 3, 6- 10.

Friedman, E. H. (1991). Bowen theory and therapy. In A. S. Gurman & D. P. Kniskem
(Eds), Handbook of family therapy (Vol. 2, pp. 134-170). New York:
Brunner/Mazel.

Fromme, K., & Ruela, A. (1994). Mediators and moderators of young adults' drinking.
Addiction, 89, 63-71.

Garbarino, C., & Strange, C. (1993). College adjustment and family environments of
students reporting parental alcohol problems. Journal of College Student
Development, 34, 261-266.

Gisske, L., & Adams, M. E. (1988). Alcohol use and drinking behaviors among college
students. College Student Journal, 22, 352-357.

Goluke, U., Landeen, R., & Meadows, D. (1983). A comprehensive theory of the
pathogenesis of alcoholism. In B. Kissin & H. Begleiter (Eds), The pathogenesis
of alcoholism: Psychosocial factors (pp. 605-675). New York: Plenum Press.

Gotham, H. J ., Sher, K. J., & Wood, P. K. (2003). Alcohol involvement and
developmental task completion during young adulthood. Journal of Studies on
Alcohol, 64, 32-42.

Grant, B. F. (2000). Estimates of us children exposed to alcohol abuse and dependence in
the family. American Journal of Public Health, 90, 112-115.

Grant, B. F ., Harford, T. C., Dawson, D. A., Chou, P., Dufour, M., & Pickering, R.
(1994). Prevalence of DSM-IV alcohol abuse and dependence: United states,
1992. Alcohol Health and Research World, I 8, 243-248.

Griffore, R. J., & Phenice, L. A. (2001). The language of human ecology: A general
systems perspective. Dubuque, IA: Kendall/Hunt.

Haemmerlie, F. M., Steen, S. C., & Benedicto, J. A. (1994). Undergraduates' conflictual

independence, adjustment, and alcohol use: The importance of the mother-student
relationship. Journal of Clinical Psychology, 50, 644-650.

103

Harford, T. C., Wechsler, H., & Seibring, M. (2002). Attendance and alcohol use at
parties and bars in college: A national survey of current drinkers. Journal of
Studies on Alcohol, 63, 726-733.

Hembroff, L. A., & Kothari, S. H. (2004). MSU student health assessment: Spring 2004.
Michigan State University, Olin Health Center and the American College Health
Association by the Office for Survey Research.

Hewes, R.L., & Janikowski, T. P. (1998). Parental alcoholism and codependency: A
comparison of female children of alcoholics and nonalcoholics in two college-age
groups. College Student Journal, 32, 140-147.

Hopkins, w. G. (2002). A scale of magnitudes for effect statistics. In: A new view of
statistics. Internet Society for Sport Science. Retrieved June 18, 2005, from
http://www.sportsci.org/resource/stats/effectmag.html

Howard, M. C. (1992). Adolescent substance abuse: A social learning theory perspective.
In G. W. Lawson & A. W. Lawson (Eds), Adolescent substance abuse: Etiology,
treatment, and prevention (pp. 29-40). Gaithersburg, MA: Aspen Publishers.

Johnson, V., & Pandina, R J. (2000). Alcohol problems among a community sample:
Longitudinal influences of stress, coping, and gender. Substance Use and Misuse,
35, 669-686.

Johnston, L. D., O'Malley, P. M., Bachman, J. G., & Schulenberg, J. E. (2004).
Monitoring the future national survey results on drug use, 1975-2003: Volume 11,
college students and adults ages 19-45 (NIH publication no. 04-5508). Bethesda,
MD: National Institute on Drug Abuse.

Jones, D. H., Harel, Y., & Levinson, R. M. (1992). Living arrangements, knowledge of
health risks, and stress as determinants of health-risk behavior among college
students. Journal of American College Health, 41, 4348.

Jones, J. P., & Kinnick, B. C. (1995). Adult children of alcoholics: Characteristics of
students in a university setting. Journal of Alcohol and Drug Education, 40(2),
58-70.

Jones, J. W. (1983). The children of alcoholics screening test: A validity study. Bulletin
of the Society of Psychologists in Addictive Behaviors, 2, 155-163.

Kaplowitz, S. A., & Campo, S. (2004). Drinking, alcohol policy, and attitudes toward a
campus riot. Journal of College Student Development, 45, 501-515.

Karwacki, S. B., & Bradley, J. R. (1996). Coping, drinking motives, goal attainment

expectancies and family models in relation to alcohol use among college students.
Journal of Drug Education, 26, 243-255.

104

Kelly, V. A., & Myers, J. E. (1996). Parental alcoholism and coping: A comparison of
female children of alcoholics with female children of nonalcoholics. Journal of
Counseling and Development, 74, 501-504.

Kline, R. B. (1998). Principles and practice of structural equation modeling. New York:
The Guildford Press.

Knight, J. R., Wechsler, H., Kuo, M., Seibring, M., Weitzrnan, E. R., & Schuckit, M. A.
(2002). Alcohol abuse and dependence among US. college students. Journal of
Studies on Alcohol, 63, 263-270.

Korcuska, J. S., & Thombs, D. L. (2003). Gender role conflict and sex-specific drinking
norms: Relationships to alcohol use in undergraduate women and men. Journal of
College Student Development, 44, 204-216.

Kushner, M. G., Sher, K. J., Wood, M. D., & Wood, P. K. (1994). Anxiety and drinking
behavior: Moderating effects of tension-reduction alcohol outcome expectancies.
Alcoholism: Clinical and Experimental Research, 18, 852-860.

Landers, D., & Hollingdale, L. (1988). Working with children of alcoholics on a college
campus: A rationale and strategies for success. In M. Rivinus (Ed.), Alcoholism,
chemical dependency, and the college student (pp. 205-222). New York: The
Haworth Press.

Lange, J. E., & Voas, R. B. (2001). Defining binge drinking quantities through resulting
blood alcohol concentrations. Psychology of Addictive Behaviors, 15, 310-316.

Larirner, M. E., Lydum, A. R., Anderson, B. K., & Turner, A. P. (1999). Male and female
recipients of unwanted sexual contact in a college student sample: Prevalence
rates, alcohol use, and depression symptoms. Sex Roles, 40, 295-308.

Larson, J. H., & Wilson, S. M. (1998). Family of origin influences on young adult career
decision problems: A test of Bowenian theory. The American Journal of Family
Therapx 26, 39-53.

Lawson, G. W. (1992). lntergenerational alcoholism: The family connection. In G. W.
Lawson & A. W. Lawson (Eds), Adolescent substance abuse: Etiology,
treatment, and prevention (pp. 41-69). Gaithersburg, MA: Aspen Publishers.

Lecci, L., MacLean, M. G., & Croteau, N. (2002). Personal goals as predictors of college
student drinking motives, alcohol use and related problems. Journal of Studies on
Alcohol, 63, 620-630. '

Lerner, R. M. (1991). Changing organism-context relations as the basic process of
development: A developmental contextual perspective. Developmental
Psychology, 2 7, 27-32.

105

Lerner, R. M. (1992). Dialectics, developmental contextualism, and the further
enhancement of theory about puberty and psychosocial development. Journal of
Early Adolescence, 12, 366-388.

Lerner, R. M. (1993). Early adolescence: Toward an agenda for the integration of
research, policy, and intervention. In R. M. Lerner (Ed.), Early adolescence:
Perspectives on research, policy, and intervention (pp. 1-13). Hillsdale, NJ:
Lawrence Erlbaum Associates.

Lerner, R. M., & Galambos, N. L. (1998). Adolescent development: Challenges and
opportunities for research, programs, and policies. Annual Review of Psychology,
49, 413-446.

Lerner, R. M., & Lerner, J. V. (1987). Children in their contexts: A goodness-of-fit
model. In B. Lancaster, J. Altrnann, A. S. Rossi & L. R Sherrod (Eds), Parenting
across the life span: Biosocial dimensions (pp. 377-404). New York: Aldine de
Gruyter.

Lerner, R. M., & Lerner, J. V. (1989). Organismic and social-contextual bases of
development: The sample case of adolescence. In W. Damon (Ed.), Child
development today and tomorrow (pp. 69-85). San Francisco, CA: Jossey-Bass.

Lo, C. C. (1995). Gender differences in collegiate alcohol use. The Journal of Drug
Issues, 25, 817-836.

Marlatt, G., Baer, J. S., Kivlahan, D. R., Dimeff, L. A., Larirner, M. E., Quigley, L. A., et
al. (1998). Screening and brief intervention for high-risk college student drinkers:

Results from a 2-year follow-up assessment. Journal of Consulting and Clinical
Psychology, 66, 604-615.

McCormack, A. S. (1996). Drinking in stressful situations: College men under pressure
(1990 and 1994). College Student Journal, 30, 65-68.

McGoldrick, M., & Carter, B. (2001). Advances in coaching: Family therapy with one
person. Journal of Marital and Family Therapy, 27, 281-300.

Noel, N. E., & Cohen, D. J. (1997). Changes in substance use during times of stress:
College students the week before exams. Journal of Drug Education, 27, 363-
372. .

O'Malley, P. M., & Johnston, L. D. (2002). Epidemiology of alcohol and other drug use
among American college students. Journal of Studies on Alcohol, Suppl] 4, 23-39.

106

O'Sullivan, C. M. (1992). Adolescents in alcoholic families. In G. W. Lawson & A. W.
Lawson (Eds), Adolescent substance abuse: Etiology, treatment, and prevention
(pp. 419-427). Gaithersburg, MA: Aspen Publishers.

Perkins, H. W., & Berkowitz, A. D. (1986). Perceiving the community norms of alcohol
use among students: Some research implications for campus alcohol education
programming. The International Journal of the Addictions, 21, 961-976.

Perkins, H. W., & Wechsler, H. (1996). Variation in perceived college drinking norms
and its impact on alcohol abuse: A nationwide study. Journal of Drug Issues, 26,
961-976.

Poon, E., Ellis, D. A., Fitzgerald, H. E., & Zucker, R. A. (2000). Intellectual, cognitive,
and academic performance among sons of alcoholics during the early school
years: Differences related to subtypes of familial alcoholism. Alcoholism: Clinical
and Experimental Research, 24, 1020-1027.

Presley, C. A., Meilrnan, P. W., & Lyerla, R (1994). Development of the core alcohol
and drug survey: Initial findings and future directions. Journal of American
College Health (double check), 42, 248-255.

Protinsky, H., & Gilkey, J. K. (1996). An empirical investigation of the construct of
personality authority in late adolescent women and their level of college
adjustment. Adolescence, 31, 291-294.

Pullen, L. M. (1994). The relationships among alcohol abuse in college students and
selected psychological/demographic variables. Journal of Alcohol and Drug
Education, 40, 36-50.

Reese, F. L., Chassin, L., & Molina, B. S. G. (1994). Alcohol expectancies in early
adolescents: Predicting drinking behavior from alcohol expectancies and parental
alcoholism. Journal of Studies on Alcohol, 55, 276-284.

Rodney, H. (1995). A profile of collegiate black adult children of alcoholics. Journal of
College Student Development, 36, 228-235.

Sadava, S. W., & Pak, A. W. (1993). Stress-related problem drinking and alcohol
problems: A longitudinal study and extension of Marlatt's model. Canadian
Journal of Behavioral Science, 25 , 446-464.

Sax, L. J. (1997). Health trends among college freshmen. Journal of American College
Health, 45, 252-262.

Schulenberg, J., Maggs, J. L., Long, S. W., Sher, K. J., Gotham, H. J., Baer, J. S., et al.

(2001). The problem of college drinking: Insights from a developmental
perspective. Alcoholism: Clinical and Experimental Research, 25 , 473-477.

107

Schulenberg, J. E., & Maggs, J. L. (2002). A developmental perspective on alcohol use
and heavy drinking during adolescence and the transition to young adulthood.
Journal of Studies on Alcohol, Supplement No. 14, 54-70.

Sher, K. J ., Gershuny, B. 8., Peterson, L., & Raskin, G. (1997). The role of childhood
stressors in the intergenerational transmission of alcohol use disorders. Journal of
Studies on Alcohol, 58, 414-427.

Sher, K. J., Walitzer, K. S., Wood, P. K., & Brent, E. E. (1991). Characteristics of
children of alcoholics: Putative risk factors, substance use and abuse, and
psychopathology. Journal of A bnormal Psychology, 100, 427-448.

Shorey, H. S., Snyder, C. R, Yang, X., & Lewin, M. R. (2003). The role of hope as a
mediator in recollected parenting, adult attachment, and mental health. Journal of
Social and Clinical Psychology, 22, 685-715.

Skowron, E. A., & F riedlander, M. L. (1998). The differentiation of self inventory:
Development and initial validation. Journal of Counseling Psychology, 45 , 23 5-
246.

Skowron, E. A., & Schmitt, T. A. (2003). Assessing interpersonal fusion: Reliability and
validity of a new DSI fusion with others subscale. Journal of Marital and Family
Therapy. 29(2), 209-222.

Smith, P. D., Wells, D. B., & Abdul-Salaam, K. (1997). Assessing alcohol problems in
student populations. In P. C. Rivers & E. R. Shore (Eds), Substance abuse on
campus: A handbook for college and university personnel (pp. 165-187).
Westport, CT: Greenwood Press.

Stewart, S. H., Karp, J., Pihl, R. 0., & Peterson, R. A. (1997). Anxiety sensitivity and
self-reported reasons for drug use. Journal of Substance Abuse, 9, 223-240.

Titelman, P. (1998). Overview of the Bowen-theoretical—therapuetic system. In P.
Titleman (Ed.), Clinical applications of Bowen family systems theory (pp. 7-49).
New York: The Haworth Press.

Titelman, P. (2003). Emotional cutoff in Bowen family systems theory: An overview. In
P. Titelman (Ed.), Emotional cutofif' Bowen family systems theory perspectives
(pp. 1-66). New York: The Haworth Clinical Practice Press.

von Bertalanffy, L. (1980). General system theory: Foundations, development,
applications (Revised ed.). New York: George Braziller.

von Eye, A., & Schuster, C. (1998). Regression analysis for social sciences. New York:
Academic Press.

108

Waite-O'Brien, N. (1992). Alcohol and drug abuse among female adolescents. In G. W.
Lawson & A. W. Lawson (Eds), Adolescent substance abuse: Etiology,
treatment, and prevention (pp. 367-3 79). Gaithersburg, MA: Aspen Publishers.

Wechsler, H., Davenport, A., Dowdall, G., Moeykens, B., & Castillo, S. (1994). Health
and behavioral consequences of binge drinking in college: A national survey of
students at 140 campuses. JAMA: Journal of the American Medical Association,
272, 1672-1677.

Wechsler, H., Dowdall, G., Davenport, A., & Rimm, E. B. (1995). A gender-specific
measure of binge drinking among college students. American Journal of Public
Health, 85, 982-985.

Wechsler, H., Dowdall, G. W., Davenport, A., & Castillo, S. (1995). Correlates of college
student binge drinking. American Journal of Public Health, 85, 921-926.

Wechsler, H., Dowdall, G. W., Maenner, G., Gledhill-Hoyt, J ., & Lee, H. (1998).
Changes in binge drinking and related problems among American college
students between 1993 and 1997. Journal of American College Health, 4 7, 57-68.

Wechsler, H., Isaac, N. E., Grodstein, P., & Sellers, D. E. (1994). Continuation and
initiation of alcohol use from the first to the second year of college. Journal of
Studies on Alcohol, 55, 41-45.

Wechsler, H., Lee, J. E., Kuo, M., & Lee, H. (2000). College binge drinking in the 1990s:
A continuing problem: Results of the Harvard school of public health 1999
college alcohol study. Journal of American College Health, 48, 199-210.

Wechsler, H., Lee, J. E., Kuo, M., Seibring, M., Nelson, T. F ., & Lee, H. (2002). Trends
in college binge drinking during a period of increased prevention efforts. Findings
from 4 Harvard school of public health college alcohol study surveys: 1993-2001.
Journal of American College Health, 50, 203-217.

Wechsler, H., Lee, J. E., Nelson, T. F., & Kuo, M. (2002). Underage college students'
drinking behavior, access to alcohol, and the influence of deterrence policies:
Findings from the Harvard school of public health college alcohol study. Journal
of American College Health, 50, 223-236.

Wetchler, J. L., & Piercy, F. P. (1996). Transgenerational family therapies. In F. P.
Piercy, D. H. Sprenkle & J. L. Wetchler (Eds), Family therapy sourcebook (2nd
ed., pp. 25-49). New York: The Guilford Press.

Williams, J. G., & Kleinfelter, K. J. (1989). Perceived problem-solving skills and
drinking patterns among college students. Psychological Reports, 65, 1235-1244.

109

Woltcrsdorf, M. A. (1997). Neuropsychological effects of substance abuse. In P. C.
Rivers & E. R. Shore (Eds), Substance abuse on campus: A handbook for college
and university personnel (pp. 23-48). Westport, CT: Greenwood Press.

Yeatman, F. R., Bogart, C. J., Geer, F. A., & Sirridge, S. T. (1994). Children of
alcoholics screening test: Internal consistency, factor structure, and relationship to
measures of family environment. Journal of Clinical Psychology, 50, 931-936.

Zucker, R. A. (1994). Pathways to alcohol problems and alcoholism: A developmental
account of the evidence for multiple alcoholisms and for contextual contributions
to risk. In R. A. Zucker, G. M. Boyd & J. Howard (Eds), The development of
alcohol problems: Exploring the biopsychosocial matrix of risk (V 01. NIAAA
Research Monograph 26, pp. 255-289). Rockville, MD: Department of Health and
Human Services.

Zucker, R A., Fitzgerald, H. E., & Moses, H. D. (1995). Emergence of alcohol problems
and the several alcoholisms: A developmental perspective on etiologic theory and
life course trajectory. In D. Cicchetti & D. J. Cohen (Eds), Developmental
psychopathology: Vol. 2. Risk disorder, and adaptation (pp. 677-711). New
York: John Wiley & Sons, Inc.

Zucker, R. A., Fitzgerald, H. E., Refior, S. K., Puttler, L. I., Pallas, D. M., & Ellis, D. A.
(2000). The clinical and social ecology of childhood for children of alcoholics:
Description of a study and implications for a differentiated social policy. In H. E.
Fitzgerald, B. M. Lester & B. S. Zuckerman (Eds), Children of addiction:
Research, health, and policy issues. New York: Garland Press.

Zucker, R. A., Reider, E. E., Ellis, D. A., & Fitzgerald, H. E. (1997). Alcohol
involvement of the life span: A developmental perspective on etiology and course.
In P. C. Rivers & E. R. Shore (Eds), Substance abuse on campus: A handbook for
college and university personnel (pp. 3-22). Westport, CT: Greenwood Press.

Zucker, R. A., & Wong, M. M. (in press). Prevention for children of alcoholics and other
high risk groups. In M. Galanter (Ed.), Recent developments in alcoholism: Vol.

I 7. Research on alcohol problems in adolescents and young adults. New York:
Kluwer Academic/Plenum.

110

Appendix A

Differentiation of Self Inventory — Revised (DSI-R)

 

 

Differentiation of Self Inventory - Revised (DSI-R)

“These are questions concerning your thoughts and feelings about yourself and
relationships with others. Please read each statement carefully and decide how much the
statement is generally true of you on a 1 (not at all) to 6 (very) scale. If you believe that
an item does not pertain to you (e.g., you are not currently married or in a committed
relationship, or one or both of your parents are deceased), please answer the item
according to your best guess about what your thoughts and feelings would be in that
situation. Be sure to answer every item and try to be as honest and accurate as possible in
your responses.” ‘

 

1. People have remarked that I'm overly emotional.

l 2 3 4 5 6
Not at all Very
2. I have difficulty expressing my feelings to people I care for.
1 2 3 4 5 6
Not at all Very
3. I often feel inhibited around my family.
1 2 3 4 5 6
Not at all Very
4. I tend to remain pretty calm even under stress.
1 2 3 4 5 6
Not at all Very
5. I usually need a lot of encouragement fi'om others when starting a big job or task.
1 2 3 4 5 6
Not at all Very

6. When someone close to me disappoints me, I withdraw from him/her for a time.

1 2 3 4 5 ' 6
Not at all Very

7. No matter what happens in my life, I know that I'll never lose my sense of who I am.

1 2 3 4 5 6
Not at all Very

lll

 

10.

11

12.

13.

14.

15.

16.

17.

I tend to distance myself when people get too close to me.

1 2 3 4 S 6
Not at all Very

I want to live up to my parents' expectations of me.

1 2 3 4 5 6
Not at all Very

I wish that I weren't so emotional.

l 2 3 4 5 6
Not at all Very

. I usually do not change my behavior simply to please another person.

1 2 3 4 5 6
Not at all Very

My spouse/partner could not tolerate it if I were to express to him/her my true
feelings about some things.

1 2 3 4 5 6
Not at all Very
When my spouse/partner criticizes me, it bothers me for days.
1 2 3 4 5 6
Not at all Very

At times my feelings get the best of me and I have trouble thinking
clearly.

I 2 3 4 5 6
Not at all Very

When I am having an argument with someone, I can separate my thoughts about the
issue fi'om my feelings about the person.

1 2 3 4 5 6
Not at all Very

I'm often uncomfortable when people get too close to me.

1 2 3 4 5 6
Not at all Very

I feel a need for approval fi'om virtually everyone in my life.

1 2 3 4 5 6
Not at all Very

112

18. At times I feel as if I'm riding an emotional roller-coaster.

l 2 3 4 5 6
Not at all Very

19. There's no point in getting upset about things I cannot change.

1 2 3 4 5 6
Not at all Very
20. I'm concerned about losing my independence in intimate relationships.
1 2 3 4 5 6
Not at all Very

21. I'm overly sensitive to criticism.

1 2 3 4 5 6
Not at all Very
22. I try to live up to my parents' expectations.
1 2 3 4 5 6
Not at all Very
23. I'm fairly self-accepting.
l 2 3 4 5 6
Not at all Very
24. I often feel that my spouse/partner wants too much from me.
1 2 3 4 ' 5 6
Not at all Very
25. I ofien agree with others just to appease them.
1 2 3 4 5 6
Not at all Very
26. IfI have had an argument with my spouse/partner, I tend to think about it all day.
l 2 3 4 5 6
Not at all Very

27. I am able to say "no" to others even when I feel pressured by them.

1 2 3 4 5 6
Not at all Very

113

28. When one of my relationships becomes very intense, I feel the urge to run away from
it.

1 2 3 4 5 6

Not at all Very
29. Arguments with my parent(s) or sibling(s) can still make me feel awful.
I 2 3 4 5 6

Not at all Very

30. If someone is upset with me, I can't seem to let it go easily.

1 2 3 4 5 6
Not at all Very

31. I'm less concerned that others approve of me than I am in doing what I think is right.

1 _ 2 3 4 5 6
Not at all Very

32. I would never consider turning to any of my family members for emotional support.

I 2 3 4 5 6
Not at all Very
33. I ofien feel unsure when others are not around to help me make a decision.
1 2 3 4 5 6
Not at all Very

34. I'm very sensitive to being hurt by others.

1 2 3 4 5 6
Not at all Very

35. My self-esteem really depends on how others think of me.

1 2 3 4 5 6
Not at all Very

36. When I'm with my spouse/partner, I often feel smothered.

l 2 3 4 5 6
Not at all Very
37. When making decisions, I seldom worry about what others will think.
1 2 3 4 5 6
Not at all Very

ll4

38. I ofien wonder about the kind of impression I create.

1 2 3 4 5 6
Not at all Very

39. When things go wrong, talking about them usually makes it worse.

1 2 3 4 5 6
Not at all Very
40. I feel things more intensely than others do.
1 2 3 4 5 6
Not at all Very

41. I usually do what I believe is right regardless of what others say.

1 2 3 4 5 6
Not at all Very
42. Our relationship might be better if my spouse/partner would give me the space I need.
1 2 3 4 5 6
Not at all Very

43. I tend to feel pretty stable under stress.

1 2 3 4 5 6
Not at all Very

44. Sometimes I feel sick after arguing with my spouse/partner.

1 2 3 4 5 6
Not at all Very

45. I feel it's important to hear my parents' opinions before making decisions.

1 2 3 4 5 6
Not at all Very

46. I worry about people close to me getting sick, hurt, or upset.

l 2 3 4 5 6
Not at all Very

115

Appendix B
Analysis of the DSI-R with College Students
Description of the DSI-R

Differentiation was measured by scores on the Differentiation of Self Inventory-
Revised (DSl-R). Subscale and firll-scale scores from the DSI-R were obtained by
averaging the six-point Likert responses for each of the respective scales afier reverse
coding the appropriate items. F ull- and subscale scores reflect a range from 1 (low
differentiation) to 6 (higher differentiation).

Descriptive Statistics

The means for the subscales ranged from 3.37 to 4.57 and the DSI-R full-scale mean was
3.89 (see Table 11, p. 117). Each of the full-scale-subscale correlations ranged from
moderate (.66) to high (.86) and were significant at p < .001 (see Table 12, p. 117).
Additionally, each of the subscale intercorrelations ranged from low (.29) to moderate
(.65) and were all significant at p < .001. The full-scale and subscale measures each
illustrated high internal consistency reliabilities based on Cronbach’s alpha values (ER ct
= .86; [P a = .77; BC on = .81; F0 on = .69; DSI-R a = .90).

Each of the subscales and full-scale showed significantly different scores for
males and females, although the relative effect sizes indicate these differences may be
relatively trivial. The Emotional Reactivity scores of females (3.46) were significantly
lower than males (3.72) (t (445) = -3.11, p = .002) with a low effect size (d = .30, 9,, =

0.02) illustrating that only around 2% of the variance in ER scores can be explained by

“6

 

Hwy—o _ _.

385m Ba Dina—Eon m2. :5 Hon: Um?” Ba mccmoaam

 

 

 

 

 

 

 

.38— mmBEn Kama «35.3
38: m0 ”mama 38.: m0 ”mama 385 m0 ”850'.
m” Pun o.oo :5 I o.oo 93...... 93 :5 I 93 93...... o.oo :m I o.oo
3 Pg 9: _.ua I 93 ALM: Paw New I 93 uba: 9% _.ua I 93
mo 93 PE Now I Foo 93...... 93 Now I abo 93...... 9.5 PS I 98
mo 93 9am _.mm I 93 who: 9% N: I Pow PM»: 93 _.mm I MS
62* who 93 NC I 93 93... 0.3 N?» I 93 Pg. ohm PS I 93

 

 

 

 

 

 

 

 

 

 

 

_._ h A .3. 398:8. .2. h A .2. 35-8%?

 

 

 

.363 _N.

mono—macaw m9. En H08. Umrw Ea madman—om

 

 

 

 

 

 

m” :U m0 m0
m” .I. .I II I
:u an: I II I
mm but. but. .I II
m0 but. am: he: I
Um?” .3: mi...» .3... Que...

 

 

 

 

 

 

3% A .2:w 35-8=8

 

 

ll7

gender differences. “1” Position scores of females (3.96) were significantly lower than
males (4.15) (t (445) = -2.9l,p = .004, d = .28, r2“, = 0.02). Emotional Cutoff scores of
females (4.68) were significantly higher than males (4.43) (t (445) = 3.56, p < .001, d =
.34, 3,, = 0.03). Fusion with Others scores of females (3.27) were significantly lower
than males (3.49) (t (445) = -3.66, p < .001, d = .35, rzpb = 0.03). Finally, the total DSI-R
scores of females (3.84) were significantly lower than males (3.95) (t (445) = -l .99, p =
.047, d = .19, 3..., = 0.01)
F our-F actor Model

Confirmatory factor analyses were run separately for males and females as well as
with both genders combined in order to gain a comprehensive understanding of the four-
factor structure of the DSI-R with college students. The methods used in Skowron and
Friedlander (1998) in the creation of the D81 were replicated for this analysis. Individual
questions fiom the scale were clustered into meta-items for each of the four subscales
(Emotional Reactivity = ER, “1” Position = IP, Emotional Cutofi‘ = EC, and Fusion with
Others = F0), with each meta-item containing a total of 3 to 4 items. For example, the 11
items in the ER subscale yielded indicators ER), ER;, and ER3, each containing 3 to 4
items. A total of 12 indicators were created (3 per subscale). Although the items in
Skowron and F riedlander (1998) were randomly assigned to the respective meta-item
cluster, this study will use the identical clusters for ER, IP, and EC used in their study
(ER; = DSI—R items 1, 14, 26, 38; ER; = DSI-R items 6, 18, 30, 40; ER; = DSI-R items
10, 21, 34; [Pr = DSI-R items 4, 15, 27, 41; [Pg = DSI-R items 7, 19, 31, 43; 1P3 = DSI-R
items 11, 23, 35; EC) = DSI—R items 2, 12, 24, 36; EC; = DSI-R items 3, 16, 28, 39; EC;

= DSI-R items 8, 20, 32, 42). As the items for the F 0 scale were revised in the DSI-R

ll8

(Skowron & Schmitt, 2003) only three of the original questions remained. Although the
wording for these items changed, their respective order in the inventory remained the
same; therefore the item numbers were entered into the same meta-items from the
original DSI with the three added questions (items 44, 45, 46) randomly assigned to the
indicator groups (F01 = DSI items 5, 17, 29, 44; F 02 = DSI items 9, 22, 33, 45; F 03 =
DSI items 13, 25, 37, 46). The intercorrelations for the 12 indicators used in the
confirmatory factor analysis are presented in Table 13 (p. 121).

Following the original protocol, each of the indicators was permitted to load
freely on its respective factor and was constrained to 0 on the other factors. Each latent
variable was scaled to the first indicator by fixing its value to 1.00. The fit indices used to
evaluate the model were also replicated from the original study: A chi-square to degrees
of freedom ratio (ledf< 2.0); goodness-of-fit index (GFI > .90); adjusted goodness-of-fit
index (AGF I > .80); and the root mean squared of the residuals (RMR < .10).

The four-factor model is illustrated in Figure 11 (p. 125).The model
showed a good fit of maintaining a four-factor structure (x2(48), N = 447) = 177.71 , p <
.001, xz/df= 3.70, on = .94, AGFI = .90, RMR = .04). Although the significant chi-
square and chi-square-to—degrees of freedom ratio values indicated a poor fit, there has
been debate about the problematic nature of these analyses as related to sample size and
has resulted in the development of various other goodness-of-fit measures including the
others in this analysis (Byrne, 2001; Shorey, Snyder, Yang, & Lewin, 2003).

The four-factor model maintained a good fit for when females were analyzed separately
(x2(48, N = 246) = 129.54, p < .001, ledf= 2.70, GFI = .92, AGFI = .87, RMR = .04);

however, it fell below the specified fit criteria for males (x2(48, N = 201) = 214.40, p <

119

.001, fi/ = 4.47, on = .86, AGFI = .77, RMR = .10). The model with differentiation as
a higher order latent factor (Figure 12, p. 126) indicated that the DSI-R was a single and
multidimensional construct for the combined sample (x260. N = 447) = 189.22, p < .001,
xz/df= 3.78, GFI = .93, AGFI = .90, RMR = .05) as well as separately for females (x260.
N = 246) = 132.19, p < .001, xz/df= 2.64, GFI = .92, AGFI = .87, RMR = .04) but not for
males (x260, N = 201) = 144.89, p < .001, xZ/df= 2.90, GFI = .89, AGFI = .83, RMR =
.06).

After analyzing the male data, one item from the Emotional Reactivity (#38), “1”
Position (#11), and Emotional Cutoff (#32) subscales were omitted afier exploring their
respective effects on their standard errors and covariance estimates. In addition to making
the necessary changes to the confirmatory factor meta-item clusters (ERr, 1P3, EC3), item
#15 was randomly chosen to move from IPr to IP; so that the later indicator would
maintain a mean of three items.

The means for the subscales ranged from 3.37 to 4.52 and the DSI-R full-scale
mean remained 3.89 (see Table 14, p. 123). Each of the full-scale-subscale correlations
ranged fi'om moderate (.68) to high (.86) and were all significant at p < .001 (see Table
15, p. 123). Additionally, each of the subscale intercorrelations ranged from low (.20) to
moderate (.66) and were all significant at p < .001. The full-scale and subscale
Cronbach’s alpha values were not altered with the deletion of the three items (ER a =
.86; [P a = .77; EC 01 = .80; F0 on = .69; DSI—R a = .90). The intercorrelations for the
revised 12 indicator used in the confirmatory factor analysis are presented in Table 16 (p.

124).

120

 

 

How—o S

09.8.30: 253x m2 :8 Um?” :2: 9888

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

m? m? m? :0. :6 .3 m9 m9 m0... v.9 _.1 ON
m? I I I I I I I I I I I
m? hm: I I I I I I I I I I
m? .31. .31. I I I I I I I I I
no. :31. but. .31. I I I I I I I I
=c~ .31. .31. .31. .81. I I I I I I I
:6 .31. .31.... .31. .31. .31. I I I I I I
ma. No.1. Mme... .31. No.1. .31. .Nm1. I I I I I
mO~ .81. .31. .31. .NN: No.1. .31. own... I I I I
mOm .81. but. be: .31. No.1. .51. .31. 6N: I I I
m0. our... .31. .81. .31. .31. .3: .31. .81. his... I I
won .81. .81. .NA: .3...» .3 .3 .31. .8 Joe 001. I
mom 3...... 31. .31. one... N01. wan... .31. No.1. NAME. 35...... .31.

 

 

208” m” n magma?— Womomfia.“ to n :1. weaned m0 0 mBonmosw— 0:83 mo n Fame: 2:: 099m.
m? u Umrw seam _. 3. mm. mm“ m? H Um?” :35 a. 3. mo. no“ m? H Um?” :35 8. 8. war a: H
Um?” :23 A. 3. Na. 3“ =u~ H Um?” :35 q. S. 3. Au“ now H Um?” :23 2. um. um“ m9 H Um?”
:23 N. 5. NA. mm“ m0“ H Um?” :35 w. 3. Mm. we“ mom n Ume :35 x. no. um. AN“ m9 n Umzw

:23 m. 3. no. .3“ m0» H Um?” :35 o. MN. mu. AM mom u Ume :23 3. Nu. uq. .3.

ate A .8. 35-8%? 1.x» A .3. gorgfia.

 

 

121

As was the case with the original analysis, the subscales and full-scale showed
significantly different scores for males and females with minimal effect sizes (see Table
14, p. 123). The Emotional Reactivity scores of females (3.54) were significantly lower
than males (3.81) (t (445) = -312, p = .002) with a low effect size (d = .30, 12,, = 0.02)
illustrating that only around 2% of the variance in ER scores can be explained by gender
differences. “1” Position score of females (3.98) were significantly lower than males
(4.19) (t (445) = -3.03, p = .003, d = 0.29, rzpb = 0.02). Emotional Cutoff scores of
females (4.63) were significantly higher than males (4.39) (t (445) = 3.24, p = .001, d =
0.31, r2", = 0.02). Fusion with Others scores of females (3.27) were significantly lower
than males (3 .49) (t (445) = -3.657, p < .001, d = 0.35, rzpb = 0.03). Finally, the total DSI-
R scores of females (3.84) were significantly lower than males (3.96) (t (445) = -2.lO6, p
= .036, d = 0.20, 12,, = 0.01).

The revised four-factor model is illustrated in Figure 13 (p. 127).The model
showed a good fit of maintaining a four-factor structure in general(x2(48, N = 447) =
171 .42,p < .001, ledf= 3.57, GFI = .94, AGFI = .90, RMR = .04) as well as separately
for females (x2(48, N = 246) = 150.79, p < .001, ledf= 3.14, GFI = .91, AGFI = .85,
RMR = .05) and males (12(48, N = 201) = 106.48, p < .001, ledf= 2.22, GFI = .91,
AGFI = .86, RMR = .05). The revised model with differentiation as a higher order latent
factor (Figure 14, p. 128) indicated that the DSI-R is a single and multidimensional
construct for the combined sample (x260, N= 447) = 181.83, p < .001, xZ/df= 3.64, on
= .94, AGFI = .90, RMR = .04) as well as separately for females (3860, N = 246) =
153.37, p < .001, xZ/df= 3.07, GFI = .91, AGFI = .85, RMR = .05) and males (38(50, N =

201) = 122.74, p < .001, ledf= 2.46, GFI = .90, AGFI = .85, RMR = .05).

l22

 

.320 E

:33 mag Oman—$58: mom :8 ”Emma .38— Umzw Ba madman—om

 

 

 

 

 

 

 

Hog: mmBEa 3&8 maaaom
2.8: m0 ”mama K8: m0 ”mama Zoe: mD ”36ml
mw o.oo 93 _.oo 1 o.oo 931. o.oo _.oo 1 o.oo 931. o.oo _.No 1 o.oo
=u Aoq o.oo To 1 o.oo PS1. ohm ~._o I o.oo o.oo»... o.ow TE 1 o.oo
mo 93 o.oo _.S 1 o.oo o.oo: o.oo _.2 I o.oo p.31. o.oo who I o.oo
m0 o.oo o.on _.um I o.oo who... o.oo N: 1 o.oo. who: o.oo _.mm 1 o.oo
Umzfl o.oo o.oo PM. 1 who o.oo... o.oo NAN I Mt o.oo. o.oo NB I who

 

 

 

 

 

 

 

 

 

 

 

... h A .om. go-§_ao. 1. h A .of 25.8%?

 

 

 

Haw—o _m

Ooflimmoam m9. :5 wn<mmoo Hog: Um?” Ea madman—om

 

 

 

 

 

 

mow av m0 m0
wow I I I I
z. .31. I I I
ma .3: :31. I I
mo .81. .No1. .31. I
Um?” .31. out. .31. .31.

 

 

 

 

 

 

 

1. h A .8? 35-8%“.

 

123

 

Hugo _o

Goad—mac: Zufix m9. :5 ”wing Um?” :9: 9588

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

m? m? m? 5 5 5 mo. mOI mo. m0. m9
m? I I I I I l l I I I I
1m? .2: I I I I I I I I I I
m? .8: .3: I I I I I I I l I
5 .8: .8: .2: l I I I I I I I
5 .3: .8: .3: .8: I I I I I I I
5 .5: .3: .X: .3: .8: I I I I I I
ma. .8: .8: .8: .8: .8: .8: I I I I I
ma. .2: .8: .8: .8: .8: .8: .8: I I I I
ma... .3: .8: .8: .3: .8: .3: .3: .3: I I I
mo. .8: .3: .8: .8: .8: .3: .8: .8: .8: I I
mo. .8: .8: .8: .:: .2 .3. .5: .8 .8 .8: I
«9 .8: .3: .3: B: .8: .8: .8: .8: .8: .3: .8:

 

124

208” =an m: 305 8:8» 0:88.. 9858 m: :5 838a mam—vim. mow u maonoaa wanna/‘5.“ :u .I. J:
V8508 mm H mBoaoaa 0:83 m0 u 353: SE. 093... a”. u Ume :93 r I: no“ mow“ H GE.”
:35 o. _w. o. how m? H Um?” :93 _o. B . of _5 H Um?” :23 9 no. .3“ :3 H Gm?” 223 q.
3. or 3“ Eva u Ume :23 _m. mm. mm“ mm. H Umzw :35 N. 5. NA. wow mm» H Ume :23 o. _o.
No. won an» H Um?” 53m x. No. AN“ m0. H Um?” roam m. 3. no. 3“ m0» H Um?” :23 o. NN. ow.
3“ wow u 69% 3an 3. Nu. mg. no.

 

.. h A .8. 25-559 1. o A .o_. EBAEFQ.

 

 

ER
.64
.40 P
.87 .42
.56 EC
.41
F0

.81

.72

.78

ER1

‘2

 

6W1

 

 

ERZ

 

 

ER3

P1

 

I'M
6 36

 

 

 

6

 

 

.70

.79

.75

P3

V
--L

 

 

E01

3" I

 

 

 

E02

I.

 

 

.82

.80

E03

I

 

 

 

F01

a)
_L

 

 

 

 

F02

 

 

.65

F03

III

 

Figure l 1. F our-Factor analysis of the DSI-R with males and females.

125

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

®
' ®
F03 ' 76 ®

 

 

 

 

 

 

 

 

 

 

n
O
“U
.a
m
m
8
mm
D

Figure 12. Differentiation as a higher order latent factor with males and females.

126

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2 98 4 10
"mm M WSW 56 6 6.9
. . . .
F 1 2
m a m p. p. P m m E F F
7
m u. 37 a
o 8 9
we. n7. m7 7
a F m m
m m. u
7
M 5.

.65

 

F03 ‘ '76 @

Figure 13. Four-factor analysis of the Revised DSI-R with males and females.

127

 

ER '
433 .30 .83 ER1 err1

°.80 ER2 .59 .
.86

ER3

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Wm
.95 .70
.72 .72 P1 ‘———@
.78 P2 .63 C
.75
.69 .
P3 .67 @
Differentiation EW
.46 .89 ‘50
.80 EC1 @
EC .73 EC2 .68 .
.77

 

.90 £03 -.63 w

-.61
.45 .
.79 Pm ®

044 F02 .90 .
.65

F03 err1

 

 

 

 

 

 

 

 

 

 

 

 

Figure 14. Differentiation as a higher order latent factor with males and females in the
Revised DSl-R.

128

Appendix C
Children of Alcoholics Screening Test (CAST)

1. Have you ever thought that one of your parents had a drinking problem? 1. Yes 2.
No

2. Have you ever lost sleep because of a parent's drinking? 1. Yes 2. No
3. Did you ever encourage one of your parents to quit drinking? 1. Yes 2. No

4. Did you ever feel alone, scared, nervous, angry or frustrated because a parent was not
able to stop drinking? 1. Yes 2. No

5. Did you ever argue or fight was a parent when he or she was drinking? 1. Yes 2. No

6. Did you ever threaten to run away from home because of a parent's drinking? 1. Yes
2. No

7. Has a parent ever yelled at or hit you or other family members when drinking? 1. Yes
2. No

8. Have you ever heard your parents fight when one of them was drunk? 1. Yes 2. No

9. Did you ever protect another family member from a parent who was drinking? 1. Yes
2. No

10. Did you ever feel like hiding or emptying a parent's bottle of liquor? 1. Yes 2. No

11. Do many of your thoughts revolve around a problem drinking parent or difficulties
that arise because of his or her drinking? 1. Yes 2. No

12. Did you ever wish that a parent would stop drinking? 1. Yes 2. No
13. Did you ever feel responsible for or guilty about a parent's drinking? 1. Yes 2. No

14. Did you ever fear that your parents would get divorced due to alcohol misuse? 1. Yes
2. No

15. Have you ever withdrawn from and avoided outside activities and friends because of
embarrassment and shame over a parent's drinking problem? 1. Yes 2. No

16. Did you ever feel caught in the middle of an argument or fight between a problem

drinking parent and your other parent? 1. Yes 2. No
17. Did you ever feel that you made a parent drink alcohol? 1. Yes 2. No

18. Have you ever that a problem drinking parent did not really love you? 1. Yes 2. No
19. Did you ever resent a parent's drinking? 1. Yes 2. No

20. Have you ever worried about a parent's health because of his or her alcohol use? 1.
Yes 2. No

21. Have you ever been blamed for a parent's drinking? 1. Yes 2. No
22. Did you ever think your father was an alcoholic? 1. Yes 2. No

23. Did you ever wish you home could be more like the homes of your fiiends who did
not have a parent with a drinking problem? 1. Yes 2. No

129

24. Did a parent ever make promises to you that he or she did not keep because of
drinking? 1. Yes 2. No

25. Did you ever think your mother was an alcoholic? 1. Yes 2. No

26. Did you ever wish that you could talk to someone who could understand and help the
alcohol-related problems in your family? 1. Yes 2. No

27. Did you ever fight with your brothers and sisters about a parent's drinking? 1. Yes 2.
No ‘

28. Did you ever stay away from home to avoid the drinking parent or your other parent's
reaction to the drinking? 1. Yes 2. No

29. Have you ever felt sick, cried, or had a "knot" in your stomach after worrying about a
parent's drinking? 1. Yes 2. No

30. Did you ever take over any chores and duties at home that were usually done by a
parent before he or she developed a drinking problem? 1. Yes 2. No

130

Appendix D
Children of Alcoholics Screening Test - Mother (M-CAST)
. Have you ever thought that your mother had a drinking problem? 1. Yes 2. No
. Have you ever lost sleep because of your mother's drinking? 1. Yes 2. No
. Did you ever encourage your mother to quit drinking? 1. Yes 2. No

. Did you ever feel alone, scared, nervous, angry or frustrated because your mother
was not able to stop drinking? 1. Yes 2. No

. Did you ever argue or fight when she was drinking? 1. Yes 2. No

. Did you ever threaten to run away from home because of your mother’s drinking?
1. Yes 2. No

. Has your mother ever yelled at or hit you or other family members when
drinking? 1. Yes 2. No

. Have you ever heard your parents fight when your mother was drunk? 1. Yes 2.
No

. Did you ever protect another family member fi'om your mother who was
drinking? 1. Yes 2. No

10. Did you ever feel like hiding or emptying your mother's bottle of liquor? 1. Yes

2. No

11. Do many of your thoughts revolve around a problem drinking mother or

dificulties that arise because of her drinking? 1. Yes 2. No

12. Did you ever wish that she would stop drinking? 1. Yes 2. No

13. Did you ever feel responsible for or guilty about your mother's drinking? 1. Yes

2. No

131

14. Did you ever fear that your parents would get divorced due to your mother’s
alcohol misuse? 1. Yes 2. No

15. Haveyou ever withdrawn from and avoided outside activities and friends because
of embarrassment and shame over your mother’s drinking problem? 1. Yes 2.
No

16. Did you ever feel caught in the middle of an argument or fight between a problem
drinking parent (mother) and your other parent? 1. Yes 2. No

17. Did you ever feel that you made your mother drink alcohol? 1. Yes 2. No

18. Have you ever thought a problem drinking parent (mother) did not really love
you? 1. Yes 2. No

19. Did you ever resent your mother’s drinking? 1. Yes 2. No

20. Have you ever worried about your mother’s health because of her alcohol use? 1.
Yes 2. No

21. Have you ever been blamed for her drinking? 1. Yes 2. No

22. Did you ever wish your home could be more like the homes of your friends who
did not have a mother with a drinking problem? 1. Yes 2. No

23. Did your mother ever make promises to you that she did not keep because of
drinking?

1. Yes 2. No
24. Did you ever think your mother was an alcoholic? 1. Yes 2. No

25. Did you ever wish that you could talk to someone who could understand and help
the alcohol-related problems of your mother? 1. Yes 2. No

26. Did you ever fight with your brothers and sisters about your mother’s drinking? 1.
Yes 2. No

27. Did you ever stay away from home to avoid your drinking mother or your father's
reaction to her drinking? 1. Yes 2. No

132

28. Have you ever felt sick, cried, or had a "knot” in your stomach afier worrying
about your mother’s drinking? 1. Yes 2. No

29. Did you ever take over any chores and duties at home that were usually done by
your mother before she developed a drinking problem? 1. Yes 2. No

133

5‘3"!”

10.

11.

12.
13.

14.

15.

Appendix E

Children of Alcoholics Screening Test - Father (F-CAST)

Have you ever thought that your father had a drinking problem? 1. Yes 2. No
Have you ever lost sleep because of your father’s drinking? 1. Yes 2. No
Did you ever encourage your father to quit drinking? 1. Yes 2. No

Did you ever feel alone, scared, nervous, angry or frustrated because your father
was not able to stop drinking? 1. Yes 2. No

Did you ever argue or fight when he was drinking? 1. Yes 2. No

Did you ever threaten to am away from home because of your father’s drinking?
1. Yes 2. No

Has your father ever yelled at or hit you or other family members when drinking?
1. Yes 2. No

Have you ever heard your parents fight when your father was drunk? 1. Yes 2.
No

Did you ever protect another family member from your father who was drinking?
1. Yes 2. No

Did you ever feel like hiding or emptying your father's bottle of liquor? 1. Yes
2. No

Do many of your thoughts revolve around a problem drinking father or difficulties
that arise because ofhis drinking? 1. Yes 2. No

Did you ever wish that he would stop drinking? 1. Yes 2. No

Did you ever feel responsible for or guilty about your father's drinking? 1. Yes
2. No

Did you ever fear that your parents would get divorced due to your father’s
alcohol misuse? 1. Yes 2. No

Have you ever withdrawn from and avoided outside activities and friends because
of embarrassment and shame over your father's drinking problem? 1. Yes 2. No

134

16. Did you ever feel caught in the middle of an argument or fight between a problem
drinking parent (father) and your other parent? 1. Yes 2. No

17. Did you ever feel that you made your father drink alcohol? 1. Yes 2. No

18. Have you ever that a problem drinking parent (father) did not really love you? 1.
Yes 2. No

19. Did you ever resent your father's drinking? 1. Yes 2. No

20. Have you ever worried about your father’s health because of his or her alcohol
use?

1. Yes 2. No
21. Have you ever been blamed for his drinking? 1. Yes 2. No

22. Did you ever think your father was an alcoholic? 1. Yes 2. No

23. Did you ever wish you home could be more like the homes of your friends who
did not have a father with a drinking problem? 1. Yes 2. No

24. Did your father ever make promises to you that he or she did not keep because of
drinking? 1. Yes 2. No

25. Did you ever wish that you could talk to someone who could understand and help
the alcohol-related problems of your father? 1. Yes 2. No

26. Did you ever fight with your brothers and sisters about your father’s drinking? 1.
Yes 2. No

27. Did you ever stay away from home to avoid your drinking father or your mother’s
reaction to the drinking? 1. Yes 2. No

28. Have you ever felt sick, cried, or had a ”knot” in your stomach after worrying
about your father’s drinking? 1. Yes 2. No

29. Did you ever take over any chores and duties at home that were usually done by
your father before he developed a drinking problem? 1. Yes 2. No

135

Appendix F
Demographic lnforrnation and Alcohol Use

Please identify your gender: 1. Female 2. Male

What is your age:
What is your height (in feet and inches)? Feet Inches
Please type in your weight (in lb.’s)? Pounds

Please identify your race:

a. Caucasian d. Asian American
b. Afiican American e. Native American
c. Hispanic/Latino f. Other

If you also identify with a second race please select that as well?

a. Caucasian e. Native American
b. Afiican American f. Other
c. Hispanic/Latino g. Does not apply to me

d. Asian American

Please type in your major (including undecided or no preference):

 

 

Please identify your grade level:
a. Freshman
Sophomore

9‘

Junior

Fourth Year Senior
Fifth Year Senior
Graduate Student

hep-.0

I36

9. What is your current GPA?

 

10. Are you a member of a non-academic fi’aternity/sorority? 1. Yes 2. No

11. Please identify your parents’ marital status:

a.

b.

Married
Separated

c. Divorced

Never Married
Widow/Widower

12. Please identify when your parents divorced:

a.

b.

C.

d.

6.

They’re still married
This year

1 year ago

2 years ago

3 or more years ago

13. Are you currently in a significant relationship? 1. Yes 2. No

14. Please identify your marital status:

a.

b.

C.

d.

6.

Married
Separated
Divorced

Never Married
Widow/Widower

15. What is your current religious/spiritual affiliation (include none)?

 

16. How many times do you go to mass or religious service in an average month?

 

137

l7.

18.

19.

20.

21.

22.

23.

24.
25.

26.

What is your primary residence during the academic year?

a. Dorm on campus

F"

Fraternity/Sorority house
An apartment/house with roommates
An apartment/house with yourself

9-9

An apartment/house with your parents

9

f. An apartment/house with your significant other
g. Other

Please write in how many occasions you consumed alcohol over the past month (30
days)? '

 

Please write in how many average drinks you consumed per drinking occasion over
the past month? (1 drink is equal to a 12-oz beer, a 4-oz glass ofwine, a 12-oz wine
cooler, or a 125-oz shot of liquor, either straight or in a mixed drink)

 

If you are a male, please write in the number of occasions within the past 2 weeks
you consumed 5 or more drinks (including 0 for never)? (If you are a female please
write in an “X”).

 

If you are a female, please type in the number of occasions within the past 2 weeks
you consumed 4 or more drinks (including 0 for never)? (If you are a male please
write in an “X”).

 

Please type in the number of occasions within the past 2 weeks you got drunk
(including 0 for never)?

 

Do you feel you have a drinking problem? 1. Yes 2. No

Do you feel your best friend at Michigan State has a drinking problem? 1. Yes 2.
No

Please type in how many occasions you believe the average Michigan State student
consumed alcohol over the past month (30 days)?

 

Please type in how many average drinks you believe the average Michigan State
student consumed per drinking occasion over the past month? (1 drink is equal to a
12-oz beer, a 4-oz glass of wine, a 12-oz wine cooler, or a 12sz shot of liquor,
either straight or in a mixed drink)

 

138

27.

28.

29.

30.

31.

32.

On how many occasions within the past 2 weeks do you think the average male
Michigan State student consumed 5 or more drinks (including 0 for never)?

 

On how many occasions within the past 2 weeks do you think the average female
Michigan State student consumed 4 or more drinks (including 0 for never)?

 

Please type in how many occasions you believe your best friend at Michigan State
consumed alcohol over the past month (30 days)?

 

Please type in how many average drinks you believe your best friend at Michigan
State consumed per drinking occasion over the past month? (1 drink is equal to a
12-oz beer, a 4-02 glass of wine, a 12-oz wine cooler, or a 125-02 shot of liquor,
either straight or in a mixed drink)

 

If your best friend at Michigan State is a male, how many times within the past 2
weeks they consumed 5 or more drinks (including 0 for never)? (If your best friend
is a female please write in an “X”).

 

Ifyour best friend at Michigan State is a female, how many times within the past 2
weeks they consumed 4 or more drinks (including 0 for never)? (If your best friend
is a male please write in an “X”).

 

I39

Please read each statement carefully and respond with how generally true it is of you on a

Appendix G
Expectancies of Alcohol Use

0 (not at all) to 4 (a lot) scale.

1.

Drinking helps me to relax.

O 1 2 3
Notatall

. Drinking helps me forget problems at work or school.

0 1 2 3
Not at all
Drinking helps me feel better about myself.
0 1 2 3
Not at all

Drinking helps me forget my worries.

0 l 2 3
Not at all
Drinking helps me feel better when I’m down.
0 1 2 3
Not at all

Drinking helps me to relax when I’m tense.

0 1 2 3
Not at all
Drinking helps me to calm down when I’m angry.
0 l 2 3
Not at all
Drinking helps me deal with boredom.
0 1 2 3
Not at all
Drinking helps me express my opinions and ideas better.
0 1 2 3
Not at all

140

A lot

A lot

A lot

A lot

A lot

A lot

A lot

A lot

A lot

10.

ll.

12.

13.

14.

15.

16.

17.

18.

19.

20.

Drinking helps me feel more relaxed about sex.

0 1 2 3 4
Not at all A lot
Drinking makes me feel more sexy.
0 l 2 3 4
Not at all A lot
Drinking makes me do some things better.
0 1 2 3 4
Not at all A lot
Drinking makes me feel less shy.
O 1 2 3 4
Not at all A lot
Drinking makes it easier to find the right words when talking to people.
0 l 2 3 4
Not at all A lot

Drinking makes me feel more romantic.

0 1 2 3 4
Not at all A lot
Drinking helps me to fit in better with people around me.
0 l 2 3 4
Not at all A lot
Drinking makes me feel cool.
0 1 2 3 4
Not at all A lot
Alcohol tastes good.
0 1 2 3 4
Not at all A lot
Drinking makes any celebration more enjoyable.
0 1 2 3 4
Not at all i A lot
Drinking adds enjoyment to a good meal.
0 1 2 3 4

Not at all A lot

141

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

Drinking makes many activities more enjoyable.

O 1 2 3 4
Not at all A lot
Drinking helps me to fall asleep at night.
0 1 2 3 4
Not at all A lot
Drinking can be exciting.
O 1 2 3 4
Not at all A lot
Drinking makes sports (like football, basketball, car races) more enjoyable.
0 1 2 3 4
Not at all A lot

Drinking makes listening to music more enjoyable.

O 1 2 3 4
Not at all A lot
Drinking is a good way to kill time.
0 1 2 3 4
Not at all A lot

Drinking makes it easier for me to play sports.

0 1 2 3 4
Not at all A lot
Drinking makes me feel more alert.
0 l 2 3 4
Not at all A lot
Drinking helps me think better.
0 l 2 3 4
Not at all A lot
Drinking helps me understand complicated things better.
0 1 2 3 4
Not at all A lot
Drinking makes me fell more coordinated.
O l 2 3 4

Not at all A lot

142

32.

33.

34.

35.

Drinking helps me have better ideas.

0 l 2
Not at all

Drinking improves my concentration.

0 l 2
Not at all
Drinking makes me more creative.
0 l 2
Not at all

Drinking helps me perform certain tasks better.

0 1 2
Not at all

143

A lot

A lot

A lot

A lot

Appendix H

Consent Form

We are interested in gaining a better understanding into what contributes to alcohol
consumption in undergraduate college students. Specifically, we are studying the effects
of familial and interpersonal relationships with alcohol consumption.

We would like you to complete a battery of 4 surveys regarding your alcohol use,
parental alcohol use, interpersonal relationships, feelings of anxiety, and demographic
variables (e. g., your birth order, gender, major). All of your responses will be anonymous
and your privacy will be protected to the maximum extent allowable by law.

Participation in this study is completely voluntary. It is estimated that it will take no
longer than 45 minutes to complete the battery. You may stop at any point without any
penalty to you.

If you have any questions about this study, please contact the investigator (Chris Latty,
432-2271 #3, 329 Olin Health Center). If you have questions or concerns regarding your
rights as a study participant, or are dissatisfied at any time with any aspect of this study,
you may contact - anonymously, if you wish — Peter Vasilenko, Ph.D., Chair of the
University Committee on Research Involving Human Subjects (U CRIHS) by phone:
(517) 355-2180, fax: (517) 432-4503, e-mail: ucrihs@msu.edu, or regular mail: 202 Olds
Hall, East Lansing, NH 48824.

After completing the battery you may have concerns regarding your consumption of
alcohol or may wish to process rekindled feelings toward your family of origin. Two
resources are available on campus for you to discuss and explore these concerns: The
MSU Family and Child Clinic (432—2272; http://www.fce.msu.edu/mft/MFTClinic.htm)
and the MSU Counseling Center (355-8270; http://www.couns.msu.edu/).

You indicate your voluntary agreement to participate by completing and submitting the
questionnaires.

144

Appendix I
Debriefing Form

At most colleges and universities, there is a normative component in the student culture
involving alcohol consumption for reasons to socialize, relax, celebrate, and for pleasure
(Demers et al., 2002; Dreer et al., 2004; Schulenberg & Maggs, 2002; Wechsler,
Davenport, Dowdall, Moeykens, & Castillo, 1994; Wechsler, Isaac, Grodstein, & Sellers,
1994; Zucker et al., 1997). Many students will normatively drink alcohol and will never
experience consequences of drinking beyond a hangover. Unfortunately, some students’
alcohol consumption may evolve into dangerous levels of consumption, result in various
negative consequences, or begin a pattern for a life-long battle with dependence.
Additionally, some estimates indicate that around 31% of college students have a
diagnosis of alcohol abuse and 6% have a diagnosis of alcohol dependence. This high
prevalence of abuse puts students at greater risk for academic problems, sexual abuse,
risky sexual behaviors, physical injury, assault, violence, and death.

I am interested in understanding what disparate influences lead some college students to
drink at problematic levels. Although a number of studies to date have looked at the
influence of parental alcoholism and the role of peer group usage, the study you have just
participated in also adds a component from the field of marriage and family therapy as a
central variable in the role of college student alcohol consumption.

Differentiation exists on inter- and intrapersonal levels. Intrapersonally, differentiation
refers to a person’s ability to remain clear-minded and objective in the face of anxiety
rather than becoming emotionally reactive. It is also an interpersonal dynamic reflecting
an individual’s ability to maintain a sense-of-self while remaining emotionally connected
in relationships. An individual with a low level of differentiation has struggled in forming
their own identity and convictions and will adopt the beliefs and values of those they are
in close relationships with. Moreover, they are dependent on others for validation.
Therefore, when there is some form of relational anxiety (e.g., a criticism) the individual
with lower levels of differentiation are not able to tolerate this stress and become
emotionally reactive as they do not have an internal foundation of who they are. When
emotional disequilibrium occurs in a dyad, poorly differentiated individuals are likely to
triangle in other people or things to diffuse this anxiety. For example, a couple on the
cusp of fighting about themselves may channel this energy into a conflict about their
children. Moreover, individuals may triangle things (e.g., work or alcohol) as a means to
diffuse the relational tension building in the relationship, resulting in a couple having
conflict over the object rather than with their relationship. Therefore, an individual’s level
of differentiation influences how they tolerate anxiety and triangling is a dysfunctional
means to alleviate the anxiety. According to this theory, the more differentiated an
individual is, the more likely they are to tolerate anxiety, resulting in a greater potential to
objectively respond to anxiety, and therefore be less likely to exhibit a need to triangle
someone or something to regulate anxiety. The central hypothesis of this study is that
there is a relationship between differentiation and alcohol consumption in college
students.

145

This study asked you to reflect on your relationship with your family of origin, your
parents’ drinking, and your typical drinking behaviors. You may have concerns regarding
your consumption of alcohol or may wish to process rekindled feelings toward your
family of origin. Two resources are available on campus for you to discuss and explore
these concerns:

0 The MSU Family and Child Clinic (432-2272;
http://www.fce.msu.edu/mfi/MFTClinic.htm)
o MSU Counseling Center (355-8270; http://www.couns.msu.edu/).

You may also find the following resources helpful in obtaining more information about
Bowen theory, as well as other marriage and family therapy theories, and alcohol:

Bowen & MF'T ‘
Bowen, M. (1966). The use of family theory in clinical practice. Comprehensive
Psychiatry, 7, 345-374.
Nichols, M. P. & Schwartz, R. C. (2004). Family therapy: Concepts and methods (6”
ed.). Boston: Allyn & Bacon.
Titelman, P. (1998). Overview of the Bowen-theoretical-therapeutic system. In P.
Titleman (Ed.), Clinical applications of Bowen family systems theory (pp. 7-49).
New York: The Haworth Press.
Alcohol
Goodwin, D.W. (2000). Alcoholism: The facts (3" ed.). New York: Oxford
University Press.
http:/lwww.collegedrinkingprevention.gov/
http://www.a1coholics-anonymous.org/
http://www.al-anon.org/

Thank you once for participation in this study. Please feel flee to contact me if you have
specific questions or would like to get a copy of the final results.

Chris Latty
lattychr@msu.edu
(517) 432-2271 #3

146

Appendix J

Transformation of Continuous Alcohol Measures Into Categorical Data

The outcome variables for this study were subjects’ general alcohol use (GAU, as
measured by Quantity X Frequency over the past 30 days) and frequency of binge
drinking (over the past two weeks). The distributions were positively skewed
distributions for both GAU (see Figure 15, p. 148, for males and Figure 16, p. 149, for
females) and binge drinking (see Figure 17, p. 150, for males and Figure 18, p. 151, for
females). As these distributions would have violated the normality assumption of
regression analysis, the data were converted into categories.

Binge Drinking Conversion

The binge drinking behavior categories for this study replicated those fiom recent
studies (Bladt, 2002; Knight et al., 2002; Wechsler et al., 1994; Wechsler et al., 1998;
Wechsler et al., 2000; Wechsler, Lee, Kuo, et al., 2002; Wechsler, Lee, Nelson, etal.,
2002). Accordingly, four categories of binge drinking were created: Abstainers (those
that did not drink), Nonbinging Drinkers (subjects that consumed alcohol, but not at
binge levels), Occasional Binge Drinkers (1 or 2 times in the past 2 weeks), and Frequent

Binge Drinkers (3 or more times in the past 2 weeks) (see Table 17, p. 152).

147

25—

   
  

8
1

Frequency
‘7"

 

o 50 100 >150 200 250
30 Day Quantity X Frequency Alcohol Consumptlon

Figure 15. General alcohol use (Quantity X Frequency) of male drinkers for the Past 30

days.

148

   

8
1

Frequency

N
o
l

 

0.00 30.00 60.00 90.00 120.00 150.00
Quantity X Frequency of Alcohol Use

Figure 16. General alcohol use (Quantity X Frequency) of female drinkers for the Past 30

Days

149

30-

Frequency

20-

 

 

 

0.0 2.0 4.0 > 6.0 8.0 10.0 12.0 14.0

The number of days in the past 2 weeks the subject
drank at binge levels

Figure 17. Frequency of male binge drinking over the past 2 weeks.

150

100—

Frequency

is
O
l

20-

 

 

 

00 2.0 4.0 6.0 8.0 10.0 12.0
The number of times the subject drank at binge levels

Figure 18. Frequency of female binge drinking over the past 2 weeks.

151

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Table 17
Frequency of Binge Drinking Categories by Gender
Total Sample Males Females
Frequency Percent Frequency Percent Frequencyl Percent
Abstainers 49 1 1.0 16 8.0 33 13.4
“133523 98 2"9 31 15.4 67 27.2
$2;ng ‘35 302 60 29.9 75 30.5
Priming ‘65 36'9 94 46.8 71 28.9
GA U Conversion

Whereas binge drinking categories were established in multiple studies, no clear
norms for creating quantity frequency categories were present in the literature. It was
decided that abstainers would naturally be one category and it needed to be determined if
there would be enough of variation between the categories to create three additional
variables (low-, moderate-, and high drinking categories by splitting the frequency
distribution of those that consumed alcohol into thirds) or only two categories (low and
high categories by a median split). As it was established from the test of association with
binge drinking that males and females were to be analyzed separately (see Appendix K),
frequency distribution points were utilized to create categories for males and females
separately. A one-way AN OVA testing the 4 category model indicated that the moderate
group did show differences from the low and high groups on comparisons to the DSl-R
full- and subscale measures. Therefore, the positively skewed Quantity X Frequency
variable was converted into a general alcohol use variable comprised of four categories

(see Table 18, p. 153).

152

 

Haw—a ;

manage! e». O>C Omnamomom 3. @232

 

 

 

 

 

 

 

H08— mmBEo Zw—om moan—om

mama—egg $2.83 mane—nave $82: :35 m0 M.,—.2533 moans» Zoe: m0

sarcasm—09m so _ _ _ a m .. l mu _ w .A 2. ..

_.ros so 2 3 3a 3.8 «a a 8.8 93 .30
9.5.83

38888 E 8 3 8.” 3: K.,: 8 was 8.3 o.oo
0:388
Em:

. as we ac Nob Sun; the K we; 3.? 3.9»
6:388

 

 

 

 

 

 

 

 

 

 

 

 

153

Appendix K
Individual Variable Descriptions

Gender

With the transformation of the skewed continuous data to categorical data, logistic
regression was immemented to test the relationship between differentiation and the
moderating variables with the outcomes of alcohol consumption. Prior to examining the
main and interaction effects with logistic regression, males and females were tested for
differences in alcohol consumption to avoid the interpretation of three-way interaction
effects in the analyses. In order to test the hypothesis that gender and binge drinking
behaviors were independent of each other (no relationship), a chi-square test of
association was implemented with a .05 or level. The results indicated that gender and
binge drinking were related 352(3) = 19.66, p < .001 and the Cramer’s phi (9c = .21)
indicated that the association was relatively small. Since there was a significant
association between gender and binge drinking, males and females were analyzed
separately for all firrther analyses.
Race

Of the 447 subjects analyzed in this study, 373 identified themselves as Caucasian
(83.4%), 45 as African American (10.1%), 5 as Hispanic/Latino (1.1%), 20 as Asian
American (4.5%), l as Native American (.2%), and 3 as Other (.7%) (for gender
comparisons see Table 19, p. 155). A chi-square test of association indicated that there
was not a relationship between gender and race for this sample x26) = 7.214, p = .205.
As this sample had such a limited numbers of non-Caucasians, this variable was excluded

from all future regression analyses.

l54

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Table 19
Frequency of Race by Gender
Total Sample Males Females
Frequency Percent Frequency Percent Frequency Percent
Caucasian 373 83.4 169 84.1 204 82.9
African
American 45 10.1 17 8.5 28 11.4
H'SPmc/ 5 1.1 2 1.0 3 1.2
Latino
Asian
American 20 4.5 10 5.0 10 4.1
Native
American 1 .2 0 0 1 .4
Other 3 .7 3 1.5 0 0
Age

The age of participants for this study was restricted to subjects who were between
18 and 22, reflecting the typical age range of traditional students. Although males were
significantly older (19.72, SD 1.24) than females (19.45, SD, 1.19) (t (445) = -2.39, p =
.017, d = 0.22, 3,, = 0.01), the small effect size indicates this difference was relatively
trivial.
BMI

Values of BMI are typically used to categorize weight by adjusting for height.
The traditional weight ranges for adults are: Underweight (BMI < 18.5), Normal (18.5 -
24.9), Overweight (25.0 - 29.9), and Obese (30.0 _>_ BMI). The categories were not
implemented in this study, as the BMI values were maintained as a continuous measure

to serve as a proxy for alcohol tolerance based on weight. Males had a significantly

155

greater BMI value (24.62, SD = 3.78, range 17.72 to 41.38) than females (22.80, SD =
4.09, range 17.16 to 47.50) (t (445) = -4.85, p < .001, d = 0.46, 12,, = 0.05).
COA

COA status was determined by subjects’ responses to created gender-specific
versions of the Children of Alcoholics Screening Test (CAST). Scores of six or more
“yes” responses on the 29—item scales of either the F -CAST (for fathers) or M-CAST (for
mothers) indicate the presence of an alcoholic parent. High internal consistency
reliabilities based on Cronbach’s alpha values were found for both the M-CAST (males 01
= .96, females or é .96) and F-CAST (males 11 = .95, females or = .96). Almost 80% of
males and females had parents who were not identified as being alcoholic (see Table 20).
A chi-square test of association indicated that there was not a significant relationship
between gender and the number of parental alcoholics (based on categories of O, 1, 2) for

this sample (12(2) = 5.24, p = .073).

 

Table 20

Frequency of Parental Alcoholism by Gender

 

 

 

 

 

Males Females
Frequency Percent Frequency Percent
No Alcoholic Parents 169 84.1 195 79.3
One Alcoholic Parent 22 10.9 44 17.9
Two Alcoholic Parents 10 5.0 7 2.8

 

 

 

 

 

 

 

Due to the low fi'equency of one and two alcoholic parent families, tests of
association were conducted to validate combining these two categories into one variable,
creating a dichotomy of being a COA or not. There was no significant relationship

between the 3 categories of parental alcoholism with frequency of binge drinking for

156

males (x2(6) = 10.31, p = .1 12) or females (12(6) = 6.17, p = .405, see Table 21 for

gender distribution). Similarly, there was no a significant relationship between the 3

categories of parental alcoholism with general alcohol use for males (38(6) = 7.71, p =

.260) or females (36(6) = 1.78, p = .939, see Table 22 for gender distribution).

 

Table 21

Frequency of Parental Alcoholism and Binge Drinking by Gender

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

. Nonbinging Occasional Frequent
Abstainer Drinker Bmge Binge
Drinker Drinker
No alcoholic parents 16 27 51 75
Males 1 Alcoholic Parent 0 4 6 12
2 Alcoholic Parents 0 0 3 7
No alcoholic parents 24 56 56 59
Females 1 Alcoholic Parent 8 8 18 10
2 Alcoholic Parents 1 3 1 2
Table 22
Frequency of Parental Alcoholism and GAU by Gender
. Low Moderate High
“5'31““ Drinker Drinker Drinker
No alcoholic parents 16 54 52 47
Males 1 Alcoholic Parent 0 7 7 8
2 Alcoholic Parents 0 2 3 5
No alcoholic parents 24 6O 51 6O
Females 1 Alcoholic Parent 8 13 11 12
2 Alcoholic Parents 1 3 1 2

 

 

 

 

 

 

 

Individuals who had either one or two alcoholic parents were combined into one

COA category. As was the case with the three categories of parental alcoholism, there

was no significant relationship between gender and COA status 080) = 1.69, p = .193).

157

 

There was no significant relationship between the two categories of parental
alcoholism with frequency of binge drinking for males (36(3) = 7.02, p = .071) or
females (36(3) = 3.22, p e .359, see Table 23 for gender distribution). Although neither
test was significant, the p values for the dichotomous COA variable
were closer to significance than were the original three category variable. There also was
no significant relationship between the two categories of parental alcoholism with general
alcohol use for males (38(3) = 7.04, p = .071) or females (78(3) = 1.12, p = .772, see
Table 24 for gender distribution). Since the COA variable slightly improved the
frequency distribution and did not alter the relationship with alcohol use, the

dichotomous variable was utilized for all future analyses.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Table 23
Frequency of COA Status and Binge Drinking by Gender
Occasional Frequent
Abstainer “winging Binge 03:5: (3
Drmk' er Dnnk' er (1-2 or
times/week) more/week)
Males Not a COA 16 27 51 75
COA 0 4 9 19
Not a COA 24 56 56 59
Females COA 9 11 19 12
Table 24
Frequency of COA Status and GAU by Gender
. . Moderate High
Abstainer Low Drinker Drinker Drinker
Males Not a COA 16 54 52 47
COA 0 9 10 13
Females Not a COA 24 60 51 60
COA 9 16 12 14

 

 

 

 

 

 

158

 

Expectation of Alcohol as a Tension Reducer

Nine questions fiom Kushner et al.’s (1994) measure of alcohol outcome
expectancies comprise the Tension Reduction subscale. The subjects’ responses on a 5-
point scale were summed with higher scores reflecting greater tension reduction
expectations. The Tension Reduction subscale had a high internal consistency reliability
based on Cronbach’s alpha value for both males (01 = .91) and females (01 = .92). Males
had higher expectations of alcohol as a tension reducer (15.07, SD = 8.73) than females
(13.15, SD = 8.81) (t (445) = -2.39, p = .022, d = 0.22, r2» = 0.01) although the point-
biserial correlation indicated that gender only accounts for about 1% of the variance in
expectations of alcohol as a tension reducer.
Expectation of Alcohol as a Social Lubrication

Eight questions from Kushner et al.’s (1994) measure of alcohol outcome
expectancies comprise the Social Lubrication subscale. The subjects’ responses on a 5-
point scale were summed with higher scores reflecting greater social lubrication
expectations. The Social Lubrication subscale had a high internal consistency reliability
based on Cronbach’s alpha value for both males (01 = .89) and females (01 = .90).
Although males had higher expectations of alcohol as a social lubricant (13.93, SD =
7.89) than females (10.97, SD = 7.74) (t (445) = -3.99, p < .001, d = 0.38, rzpt, = 0.03),
only 3% of the variance of expectations of alcohol as a social lubricant can be accounted
for by gender.
Perceptions of A verage MSU Student 's Alcohol Use

A total drinks-per-month variable was calculated by multiplying the number of

days by the average number of drinks subjects believed the average Michigan State

159

 

 

 

student consumed in the past 30 days. Females (57.57 drinks-per—month, SD = 38.75) and
males (50.89 drinks-per-month, SD = 37.22) did not significantly differ from one another
in their perceptions of MSU student alcohol consumption (t (445) = 1.85, p = .065).
Perceptions of Subjects ’ Best Friend at MS‘U ’s Alcohol Use

A total drinks-per-month variable was calculated by multiplying the number of
days by the average number of drinks subjects believed their best friend at Michigan
State consumed in the past 30 days. Females had significantly lower perceptions of their
best fiiends alcohol usage (36.88, SD = 41.48) than males (68.16, SD = 72.43) (t (303.78)
= -5.44, p < .001, d = 0.54, 12,3, = 0.09), although only 9% of the variation in perceptions
could be attributed to the subjects’ gender.
Tests of Multicollinearity

A summary description of means and standard deviations is provided in Table 25
(p. 160) After transforming the respective independent variables, correlation analyses

were done separately for males (see Table 26, p. 162) and females (see Table 27, p. 163).

 

Table 25

Frequency of Independent Variables by Gender

 

Males Females

 

Mean SD Range Mean SD Range

 

Age 19.45"I 1.19 18-22 1972" 1.24 18—22

 

BMI 24.62" 3.78 17.72 - 41.38 22.80“ 4.09 17.16 - 47.5
Tension reduction 1508* 8.73 0 - 33 1315* 8.81 0 — 36
Social Lubrication 13.93“ 7.89 0 — 30 10.97" 7.74 0 — 30
MSU Student Use 50.89 37.22 6 — 176 57.58 38.75 4 — 180
Best Friend Use 68.16” 72.43 0 — 387 36.88" 41.48 0 — 180

 

 

 

 

 

 

 

 

 

 

 

* p < .05, two-tailed. ** p < .01, two-tailed.

 

l60

 

Ignoring the correlations between the DSI-R full scale. with the 4 subscale measures (as
the full- and subscales will not be analyzed in the same equations) the correlations ranged
from .00 to .79 for males and .00 to .71 for females. As logistic regression does not have
a measure directly equivalent to the R2 value of OLS regression, tests of multicollinearity
could not be directly computed using the variance inflation factor formula (V IF) (von
Eye & Schuster, 1998). In order for an exploratory analysis of multicollinearity to be
conducted, liner regressions with the positively skewed continuous GAU and Binge
Drinking variables were analyzed using the VIP. The VIP values were identical for GAU
and Binge Drinking outcomes.

None of the respective regression equations (differentiation only, five equations
of differentiation with interaction terms, and the comprehensive equation including all
variables) had VIF values reaching a value of 10.0 — a rule of thumb for multicollinearity
(von Eye & Schuster, 1998) (see Table 28, p. 164, for VIP values for the Comprehensive
Model which contained all of the variables). Therefore it was concluded that

multicollinearity was not a concern for the remaining analyses.

161

mono—macaw 31330363 55628 1 3m? masooa

mw wagon—o Amwv
:0 wagon—o :3
m0 wagon—o Ame
m0 mccmoao GHQ
w7=
>mo
00>
Ema 3538
wow" Rona
H088: non—Como:

mega r5108:

0 .1. :53 n a: £62m. n 363m."
w.” n 330303 cm com” $256 :8. a.” u oxeoosmoam cm Same: 898:0? 9.. n 9888225 0». 8&3 _cclomnoz.

_._ A "Serge—ma. _._... A .o go-§_oa.

 

162

mono—mnonm men Engages» 556.8 I moan—o mcgmoa

mw wagon—o

:0 mccmoao . .
m0 mccmofla

m0 wagon—o

get:

163

>mn
00>
ch men—0:8
wow» Rosa
H088: 398:0:

moon: 5&3.er
u 0 38m 0 a: mBEm. u madame
mm. N 380303 em com» «lo—aw :mo. H” u 06008203 cm 838: 898:0? mr n 0633203 om moan. 3.61820?

_._ A 25-828. .1. A.A: 25-859

 

 

Table 28

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

VIP for Independent Variables
Males Females

DSI-R 1.54 1.67
Age 1.11 1.05
Body Mass Index 1.09 1.09
Perceptions of MSU student use 1.21 1.12
Perceptions of best friend at MSU’s use 1.45 1.35
Expectations of alcohol as a tension reducer 2.97 2.17
Expectations of alcohol as a social lubricant 2.93 2.22
COA status 1.07 1.08
DSI-R X Perceptions of MSU student use 1.11 1.11
DSI-R X Perceptions of best fiiend at MSU’s use 1.52 1.44
DSI-R X ExEctations of alcohol as a tension reducer 3.47 2.79
DSI-R X Expectations of alcohol as a social lubricant 3.21 2.70
DSI-R X COA status 1.35 1.33

 

 

Note: All variables have been centered around the mean, with the exception

of the dummy variable of COA status

 

I64

 

Appendix L
Analyses of General Alcohol Use

As previously described, the continuous General Alcohol Use (GAU) measure (30
day Quantity X Frequency) was converted into a categorical variable consisting of 4
levels: Abstainers, Low Drinkers, Moderate Drinkers, and High Drinkers (see Appendix
I). As the four categories reflect increasing degrees of alcohol consumption, the nature of
this study’s originally stated hypotheses for a continuous dependent variable (e. g.,
students with lower levels of differentiation will be more likely to have greater amounts
of general alcohol consumption) was maintained with the categorical analyses (e. g.,
students with lower levels of differentiation will be more likely to have a greater
relationship with High Drinkers than Low Drinkers when controlling for BMI and age).

As logistic regression is based on a dichotomous dependent variable, the four
GAU categories were contrasted in ways such that four unique outcome dichotomies
were created:

1. Drinkers (Low, Moderate, High) versus Abstainers

2. Moderate Drinkers versus Low Drinkers

3. High Drinkers versus Low Drinkers

4. High Drinkers versus Moderate Drinkers
In each instance, the variable with the greater amount of alcohol consumption was coded
as a “1” in SPSS 13.0 and the lesser category was coded as a “0.” For example, in the
Drinkers versus Abstainers analyses, all persons in the Low, Moderate, and High

Drinking categories were coded as a “l” and Abstainers were coded as a “0.”

165

Drinkers versus Abstainers

Diflerentiation and GA U

Males. The Basic Model was significant according to the model chi-square (x2(3)
= 18.00, p < .001, Ric,“ 5,," = .09, PM...” = .20). The model illustrated a 92%
success rate in predicting the correct drinking classifications, however, this is the same
value if someone were to predict an individual was a drinker with every case (a model
with no predictors), indicating the Basic Model adds nearly nothing to the ability to
predict individuals that drink from those that abstain. When an individual is at an average
BMI level and age, differentiation had a significant negative coeflicient indicating that
the odds of drinking are multiplied by .190 (95% CI = .068, .530) with a one unit
difference in differentiation, which is a 81% (1 - .19) decrease (see Table 29, p. 167).
Additionally, when differentiation and age are at average levels, an individual is 2.24
times more likely to drink as they gain one year of age (e.g., 19- to 18-year olds, 20- to
19-year olds).

Females. The Basic Model was not significant according to the model chi-square
(£0) = 6.91, p = .075), indicating that there was no effect of the independent variables,
taken together, on the dependent variable (see Table 29, p. 167).
F our factors of the DSI-R (Emotional Reactivity, “I ” Position, Emotional C utofl and
Fusion with Others) with GA U

Males. As was the case in the Basic Model, the overall four factor model was
significant (36(6) = 18.21, p = .006, R20», & s...“ = .09, 82mm. = .20). This model had

the same successful prediction rate of 92% that the Basic Model and having no model

 

flea—a No

rommmno Womaommmo: wmnamnzsm 9338 an: 013w I Omanqoszmaoa

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Zea—aw maaaom
a mm €65 OW 8.x. 9 e mm $35 or 8.x. 9
Um?” -_ .3 93 8.3.... 93o o.oam ouwo -o.$ 93 #8 a PM; 0.3x o.ouo
95 .92 o.ou Pow o.omc Pmfi : .3 .98 obs mud 993 0.30 foo.
>ma of on: o.oo: wwwo fwww are: o.oo 93 ohm _.owm 93m _.hmo
ate A bu. 23-8:09 rate A .2. 396:3.
Haw—o we
homage ”3388: 32:on mEaga 9m: 9.5x i mean 1888
:33 moan—om
w mm (Sea OW 3.x. 9 e mm (5.5 OW 3.x. 9
wages”— Woeomia. .93 ohm _.ws chum PNNN T3» 93 ohm PI :8 993 _.08
:1 $830: -93 93 93 0.95 obmw _.uua -93 9.63 93 9me 933 ram
mangoes. OER“ -o.~o 9.3 9; Pawn 930 News 93 ohm o.oo _.cow cud. _.qfi
358: SE. 0563 -obm PS cow oboe chow Puma ion; owe Pam 93¢ obB _.cmo
9,3 -93 0.3 98 923 Pug :3 -93 PE Nd 98.“ 33.6. _.Sw
>mo Pmo out 93: NNNN _.Nuc are: o.oo 93 ohm _.oE chem _.Aom
e m A .3. 236:8. 1g A .3. 298:8.

 

 

167

had. None of the four factors significantly changed the odds of group membership (see
Table 30, p. 167).

Females. As was the case in the Basic Model, the overall four-factor model was
not significant (12(6) = 9.21,p = .162, see Table 30, p. 167).
Dijferentiation & Perception of A verage MSU Student ’3 Alcohol Use

Males. The Additive Model with perceptions of the average MSU student’s
alcohol consumption was significant overall (38(4) = 18.20, p = .003, chox & 5m" = .09,
Rzumm = .20) and had a successful prediction rate of 92%. The model did not show a
block improvement over the Basic Model (x20) = .03, p = .862). The change in odds of
drinking remained virtually the same as the Basic Model for the significant individual
predictor of differentiation (OR = .190, 95% CI = .068, .530, see Table 31, p. 169).

The Interaction Model also was significant overall (78(5) = 18.20, p = .003, R20».
& 5nd1= .09, from“, = .20) and had a successful prediction rate of 92%. However, it did
not show a block improvement over the Basic Model (x2(2) = .20, p = .900). Although
the individual predictors of differentiation (OR = .181, 95% CI = .063, .525) and age (OR
= 2.242, 95% CI = 1.227, 4.096) remained significant, the interaction term failed to reach
a level of significance (see Table 31, p. 169).

Females. The Additive Model with perceptions of the average MSU student’s
alcohol consumption was not significant as an overall model (x2(4) = 7.004, p = .136, see
Table 31, p. 169). The Interaction Model also was not significant as an overall model

(38(5) = 7.480, p = .137, see Table 31, p. 169).

I68

 

Hugo 2

rommwmo Womqommmo: $3833 mamas; 93 01:.» l $28303. om ch mamas» Cmo

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

_ 3E8 _ medium
_ a _ mm _ $5 _ ox _ 8.x. 9 _ o _ mm _ gen 9 ow 2 3.x. 9
wwmmo zona—
Umrw 2.3 9% 293...... 97.5 93m 93¢ -93 98 92... 92c 9mg 93c
22 -92 93 98 923 922 :3 -93 o.oo. Pd 93a 93a _.o2
>mo 92 9.2 93...... 980 _.NwN arc-2 o.oo 93 9mm _.cmm 93m _.Axo
>552?“ Zena
02-x 2.3 93 _98: 93c 98m 9H: -93 9mm “rub... 93o chow 90%
W73 -92 93 98 923 930 :3 -98 92. who o.owu 93m _.ooo
>mo 92 92 98.... NNE 2N: s: 3 98 93 93 _.oma 9qu _.Amm
Emc C8 98 92 98 :5. 903 _.o; o.oo 92 o.oo 98c 90% _.oom
383030: 30%;
02-x Ln: 9% 92...... 922 98m 93m 9% 98 Puma 9M2 9N2 902
92 -98 93 93 90mm 93A :3 -93 92 no.2 9on 93m _.oom
>mo 92 92 93...... when _.NN-N Poem 93 93 93 :8 98A _.2.»
ch Cmo o.oo 92 95 _.ooe 93w _.ONA 98 92 92 o.ooo 93o _.ooo
Um?” X ch CR -92 98 93 .09.. chow 2.8m -92 92 93 985 93a _.o:
a. NA .8. ”So-8:8. .I. to A .o . "so-8:8.

 

 

 

169

Dzflerentiation & Perception of Best Friend at MS U ’3 Alcohol Use

Males. The Additive Model with perceptions of their best friend at MSU’s
alcohol consumption was significant overall (38(4) = 24.12, p < .001, R29,” 5n,“ = .11,
[(2)-agent, = .27) with a significant block improvement over the Basic Model (x20) =
6. 12, p = .013), and had a successful prediction rate of 92.5%, an increase .5% over the
Basic model. The change in odds for being a drinker remained a significant individual
predictor of differentiation (OR = .225, 95% CI = .081, .624), when the student has an
average level of age, BMI, and perception of their best friend at MSU’s alcohol use (see
Table 32, p. 172). Similarly, when subjects had average levels of the other predictors,
they were 1.02 times more likely to be a drinker with every increase of 1 unit in
perceived monthly Quantity X Frequency drink they believed their friends drank. As the
range for perceptions is quite large, a 1 unit change is relatively niiniscule. To better
illustrate this effect, if a subject’s perception were to increase by the equivalent of one
standard deviation for males (72.43), they would be 4.26 times more likely to be a
drinker. Additionally, when differentiation and perceptions are at average levels, an
individual is 2.27 times more likely to drink as they gain one year of age.

The Interaction Model also was significant overall (38(5) = 24.64, p < .001, RIC”
& Sue" = .12, Rznmm = .27) and had a successful prediction rate of 92.5% (see Table 32,
p. 172); however, it did not show a block improvement over the Additive Model (xz( 1) =
.52, p = .470). Although the interaction term itself was not significant, by adding it to the
equation differentiation no longer significantly changed the odds of being a drinker.
When the subject had average levels of the other predictors, they were 1.02 times more

likely to be a drinker with every 1 unit increase of best friend perception and 1.75 times

170

for one standard deviation change. Additionally, when differentiation and age are at
average levels, an individual is 2.28 times more likely to drink as they gain one year of
age.

Females. The Additive Model with perceptions of their best friend at MSU’s
alcohol consumption was significant (38(4) = 28.53, p < .001, R20,x & Sm" = .11, Rznmmc
= .20) with a significant block improvement over the Basic Model (x20) =
21.62, p < .001), and had miniscule improvement in successfirl prediction rate to 87%, an
increase .4% over the Basic Model. Perceptions of best friend use changed the odds of
being a drinker by 1.044 (95% CI = 1.018, 1.070) for every one unit change in
perceptions (see Table 32, p. 172). Individuals who had higher best friend perception
equivalent to one standard deviation for (41.48) were 5.95 times likely to be a Drinker
versus Abstainer. Additionally, the presence of best fi'iend perceptions eliminates the
significance of difl‘erentiation found in the Basic Model.

The Interaction Model also was significant (36(5) = 29.48, p < .001, R29“ & Sm" =
.1 1, [(2)-make = .21), but did not show a block improvement over the Additive Model
(12(1) = .95, p = .329) and actually had a worse successful prediction rate (85.8%) than
all of the other models—including not having a model at all. The interaction term was not
significant, and the effect of perceptions of best fi'iend use remained equivalent to the

Additive Model (OR = 1.044, 95% CI = 1.018, 1.070, see Table 32, p. 172).

171

 

333—3 3

homage womaommmos 383333 9:338 98 Oman .- woqoavmosm om mam” 35:: :8

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

3E3 woBEom
@ rmm _ éma _ ow _ 8.x. 9 e _ mm _ €95 _ ow _ 3.x. 9
wwmmo Zona—
UmH-W -_ .33 933 333...... 9 2 3 9333 93.3 -93 93 PB _. 93 3 9333 933
E": -92 93 93 933 9332 :3 -93 93... 3.33 933 9333 _.32
>mo 92 93 3.3.1. ~33 _.N3 P3— 93 93 933 _.33 93cm _.Amo
>333<o gono—
Ume -_ .3 93 92 _._... 933 932 933 -93 93 3.3 9333 933 :3
BE -92 93 92 903 9333 :3 -93 93 _ .3 933 933A _.33
>3 93 93 3.33... 3.33 _.NZ N33 933 93 93 _.333 9333 _.3u
wow” magma Cmo 93 92 P3... _.23 _.o2 2.33 93 92 2.3.... _.oi _.23 _.33
38398: 7330.
6mm.” -23 933 N2 . 9qu 933 2.33 -933 93 92 903 933 3.33
E": -92 93 92 903 9333 :3 -933 93 _.NA 9033 9mm— _.33
>mo 93 93 3.3... ~33 _.N: 3.33 933 9_ 3 93 _.333 9333 _.33
wow" magnum Cma 93 92 93... 2.23 930 2.33 93 92 2.3 _ a... 2.33 _.2 w _.o3c
Um?” X mama mania Cmo 92 92 93 2.2— 933 2.03 93 93 _.oo _.33 .2: .32
HA .3. "so-82am. _._... 3 A .2. 25-833.

 

 

 

172

Diflerentiatian & COA Status

Males. The Additive Model with COA status was significant overall (38(4) =
24.09, p < .001, R20»- & 5...," = .11, Rzumumc = .27) with a significant block improvement
over the Basic Model (380) = 6.07, p = .014).The model had a successful prediction rate
of 92%, the same'as the Basic Model. The individual predictor of differentiation
remained significant (OR = .162, 95% CI = .052, .507, see Table 33, p. 174) when the
student has an average level of age, BMI, and was a COA. Being a COA by itself did not
increase the odds of being classified as a drinker; however this may have more to do with
the sample. The results of these analyses should be interpreted as controlling for COA
status, as none of the Abstainers had an alcoholic parent.

The Interaction Model was also significant as an overall model (12(5) = 24.07, p <
.001, R20,” 5...,“ = .11, Rzumm = .27) and had a successful prediction rate of 92%;
however, it did not show a block improvement over the Additive Model (x2( I) = .000, p
= 1.000). The interaction term was neither significant nor did it change the significant
odds for differentiation (OR = .162, 95% CI = .052, .507, see Table 33, p. 174).

Females. The Additive Model with COA status was not significant as an overall
model (76(4) = 8.064, p = .089, see Table 33, p. 174).

The Interaction Model also failed to achieve a significance as an overall model
(36(5) = 8.200, p = .146, see Table 33, p. 174).
Dlfierentiation & Expectations of Alcohol Serving as a Tension Reducer

Males. The Additive Model with expectations of alcohol as a tension reducer

was significant overall (x2(4) = 55.06, p < .001, R2 Cox & Snell = .24, Rznmmc = .56) With a

173

 

H35 8

rommmno ”08638: 33338 £338 92 015* I 00> mSEm

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

_ K38 _ moBEom
_ u mm H 53 _ ox _ 3.x. 9 _ w _ mm _ 2% _ ox _ 8.x. 9
W83 goua—
Ume Lao 98 3.8.... 93¢ .88 98¢ -93 98 PM: 93o 9n8 93o
ESH -93 93 98 980 .83 :3 -98 98 8.3 93a 93o _.oS
>mo 9M: 9”: o.oo: 9N3 _.Num “#3— 98 93 98 :88 93m #36
253$” gono—
Umzw Law 98 93.... 93w bum 9m3 -93 98 9mm... 93m 98o 993
El: o.oo 93 o.oo 98a .83 :3 -98 98 9% 908 983 _.oou
>ma 9m— 93 3.8a: Nbam _.Num Nr30 98 93 9mm _.38 938 #33
00> L 98 8.3.3 98 o.ooo o.ooo . 98 9.3 _.B _.3 o 98¢ Pm _ 8
388030: 30%;
Um?” 93 98 93...... 93m 98m 9m3 -93 98 .28... 93m 9~Nw o.ooo
wz: 98 93 o.oo 998 983 :3 -98 92 PE 93a 93a _.oom
>mo 9M: 93 3.3: NNAM _.Nuw Kr3o 98 9 3 9w; _.8_ 938 _.38
00> - _ 93 83.8 o.oo o.ooo o.ooo . 93 9.3 #8 fat 93A 983
Um?” x 00> 53 33.3 o.oo Pm; o.ooo . 98 93 9E _.uuc 9N3 93o

 

 

 

 

 

 

 

 

 

 

 

 

 

 

.6 A .8, «so-E39

IN A .3. arc-8:3.

 

 

174

significant block improvement over the Basic Model (12(1) = 37.06, p < .001), and had
successfill prediction rate of 95.5%, an increase 3.5% over the Basic model. The tension
reduction variable was significant as subjects were 1.388 (95% CI = 1.175, 1.639) times
more likely to be a Drinker with a 1 unit increase in tension reduction expectations (see
Table 34, p. 176). Subjects whose expectations were by one standard deviation greater
(8.73) are 17.53 times more likely to be drinkers.

The Interaction Model also was significant overall (36(5) = 55.73, p < .001, R20.
& 5..." = .24, Rznmlkm = .57) and had a successful prediction rate of 95.5%; however, it
did not show a block improvement over the Additive Model (x20) = .67, p = .412). The
interaction term did not significantly change the odds of being a drinker and the tension
reduction variable maintained the same level of influence (OR = 1.375, 95% CI = 1.169,
1.618, see Table 34, p. 176).

Females. The Additive Model with tension reduction expectations was significant
as an overall model (12(4) = 81 .39,p < .001, R20,” 5..." = .28, [Pp-mm.“ = .52) and had a
significant block improvement over the Basic Model (x2(1)= 74.47, p < .001). The
model had a successful prediction rate of 91 .5%, an increase of nearly 5% over the Basic
model. Expectations of alcohol serving as a tension reducer changed the odds of being a
drinker by 1.374 (95% CI = 1.225, 1.542) for a one unit difference in expectations (see
Table 34, p. 176). By changing the difference in expectations to a unit equivalent to one
standard deviation for expectations (8.8.1) the likelihood of being a Drinker versus
abstainer is 16.47 times greater. Additionally, the presence of tension reduction

expectations eliminated the significance of differentiation found in the Basic Model.

175

 

Hugo wk

rommmzo flamingo: 3.33:5 magma :5: Oman I mxvgfimmoaw cm H258: flamenco:

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

3&8 _ moBEam
a _ mm _ éaa _QC 89.9 a a 4 mm_ 55 _ow_ 89.9
mean 7.3%.;
cm; -98 9a _92: 93¢ 98“ 93° -93 98 92.. 9m; 98“ 93°
wz: -92 93 98 926 $2 :3 -98 92 9d 98a 93a _.9:
£0 92 93 98: Sue Cu» .8: 98 9; 9g 5% 38 _.ae
>858 Zena
cm; 9% 9% 98 933 9:: 93a -9; 9; 9: 9me 9§ $3
wz: 93 98 98 _.So 93... 3% -98 9E 93 93“ 38 SS
2% 93 93 38.. 93.. _.Se 93... 9; 98 9.5 _.E 95 _.Sa
38%: 98:88 9% 98 2.8: 5% Ed 88 98 98 N93: 3: _.BM 33
58338: Keno—
Um; 9s GM 98 ~95 98” 3.83 9% _.8 93 38 9M: .38
BE 92 98 9o. 28 93° :3 -98 99. 9.; 9e: 935 2%
£0 9% 9% 98°. 83 _._: SB 9: 98 93 _.Ge 9§ 3%
#8389888: 93 98 :3: Cd 2% Z; 93 98 N92: 33 SB 9%..
3:; .318 98:38 9: 9; 9E _._; 993 it 98 98 9.3 2% 933 _.qu

 

 

 

 

 

 

 

 

 

 

 

 

 

 

.. an A .8. 30.823. 2.. Ne A .2. "Ea-8:8.

 

 

176

The Interaction Model also was significant overall (12(5) = 81.82, p < .001, R20...
& 5nd1= .28, Rznwde = .52) and had a successful prediction rate of 91 .1%; however, it
did not show a block improvement over the Additive Model (780) = .43, p = .511). The
interaction term was not significant and the significant odds ratio for tension reduction
remained equivalent to the Additive Model (OR = 1.370, 95% CI = 1.223, 1.534, see
Table 34, p. 176).
Ditfirentiation & Expectations of Alcohol Serving as a Social Lubricant

Males. The Additive Model with expectations of alcohol as a social lubricant was
significant overall (38(4) = 53.38, p < .001, 12%,, s s..." = .23, [(2ch = .55) with a
significant block improvement over the Basic Model (x20) = 35.28, p < .001). The
model had a successful prediction rate of 95.5%, an increase 3.5% over the Basic model.
The social lubrication variable was significant for subjects as individuals with 1 unit
higher social lubrication expectations were 1.366 (95% CI = 1.177, 1.586) times greater
odds of being a Drinker (see Table 35, p. 179). Subjects whose expectations were one
standard deviation (7.89) higher were 11.73 times more likely to be a drinker.

The Interaction Model also was significant overall (38(5) = 53.96, p < .001, R2c°x
& 5.," = .24, from = .55) and had a successful prediction rate of 95.5%; however, it
did not show a block improvement over the Additive Model (x20) = .68, p = .410). The
interaction term did not significantly change the odds of being a Drinker versus Abstainer
and the social lubricant variable maintained the same level of influence (OR = 1.361,
95% CI = 1.174, 1.577, see Table 35, p. 179)

Females. The Additive Model with social lubrication expectations was significant

as an overall model (36(4) = 66.75, p < .001, R2Cox ,, s..." = .24, mm“, = .44) and had a

177

significant block improvement over the Basic Model (380) = 59.84, p < .001). The
model had a successful prediction rate of 89.4%, an increase of nearly 3% over the Basic
Model. Expectations of alcohol serving as a social lubricant changed the odds of being a
drinker by 1.387 (95% CI = 1.227, 1.567) for every one unit change in perceptions (see
Table 35, p. 179).1By altering the unit of change to be equivalent to one standard
deviation for expectations (7.74) students were 12.58 times more likely to be a Drinker
versus an Abstainer Additionally, the presence of social lubrication expectations
eliminates the significance of differentiation found in the Basic Model.

The Interaction Model also was significant overall (386) = 68.81, p < .001, R29»,
& s..." = .24, 1?sz = .45) and had a successful prediction rate of 90.7%; however, it
did not show a block improvement over the Additive Model (x2(1)= .43, p = .511). The
interaction term was not significant and the significant odds ratio for social lubrication
became 1.418 (95% CI = 1.239, 1.623) for a one unit change in expectations (see Table
35, p. 179).
The Comprehensive Model

Males. The Comprehensive Model was significant according to the model chi-
square (1203) = 66.18, p < .001, RICO... 3,... = .28, RIM“, = .66) and showed a
significant improvement over the Basic Model (x2(10)= 48.18,p < .001). The model
illustrated a 96% success rate in predicting the correct drinking classifications, an
improvement of 4% over the Basic Model. Interestingly, taken as a whole the model was
significant; however, no individual predictors in this model significantly changed the

odds of being a drinker (see Table 36, p. 181).

178

 

Haw—n um

rommmao Womaommmoa vamamoaam mammam 9m: 01:.» l men—688325 om momma. rewind—mo:

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

_l 3&8 _ man—5&8
_ a _ mm _ 9...: _ ow _ 89.2 _ a :m _ fie: _ ow _ 39.9
woman Zeno.
Umzw -_ .3 9mm _931 9 _ co 93m 9mg -93 93 PB a 9m _ o 92% 930
W7: -93 93 98 90% 9mg :3 -93 9g and 98a 93a 92:
>mo 9M: 9“: 98...... wk? 9an Po: 9% 9 3 9mm _.omm 9.3m _.Amo
>335 zona—
UwEw -93 9mm 93 9mg 9N8 _.owm 93 93 _.ow 93. 998 No?
95 98 9cm 9 _ m 995 93o :8 -98 93 9B 928 98A _.oow
>9». 9% 9mm moo... thw :3 Memo 93 9 _ m “a: _.Nfi 93w 9qu
mean: 931820: 9B 98 3.9»: _.uam :3 _.mma 9mm. 9cm 3.3.... _.wmq _.qu _.mou
583830: Zona—
UmTW 92 fun 98 _.muo 95k 398 _.mo 93 Nam Agoo 9.2: .365
$5 98 9on 9_ m 989 93m :8 -93 92 93 923 90: _.3o
>mo 93 93 mama... NS— :mo mp3 93 9; :5 _.NS 93m _.quo
mega b.5183: 9w— 9om 360...... _.wE _._.E _.m: 93 93 ~32: _.fim _.Nwo 9an
Um?” X mega 951838 9: 95 9d :3. 93m _.3m 93 93 NS _._ma 902 _.AS

 

 

_._ me A bu. «Seaman. _._... N» A .2. 35-69%?

 

 

179

Females. The Comprehensive Model was significant according to the model chi-
square (11203) = 90.70, p < .001, chmm = .31, Raw...“ = .57) and showed a
significant improvement over the Basic Model (x2(10)= 83.79, p < .001). The model
illustrated a 91 .5%'success rate in predicting the correct drinking classifications, an
improvement of almost 5% over the Basic Model. Only the tension reduction variable
showed an individual level of significance in changing the odds of being a Drinker (OR =
1.237, 95% CI = 1.078, 1.419, see Table GAU 36, p. 181).

Moderate Drinkers versus Low Drinkers

Difi'erentiation and GA U

Males. The Basic Model did not reach a level of significance (£0) = 2.92, p =
.405, see Table 37, p. 182).

Females. The Basic Model failed to reach a level of significance (12(3) = 2.62, p
= .453, see Table 37, p. 182).
F our factors of the DSI-R with GA U

Males. As was the case with the Basic Model, the overall four-factor model failed
to reach a level of significance (12(6) = 4.94, p = 551, see Table 38, p. 184).

Females. As was the case in the Basic Model, the four factor model failed to
reach a level of significance as an overall model (78(6) = 2.97, p = .813, see Table 38, p.
l 84).
Dmerenfiafion & perceptions of the average MS U student ’s alcohol use.

Males. The Additive Model with perceptions of the average MSU student’s
alcohol consumption was not significant (78(4) = 8.655, p = .070, see Table 39, p. 184).

The Interaction Model also failed to reach a level of significance (38(5) = 9.94, p = .077).

180

 

H33 ”3

roman." womnommmoa vaamnnmsm manoaa 9m: Claw 1 OoBuansmmé Zona—

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Zaom mesa—am
3 mm 935 O” 3.x. Q 3 mm «Sea 0% 8.x. 0:
Um?” -93 9va _.co 93m 9 _ 3 :38 _._u _._ _ _ .3 u. _ a 933 N930
ES: 9 a 93 N3 :3 93m _.Aco -93 93 93 903 oboa _.omm
>mo 93 93 9 _ q _.o3 93mm _.oNN 9 _ N 9: 93 :3 933 :63
Km: $083308 o.oo 93 93 9000 903 _.SN 98 9E 93 98m 93m _.o:
0.2% X 25:: $28283 -93 9o— wbw 93a 98o roam 98 92 93 _.oow 93.). _.SN
mom. mini 98 9c: :93: roma _.SA _.oum 93 90: NS _.3 w 963 _.oS
Ume X mam: 339$ -98 92 93 98m 93m _.oa 9c: 93 9G _.o3 903 2:3
00> ._ . _ a 93 who 9w : u 93: _.3a 93 9% 9mm _.ucu 9.39 A.Aoa
Umrw X 00> _.3 _._m #3 9.33 933 3.58 93 90A 98 _._: 95m NAB
H258: Wane—nae: 93 9cm 9% _.o3 98m :3 9: 93 93...... _.N3 _.3m _.Sc
Um?” X #958: £89030: 9: 9: 90¢ _._ _m 93: f3» 93 95 9o: _.o3 9qu _.NNN
moon: reggae: -93 93 93 92m 93m _.omm 9 _ w 93 93 _._AN 923 _.wwa
Um?” X moni— rccaomaoa -93 93 9mm 93. 93a _.NOu 93 93. 9.8 :3 93¢ _.wf

 

 

 

 

 

 

 

 

 

 

 

 

 

 

_._ h A .3. 23-823. .1. E A .3. 35-823.

 

 

l8l

 

.735 3

rommmmn ”$835: 33533 952.8 9% m3 3053.8 585 hos Una—83 I Diana—Emacs

 

 

 

 

 

7.5.8 mama—am
5 mm 9&5 OX 8:. Q 5 mm 9&5 OW 3.x. 0—
0m?” -9 _ a 9X 98 9mm~ 9&3 _.58 -98 9mm 9mm 9.33 93o _.Aou
E5 -98 98 N8 908 98c _.8o -98 98 95 93m 908 :5;
>ma -98 9 _ m 93 90mm 938 _.88 9- 9 _ m N. _ 3 .9: 98: ~58

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Co A .8. 35-8555. _._... h A .2. 3.0-855.

 

 

 

.365 um

rommmao ”£835: 5.35535 9525.. 9m” 86 2559.98 338 has Una—83 I mos. $888

 

 

 

 

 

 

 

 

 

 

 

755m moaaom

5 mm 9&5 OX 8.x. 9 5 mm 9&5 OW 8.x. 9
magma—5. ”83758. -98 98— 9; 93.3 .39 .58 -98 9N5 9S 93m 98m _.53
J: Page: 98 98— 93 _.u _o 9.33 958 98 93 98 room 9mm~ ~.3o
magmas. 028mm -93 95 _.3o 938 93...” Sec -93 9N9 93 985 9me _.fiS
F555 2:: 0924. 9; 93 93 :3 938 N35 -93 98m .33 98m 953. .58
Eé AGE 98 N. _ o 908 935 _.omc -98 98 .80 98A 908 _.3N
>mo 38m _.mo 98 908 938 _.u 3 9- 9 _ m N. _ om _.Nau 98w _.amm

 

 

 

 

 

 

 

 

 

 

 

c. A .8. 25-5.3. _._. a A .8. 3928.

 

 

182

Females. The Additive Model with perceptions of MSU students’ alcohol
consumption failed to reach a level of significance (12(4) = 3.77, p = .43 8, see Table 39,
p. 184). The Interaction Model also failed to reach a level of significance (36(5) = 4.63, p
= .463, see Table 39, p. 184).

Diflerentiation & Perception of Best Friend at MSU 's Alcohol Use

Males. The Additive Model with perceptions of their best friend at MSU’s alcohol
consumption was significant (12(4) = 19.894 p = .001, R2“, & 5,... = .15, RIM“, = .20)
with a significant block improvement over the Basic Model (12(1) = 16.98, p < .001). The
model had successful prediction rate of 68.8%, an increase of greater than 18% over the
Basic Model. Perception of best fi'iend alcohol use was a significant individual predictor
in that the odds of being a Moderate Drinker were 1.024 (95% CI = 1.011, 1.037) times
greater (see Table 40, p. 186) for every one unit change in perception and 5.69 times
greater for a change of 1 standard deviation (72.43).

The Interaction Model also was significant overall (x2(5) = 21.2l, p < .001, R20”,
& 5...," = .16, RZNW = .21) and had a successful prediction rate of 68.0%, however, it
did not show a block improvement over the Additive Model (380) = 1.32, p = .251).
Perception of best friend alcohol use was a significant individual predictor in that the
odds of being a Moderate Drinker were 1.025 (95% CI = 1.011, 1.039, see Table 40, p.

186).

183

 

Hugo we

rommmao anaommmoa 588:8 mange 32 m8 Zone-38 <02; r02 013.83 I $908325 o»- ch wagon:

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Cmo
f 3&8 _ moan—om
_ a _ mm _ sea _ ow _ ease _ a _ mm 2 so: ow _ ease
2520381
Umzw -9 3 9?» 98 938 9&8 _.afi. -98 98 98 9.38 980 2&8
wz: -98 98 8.8 98m 98o 28¢ -98 9?. 98¢ 93m o.ooo _.oma.
>mo -98 93 98 98m 9.5m _.qu 98 9G 8.: _.83 98— _.mmw
>aa5<n Zona—
Umrw -93 9mm 93 928 98a 2.3m. -98 98 98 93a 9A8 _.A-NN
wz: -98 98 :3 98a 98o 28¢ -98 98 9 3 908 908 2.80
>mo -98 9 _ a 98 903 9a _ w _.wi 92 9 2 m 2.0m _.8m 98c _.80
Ewe Cmo 98 92 P8... 2.2m .28 _.8o 92 92 2.8 _.ooa 93m _.2m
58328: Zona—
DMEN -9 _ a 93 9 _ c 93m 98; 2.8a. -98 98 9.3 93o 98A :28
get: -98 98 _.3 925 988 2.80 -98 98 9 I 923 9ch _.8N
>ma -92 9 a o.oo 908 98. 2.3» 98 9_ m who _.NAm 908 _.od
ch CR 92 92 +3... _.25 _.o2 _.8w 92 92 _.Nm _.ooa 98m _.2m
02-x x ch 92 92 _.Nm _.oE 93c _.80 -92 92 93 908 93a _.ooc
Cma

 

 

 

 

 

 

 

 

 

 

 

 

 

 

_._ E A bus "So-8:3.

I. w A .2. sec-£59

 

 

Females. The Additive Model with perceptions of their best friend at MSU’s
alcohol consumption was significant (x2(4) = 19.29, p = .001, RZCM a 5...,“ = .13,
[Pram-1km = .17) with a significant block improvement over the Basic Model (12(1) =
16.67, p < .001) and a successful prediction rate of 68.3%, an increase of 6.3% over the
Basic model. The change in odds for being a Moderate Drinker versus a Low Drinker
was 1.025 (95% CI = 1.010, 1.040) for every 1 unit change in perception of best friend
drinking, and 2.71 times for an increase equivalent to one standard deviation (41.48) (see
Table 40, p. 186).

' The Interaction Model was also significant overall (38(5) = 19.55, p = .002, 18%,,
& 3m" = .13, [gum = .18) and had a successful prediction rate of 68.3%; however, it
did not show a block improvement over the Additive Model (x20) = .26, p = .611). The
only significant individual predictor was perception of best fi'iend use, which remained
identical to the value of the Additive Model (see Table 40, p. 186).
Dyfirentiation & C 0A Status

Males. The Additive Model with COA status did not reach a level of significance
(78(4) = 2.92, p = .571, see Table 41, p. 187). The Interaction Model also failed to reach
a level of significance 06(5) = 2.94, p = .709).

Females. The Additive Model with COA status failed to reach a level of
significance 08(4) = 2.88, p = .578, see Table 41, p. 187). The Interaction Model also

failed to reach a level of significance 086) = 3.218, p = .666).

185

 

.320 .5

rommmzo womaammmoa 082038 250:8 Em" man 38038 588 P03 Can—83 1 0280225 ow wow” magma Cmo

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

_ 330m _ moan—ow
b a 2 mm _ an: _ ow _ 8.x. 9 _ a _ mm _ fin: _ ox _ 3.x. 9
mean Kong
02-” -9 3 98A 98 9w8 9&8 r98 -98 98 98 908 98c _.A8
ES— -98 90¢ 8.8 98m 980 28¢ -98 98 98a 908 908 _.oaA
>mo -98 93 98 90mm 908 _.woM 98 93 8.3 2.82 902 _.amu
>a&:<n Zona—
Umm-w -9; 9mm 9; 98m 98m _.08 -98 9mm 92 908 9800 _.80
wz: .98 98 8.8 908 985 _.8A -98 98 98 903 98— _.omm
>8 -9 8 9 2 0 93 98A 980 _.88 93 9 _ a T3 _.Noa 980 _.8a
mom” manna; Cmo 98 92 8.88.. _.8A 2b: _.80 98 92 5.8.... _.8m 22 o _.vo
383030: 30%;
62.” 98 93 9.: TBA 93m ALE -98 93 9E 908 98o _.mm0
22 -98 98 2.3 98o 928 :08 -98 93. 93 90mm 9me :50
>8 -98 9; 93 980 982 _.88 9; 93 83 :8 988 _.80
wow» man—ad Cmo 98 92 8.8.... 2.98 _.o: 2.80 98 92 0.8: _.8m _.oo0 no.5
UmH-W x wow» 3.225% Cmo 92 92 _.N-N 2.2a. 900o _.ouw -92 92 98m 908 908 _.8o
... h A .8. "So-$259 3.8 A .2. "So-8:09

 

 

186

 

0.033 t

rommmmo womnommmos 3003005 mange 0.5” Ed 7003088 338 has Una—88 I 00> mgcm

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

_ 3&8 _ 02:38
r 0 _ mm _ €33 _ 2w _ 03$ 2 _ 3 _ mm _ng _ 0% F 03.x. 2
wmmmo Kama—
Umzw -0. 3 0.80 0.8 0.08 9.38 _.38 -98 0.8 0.8 0808 .80 :38
ES. -98 0.03 8.8 0.083 0.80 _.80 -98 0.00 983 0.08 .000 203A
>ma -98 0.3 0.0-\ 0.030 088 _.803 0.8 0.3 8.: _.Nfi .0”: _.38
>035<o Zona—
Umzw -9 3 98¢ 98 9038 9&8 2.353 -98 0.8 0.33 9.33 0.000 _.033
22 -98 0.03 _.00 9083 0.80 _.82 -98 900 0.8 0.03 0.000 _.038
>mo -900 9 2 3 98 0.030 088 _.803 0.8 9 _ 3 8.8 _.83 980 2.33—
OO> -900 0.3. 92 0.033 938 8.303 98 9.3 983 _.83_ 0.83 8.30
258—Samoa 3002
szw -93 0.8 0.8 908 933 2.80 -98 0.8 0.03 .300 0.80 0.93
wz: -98 0.03 2 .03 0.083 0.80 280 -98 90¢ 900 .08 0.00» 2 .030
>mo -98 0.3 0.8 0.030 088 _.803 0.8 0.3 8.”: _.N3_ 908 2.33
00> -98 932 900 0.000 988 N38 93 0.03 900 _.3_ 0.33 8.08
Um?” x 00> -9: 0.2 0.8 9000 9:8 0.08 0.3 0.8 0.8 0.300 0.88 0.033

 

 

 

 

 

 

 

 

 

 

 

 

 

 

*8 A .03v 25.880. 1.0 A .2. "so-8:03.

 

 

187

Difizrentiation & Expectations of Alcohol Serving as a Tension Reducer

Males. The Additive Model with expectations of alcohol as a tension reducer was
significant (78(4) = 16.17, p = .003, RZCM 5,... = .12, mm = .16) with a significant
block improvement over the Basic Model (38(1) = 13.26, p < .001), and had successful
prediction rate of 69.6%, an increase 19.2% over the Basic Mode]. The tension reduction
variable was significant indicating subjects were 1.101 (95% CI = 1.042, 1.163) times
more likely of being a moderate drinker versus a low drinker with a 1 unit increase in
tension reduction expectations (see Table 42, p. 189). Subjects whose expectations
changed by one standard deviation (8.73) were 2.31 times more likely to be drinkers.

The Interaction Model was also significant overall (78(5) = 16.24, p = .006, 12%,,
& 5.,“ = .12, RZNM = .16) and had a successful prediction rate of 70.4%; however, it
did not show a block improvement over the Additive Model (x20) = .07 , p = .798). The
significant tension reduction odds ration was identical to the Additive Model (see Table
42, p. 189).

Females. The Additive Model with expectations of alcohol as a tension reducer
was significant (36(4) = 24.47, p < .001, R2“, ,1 s..." = .16, hm, = .22) with a
significant block improvement over the Basic Model (36(1) = 21 .84, p < .001), and a
successful prediction rate of 66.9%, an increase of 5% over the Basic model. Age (OR=
1.386, 95% CI = 1.007, 1.90) and Tension reduction (OR = 1.123, 95% C1 = 1.064, 1.184
for a 1 unit change, and OR = 2.78 for a change of one standard deviation of 8.81 units,

see Table 42, p. 189).

 

.320 AM

rommmmo Wow—dame: vanamoasm 953$ 92 m3 30%-.88 588 float 013.83 I wax—808205 0». H253: flamenco:

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

T 3&8 mania
_ o _ mm _ 55 _ ox _ 8&9 s m a: 2% _ ow. _ 3:6.
momma Zone.
Ume -9 3 9.3 98 93m 9th. _.afi -98 93 93 933 93o _.Aou
W75 -93 93 Now 993 93c _.owo -98 93 93 93m 925 93¢
>mn -93 93 93 93w 93w :3 9mm 93 N: _.Nfi 9o”: _.3u
>338 Zona—
Ume 92 93 93 _.N3 920 Name 93 93 9B :3 93-3 ~33
wz: -93 93 93 93; 93c _.o-Z -93 93 o.oo 903 933 2:3
>mo -93 9 m q 93 9on 933 awe 93 9 _ a Pom... :3 foo-V _.oom
Adamo: W885: 95 93 Susan :3 _.ofi :3 9E 93 5.3:. :3 r3» :3
388030: 30%:
Ume 98 9mm 93 _.Nwo 933 ~68 93 93 9: :3 93w 9wa
W7: -93 93 93 93c 93c _.oqo -93 93 93 923 93m roam
>mo -93 9 _ a 93 930 9me _.ufio 9mm 9 H m Pow... _.umu “boa _.cou
4253: W883: 98 93 2.: a... :2 _.oAN :3 9: 93 3.8.3.. :8 r3: :3
02% x Anzac: x8830: -93 93 93. 903 93c #23 -98 93 9N5 93a 93— _.o:

 

 

 

 

 

 

 

 

 

 

 

 

 

 

am A .3. ”fie-gnu. _._... No A .3. 25-839

 

 

The Interaction Model also was significant overall (36(5) = 24.7l,p < .001, RZCO,
& 5....- = .16, 32,-W.,, = .22) but failed to yield a significant block improvement over the
Additive Model (x20) = .24, p = .622), maintaining the identical prediction rate of the
Additive Model 66.9%. Age (OR= 1.383, 95% CI = 1.006, 1.903) and tension reduction
(OR = 1.121, 95% CI = 1.061, 1.183 for al unit change) remained nearly identical in
their respective values of categorical prediction between Moderate versus Low Drinkers
(see Table 42, p. 189).
Diflerentiation & Expectations of Alcohol Serving as a Social Lubricant

Males. The Additive Model with expectations of alcohol as a social lubricant
approached, but did not obtain, significance (12(4) = 9.37, p = .052, see Table 43, p.
193). Additionally, the Interaction Model also was not significant (x2(5) = 9.386, p =
.095).

Females. The Additive Model with expectations of alcohol as a social lubricant
was significant overall (38(4) = 14.01, p = .007, R’CM 5....- = .10, RZNW, = .13) with a
significant block improvement over the Basic Model (38(1) = 11.38, p = .001), and a
successful prediction rate of 61 .2%, which was .7% less than the Basic Model. The social
lubrication variable was a significant individual predictor, indicating subjects had a 1.094
(95% CI = 1.036, 1.155) times greater chance of being classified as a Moderate Drinker
versus Low Drinker with a 1 unit increase in social lubrication (see Table 43, p. 193).
Subjects whose expectations change by one standard deviation (7 .74) were 2.01 times

more likely to be Moderate Drinkers.

190

The Interaction Model was also significant (x2(5) = 14.18, p = .015, RICO“ SM“ =
.10, 1?sz = .13) and had a successful prediction rate of 61 .2%. The model did not
have significant block improvement over the Additive Model (380) = .171 , p = .679).
Social lubrication remained the only significant predictor and the values remained
equivalent to the Additive Model (OR = 1.095, 95% CI = 1.037, 1.157, see Table 43, p.
193).
The Comprehensive Model

Males. The Comprehensive Model was significant according to the model chi-
square (38(13) = 34.99, p = .001, RIC,” s..." = .24, RIM“, = .33) and showed a
significant improvement over the Basic Model (3800) = 32.07, p < .001). The model
illustrated a 73.6% success rate in predicting the correct drinking classifications, an
improvement of 13.2% over the Basic Model. Perceptions of best friend use (OR =
1.020) and expectations of tension reduction (OR = 1.109) were the only individual
variables to significantly improve the odds of being classified as a Moderate versus a
Low Drinker (see Table 44, p. 194).

Females. The Comprehensive Model was significant according to the model chi-
square (x203) = 45.72, p < .001, RICO, & s...“ = .28, mm = .38) and showed a
significant improvement over the Basic Model (7800) = 43.1, p < .001). The model
illustrated a 75.5% success rate in predicting the correct drinking classifications, an
improvement of 13.6% over the Basic Model. Age (OR = 1.461, 95% CI = 1.014, 2.107),
perceptions of best friend use (OR = 1.028, 95% CI = 1.009, 1.047), and expectations of

tension reduction (OR = 1.118, 95% CI = 1.041, 1.201) were all significant individual

19]

predictors in the likelihood of being classified as a Moderate versus Low Drinker (see
Table 44, p. 194).
High Drinkers versus Low Drinkers

Differentiation and GA U

Males. The Basic Model did not reach a level of significance according to the
model chi-square (38(3) = 2.208, p = .530, see Table 45, p. 195).

Females. The Basic Model failed to reach a level of significance (38(3) = 2.057, p
= .561, see Table 45, p. 195).
F our factors of the DSI-R with GA U

Males. As was the case in the basic model, the overall four-factor model failed to
reach a level of significance (x2(6) = 10.841, p = .093, see Table 46, p. 195).

Females. The four factor model failed to reach a level of significance as an
overall model (36(6) = 3.405, p = .757, see Table 46, p. 195).
Dlflerentiation & perceptions of the average MSU student ’s alcohol use

Males. The Additive Model with perceptions of the average MSU student’s
alcohol consumption was significant as an overall model for predicting High versus Low
Drinkers (38(4) = 24.11, p < .001, 1&0, ,, 3,... = .18, 12sz == .24). The Additive
Model had a successful prediction rate of 65%, a 9.7% improvement over the Basic
Model. Only perceptions of MSU student use was a significant individual predictor in the
Additive Model indicating that for every 1 unit increase in perceptions they were 1.032
(95% CI = 1.016, 1.049) times more likely to be classified as a High Drinker (see Table
47, p. 197). With a one standard deviation change (37.22), the likelihood for being a

Moderate Drinker was 3.29 times greater.

192

 

.320 3

rommmmo Wamammmo: 3.383% 9:938 52 m8 gonna?” 53% was 01288 I mxnoogmoam o». magm— 551330:

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Zion _ mango—am
e _ mm _ fie: ow _ 8:9 _ e _ mm _ éea _ ox _ 8&9
wmmmo Zona—
Umrw -o. _ o out. 93 Paw chew _.fiw -obu Pug. ohm 9ch 93¢ _.Aow
W7: .98 o.oo Now 908 93¢ _.owc .98 PE oho 0.3m o.ooo _.oi
>mo .9?“ o. _ a 93 Paw Paw _.wom ohm o. _ m N. _ .V _.N.: 0.8. raw
>aa5<a Zona—
szw 9: 9: Pee : 5 9mg ~.Nmu PI 93 PG :3 93¢ Nb _ q
95 .98 o.oa _.oq 0.03 chew _.cmA o.oo 98 2: 98¢ 92 _ _.ooo
>mo .92 93 93 993 92: _.wuc PS 93 Pow _.w8 o.oam ruqm
mesa r5185: o.ou o.ou «o.oo... _.omo 2:.“ :3 o.oo 98 3.3.1. _.ooA _.Sa :3
58830: Kong
62% 9.5 0.3 93 _._NA 003 Pwuo PS Pun ohm _.Noa on: NUS
9’4: .98 o.oo _.3 903 SWAN _.omm -93 93 PS Poem o.ooo rowm
>Mo -obp o. _ a 98 o.oo”. o.oo» _.wwo 93 o. m m Pow _.w _ o 903 _.SN
moan. 551035: 93 o.ow 93%.. _.ome mo: :8 o.oo 0.8 3.3.... _.oom rowq :3
Ume x moor: 551820: 0 3 93 o.om _.ooa o.owo _._oo PS 93 93 EoN— o.owo :NA

 

 

 

 

 

 

 

 

 

 

 

 

 

 

.. e A .8. gamed. the A .8. 398:8.

 

 

I93

 

32o t

homage womaommmos 32:03.8 meqoam :5” Rd 33088 33% mos Una—81o. l OoBvarasmmé Zona—

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Zm—nm moan—am
3 mm £35 3w 8.x. 3 @ mm («SE 3w 8.x. 3
03-x 98¢ 98 93 _.83 98.3 98¢ -98 9mm 98 980 983 Nmae
3": -98 98 9mg 98o 98w _.omm -98 98 9% 98¢ 98A _.omw
>ma -98 9 _ c 98 988 98m 3.88 98 9 3 c a. _ 8.. _.AS _.35 N. _8
78¢ $28388 93 93 PB 3.3u 93a #80 93 93 98 _.oco 908 _.3m
U3-” X ZmC 30283303 93 93 9~a _.ooq 93c _.80 -93 93 _.3 98m 90% _.3o
mam” man?— 98 93 93...... _.8c 3.8 _.8m 98 93 93... _._ _.owm _.ooo Tog
Um?” x mam" manna 93 93 98 _.ooa 903 _.8_ -93 98 9%. 98¢ 93o _.o3
00> -9 _ N 93 98 9me 980 u. 3 mm -98 98 98 90mm 983 Na 3 m
U3-” x 00> 980 3.8 95 _.Sm 93w $03 93 98 9AM _.mcu 935 5.03
Adamo: ”88:0: 93 98 93... :8 he: _.Noc 9: 98 98...... _._ _m 3?: _.N3
Um?” x H258: ”3:38: -98 98 93 98a 938 :8 -95 93 8.3 A53 9.3m _.3m
mega gang—no: -93 98 98 983 908 _.oma 98 98 98 _ .8 _ 98o :8
Um?” x moor: rains—mo: 98 o.oo 98 E83 938 _.Bm 9: 98 _.8 _._ 3 98m _.88

 

 

 

 

 

 

 

 

 

 

 

 

 

 

_._ E A .8. 35.823. _._... w A .3. "so-8:3.

 

 

194

 

Hugo 3

homage Womnommmoa wnoamoasm mzaoam 93 m8 Em: 83% roé 033.82 I baggage:

 

 

 

 

 

 

2E8 mafia—8
@ mm éma OW 8.x. Q v mm $35 0% 8.x. 0—
Umzw -98 93 _.mm 93m 98.; _.NoM -98 93 93 993 939 ..A8
W7: o.oo 98 93 foo» 98o _.oea -98 9i _.8 98¢ o.ooo _.oaA
>mo o.oo 9 3 98m _.oou 93A :30 9mm 9 3 m 93 926 98_ _.omu

 

 

 

 

 

 

 

 

 

 

 

 

 

_...u A .8, "So-Swag. {Nu A .3, "so-8:8.

 

 

 

Haw—a .3

roman.“ Womammmo: qumomsm magma 92 Em: 585.. r92 Una—88 .- «.2: 3883

 

 

 

 

 

 

 

 

 

 

3E8 maafiom

v mm «.33 OW 3.x. 3 v mm 9&5 OW 83 Q
maomosa ”$035 9% 98 9mm _.wao 98» PM: 93 9mm 98 3.8” 9.38 N28
.4: $850: -93 98 98...... 93» 9qu 9mg 93 98 98 #80 93“ _.m:
mac-585E 053% -98 9mm 93 93c 9&3 83 -98 93 93 93c 9m£ _.mmo
353: SE. 0908 -93 93 93 923 936 :8 .93 98 3.8 93.» 9on Home
wz: 98 98 9S 33¢ 930 _._ 5 -93 98 3.3 9?: chug _.oww
>mo 9cm 9 _ u 98 _.oqw 9m3 3.3. -93 9 _ A o.oo 93; 9d” 88

 

 

 

 

 

 

 

 

 

 

 

 

_._ h A .omv "So-3:8. _I w A .3, "So-8:3.

 

 

195

The Interaction Model also was significant as an overall model for predicting high
versus low drinkers (36(4) = 24.48, p < .001, RZCM 5,... = .13, mm“, = .24) and had a
65.9% successful prediction rate. However, the Interaction Model did not show a block
improvement over the Additive Model (38(2) = .37, p = .545). The only significant
individual predictor was the perception of MSU students use, which remained almost
identical to the OR from the Additive Model (1.032, 95% CI = 1.015, 1.049, see Table
47, p. 197).

Females. The Additive Model with perceptions of their best friend at MSU’s
alcohol consumption was significant (12(4) = 13.42, p = .009, 1620,x & 5...," = .09,
RZNW = .11) and a significant block improvement over the Basic Model (12(1) =
11.36, p = .001), and had successful prediction rate of 60.7%, an increase of greater than
8% over the Basic Model. Perception of MSU student use was a significant individual
predictor in that the likelihood was 1.015 (95% CI = 1.006, 1.025) times greater of being
a High versus Low Drinker for every one unit change in perception (see Table 47, p. 197)
and 1.79 times greater for every one standard deviation (38.75).

The Interaction Model also was significant overall (12(5) = 18.01, p = .003, chOx
& 5m" = .11, 1(2an = .15), had a successful prediction rate of 62.7%, and had a
significant block improvement over the Additive Model (12(1) = 4.59, p = .032).
Perception of MSU student use remained a significant predictor of High versus Low
Drinkers (OR = 1.015, 95% CI = 1.006, 1.025, see Table 47, p. 197). Additionally, the

interaction of differentiation with perception of MSU student use was significant with the

196

 

Hugo e3

rommmzo Womamwmos 309035 manam 5m: Em: <oacm r02 Una—83 I voqoounosm om ch mamas. Cmo

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

3&8 mean—am
w mm Irma OW 8.x. 9 w mm $35 OW 3.x. 9
wwmmo Zona—
Umzw 93 93 5mm 93w 9:“ _.Nou .98 9mm 92 90.3 93o _.Aow
ES— 98 93 9S foo; 98o _.oeo -98 9i :3. 980 o.ooo _.cg
>mo o.oo 9 3 9wm _.oou 93A :39 9mm 9 _ m 99 930 98— _.amw
>338 Kong
09% -98 9mm 93 9m; 9.:— _.SA 93 9N0 o.oo _.SN 9%: r3;
9’: -98 98 93 903 9a.: _.od -93 93 _._w 993 93m fog
>mo 93 93 95 _.omq 93o _.Amw 9cm 9: 9S _.oB 93m _.wfl
zwc Can 98 93 3.3.1. _.owm _.Sa fog 98 92 .98... _.Sm _.oom _.omm
58393: gono—
Umrw -93 93 93 9.39 9qu _.mmm -98 98 9S 9an 9H3 _.qfi
EE -93 98 93 90am 93o _.oqm -93 98 9% o.ofl 9me _.owm
>ma 9g 9 fl m 9 H w _bS 9.3M _.Amo 9cm 9 3 9 E _.oma 939 _.wom
ch C8 98 93 5.3.... _.SN _.Sm _.oao 98 93 98: _.S o. _.ooo _.owm
Um?” x ch Gun -92 92 93 obs 903 _.So -98 93 .35... 903 903 o.ooo

 

 

 

g A .3. ”So-8:3. .2. h A .3. 35-8w

 

_na.

 

 

197

odds ratio of .983 (95% CI = .967, .999) for every unit increase in the interaction term
(see Figure 4, p. 62, for illustration of the interaction effect).
Difierentiation & Perception of Best Friend at MS U '5 Alcohol Use

Males. The Additive Model with perceptions of their best friend at MSU’s alcohol
consumption was significant overall (38(4) = 70.40, p < .001, R20”, & Sm" = .44, Rznme
= .58) with a significant block improvement over the Basic Model (x20) = 68.19, p <
.001), and had successful prediction rate of 84.6%, an increase 29.3% over the Basic
model. The only significant individual predictor in the model was perception of best
friend use, which illustrated a difference of 1 unit of perception made it 1.04 times more
likely the subject was a High versus a Low Drinker (see Table 48, p. 199). A difference
in perceptions by a value of one standard deviation (72.43) improved the odds of being
classified as a high drinker by 13.56 times.

The Interaction Model also was significant overall (38(5) = 72.77, p < .001, R20,,(
& 3...... = .45, [fink-fining“,e = .60) and had a successful prediction rate of 84.65%; however, it
did not show a block improvement over the Additive Model (x20) = 2.37, p = .124).
Again, the only significant individual variable was perceptions of best friend use which
remained similar to the Additive Model (OR = 1.041, 95% CI = 1.026, 1.056, see Table

48, p. 199).

198

 

flea—o Am

rommmno Wow—.038: 39:35 930:8 Em: Em: <29; r02 013.83 I hen—.8303 cm 88. 353% CR

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

_ 3&8 _ deE8
9 e _ mm _ as: _ ow _ 8.x. 9 _ s _ mm _ sea _ ow _ 8.x. 9
mean zona—
Ume -98 98 5mm 99..» 98.8 whom -98 98 93 993 93¢ #88
El: o.oo 98 9S foo.» 98o fog -98 98 _.8 980 98c _.omA
>mo 93 9: 98 room 93A _.fio 98 93 92 996 98; _.88
>853 Zena
Um?” -98 93 93 98o 9.33 Mb; 9; 98m 98 :8 9%: Pmmm
BE -98 93 93 903 980 _.oow -98 98 9N9 930 98w. _.o3
>mo o.oo 9E o.oo o.ooo 93m _.mo_ -93 9 _ m 93 98c 98m _.88
mom" 333% can 93 93 89.3: #83 _.oNA 28¢ 98 93 No.51. _.oka _.omc _.o8
58330: Zona—
Umrw 9mm 93 9B EMA”... 953 w. _ m; -9: 93 93 98m 98o _.mao
W72 -98 93 98; 908 980 :3 -98 98 98 93c 98o room
>mo 98 9B 98 _.8m 9me _.mmu -93 9 _ o 98 903 933 _.wmm
wow" man-baa :8 98 93 892 _._... foe: _.owa bag 98 93 8.8.1. _.OAo _.8_ foam
UmH-W X wnamnosaa 96 98 93 N8 #33 98m #95 -98 93 8.3 903 908 _.o8
_. w A .8. 30.823. _._... h A .2. 25.3209

 

 

 

Females. The Additive Model with perceptions of their best friend at MSU’s
alcohol consumption was significant overall (352(4) = 55.99, p < .001, RIC“ & 5m" = .31,
1?sz = .42) with a significant block improvement over the Basic Model (780) =
53.93, p < .001), and had successful prediction rate of 80.7%, an increase 28% over the
Basic model. Perceptions of best friend’s alcohol use significantly changed the odds of
being a High versus Low Drinker by 1.046 (95% CI = 1.029, 1.063) times for a one unit
change in perception (see Table 48, p. 199) and 6.47 times greater for a change
equivalent to one standard deviation change (41.48).

The Interaction Model also was significant overall (38(5) = 59.09, p < .001, RZCO,‘
& gm" = .33, Rznmnmke = .43) and had a successful prediction rate of 79.3%, but did not
show a block improvement over the Additive Model (352(1) = 3.11, p = .078). The
perception of best friend use remained the only significant individual predictor with the
values remaining nearly identical (OR = 1.049, 95% CI = 1.031, 1.068, see Table 48, p.
199).

Differentiation & C 0A Status

Males. The Additive Model with COA status failed to reach a level of
significance (12(4) = 3.20, p = .525, see Table 49, p. 201). The Interaction Model also
failed to reach significance (38(5) = 3.29, p = .655).

Females. The Additive Model with COA status failed to reach a level of
significance (78(4) = 2.09, p = .719, see Table 49, p. 201). The Interaction Model also

was significant overall (38(5) = 4.959, p = .421).

200

 

.930 he

rommmmo ”3338: 32:33 mama-ha 9% 20 En: 535 hog 01383 I 00> wSEm

 

_ 3&8 h moaaom
_s _mm _ fiaa _ on: 839 _ a _mmnéa: 9: 8:9

 

momma Zona—

 

UmH-W -9Aw 9.3 :3 9?; 9w? _.Nom .98 93 93 993 93o _.Acw

 

wz: o.oo 93 93 focus 930 _.ccm -9o~ 93 _.mu 980 o.ooo _.oaA

 

 

 

 

 

>mo o.oo 9: 9mm _.oow 93A :39 98- 9G 93 933 903 _.mmu

 

>355" 383

 

Uwzw -93. 93 _.3 93m 93— _.NE -93 9Nm 90A 993 9m: 3.3x

 

E’s: 93 93 98 _.oom 908 _.ccc -9om 9cm _.3 9vo 93V memo

 

>mo 9cm 9: 93 .93 9x: _.Awa -93 93 93 98» 9.3m when

 

 

 

 

 

 

 

 

 

00> -9? 9A0 93 930 9N5 _.moo 93 93 9E _.omw 93o Nara

 

 

58330: Zeal

 

UwH-W -9“; 9.3 _.qm 93o 9mg rumm .93 9mm 9mm 9.3M 93m _.umN

 

W73 93 93 98 _.ooo 98M _.oco -98 93 _.mm 98m 93m _.owm

 

>mn 93 93 93 _.oww 93w TEN -93 9: 93 98c ohmu _.Noq

 

00> .9A0 93 _.ow 93w 9N3 rum: o.oo 93 98 HbOu 93— ~08

 

UMH-WX 00> 98. 9a o.oo _.Noo 9~mw whoa f; 9.3 Now PM: 93m Sham

 

 

 

 

 

 

 

 

 

 

 

 

 

 

a. so A .3. 25-8209 3.. h A .3. Eco-EFF

 

 

201

Difierentiation & Expectations of A lcohol Serving as a Tension Reducer

Males. The Additive Model with expectations of alcohol as a tension reducer was
significant overall (12(4) = 27.64, p < .001, RICO, & 3....- = .20, RzNwm -—- .27) with a
significant block improvement over the Basic Model (x20) = 25.43, p < .001), and had
successful prediction rate of 67.5%, an increase 12.2% over the Basic model. The tension
reduction variable was significant for subject’s in that they were 1.137 (95% CI = 1.075,
1.203) times more likely to be a High versus a Low Drinker with a 1 unit difference in
tension reduction expectations (see Table 50, p. 203). Subjects with one standard
deviation (8.73) higher expectations are 3.08 times more likely to be High Drinkers.

The Interaction Model also was significant overall (36(5) = 27.64, p < .001, RZCO,
& 5m" = .20, Rznmm, = .27), had a successful prediction rate of 67.5%, but did not
significantly show a block improvement over the Additive Model (12(1) = .00, p = .993).
The only significant individual variable in the model, expectations of tension reduction,
maintained the same values as the Additive Model (OR = 1.137, 95% CI = 1.075, 1.203,
see Table 50, p. 203).

Females. The Additive Model with expectations of alcohol as a tension reducer
was significant overall (78(4) = 35.99, p < .001, we,” 5...“ = .21, R’Nm-m = .28) with a
significant block improvement over the Basic Model (780) = 33.93, p < .001). The
model had a successful prediction rate of 69.3%, an increase 16.6% over the Basic model.
The tension reduction variable was a significant individual predictor for the increased
odds of being classified as a High versus Low Drinker (OR = 1.146, 95% CI = 1.087,
1.209) for a one unit change in tension reduction (see Table 50, p. 203) and (OR = 3.31)

for a change equivalent to one standard deviation (8.81).

202

 

Hugo mo

rommmno Womwommmo: 32:38 930:8 92 Em: 325 has Gianna I mxvoosmoam om H255: Woacnaos

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

_ 3&8 _ moan—om
_ a _ mm _ as: _ ow # 8.x. 9 _ a 4 mm _ fies _ ow _ 3.x. 9
wmmmo zona—
Uwrw -98 98 #8 98¢ 98m ~88 -98 98 98 90.3 980 #38
wz: o.oo 98 9S foo; 98c #08 -98 98 _.8 980 o.ooo _.oan
>mn 98 93 98 _.o8 98A :13 98 98 92 926 993 #88
>335 381
Um?” -98 93 98 998 930 No8 98 98 _._o _.80 9.3a 9qu
95 98 98 98. _.8A 988 :8 -98 98 93 98c 98o _.oum
>mo 93 93 98 :8 98¢ H6: 98 93 98 r8_ 98c 88
Adamo: ”8:38 98 98 898...... :3 _.o-G ~88 93 98 8.8.3.. :3 _.ow-N _.Noc
58830: Keno—
Umzw -98 93 o.oo 98A 9A8 8.8m 98 98 :8 _.Sq 9.3a ~un
95 98 98 9B 58; 988 _.88 -98 98 93 98o 98a roux
>mo 93 95 98 :8 980 _.m: 98 93 98 _.88 9§o ~88
H088: Waggon 98 98 3.3.... :3 _.o-G _.88 9E 98 8.8.... :3 _.omu r88
Um?” X H958: Waggon o.oo 98 98 race 90: roe-N 98 98 92 903 903 _.8A
_._ h A .8. "so-8:09 sale A .o . go-§_oa.

 

 

 

203

The Interaction Model also was significant overall (£6) = 36.00, p < .001, R20“
& 5...," = .21, [CZ-imam = .28) and had a successful prediction rate of 70%, but did not
show a significant block improvement over the Additive Model (12(1) = .01, p = .93 8).
The tension reduction variable remained the only a significant individual predictor for the
increased odds of being classified as a High versus Low Drinker (OR = 1.146, 95% CI =
1.087, 1.209, see Table 50, p. 203).
Ditfirentiation & Expectations of Alcohol Serving as a Social Lubricant

Males. The Additive Model with expectations of alcohol as a social lubricant was
significant (38(4) = 15.83, p = .003, RZCOX-g 5...," = .12, Rzuw = .16) with a significant
block improvement over the Basic Model (x2(1)= 13.63, p < .001) and had a successful
prediction rate of 67 5%, an increase 12.2% over the Basic model. The social lubrication
variable was the only significant individual variable that changed the odds of drinking
classification. Having a social lubrication value change by 1 unit increased the likelihood
ofbeing a High Drinker by 1.114 (95% CI = 1.047, 1.184) times (see Table 51, p. 205).
Additionally, a difference in social lubrication scores equivalent to one standard deviation
(7.89) improved the odds of being a high drinker by 2.35.

The Interaction Model also was significant overall (12(5) = 15.84, p = .007, RICO,-
& Sm" = .12, Rznm = .16) and had a successful prediction rate of 66.7%; however, it
did not show a block improvement over the Additive Model (x20) = .01, p = .923). The
social lubricant variable maintained the same level of influence 1.114 (95% CI = 1.047,

1.185) as the Additive Model (see Table 51, p. 205).

204

 

.320 m.

rommmno ”3333: 32:35 $836 9:» Em: «688 res: 0138?. I 9685303 om meow”: 95183::

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

7 73:8 _ mafia—om
_ 8 _mm _ 9:»: _ o: _ 8:9 _ :c _ mm_ 9:»: _ o: _ 39.9
awmmo Kean—
:m; -93 93 _.3 9%: 9:: 9:8 -98 9:: 9% 9:3 93: :8
:7: 98 98 9S :9: 9:8 988 -98 9i 2: 98: 9:8 5%
>9” 9% 9:. 9% S8 93.. 9.3: 9:: 9a 92 923 98: 93:
>maa<o Kong
:2: -93 9a 98 98: 9.3: 2:: 9.: 9w: 9% $8 98: 2.:
:7: 98 98 9a 2:: 98: 9:: -93 98 3.: 9:: 98: :5
> o 9; 93 SN :3 9:: 98m 92 9a 98 :5 9:; IE
moaaEEsaS 9: 98 :3: I: :3: 9:: 9: 98 3:... :8 SS _._:
583030: Zeno—
:m; -98 9.3 98 9B: 93” 28 9.: 9:: Z: 38 9:9. 5%
:3: 98 98 9:. 2:: 9S: 9:: -9o: 98 5o 98: 98. 2.:
2a 9; 9a _.3 :3 92: 9%: 98 9: 98 993 9:: 98:
mos». €388: 9: 98 :8”: Z: 2:: :3 9: 98 3.2: 9:: 22 9::
SEQ mega—.3188: 92 98 92 28 9:8 :8 98 99: 98 98: 9:; 993

 

 

 

 

 

 

 

 

 

 

 

 

 

 

_._-u A .8. arc-S89 3...: A .2. «So-8:69

 

 

205

Females. The Additive Model with expectations of alcohol as a social lubricant
was significant overall (36(4) = 22.34, p < .001, RICO, & 5,... = .14, RIM“, = .19) with a
significant block improvement over the Basic Model (380) = 20.29, p < .001). The
model had a successful prediction rate of 67.3%, an increase 14.6% over the Basic model.
Only social lubrication was a significant individual predictor for High versus Low
Drinker, with individuals having a 1 unit change being 1.120 (95% CI = 1.061, 1.181)
times more likely to be a High Drinker (see Table 51, p. 205), and 2.40 times more likely
for a difference equivalent to one standard deviation (7.74).

The Interaction Model also was significant overall (x26) = 22.35, p < .001, chm
a 3...," = .14, Rzumc = .19) and had a successful prediction rate of 67.3%; however, it
did not show a block improvement over the Additive Model (x20) = .00, p = .950).
Social lubrication remained the only significant individual predictor for high versus low
drinkers, with individuals having a1 unit change being 1.119 (95% CI=1.061,1.181)
times more likely to classified as a high drinker (see Table 51, p. 205).
The Comprehensive Model

Males. The Comprehensive Model was significant according to the model chi-
square (38(13) = 86.95, p < .001, Ric... & 3...... = .51, RZNW. = .68) and showed a
significant improvement over the Basic Model (x200) = 45.97, p < .001). The model
illustrated an 87% success rate in predicting the correct drinking classifications, an
improvement of 3 1 .7% over the Basic Model. Perception of best fi'iend usage improved
the likelihood of being classified as a high drinker (OR = 1.037, 95% CI = 1.021, 1.053)
while being a COA actually decreased the likelihood of being a high drinker versus low

drinker (OR = .188, 95% CI = .037, .947, see Table 52, p. 207).

206

 

.8:—o 8

:ommmmo womammmoz 3.38:8 maaoam 9m. Em: 588.. :02 Una—83 I 0950883?" Zeno—

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

?88 8388

8 mm ’58 OW 88 Q 8 mm «<95 0w 8.8 Q
Um?” -98 9mm 98 90:o 9w _ N 8.80 93 98 9 _ m :88 988 98:
9,3 -98 98 98 908 980 :Go -98 98 9 8 98: 980 :88
>8 98 98 98 :8o 98: ::8 9o: 98 9 _ N :o:~ 9:8: :88
ch 09.80308 98 92 mm: :8N 908 :8. 98 93 888 :28 :o8 :80
CE.” x 3m: 82.8083 -93 98 98 908 908 :80 -98 93 N8 908 98c :28
wow. m 125 98 93 89.3... _._ :8: :8_ :08 98 93 ~98: :8m :8: :80
02% x wow» 3.8.5 98 9S :8 :3m 900a :88 -98 98 8.88... 90:0 908 900m
00> :8: 98 8.5... 9:8 98: 98: 9:0 980 95 98: 98c N80
Umvw X 00> :8 :8 90m 883 98.8 ::.:8 98 98 98 :80 9N0o _:m8
883: ”3:08: 93 98 N8 :8 908 :88 98 98 _98: :8 :8~ :88
Ume x 8:80: ”8:30: -92 93 98 908 93c :Noo -98 98 98 908 98: :8
meow».— ::c:omao= -98 98 98 908 988 :_ _o 98 98 9N0 :o8 908 :— ;
Um?” x moan. reggae: 98 98 98 :88 938 :N0_ 98 98 9.0 :o8 988 :83

 

 

 

 

 

 

 

 

 

 

 

 

 

 

ca A .8. 23-889 _._... .a A .2. 2.98:8.

 

 

207

Females. The Comprehensive Model was significant (3803) = 88.84, p < .001,
RZCO,‘ & gm“ = .45, Rszlmkc = .60) and showed a significant improvement over the Basic
Model (x2(10)= 86.81, p < .001). The model illustrated an 84.7% success rate in
predicting the correct drinking classifications, an improvement of 32% over the Basic
Model. Perceptions of MSU student use (OR = 1.016, 95% CI = 1.002, 1.029).
perceptions of best fiiend use (OR = 1.048, 95% CI = 1.027, 1.069), the interaction
between differentiation and the perception of best fi'iend use (OR = .970, 95% CI = .942,
.998), and tension reduction (OR = 1.135, 95% CI = 1.052, 1.225) (see Table 52, p. 207).

High Drinkers versus Moderate Drinkers

Differentiation and GA U

Males. The Basic Model failed to reach a level of significance as an overall model
(36(3) = 5.851 p = .119, see Table 53, p. 209).

Females. The Basic Model failed to reach a level of significance as an overall
model (78(3) = 4.095, p = .251, see Table 53, p. 209).
F our factors of the DSI-R with GA U

Males. The four-factor model was significant as an overall model (78(6) = 16.35,
p = .012, chmg Smu= .13, Rznwm = .17) and had a 61.5% success rate in predicting
the correct drinking classifications, an improvement of 5.8% over the Basic Mode].
Interestingly, a 1 unit increase in Emotional Reactivity (being less emotionally reactive)
increased the likelihood of being a High Drinker (OR = 1.928, 95% CI = .999, 3.720),
whereas increasing the “I” Position value by 1 unit decreased the likelihood of being a

High versus a Moderate Drinker (OR = .374, 95% CI = .198, .706, see Table 54, p. 209).

208

 

4min mm

rommmmo womammmo: wammonam macaoam EB Em: 535.. 2.3230 Una—22m I Gama—ammo:

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

3&8 man—£8
w mm 9&5 0w 3.x. 9 w mm €95 OW 8.x. 9
Um?” -93 93 9S 9§¢ 930 Home 93 93 98 _._oo 93w Poem
Eé 93 93 who _._.3 93A _.Nou -93 o.oo _.qn 995 993 _.owm
>mo 93 93 9mm :3 93a _.mua -98 93 Num 9.3a 9mg room
_._ h A .8. 35-3%? .2. E A .3. 23-6%?
HmEo m;
hommmmo ~398me vanaaasm mannaa Em: Em: 585 33968 Una—88 I moE $883 w
2
3&8 moafiom
v mm 9:3 or 3.x. 9 w mm €95 OW 8.x. 9
magmas»— WomoaiQ 99w 9m» 93... #on 980 VHS 95 9mm 98 _._Na 9mg _.mNm
.4: $830: -93 9mm 93.... 93A 93% 93a -98 9X 98 9vo 93» _.m8
magmas». 050m. 98 93 93 _.NB 93o Nb; 9: 98 92 Z 3 .23 :2»
Tie: 2:: 092m 9% 93 _.3 93c 98. :3 -93 9.3 93 9me 93m _.uom
EE 9 _ w 93 P: :3 99mm _.won -93 98 _.2 98m 93m #93
>mo o.oo 93 9w» fog 93m _Lom -98. 9; Nu— 9q3 9mg _.oam
.. h A bu. 35-8%? .2. V A .3. 25-8%?

 

 

Females. As was the case in the Basic Model, the four-factor model failed to
reach a level of significance as an overall model (352(6) = 4.49, p = .61, see Table 54, p.
209).

Dmerentiation & perceptions of the average MSU student ’s alcohol use

Males. The Additive Model with perceptions of their best friend at MSU’s alcohol
consumption failed to reach a level of significance (12(4) = 8.455, p = .076, see Table 55,
p. 212). However, the Interaction Model was significant overall (386) = 13.21, p = .022,
RICO, & 5...,“ = . 10, Rznmmm = .14) and had a successful prediction rate of 60.7%,‘and
showed a significant a block improvement over the Basic Model (38(2) = 7.36, p = .03).
The interaction of differentiation and perception of MSU student use was a significant
individual predictor with individuals being .978 (95% CI = .957, .999) times as likely to
be a High versus a Moderate Drinker (see Table 55, p. 212, and see Figure 3, p. 60, for
illustration of the interaction efl'ect).

Females. The Additive Model with perceptions of MSU student alcohol
consumption failed to reach significance as an overall model (x2(4) = 9.05, p = .060, see
Table 55, p. 212). However, the Interaction Model was significant overall (13(5) = 11.52,
p = .042, R29” 3nd. = .08, Rzrmeumke = .11) and had a successful prediction rate of
65.0%, however, it did not show a block improvement over the Basic Model (38(2) =
7.43, p = .024). Although it was not significant, the interaction between differentiation
and perception of MSU student use removed the significance of MSU student use as an

individual variable (see Table 55, p. 212).

210

Dijferentiation & Perception of Best Friend at MS U ’s Alcohol Use

Males. The Additive Model with perceptions of their best friend at MSU’s alcohol
consumption was significant overall (12(4) = 55.99, p =< .000, R29“ & gm“ = .31,
from“: = .42) with a significant block improvement over the Basic Model (380) =
35.81, p = .000), and had successful prediction rate of 76.2%, an increase 20.5% over the
Basic model. The perception of best friend use increased the likelihood of being a
High versus a Moderate Drinker by 1.024 (95% CI = 1.013, 1.034, see Table 56, p. 214)
times for a one unit change in perception and 5.29 times for a difference equivalent of
one standard deviation (72.43).

The Interaction Model also was significant overall (76(5) = 59.09, p < .001, k’m
& 3.," = .33, Rzumu‘m = .43) and had a successful prediction rate of 76.2%, but did not
show a block improvement over the Additive Model (360) = .45, p = .504). Perception
of best fiiend use remained the only significant independent variable in the categorization
of High versus Moderate Drinkers and remained equivalent to the values in the Additive
Model (OR = 1.023, 95% CI = 1.013, 1.033, see Table 56, p. 214).

Females. The Additive Model with perceptions of their best friend at MSU’s
alcohol consumption was significant (78(4) = 12.05, p = .017, R20», & s..." = .08,
Rznwm = .11) with a significant block improvement over the Basic Model (x20) =
7.96, p = .005). The model had a successful prediction rate of 61 .3%, an increase 2.2%
over the Basic model. Perception of best friend use was a significant individual variable
with individuals that had a one unit higher value on best friend perception were 1.012
(95% CI = 1.003, 1.021) times more likely to be a Higher versus a Moderate Drinker (see

Table 56, p. 214). The change in odds for being a high drinker increased to 1.65 times

211

 

H.920 mm

rommmao Womaommmos 333:5 mannam 98 Em: <oa=m 30338 9.3.88 I moo—.8303 om 25C manna Cmo

 

 

 

 

 

 

 

_ 3&8 _ 10323
_ e :3 sea _ on: 8&2 _ o :m _ $3: on: 8:9
W83 zona—
em; -98 93 9% o.ooo 8m... 38 So 98 98 :8 5: “So
we: 9:. 93 was _._: 99.: 58 -93 o.oo EN 98° 33 23
>% 9: 9; Sn :3 Sam 33 -98 9; N3 o.ooo o.ooo 2%
>338 gono—

 

Umm-W -9~m 9g 93 9am 9ch rm-R 95 9: 93 _.N_N 93m Nun—o

 

Ext: 9: 93 NS _.Sn 90.3 _.N-S -93 9ko _.3 98¢ 98M #03

 

>Mo 95 93 9am _._wc 930 _.MB -95 93 _.Am 93o 93m _._3

 

ZmCCmo 93 93 New :5» 908 _.o: 93 93 PS... 53o _.oS _.owo

 

383020: Zona—

 

Uwzw -98 93 9: 93m 93m ~68 9mm 9: 93 _.NAM 93a Puma

 

wz: 9: 93 me 92a 9on _.wa -98 9cm _.2 995 93m _.oma

 

>mo 9B 93 9mm :2 9qu _.mwm -93 93 _.E 9mAc 93m _.Iq

 

chcmo 93 9o— TS _.coa 90cm _.03 9o— 92 98 _.So _.ooo _.Sc

 

 

 

 

 

 

 

 

 

 

 

 

 

Umzwx ch can -98 92 Puma 93w ohm-N 908 -98 93 who 90% chow _.oom

 

 

a A .3. 35.823. _._... .u A .S. "So-8:8.

 

 

212

more likely when the difference between perception scores was equal to one standard
deviation (41.48).

The Interaction Model also was significant overall (x2(5) = 12.45, p = .029, RZCO,‘
& 5...," = .09, [ills-ageing, = .12) and bad a successful prediction rate of 61 .3%, but did not
show a block improvement over the Additive Model (12(1) = .40, p = .527). Perception
of best fi'iend use remained the only significant individual variable in changing the
likelihood of being a High versus Moderate Drinker (OR =1 .013, 95% CI = 1.003, 1.021,
see Table 56, p. 214).

Differentiation & COA Status

Males. The Additive Model with COA status failed to reach a level of
significance (78(4) = 6.98, p = .137, see Table 57, p. 215). The Interaction Model also
failed to reach a level significance (12(5) = 7.55, p = .183).

Females. The Additive Model with COA status failed to reach a level of
significance as an overall model (38(4) = 4.10, p = .393, see Table 57, p. 215). The
Interaction Model also failed to reach a level of significance (x26) = 6.63, p = .249).
Dijferentiation & Expectations of Alcohol Serving as a Tension Reducer

Males. The Additive Model with expectations of alcohol as a tension reducer
failed to reach a level of significance (38(4) = 8.77, p = .067, see Table 58, p. 216).

Females. The Additive Model with expectations of alcohol as a tension reducer
failed to reach significance (38(4) = 7.34, p = .119, see Table 58, p. 216). The interaction

Model also failed to achieve a level of significance (38(5) = 8.202, p = .145).

213

 

flea—a 3

rommman womammmoa 333:5 wagons =5” ”8 an: 388 33038 01383 I $28303 3- W8" magma Cwo

 

 

 

 

 

 

 

_ 3&8 _ moan—Em
_ a _mm_ are: _09_ 8x9 _ a _mm_ fin: _ 2: 39.9
wmmmo Zona—
cm; -98 93 93 9s; 8% 38 98 98 98 :8 9m: No8
my: 9: 98 use 2w... 98a 58 -93 98 SN 98o 98m _.8...
>ma 9: 9; 93 :3 935 _.aa -98 9; 93 93a 9M8 23
>a&m<o Keno.

 

OMEN -9“: 9.3 93 9qu 92% 33¢ 93 9“: 93 _.qu 93me N38

 

wz: 93 o.oo Pmm 573 93c _.wmq -93 93 3: 989 9mg :50

 

>mo o.oo 93 9M... _.ocw 93m #3; -93 93 who 93o 93mm _.33

 

womfimaosadcma 98 92 3.3.... _.ONA _.Su row.“ 93 o.oo 93...... _.SN _.oow rom—

 

EfiBoaos Zona—

 

szw -93 93 93 93m 930 ~33 93 93 93 _._-3 9m-NN 9A3

 

wzm 93 98 93 _.3o 93w _.wAc -93 93 _.3 93m 930 _.95

 

>mb. 90¢ 93 9mm _.ocw 93M _.mow -93 93 N3 9.2: 933 _.omm

 

wamfimloamw :8 98 9o— Nouo: fiowu _.Sw _.Owu 93 93 Soon... fo—w _.oow _.omw

 

 

 

 

 

 

 

 

 

 

 

 

 

02% X wow» magma C3 -93 93 93 908 9on _.oE 93 9S 9% foom 90cc _.owo

 

 

_._-u A .3. "so-$39 1.9 A .2. 3.0-4.359

 

 

214

 

.320 mu

Famine Womammmo: 33835 magnum 92 Ed Em: 385 gonna—8 U118; I 00> mSEm

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

_ 3&8 _ 122:8
_ u _ mm _ $5 3 oz _ 8.x. 9 3 w _ mm 75:: oz 9 8.x. 9
P50 zona—
Ume .93 93 93 9:6 93o _.mmo 93 9mm 98 :8 93a Nova
93 9 _ u 93 who _._: 903 _.Now -93 93 _.qw 98o 93m :35
>mo 93 93 93 :3 93m _.mwm -98 93 Mom 9qu 93m _.oao
>aa§<o Zena
Um?” -93 93 9K 993 93.» _.muo 95 9mm 93 :oo 9%.» Poem
3": 9 _ w 93 93... :3 _.c3 _.wom -93 98 _.qu 98c 93; _.ouq
>mo 9G 93 9mm :3 9%?“ _.mm_ -98 93 93 9qu 9mg :33
00> -93 93 _._N 9mg 98o 9mm». -98 93 93 908 93c Nuum
38330: 380.
03.” -93 93 _.ou 93a 93.‘ rmom -9 E 93 9 E 93; 93w 9.3m
EE 9 3 A 93 Brow“. :3 _.OOA 3b _ m -93 93 _.oo 9on 93m _.Ouw
>mo 9G 9G 93 :3 93¢ rum”. -9? 93 who 9.23 9qu _.ome
00> -9mm 93 _.wo 9&0 9B~ 9A: -93 9g o.oo 93o 9min N»;
03.9 x 00> 93 93 9% 3b: 93; _9w3 :6 98 Pun A.A”: 93m w9m3

 

 

 

 

 

 

 

 

 

 

 

 

 

 

_.. .u A .3, 25-8209

1% A .3 . 25-559

 

 

215

 

HmEo mm

rommmzo Womaommmo: $9538 95an 92 98 En: 835 30838 013.88 I mxvaofiaoam om Hoswmoa flamenco:

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

_ 3&8 _ mesa—om
_ v _ mm 9 a»: 9 ow _ 8.x. 9 _ w _ mm _ 55 _ ox _ 8.x. 9
wmmwo Zona—
Umvw -98 93 98 .Eo 98¢ 83 93 98 98 :8 983 No8
EE 9 8 w 93 who _._“: 903 _.N8 -93 98 _.qm 98o 988 58a
>mo 9: 9G 98 :3 98a _.8m -98 93 N8 93¢ 988 fog
>833 Zona—
Uwvw -98 93 98 905 920 905 98 9mm 98 88 9mm: 9me
EE 9 _ m 98 8.8 2.: 993 83 -98 98 _.uo 98m 980 _.ofi
>ma 9 H m 9 _ 8 _.AN _.Nom 98m _.mwm -98 9 _ m 9.3 93m 9ch rot
Anzac: won—8:0: 98 98 NM: rot 98w :8 98 98 8.: r8_ 98A _._ I
58828: 38¢.
62% -93 9t 98 908 92: 9on 9B 98m 93 _.N8 988 N58
wz: 9 B 98. 8.2 :2 98a 88 -98 98 _.8 98a 98o _.8m
>mo 9_ m 9 _ a 8g :8 988 ram; -98 9 _ o 8.8 98¢ 98m “.88
Adamo: ”8:30: 98 98 N? _.onm 98w :8 98 98 8.8 #8; 9o3 _._ _m
Ume x Anzac: romance: 92 98 98 _.oom 98m :8 98 98 98 reg 98c :8
_._fi A .8u 25-889 _._... u A .3. 25.823.

 

 

216

Difierentiation & Expectations of Alcohol Serving as a Social Lubricant

Males. The Additive Model with expectations of alcohol as a tension reducer
failed to reach a level of significance (38(4) = 6.933, p = .139, see Table 59, p. 218). The
Interaction Model also failed to reach a level of significance (38(5) = 6.939, p = .225).

Females. The Additive Model with expectations of alcohol as a social lubricant
failed to reach significance (36(4) = 5.45, p = .244, see Table 59, p. 218). The Interaction
Model also failed to achieve a level of significance (76(5) = 5.58, p = .350).
The Comprehensive Model

Males. The Comprehensive Model was significant according to the model chi-
square (x203) = 51 .44, p < .001, RICO, a s..." = .34, Rimm, = .46) and showed a
significant improvement over the Basic Model (x2(10)= 45.59, p < .001). The model
illustrated a 74.6% success rate in predicting the correct drinking classifications, an
improvement of 18.9% over the Basic Model. Perception of best friend use was the only
significant individual variable in the Comprehensive Model (OR = 1.026, 95% CI 1.014,
1.038, see Table 60, p. 219).

Females. The Comprehensive Model was significant according to the model chi-
square (x203) = 23.41, p = 037,189,... 5,... = .16, mm... = .21) and showed a
significant improvement over the Basic Model (x2(10)= 19.32, p = .04). The model
illustrated a 65% success rate in predicting the correct drinking classifications, an
improvement of 5.9% over the Basic Model. Interestingly, taken as a whole the model
was significant, however, no individual predictors in this model significantly changed the

odds of being a High Drinker versus a Moderate Drinker (see Table 60, p. 219).

217

 

.320 we

rommmzo ”$388: 38338 mange 93 m3 Em: 58% 38088 01388 I mxuooSaoa om moor: raw—mouse:

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

fl 3&8 _ moBEom
_ a _ mm _ 2o; _ o: _ 8:9 _ a _ mm H ,5: o: _ 8.x. 9
wmmmo Zena
0mm-” -93 93 9¢N 938 98c _.8o 95 98. 98 :8 93.3 No8
ES— 98 93 who _._: 908 _.N8 -93 98 _.3N 98o 9m8 _.8¢
>8 9: 98 98 :3 98¢ _.8¢ -98 98 N8 93¢ 988 _.8¢
>835. 333
Ume -9 fl 8 9.8 9 I 98c 98; _.30 9N¢ 98 98 _.Noo 9¢s5 ~.¢~¢
m3: 9 _ A 93 98¢... _.TK moon _.83 -98 98 #8 98m 9mg _.oao
>mo 98 98 :A _.3¢ 938 .88 -98 98 NE 98; 983 _.33
moan— rccnowmo: 98 98 _.3 _.8_ 938 room 98 98 _.8 _.8m 93¢ _.08
58:83:: :83
Um?” .93 93 98 98¢ 98m #80 98 98¢ 9mm #83 9¢3 ~88
EE 9 I 93 u .93... _._: _.oow _.w3 -98 98 93 9ko 9?: :8“
>8. 9 _ ¢ 9_ u _.E _.3¢ 938 ~88 -93 9 _ m _ .3 98¢ 9¢8 _.omm
mooi r8183? 98 98 _ .3 _.8w 93m _.ooA 98 98 _.3 #8; 93¢ 58¢
09-x X moan. C5188: 98 98 9S _.ooa 908 :3 -98 98 98 908 98¢ _.omo

 

 

 

 

 

 

 

 

 

 

 

 

 

 

am A .8. 25-889 .3. .: A .3. 25-889

 

 

218

 

Ania ¢o

homage wamaammmo: 0898988 850:8 9m» m8 Em: 885 80888 03:88 I 0030330830 Zona—

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Km—om moan—am
8 mm «$8 CW 88 Q 8 mm 8:5 CW 88 Q

02% -93 980 _.03 98 9 EN ..3¢ 9 _ m 98 9 3 :8 98.. 8.88

2,5 9 _ ¢ 930 N8 :8 908 _.3¢ -98 98 98 903 98— ~38

>ma 98 98 9 _ 8 .38 988 _.¢- -98 9 _ 8 _.8 980m 988 _.38

Km: 88388 98 93 98 9000 908 _.SN 92 9S _.8 _.38 908 .38

Umzw X 78: 880083 -98 93 8.8 98¢ 90% .38 -98 9S _.8 908 983 ~38
mam. manna 98 93 8.83.. :08 _.o _ A _.38 92 92 P8 #80 9000 _.83

08% X mam. $82. -92 93 98 900m 908 ~38 93 92 98 _.38 908 ~38

00> L . _ ¢ 98 8.3 9w _ w 930. _.8¢ -98 9mm 98 9un 98A No:
Umzw X 00> _.8 r8 _.8 .988 988 8.33 #8 _.30 _.8 ”80¢ 93¢ 8988

838: 8988: 98 98 93 _.88 98¢ :3 98 98 Nun _.88 908 _._”:

02-x X 838: 8908: 9: 9: 90o : 8 980. .80.» 98 98 _.8 _.8_ 908 #85
moan— rcgomao: -98 98 90¢ 908 98m _.88 98 98 98 _.8_ 903 :08

Um?” X momma 55188: -98 98 98 90.: 98¢ 83.» -98 98 98 908 98.. _.08

 

 

 

 

 

 

 

 

 

 

 

 

 

_._ h A .8. 25-328. .2. h A .3 . 23-889

 

 

219

Appendix M
Analyses of Binge Drinking Use

As was done with the General Alcohol Use measure, the continuous Binge
Drinking variable (number of days in the past 14 days the subject drank at binge levels - -
5 drinks per occasion for males, 4 drinks per occasion for females) was converted into a
categorical variable consisting of 4 levels of binge drinking: Abstainers (those that did
not drink), Nonbinging Drinkers (subjects that consumed alcohol, but not at binge levels),
Occasional Binge Drinkers (1 or 2 times in the past 2 weeks), and Frequent Binge
Drinkers (3 or more times in the past 2 weeks) (see Appendix J). As the four categories
reflect increasing degrees of alcohol consumption, the nature of this study’s hypotheses
which was stated for a continuous dependent variable (e. g., students lower levels of
differentiation will be more likely to have greater amounts of binge drinking) was
maintained with the categorical analyses (e.g., students lower levels of differentiation will
be more likely to have a greater relationship with Frequent Binge Drinkers than
Occasional Binge Drinkers when controlling for BMI and age).

The four binge drinking categories were contrasted in ways such that four unique

outcome dichotomies were created:

H

. Binge Drinkers (Occasional & Frequent Bingers) versus Nonbinging Drinkers

2. Occasional Binge Drinkers versus Nonbinging Drinkers

E”

Frequent Binge Drinkers versus Nonbinging Binge Drinkers
4. Frequent Binge Drinkers versus Occasional Binge Drinkers

The same coding procedure fiom the General Alcohol Use variables was implemented

with the Binge Drinking variables with the greater prevalence of binging was coded as a

220

 

“1” in SPSS and the lesser category was coded as a “O.” For example, in the Bingers to
Nonbinging Drinkers analyses, all persons in the Occasional and Frequent Binging
categories were coded as a “1” and Nonbinging Drinkers were coded as a “O.”
Bingers (Occasional & Frequent Bingers) versus Nonbingers

DWrentiation and binge drinking

Males. The Basic Model was not significant according to the model chi-square
(38(3) = .75, p = .861, see Table 61, p. 222).

Females. The Basic Model failed to reach a level of significance according to the
model chi-square_(x2(3) = 3.38, p = .337, see Table 61, p. 222).
F our- factors of the DSI-R (Emotional Reactivity, “1” Position, Emotional Cutofl,‘ and
Fusion with Others) with binge drinking

Males. As was the case in the Basic Model, the overall four-factor model failed to
reach a level of significance 08(6) = 1.327, p = .970, see Table 62, p. 222).

Females. As was the case in the Basic Model, the overall four-'factor model failed
to reach a level of significance (x2(6) = 4.987, p = .546, see Table 62, p. 222).
Di/firentiation & perception of average MSU student ’s alcohol use

Males. The Additive Model with perceptions of the average MSU student’s
alcohol consumption was not significant as an overall model (36(4) = 5.616, p = .230, see
Table 63, p. 223). The Interaction Model also was not significant as an overall model

(38(5) = 5.923, p = .314).

221

 

Hugo 3

rommmmo Nowaommmoa 32:35 manna—a 93 93 Ooommmoaa gamma <22; 2039508 I 059.3323:

 

 

 

 

 

 

25.8 mafia—om
w mm $33 OW 3.x. 9 w mm «Sm:— Ow 8.x. 0—
Umzw -98 93 93 9on 93m H28 .98 93. 93 93m 93m #me
9’: -98 93 9B 993 9mm~ _.omw -93 93 New 9ch 930 fiooo
>mo 9B 93 98 _._: 93m _.mmo 9cm 9; 9; _.omm 93m :30

 

 

 

 

 

 

 

 

 

 

 

 

 

_._ .m A .8. go-8=om. .15 A .2. "So-Smog.

 

 

 

Hug—o am

rommmao womaommmoa mammogm 9:936 92 3a wmamaa <oa=m Zoagsmoa I mean v.88;

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

3&8 mania-a.
w mm Qua OW 8.x. 9 w mm Irma OW 8.x. 9

wagon»: WomomiQ 9 _ A 93 9 H .V :3 o.ooo N. _ S 98 9B 9 _ A _.cmA 9g _ o _.mmm
is peace: 93 93 98 _.oum 9mg _.emq .9? 93 93 93c 93o _.w _ a
5530:». 0503. -9: 9wo 93 9mg 9&3 _.aoc 93 9B 93 :3 9d~ _.dw
«ammo: 2:: 099m 59¢ 93 93 93o 93m _.wwm -98 93 93 9.30 93o _.ANM
BE -98 93 9M» 93a 9mm— :x: -93 9E 98 993 9mg _.ooo
>mo 93 93 9% _._-3 93m _.mam 93 93 9mm _.omo 930 _.wma

_._ E A .3. "So-€39 I. k A .2 . Hie-8:8.

 

 

 

222

 

H920 3

homage ”£8.88: 32:35 mama-=8 53 98 gamma «685 Zoacmswoa I 383303 om zwc manna Cmn

 

 

 

 

 

 

 

T :83 _ maamfim
_ N. Hmm A 2% _ 2: 882 _ 8 _mm :59: 9: 889
mum? Zona—
cm; :98 98 98 988 988 SS -98 93 98 9% 938 988
$2 -98 98 9B 98. 98M :5 -93 98 98 903 98c 28
>9. 9: 9a 98 _._: 938 38 98 9; 9; 988 938 3.3
>ma3<o gona—

 

Umzw -98 93 9o— oboo 9A3 “.03 -9~o 93 9mA 9m§ 9AmA _.wow

 

wZH -9OA 9cm 9% 93w 9mg #08 -9oa 99A Num 90AM 9wa be;

 

>mn 9B 9: 9mm EBA 9m; _.m-Nm o.oo 9; ohm room 9mm~ _.waq

 

ch Cmo 98 9S PmA _.oG _.ooo _.owo 9o— 98 NAN _.ooq 98w Ab;

 

588030: Zona—

 

Ume 90m 93 98 _.omo 9A3 whom .98 chm 93 9a.: 9AAw #qu

 

wz: -90A 98 9Aa 93A 930 _.o: -93 90A :3 903 9a: HbMA

 

>mo 93 9: 9mm #33 9m: _.mww o.oo 9S 9AM _.ooo 9wAm _.AS

 

ZmC Cmo 98 9o. Ab? _.Sa _.oco Ab”: 93 o.oo MUN _.ooq 993 _.So

 

 

 

 

 

 

Umzwx ch C8 93 9S 9”: _.ooq o.oww _.owA -92 9S 9; 98a 99: _.oo_

 

 

 

 

 

 

 

 

 

_._ w A .8. «so-8:09 .3 h A .9. "So-8:09

 

 

223

Females. The Additive Model with perceptions of the average MSU student’s
alcohol consumption was not significant as an overall model (38(4) = 6.00, p = .200, see
Table 63, p. 223). The Interaction Model also was not significant as an overall model
(36(5) = 9.293, p = .093).

Dijferenfiation &Perception of Best Friend at MS U ’s Alcohol Use

Males. The Additive Model with perceptions of their best friend at MSU’s alcohol
consumption was significant (38(4) = 25.61, p < .001, RIC“ & gm“ = .13, Rznmu‘m = .22)
and had a successful prediction rate of 82.7%, which was 5% less than having no model.
Students that had one unit higher in perceptions of best friend use were 1.028 (95% CI =
1.013, 1.043) times more likely to be a Binge Drinker versus a Nonbinging Drinker (see
Table 64, p. 225). Students that had a higher level of perception of best friend use
equivalent to one standard deviation (72.43) were 7.067 times more likely to be a Binge
Drinker.

The Interaction Model also was significant (78(5) = 25.68, p < .001, RIC... & s...“ =
.13, Rznmmm = .22) and had a successful prediction rate of 82.7%, but did not show a
block improvement over the Additive Model (380) = .16, p = .693). The only significant
individual predictor in the model was perception of best friend use (OR = 1.028, 95% CI

= 1.013, 1.043, see Table 64, p. 225).

224

 

Haw—a i

rommmmo Womnnmmmo: wagon—gm 958$ 92 m8 gamma 588 232508 I men—.8393 cm wow" «.1053 Can

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

_l Zaom _ moBEom
_ a _ mm _ éaa fl ox _ 3.x. 9 _ a _ mmF éoa _ ox _ 3.x. 9
W83 30%.
Ume -98 9g 93 9on 93m _.mmo .93 93 93 9mg 93m _.wma
95 -98 93 9B obi 92$ roam -93 93 N8 903 93o _.ooo
>mo 98 93 98 _._HX 93m _.mao 93 95 9; _.omm 93m _.wao
>355 Zona—
Umzw 93 93 9S _.owq 980 N. _ mo -9 _ .\ 98 9“: 9%3 93o _.MNA
95 -93 o.oo 9% 903 9me _.o.: -98 92 _.No 903 93w _.8.»
>mo o.oo 9; 98 room 9qu _.u: 93 9E o.oo #ko 9.3m _.uau
wow” «125$ C3 98 93 3.3.... _.owm _.Sw roam 93 93 ~93: rouo _.omu _.omm
58330: zona—
Umzw 9mm 93 9 D m _.Nmm 93o Pug -93 93 93 9.25 93% _.3a
95 -93 o.oo 93 903 9mg _.Sm -93 93 _.No 99% 93m _bAo
>mo o.oo 9 _ a 9g _.oom 9.5a rm: 9% 9 _ A 98 $vo 9.3m #89
wow» 3.5:: :3 98 93 5.21... _.oum _.Su _.ofi. 93 92 3.31. _.OAo _.owu _.omu
Um?” x womamnosaw C8 93 9...: 93 roam 928 _.9: -93 92 9.3 90$ 93» _.oz

 

 

 

 

 

 

 

 

 

 

 

 

 

 

a. to A .3. 25-8:09 .1 No A .3. EGAN—=09

 

 

225

Females. The Additive Model with perceptions of their best friend at MSU’s
alcohol consumption was significant (12(4) = 45.28, p < .001, R2“, & Sm" = .19, Rzuwkflh
= .27) and had a successful prediction rate of 75.6%, an increase 5.6% over the Basic
Model. Students that had one tulit higher in perceptions of best friend use were 1.039
(95% CI = 1.023, 1.055) times more likely to be a Binge Drinker versus a Nonbinging
Drinker (see Table 64, p. 225). Students that had a higher level of perception of best
friend use equivalent to one standard deviation (41.48) were 4.84 times more likely to be
a Binge Drinker versus a Nonbinging Drinker.

The Interaction Model also was significant (752(5) = 46.02, p < .001, 14220,,( a 5.," =
.19, RZNW = .27) and had a successful prediction rate of 76.1%, but did not show a
block improvement over the Additive Model (12(1) = .74, p = .39). The only significant
individual predictor in the model was perception of best friend use, which was nearly
identical in value as the Additive Model (OR = 1.040, 95% CI = 1.023, 1.057, see Table
64, p. 225).

Dijfirentiation & C 0A Status

Males. The Additive Model with COA status failed to reach a level of
significance (76(4) = 1.23, p = .873, see Table 65, p. 227). The Interaction Model also
failed to achieve a significance as an overall model (x26) = 1.359, p = .929).

Females. The Additive Model with COA status failed to reach a level of
significance (38(4) = 4.200, p = .380, see Table 65, p. 227). The Interaction Model also

failed to achieve a significance as an overall model 086) = 5.033, p = .412).

226

 

.320 am

rommmmo Womaommmos wRamozam £595 93 8d wmamoa <oa=m Zosgsmoa l 00> mSEm

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

_ 3&8 _ moaaom
_ a _ mmP drama 4 ow _ 8.x. 9 _ a # mm m 55 _ ow _ 8.x. 9
momma 30%:
62% -93 93 98 9on 93m _.mmo .93 93 9% 98M 93m _.wmm
$5 -98 93 9B 903 9me _.omm -93 93 New 9er 9on _.ooo
>mo 9 _ u 9 _ m o.oo :3 93m _.mmo 9cm 9 _ u 9 _ a _.omm 93m :3
>335 gono—
Uwvw -93 93 93 93A 93a _.05 -98 93 9% 9m 2 m 9.28 _.wqo
E’s: -98 93 9 _ a 928 9m? #23 -93 9E 93 93A 9me _.oom
>mo 9B 93 9% _.SA 9me 32 9am 9B 9B _.oao 93m 53
00> -93 9mm 93 93c 930 930 -93 98 93 93; 93a 2.9.8
58398: Kong
GB.” -9 _ u of 9 _o 9me 9qu 2.03 -93 98 :m 9S” 9.5m _.Noo
Ex: -98 93 9 _ m 926 953 2:3 -93 93 98 98h 9m? _.ooa
>mo 93 93 o.oo :um 9Su 32 93 9G 98 _.oou 930 _.wS
00> -93 9% 93 98w 93m N. H cm -93 9a: _.3 93A 9MB :3;
Ume x 00> 9X 93 9S 2.3m 9»: PM: 93 93 93 2.9.3 93m 98a

 

 

 

 

 

 

 

 

 

 

 

 

 

 

_._.u A .3. 30-3559 3h A .S. 95-8:09

 

 

227

Dzfl'erentiation & Expectations of Alcohol Serving as a Tension Reducer

Males. The Additive Model with tension reduction expectations was significant as
an overall model 08(4) = 20.74, p < .001, RICO, & s..." = .11, RIM“... = .18). The model
had a successful prediction rate of 83.2%, which was identical to having no model.
Expectation of alcohol serving as a tension reducer was the only significant individual
variable. Students that had a one unit higher level of expectation were 1.133 (95% CI =
1.068, 1.202) times more likely to be a Binge Drinker versus a Nonbinging Drinker (see
Table 66, p. 229). Students that had a greater perception equal to one standard deviation
(8.73) were 2.98 times more likely to be Binge Drinkers.

The Interaction Model also was significant overall (38(5) = 22.13, p < .001, RICO,
& 5m“ = .11, Rsz = .19) and had a successful prediction rate of 83.2%; however, it
did not show a block improvement over the Additive Model (12(1) = 1.39, p = .24).
Expectations of tension reduction remained the only significant individual predictor (OR
= 1.142, 95% CI = 1.073, 1.216) of Binge Drinking (see Table 66, p. 229).

Females. The Additive Model with tension reduction expectations was significant
as an overall model (38(4) = 43.71, p < .001, RZCO,‘ & 5nd1= .19, Rznwak, = .26). The
model had a successful prediction rate of 74.2%, an increase of 5.7% over not having a
model. Expectations of alcohol serving as a tension reducer changed the odds of being a
Binge Drinker versus a Nonbinging Drinker by 1.147 (95% CI = 1.094, 1.202) for a one
unit difference in expectations (see Table 66, p. 229). By changing the difference in
expectations to a unit equivalent to one standard deviation for expectations (8.81) the

likelihood of being a Binge Drinker is 3.34 times greater.

228

 

Hugo mm

roman.” Wannommmoa 385:8 mzanam EN: Ba 9803 <o~mcm 225583 I mxvoofizozm ow Anzac: woacomo:

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

3&8 moan—om
u mm 9S5 OW 8.x. 9 w mm 9S5 OW 8.x. 9
momma zona—
Ume -98 9mm 98 98m 93m _.wwc -98 93 9% 98m 93m _.uma
w?= -98 98 9B 903 9mg _.oww -93 98 N8 903 989 _.ooo
>mo 9 _ u 9 H a 98 :85 98m _.mg 98 9 m u 9 _ m _.omm 98m _.uAc
25:78 30%;
Um?” 98 9mm _.Nq _.muo 98“ 8. 3w 98 9.3 98 _.NAM 98— MES
95 92 98 98 _.oou 988 :8 -98 98 :N 98m. 93m _.OAN
>mo 9B 9; T: _.8m 9m.: _.qmw 93 9: 93 :8 980 _.Ama
H255: Waggon 9; 98 3.8.1. :8 room _.Nom 9: 98 8.8.... :3 _.8» _.Nou
388030: Zona—
Umzw 9.: 93 Na; Nona 9?: +me 98 9mm 98 _.3.» 93¢ No8
BE 93 98 98 _.80 908 _._NA -98 98 9mm 98» 93A fog
>mo 98 9; _.No _.NNN 9mg _.dm 9: 9E 98 :8 9mm_ _.gm
H258: wagons 98 98 3.8%.... :3 _.oqu _.Ba 93 98 89mm: :3 room Ewen
Umzw X Adamo: Woacomoa 98 98 _.8 _.oao 93A :8 -98 90; N: 995 9mg. _.owo

 

 

 

 

_._ m A .8, 9293—8. .1. .u A .3. go-§_oa.

 

 

229

The Interaction Model also was significant overall (38(5) = 45.96, p < .001, RICOx
& 5...,“ = .19, Rznmlkm = .27) and had a successful prediction rate of 71.8%; however, it
did not show a block improvement over the Additive Model 080) = 2.25, p = .134).
Expectations of tension reduction remained the only significant individual predictor (OR
= 1.146, 95% CI = 1.092, 1.202, see Table 66, p. 229).
Diflerentiation & Expectations of Alcohol Serving as a Social Lubricant

Males. The Additive Model with social lubrication expectations was significant as
an overall model (38(4) = 19.43, p = .001, R20”, & Sod. = .10, Rzmgcumk. = .17). The model
had a successful prediction rate of 83.8%, an increase of .6% over the not having a
model. The only significant individual predictor was expectations of alcohol serving as a
social lubricant. Students that had a one unit higher level of expectation were 1.145 (95%
CI = 1.071, 1.224) times more likely to be a Binge Drinker versus a Nonbinging Drinker
(see Table 67, p. 233). Students that had a greater perception equal to one standard
deviation (7.89) were 2.93 times more likely to be Binge Drinkers.

The Interaction Model also was significant overall (38(5) = 19.58, p = .001, R20...
& 5...," = .10, Rznmu‘m = .17) and had a successful prediction rate of 82.7%; however, it
did not show a block improvement over the Additive Model (380) = .15, p = .697).
Expectations for social lubrication remained the only significant individual predictor (OR
= 1.147, 95% CI = 1.072, 1.227) in predicting Bingers versus Nonbinging Drinkers.

Females. The Additive Model with social lubrication expectations was significant
as an overall model (38(4) = 17.33, p = .002, R20, .1 5...... = .08, mm... = .11). The
model had a successful prediction rate of 70%, an increase of 1.5% over the not having a

model. Expectations of alcohol serving as a social lubricant changed the odds of being a

230

drinker by 1.090 (95% CI = 1.040, 1.142) for every one unit change in perceptions (see
Table 67, p. 233). Students with a difference in expectations equal to one standard
deviation (7.74) were 1.95 times more likely to be a Binge Drinker versus a Nonbinging
Drinker.

The Interaction Model also was significant overall (78(5) = 18.07, p = .003, RZCO,‘
& 5...," = .08, fireman“,e = .11) and had a successful prediction rate of 71 .8%; however, it
did not show a block improvement over the Additive Model (36(1) = .74, p = .3 s9).
Expectations for social lubrication remained the only significant individual predictor (OR
= 1.089, 95% CI = 1.039, 1.141) in predicting Binge Drinkers versus Nonbinging
Drinkers (see Table 67, p. 233).

The Comprehensive Model

Males. The Comprehensive Model was significant according to the model chi-
square (21203) = 41.1 1, p < 001.11%,” 3.....= .20, Rimm. = .34). The model
illustrated an 85.4% success rate in predicting the correct drinking classifications, a 2.2%
increase over having no model. Perceptions of best fiiend use (OR = 1.019) was the only
individual variables to significantly improve the odds of being classified as a Binge
Drinker versus a Nonbinging Drinker (see Table 68, p. 234).

Females. The Comprehensive Model was significant according to the model chi-
square (3603) = 85.81, p < .001, 112.2,... 5,... = .33, RIM...” = .47). The model
illustrated a 70% success rate in predicting the correct drinking classifications, the
identical value to the Basic Model. Perceptions of best fiiend use (OR = 1.037),

expectations of tension reduction (OR = l. 166), and the differentiation X expectations of

231

tension reduction (OR = .862) were the only individual variables to significantly improve
the odds of being classified as a Binge Drinker versus a Nonbinging Drinker.
Occasional Binge Drinkers versus Nonbinging Drinkers

Diflerentiation and binge drinking

Males. The Basic Model was not significant according to the model chi-square
(38(3) = .50 p = .919, see Table 69, p. 236).

Females. The Basic Model did not reach a level of significance (38(3) = 1.40, p =
.705, see Table 69, p. 236).
F our- factors of the DSI-R with binge drinking

Males. As was the case in the Basic Model, the overall four-factor model failed to
reach a level of significance (38(6) = 2.77, p = .838, see Table 70, p. 236).

Females. The overall four-factor model failed to reach a level of significance
(36(6) = 4.58, p = .599, see Table 70, p. 236).
Dijferentiation & perception of average MS U student ’s alcohol use

Males. The Additive Model failed to reach a level of significance (36(4) = 1.547,
p = .818, see Table 71, p. 237). Similarly, the Interaction Model also was not significant
as an overall model (38(5) = 2.980, p = .703).

Females. The Additive Model with perceptions of the average MSU student’s
alcohol consumption was not significant as an overall model (36(4) = 3.36, p = .500, see
Table 71, p. 237). The Interaction Model also was not significant as an overall model

(38(5) = 4.273, p = .511, see Table 71, p. 237).

232

 

Hmzm 8

homage Womanmmmoa 32:35 9338 93 m3 Esme; <22; Zoacmsmaa l mxvooamaoaw om moor: 9.51830:

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

fl Kan—om _ mans—mm
_ o _ mm _ éoa _ ow _ been. 8 a fi mm _ sin _ Q: 3:. Q
momma Zone
Um?” -98 98 98 908 988 _.mwo -98 98 9am 9m8 9&8 _.uma
9,3 -98 98 9B 903 9me #88 -98 98 8.8 908 980 _.oo0
>8 9 _ u 9 3 98 :8 988 _.mao 98 9 8 w 9 H m _.08 988 830
>338 Zona—
Uwzw 98 98 _.Nm _.m-fi 98; 93 q 9 _ w 980 98— :fi 98_ _ .000
BE 98 98 98 _.S m 908 :80 -98 98 8.3 902 980 H .8 _
.98 98 9; _.8 _.88 930 _.qwq 93 9B 98 :9 988 _.80
moan: r5185: 9 3 98 8.8.1. :8 #0.: _.8A 900 98 8.31. _.o0o _.80 _.EN
58388: Zona—
Umzw 98 98 _.8 _.23 9o00 98m 98 980 98 _.ofi 98.; ~80;
Eé 98 98 98 _.S m 908 :80 -98 98 _.~_ 908 9%: :08
>8 98 9 _ x _.mm _.88 980 _.08 9 _ o 9 _ u 9%. C _o 98¢ :88
mega rccnomnos 9 I 98 8.0%.... :3 _.08 _.NN-\ 900 98 8.3.3.. _.80 _.80 :3
Ume x moan—reggae: 98 98 98 _.8_ 90M— :8 -98 98 98 908 9w0¢ _.og.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

_.. h A .8. "fie-82mm. _._... m A .or ”So-3:3.

 

 

233

 

Hugo 8

rommmao wamammmo: 32:38 manna—8 :5. Ed Emma? <22; 205E803 I 0036832588 Zona—

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Zaam moBEom

8 mm $55 0w 8.8 Q . 8 mm ’58 OW 088 Q
Um?” 98 98 _.8 N. _ 8 980 m. _ 8 -9 _ u 98 98 988 988 N. ~ 08
9’: 9o— 98 9o. _.o8 9m0~ :8 -98 98 _.N0 908 988 _.8N
>ma 9 _ a 9B 98 _.NS 0808 _.m fl 8 9 _ 0 9 _ q _.8 _.N8 988 _.80
78C 00800325 92 92 _.N0 _.o: 908 _.ouo o.oo 93 9.8 _.08 908 :08
Um?” x 286 0380303 92 98 98 :08 98c :08 -98 92 who 908 908 _.08
mam“ manna 98 93 8.8... 8:0 :08 :08 98 92 8.8.... _.08 _.8o _.88
Um?” x mom" wing 98 92 98 _.oou 980 #88 -98 9S _.8 908 9080 _.08
00> -_.8 9-2 who 98w 988 ~08 -98 98: N8 988 980 :3
Um?” X 00> 98 _.No 980 _.08 988 N988 -98 98 98 908 98_ 8.88
H250: ”@9830: 98 98 _.8 #88 908 :8 98 98 8.8.3.. :8 _.08 .88—
szw x 838: ”@9830: 98 98 9.5 _.80 908 _.80 -98 98 8.8... 988 9.88 908
moor: r5188: 98 98 r00 #88 98c :00 -92 98 98 908 930 78¢
Um?” x moaa rccaomao: -92 98 9S 900a 988 _.80 98 98 98 :08 908 _.N8

 

 

 

 

*5 A .8. Eco-8:3. _._... h A .3. "So-8:3.

 

 

234

Dijferentiation & Perception of Best Friend at MSU ’3 Alcohol Use

Males. The Additive Model failed to reach a level of significance (38(4) = 6.130,
p = .190, see Table 72, p. 238). The Interaction Model also failed to reach a level of
significance (38(5) = 6.219, p = .286, see Table 72, p. 233).

Females. The Additive Model with perceptions of their best friend at MSU’s
alcohol consumption was significant (78(4) = 15.09, p < .001, RIC“ & 5,," = .11, RZNmumc
= .15) and had a successful prediction rate of 66.9%, an increase 16.2% over the Basic
Model. Students that had one unit higher in perceptions of best fi'iend use were 1.026
(95% CI = 1.009, 1.042) times more likely to be an Occasional Binge Drinker versus a
Nonbinging Drinker (see Table 72, p. 238). Students that had a higher level of perception
of best fi'iend use equivalent to one standard deviation (41.48) were 2.82 times more
likely to be an Occasional Binge Drinker versus a Nonbinging Drinker.

The Interaction Model also was significant (78(5) = 16.62, p = .005, RIC... & 5,... =
.11, RZNWe = .15) and had a successful prediction rate of 66.9%, but did not show a
block improvement over the Additive Model (12(1) = .12, p = .724). The only significant
individual predictor in the model remained perception of best friend use, which was
nearly identical in value as the Additive Model (OR = 1.025, 95% CI = 1.009, 1.042, see
Table 72, p. 238).

Dlfierentiation & COA Status

Males. The Additive Model failed to reach a level of significance (38(4) = .59, p =

.96, see Table 73, p. 239). The Interaction Model also failed to reach a level of

significance (76(5) = .66, p = .985, see Table 73, p. 239).

235

 

 

Hugo ac

rommmzo Womammmos 39:35 wficmmim :5" can 088ng gamma <22; 2039508 I 92083330:

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Zaom «nae—om
6 mm find—E OW ome\o Q 6 mm «Sea 0” 8X 9
Ume chm obo of» _.wNA 0.2: NSN .95 cm; 9.3 onao ob: _.wmo
wZH o.oo o.oM o.oo 99% 9mg 2 I -98 o.oc ouw o.owo 93w fog
>mo .98 93 0.3 903 9ko _.woo o.om 93 PM; _.owm PmNm _.ANa
ate A .8. "$5-8:me in A .3. 23-x: on.

 

 

 

How—o .5

bommmao anqommmoa 383:5 manners. :5» m8 08853. wmsmoa 585 Zoacmnmoa l mos. 1883

 

 

 

 

 

 

 

 

 

 

 

 

 

3E8 moan—flow

6 mm (<39 OW 3.x. 9 9 mm fins—E OW 3.x. Q
magmas». wowomSQ -93 of 9: 9w: chow #03 Pom obo PS #on ob: _.oom
J: menace chm coo TS #qu 9qu “$3 -93 Pm» New PM: oboe _.oua
wagon». 0:8? PNN 93 93 #th ob: Nuwm ob» ohm 9mm rwaa @qu No:
558: in. 0963 -93 chi o.oo 9on obs PAS .98 93 Pom 903 ohm“ _.oa
w?= -obw o.oo o.oo 903 Pm: room .93 o.oo _.3 90% Puma— #80
>ma -obA obo o.om o.oum Pat. {bu o. fl o o. 3 ohm :8 PEG _.Aua

 

 

 

 

 

 

 

 

 

a .u A bu. 35-3.8. IbA .2. 35-3159

 

 

236

 

flmEo .2

rommmao Womammmoa 32:35 £525 :8: Ed 0888:»— wwsmoaw <28; Zoaumsmoa 1 wonoouaoam om ch

manna Cm»

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

:88 _ won—£8
8 :3 €65 _ ox _ 89.9 _ 8 _mm _ £2: ox _ 889
wwmmo gong
em; 98 98 9.3 $8 99: $8 -98 9: 98 98o 9a: 58
we: 98 98 98 988 88 _._: -98 98 98 98o 88 :5
>8 -98 98 98 99: 9an C8 98 9: 98. S8 988 :8
>aa5<n 30%.;
em; 98 9.: 9t 38 988 ~88 -98 9: 98 92a 98a 58
we: 98 98 92 98 38 :8 -98 98 9.: 988 92: 98..
>8 -98 9B 98 98a 98» $8 98 9: 98 23 9x; :8
78¢ can 92 92 98 S8 998 :8 92 9S _.8 S8 9er S;
38330: 325
am; 93 98 _.8 38 993 988 -9: 9: 98 98a 988 58
we: 98 98 9S 98o 98¢ 28 -98 98 98 9o: $8 98““
>8 -98 98 98 98“ 986 C8 98 9: 98 _.8_ 986 2.:
am: can 92 92 9t 28 923 _.8a 92 92 _.2 S8 98 28
3:; 78: can 98 98 _.8 :5 923 IE -92 92 98 98 9c: :8

 

 

 

 

 

 

 

 

 

 

 

 

 

 

_._fi A .3. 35.823. .1. .8 A .3. "So-8:8.

 

 

237

 

Hugo .5

rommmzo ”3385: 38838 9338 98 m8 088338 gamma 885.. 20:28on I $80388 om mom” man—aw can

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

_ 3&8 _ mange—am
_ 8 3mm _ sea _ ox _ 889 a 8 _ mm _ éoa _ ow 9 8.89
m8? Zona—
UmH-W 98 93 9.6 _.wNA 9m3 905 -98 93 98 9.8a 9.2; 3.80
35 o.oo 98 o.oo 908 9mg 3 Z -98 98 98 93c 98m 90%
>8 -98 98 98 983 988 88 98 9 3 A 9: _.omm 98m 3.3m
.2535 Zeal
UmZN 93 9A— 98 _.Am-N o.oon 8.8.: -93 98 o.oo 98A 988 _.wow
2": -93 o.oo 93 98a 9me room -98 98 9% 93¢ 98o _.otw
>8 -93 93 93 93c 93a _.38 98 9G 93 reg 93w {Ga
wow” magma Gun 98 93 PK... 33 m _.o3 98¢ 98 93 9: a... _.ono _.ooo 995
588030: Zona—
UmH-W 9% 9.3 o.oo _.ucx 928 No.3 -98 980 98 9%: 9w _ m _.Aum
E": -93 98 98 998 98a :8 -98 98 9?. 93¢ 9ch 903
>mo -9: 9B 93 980 93a _.w: 98 9 _ m 98 #28 9x _ m _.aw
wow» magma 9.6 98 93 Ana... 93¢ _.o3 _.8_ 98 93 951. 9on race _.OAN
03-x x wow» magma Cmo o.oo 93 o.oo :5; 993 H.8N -93 93 9S 93m 90% _.cww

 

 

 

 

 

 

 

 

 

 

 

 

 

 

sh A .8. 36.828. _._... N» A .3. 35.8on.

 

 

238

 

Ania d

rommmao wamaammmoa 3.23953 3:538 35» man 008925. wmsmaam 888 20:33:33 I 00> 388m

 

_ 3&3 _ maBEom
_ 3 _mm _€m_a90x_ 3:9 _ @ _mm_€%_ 9: 3:9

 

wgmo gono—

 

UMZN 9mm 93 93 _.uNA 938 PEN -93 9B 98 983 93A #33

 

wzm 933 98 93 983 933 _._: -98 93 933 933 933 9.33

 

>mo -98 93 98 903 933 “.303 98 9: 93; 9.38 933 #33

 

>335 Kean.

 

Um?” 93 93 933 _.83 938 No3 -93 93— 93 938 933 9.30

 

w?= o.oo 933 o.oo _.333 933 :5 -98 93 93 9033 9303 .33

 

>mo -98 93 933 903 933 _.303 93 9: 98 _.33 93: _.AE

 

00> -93 933 93 933 983 N33 -92 93 _.33 938 983 _.33

 

583030: gono—

 

UmEN 9m» 93 93 3.33 9H3 3.33 -95 98 933 935 933 3.38

 

W73 93 933 o.oo 903 933 _._: -98 93A 933 9030 9303 _.33

 

>mo -98 93 93 983 933 _.3~ 93 9: 98 _.3N 93: _.AE

 

00> -98 933 93 93: 9N3 N30 -93 93 3.3 933 9qu _.Aoo

 

 

 

 

 

 

 

 

 

 

 

 

 

UMH-WX 00> 9mm _33 93 _.38 93.3 3.3: -98 98 o.oo 983 9:6 A33

 

 

*5 A .8. "so-8:8. 2.. w A .or «fie-8:09

 

 

239

Females. The Additive Model with COA status failed to reach a level of
significance (12(4) = 2.792, p = .593, see Table 73, p. 239). The Interaction Model also
failed to achieve a significance as an overall model (38(5) = 2.795, p = .732).
Difierentiation & Expectations of Alcohol Serving as a Tension Reducer

Males. The Additive Model failed to reach a level of significance (x2(4) = 8.94, p
= .063, see Table 74, p. 242). The Interaction Model was significant as an overall model
(x2(5)=11-19,p = .048, slams“: .12, Rimm, = .16) and had a successful
prediction rate of 75.8%. Although differentiation approached significance (p = .052),
expectations of tension reduction remained the only significant individual predictor (OR
= 1.108, 95% CI = 1.030, 1.191) of Occasional Binge Drinkers versus Nonbinging
Drinkers (see Table 74, p. 242).

Females. The Additive Model with tension reduction expectations was significant
as an overall model (38(4) = 21.62, p < .001, RZCOuSndl = .14, R214“: = .19). The
model had a successful prediction rate of 66.2%, an increase of 14.8% over not having a
model. Expectations of alcohol serving as a tension reducer changed the odds of being an
Occasional Binge Drinker versus a Nonbinging Drinker by 1.117 (95% CI = 1.060,
1.177) for a one unit difl‘erence in expectations (see Table 74, p. 242). By changing the
difference in expectations to a unit equivalent to one standard deviation for expectations
(8.81) there is a 4.13 times greater likelihood of being anOccasional Binge Drinker.

The Interaction Model also was significant overall (38(5) = 24.15, p < .001, RIC“
& 5...," = .19, Rznmn‘m = .27) and had a successful prediction rate of 64. 1%; however, it

did not show a block improvement over the Additive Model (22(1) = 2.52, p = .112).

240

Expectations of tension reduction remained the only significant individual predictor (OR
= 1.114, 95% CI = 1.056, 1.176, see Table 74, p. 242).
Diflcrentiation & Expectations of A lcohol Serving as a Social Lubricant

Males. The Additive Model with social lubrication expectations was significant as
an overall model (78(4) = 10.12, p = .039, ale,” 3...“ = .11, Emma... = .15). The model
had a successfill prediction rate of 70.3%, an increase of 4.4% over the not having a
model. The only significant individual predictor was expectations of alcohol serving as a
social lubricant. Students that had a one unit higher level of expectation were 1.118 (95%
CI = 1.036, 1.206) times more likely to be an Occasional Binge Drinker versus a
Nonbinging Drinker (see Table 75, p. 243). Students that had a greater perception equal
to one standard deviation (7.89) were 2.40 times more likely to be Occasional Binge
Drinkers.

The Interaction Model failed to reach a level of significance (13(5) = 10.21, p =
.070) and had a successful prediction rate of 71.4%. Expectations for social lubrication
remained the only significant individual predictor for the sample (OR = 1.118, 95% CI =
1.036, 1.206) in predicting Occasional Binge Drinkers versus Nonbinging Drinkers (see
Table 75, p. 243).

Females. The Additive Model failed to reach a level of significance as an overall
model (38(4) = 6.094, p = .192, see Table 75, p. 243). The Interaction Model also failed
to reach a level of significance (38(5) = 7.056, p = .216).

The Comprehensive Model
Males. The Comprehensive Model failed to reach a level of significance (x203) =

19.269, p = .115, see Table 76, p. 245).

241

 

Hugo .3

rommmno ”anaemic: 32:35 magmas. 5m» man Oonmmmosfi gamma <22; Zoacmsmonm l mxuoofimosm om. Adamo: Won—come:

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

_ 2&8 _ moBEam
_ a _ mm _ sea _ ow _ 8&9 _ a _ a: 55 _ ow _ 8&9
wean Zena
DE.” 98 9.8 93 8mm 92: 905 -98 9“: 98 98m 9A: 88
El: o.oo 98 o.oo 93m 98.» _._ E -98 98 9mm 99.5 998 _.oao
>ma -98 93 98 993 988 83 98 9: 98A _.omm 93m _.ama
>355 332
02% 98 98 98 Emma 9.3m 988 -98 98 92 903 9m8 _.mmo
ES— 98 98 98 _.8A 903 :8 -98 98 9 3 998 92: _.cum
>mo -93 9B 98 93m 92m _.Aoo 9; 9G _.qu :3 9mg rm:
Anammo: Waggon 93 98 98...... :8 _.ouq :8 o; _ 98 3.81;. _._ 3 _.oao :3
58398: Zona—
Umzw _.oo 9% 8.3 NQNN 90m: 93¢ -9»; 9mm 98 9:3 983 _.aoo
ave: 98 o.oo 98 _.80 903 :8 -93 98 93 90mm 98w _.owm
.98 -95 92 98A 923 9m8 _.wt 9; 9G _.3 _.N8 9mg _.omn
Hosea: ”3:98: 98 98 98...... :8 E80 :3 9: 98 3.3.... C E memo :uo
Um?” x fiasco: Waggon 98 98 PM» font: 93A _.NNN -93 98 who 98o 9mg _.Sm

 

 

 

 

 

 

 

 

 

 

 

 

 

 

_._ me A .8. 26-889 _._... E A .o 7 25-889

 

 

242

 

Haze 38

hommmao woman-$3: 33838 93038 33 80 08382: 959.5 58% 22.3803 1 mxvooSmoam em wooE reggae:

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2E8 a 12:38
a a :m _ éaa _ owfl axe _ e _mm _ fie: 9: case
888 Zona—
Ume 98 93 9A0 _.uNA 92: N05 -98 9”: 98 938 9A: _.wmo
W72 o.oo 98 o.oo 900m 9m3 _._ _ _ -98 98 9% 93¢ 9m0m _.80
>8 -98 9.8 98 903 998 _.88 98 9 E 980 _.08 988 ~08
>833 Zona—
UmH-W 900 98 Pt _ .000 98; A308 -900 93 93 90 _ 0 93m _.33
m2: 98 98 98A #8.. 908 :3 -98 98 9N0 93m 908 .08
>mo -92 9B 98 903 988 ~03 98 9: 9M: :80 980 _.800
mega Enigma: 9: 98 98...... Z 5 _.8o _.Nca 98 98 #8. _.8m _.ooa _._ 2
38898: Zone
Gm?” 93 93 N3 MLE 930.. 8.38 -98 98 9.: 9303 980a _.aow
ES— 98 98 98 _.08 908 :8 -98 90A 9 _ 3 908 903 _.08
>8 -98 9B 93 90M: 980 _.Awo 9: 9E 9m0 :8 98A _.m_0
meow». 551830: 9: 98 9mm: : a 0.8m _.N3 98 98 Fun... _.88 _.oou _._ _o
Umrw x wage—haulage: 98 98 98 _.30 908 :8 -98 98 98 98A 980 roam

 

 

 

 

 

 

 

 

 

 

 

 

 

 

an A .8. 3.5-8:8. new A .8. vac-8:8.

 

 

243

Females. The Comprehensive Model was significant according to the model chi-
square (x203) = 45.05, p < .001, R2c°x & 5...,- = .27, RIM-m. = .36) and showed a
significant improvement over the Basic Model 0800) = 43.65, p < .001). The model
illustrated a 74.6% success rate in predicting the correct drinking classifications, an
improvement of 23.9% over the Basic Model. Perceptions of best friend use (OR =
1.030), expectations of tension reduction (OR = 1.147), and the difl'erentiation X
expectations of tension reduction (OR = .861) were the only individual variables to
significantly improve the odds of being classified as an Occasional Binge Drinker versus
a Nonbinging Drinker (see Table 76, p. 245).

Frequent Binge Drinkers versus Nonbinging Drinkers

Dlfierentiation and binge drinking

Males. The Basic Model was not significant according to the model chi-square
(78(3) = 2.48, p = .479, see Table 77, p. 246).

Females. The Basic Model was not significant according to the model chi-square
(x20) = 5.81, p = .121, see Table 77, p. 246).
F our- factors of the DSI-R with binge drinking

Males. As was the case in the Basic Model, the overall four-factor model failed to
reach a level of significance (12(6) = 3.78, p = .706, see Table 78, p. 246).

Females. The overall four-factor model failed to reach a level of significance

(x2(6) = 7.39.12 = .286).

244

 

H35 .3

homage Womamao: wnommoaam mamas? Em» 30 0883:»— wEMo—a <99; Zoacmnmanm I OoBEoroammé Zona—

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Zaom Edam—om
u mm €95 OW 8.x. 9 w mm 955 CW 3.x. Q
Ume -93 9% “.8 93a 9 a AN flhoa -93 9mm 93 923 9mg .2 u
wz: 9 # a 98 NS :3 93m _.Aoa -93 93 93 993 9%: rod
>mo 98 93 93 _.omq 9q~m _.aww 9B 9: r3 _.an 93w _.‘Do
Em: 38828.4. o.oo 93 92 980 923 _.SN 93 93 93 _.com 93; re;
02% x ch waaobaosm -98 92 Wow 993 98c _.oou -93 93 98 98. 903 _.So
wow» «123 98 9o— 5?“ _.owa 5o: _.oww 98 9S 9N5: _.owo I: o fog
‘39
Ume x mam" $55 -93 93 93 93M 98m re; -93 98 93 93m 90% _.So
00> -Za 99). who 9w; 993 _.oqa -93 9mm _.ao 9%: 93a _.ANN
02%, X 00> _.ww _._m _.uw Put 9mg $.25 -93 93 93 9%: 9:6 .33
Adamo: ”098:0: 90¢ 93 93 rowq 98a :3 9E 93 3.8... :3 _.oau _.Nuw
§
02% x Anzac: ”3.830: 9: 9: o.oo _._ G 9m£ 7qu -9; 93 a; _._ 9M3— 9.Rm 98m
mega r5183: -93 93 98 903 93w _.omm -98 9E 9: 93m o.ooo fog
Ume x moomm— r5183: -93 93 9B 9?: 93¢ _.Now 98 98 9mm _.oma 9o: rmmm

 

 

 

 

 

 

 

 

 

 

 

 

 

 

_._ u A .8. 25-8:8. _._... E A .2. 25-823.

 

 

245

 

ngo .3

rommmao Womammmo: 32:33 mange :5" m3 33:9: wmamaa 835m Zoacmsmoa I 0520:3303

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

3&8 moaaam
w mm 9&5 OW 3.x. 9 v mm «SHE OW 8.x. 9
Umzw -93 9mm 93 9‘2 9?: _.MNA -95 9mm 93 9me 93A _.omw
ES“ -93 o.oo 9mm 93m 93w rouo -93 93 PB... 9on 93m 90:
>mo 9- 9: _.3 _.Ng 9mg ruww 98. 9G 93 _.Sm 9am _.wmo
_._ A bu. 25-8:3. 2;» A .3. 2:13—09

 

 

 

Haw—a uw

homage Womnommmos vammomsm mange 9E 98 man—52 9508 838 Zea—gmamoa I «9: 3888

 

 

 

 

 

 

 

 

 

 

3&8 moans—om

w mm €95 OW 8.x. 9 v mm £35 ow 3.x. 9
mBomoaa WomomiQ 9mm 9mm 9mm rwum 93m ~03 9G 98 9N” :fi 93m _.mfi
2H: vomaos .98 93 9mm 9mm. 93» _.mqa 93 9mm o.oo re: 9903 _.EN
magmas»— OEom. -93 9u~ 93 9H3 93a Emma 98 98 98 room o.ouo _.ud
mama: £9 0:83 .95 9.8 9? 93w 93o rcmm -93 9: _.3 9?: 93a ”bum
BE -93 98 9% 93m 9mm.» 2:: -9 _ w 93 a. _ m. 93m 93; 93m
>mo 98 9: E8 _.Nwm 923 _.qmw 98 9G 98 _.omo 9&0 3,:

 

 

 

 

 

 

 

 

 

 

 

 

_._ h A .omv 25-8%9 IE A .3. 25.4359

 

 

246

Dijflrentiation & perception of average MS U student ’3 alcohol use

Males. The Additive Model was significant as an overall model (x2(4) = 10.96, p
= .027, R20”, & 3...,“ = .08, [€sz = .13). The model had a successful prediction rate of
76%, an improvement of .8% over having no model. Perceptions of MSU student use was
the only significant individual predictor. Students that had a one unit higher level of
perception were 1.022 (95% CI = 1.004, 1.040) times more likely to be Frequent Binge
Drinkers than Nonbinging Drinkers (see Table 79, p. 248). Students that had a greater
perception equal to one standard deviation (37.22) were 2.27 times more likely to be
Frequent Binge Drinkers.

The Interaction Model was significant as an overall model (12(5) = 11.20, p =
.048, R2“x & 5...," = .09, Rznwm = .13), had a successful prediction rate of 74.4%;
however, it did not show a block improvement over the Additive Model (380) = .24, p =
.624). Perceptions of MSU student alcohol use remained the only significant individual
variable in predicting Frequent Binge Drinkers versus Nonbinging Drinkers (OR = 1.023,
95% CI = 1.005, 1.043, see Table 79, p. 248).

Females. The Additive Model with perceptions of the average MSU student’s
alcohol consumption was not significant as an overall model (12(4) = 7.850, p = .097, see
Table 79, p. 248).

The Interaction Model was significant as an overall model (13(5) = 14.01, p =
.016, RICO, & 5m“ = .10, Rzumn‘m = .13) and had a significant block improvement over
the Additive Model (x2(1)= 6.16, p = .013). The model had a successful prediction rate
of 63%, which was a 4.3% improvement over the Basic Model and 11.6% over having no

model. The differentiation X perception of MSU student use was the only significant

247

 

#63 ‘3

rommmao Womammmoa vaoamflim $5.38 :5” 2.0 33:02 wmnmoa <oB=m Zoacmsmonm I $28325 om ch magma

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

C3
258 _ manHam
w _ mm H $35 _ OW _ 8.x. 9 _ w _ mm _ €95 _ OW _ 8.x. 9
mama Zeno.
Umzw -93 9mm 93 9w- 9?: fun; -95 93 93 9wa 93A _.amm
ES— -92 9cm 9% 90mm 93w _.oqa -9 _ u 98 Ab— . 930 9.3m 98A
>mo 9- 9 H u _.i _.Nao 9wo~ rumw 9S 9 _ m 93 _.S m 9.3m _.umo
25:76 Zona—
Umzw -98 9.5 9.3 930 93a _.Ew 9% 9w» 93 93 x 93m _.uwm
mg“ -98 93 :3 90: 9mg _.ouc -9 _ N 98 93 9a: 9.30 :5“.
>mo 93 9; NS _.Noo 98m _.mum 98 9G 9? 2H: 93m _.wmu
ch C8 98 9S o.oo. _.oum _.ooa fog 93 9S _.8 _.ooq 903 _.o:
38898: :32
Um?” -9 _ m 93 9: 93; 9a.: No: -9NA 9w» 9km 93o 9.8m #qu
95 -98 93 :3 93 m 9mg _.oao -9 B 93 Pun 933 9.3a room
>ma 9g 9; NE “New 903 _.wwq 98 93 93 room 93o TEN
ch C3 98 9E 93... fonu _.oom #96 93 93 Now _.oom 903 _.So
Ume X ch Gun 93 98 9~A rcow 903 _.owo -98 93 9%... 93m 98m 908

 

 

 

 

 

 

 

 

 

 

 

 

 

 

_._Lu A .3. 25-859 _._... w A .S. go-$=om.

 

 

248

individual predictor (see Table 79, p. 248, and Figure 6, p. 73, for interaction effect).
Students with a one unit higher interaction value was .975 (95% CI = .955, .996) times
more likely to be a Frequent Binge Drinker than a Nonbinging Drinker. Students that had
a difference in the interaction term equivalent to one standard deviation (21 .57) were .58
times more likely to be a Frequent Binge Drinker than a Nonbinging Drinker.
Dijferentiation & Perception of Best Friend at MSU '3 Alcohol Use

Males. The Additive Model with perceptions of their best friend at MSU’s alcohol
consumption was significant (38(4) = 25.61 , p < .001, RZCOu Sm“ = .13, Rzumm = .22)
and had a successful prediction rate of 88.8%, which was 13.6% greater than having no
model. Students that had one unit higher in perceptions of best fi'iend use were 1.043
(95% CI = 1.024, 1.062) times more likely to be a Frequent Binge Drinker versus a
Nonbinging Drinker (see Table 80, p. 251). Students that had a higher level of perception
of best friend use equivalent to one standard deviation (72.43) were 20.94 times more
likely to be a Frequent Binge Drinker.

The Interaction Model also was significant (33(5) = 47.43, p < .001, RZCO,‘ & 5nd1=
.32, RZNW = .47) and had a successful prediction rate of 88.8%, but did not show a
block improvement over the Additive Model (x20) = 2. 10, p = .148). The only
significant individual predictor in the model was perception of best friend use (OR =
1.046, 95% CI = 1.026, 1.067, see Table 80, p. 251).

Females. The Additive Model with perceptions of their best fiiend at MSU’s
alcohol consumption was significant (38(4) = 61.52, p < .001, RZCO,‘ & 5n,“ = .36, Rznmmm
= .48) and had a successful prediction rate of 82.6%, an increase of 23.9% over the Basic

Model. Students that had one unit higher in perceptions of best friend use were 1.052

249

(95% CI = 1.032, 1.071) times more likely to be a Frequent Binge Drinker versus a
Nonbinging Drinker (see Table 80, p. 251). Students that had a higher level of perception
of best fiiend use equivalent to one standard deviation (41.48) were 7.96 times more
likely to be a Frequent Binge Drinker versus a Nonbinging Drinker.

The Interaction Model also was significant (38(5) = 65.35, p < .001, 122“” gm" =
.38, RZNmIMC = .50) and had a successful prediction rate of 81.2%, but was only on the
hinge of showing a block improvement over the Additive Model (38(1) = 3.84, p = .050).
The only significant individual predictor in the model remained perception of best fiiend
use, which was nearly identical in value as the Additive Model (OR = 1.057, 95% CI =
1.036, 1.079, see Table 80, p. 251). Additionally, the interaction term nearly reached a
level of significance (p = .054).

Difirentiation & COA Status

Males. The Additive Model failed to reach a level of significance (36(4) = 3.26, p
= .516, see Table 81, p. 252). The Interaction Model also failed to achieve a significance
as an overall model (38(5) = 3.477, p = .627).

Females. The Additive Model with COA status failed to reach a level of
significance (38(4) = 5.809, p = .214, see Table 81, p. 252). The Interaction Model also
failed to achieve a significance as an overall model (78(5) = 9.296, p = .098).
Dzficrentiation & Expectations of Alcohol Serving as a Tension Reducer

Males. The Additive Model with tension reduction expectations was significant as
an overall model (38(4) = 28.32, p < .001, R2“, at s..." = .20, 182an = .30). The model
had a successful prediction rate of 80.8%, which was a 5.6% improvement over having

no model. Expectation of alcohol serving as a tension reducer was the only

250

 

Hue—o we

rowmmmo WomRmmmo: 3.8335 9538 93 m8 33:2: gamma 332m 2032.508 .- womoowaosm 0». wow” magnum Cmo

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

_ 3&8 _ mnBEom
_ w _ mm _ fima _ OW _ 8.x. 9 _ w _ mm _ $35 _ OW _ 8.x. 9
momma Kama
Um?” -98 93 9S HEB 9?: fun.“ -95 93 93 9mm.» 935 ram
W7: -93 o.oo 93 90mm 9wa fio-R -93 9cm PE... 930 9.3m 98A
>mo 9M“ 9 _ a :3 _.Nao 9mo~ TEN 9S 9 _ m 93 _.S m 93m _.umo
>aa§<a Zona—
Ume .93 93 _.mm 93.. 930 rug-N 93 9t 93 :3 9MB Name
ES. -98 93 _.Nm 93 m 93w _.oaa -93 93 9% 903 93 _ _.oco
>mo 93 9NN Nd _.fic 98m Nb? -9 _ m 9 _ c 9% 93— 9meo _.Nfi
mam» magma Cmo 9o; 93 ~93: _.OAu _.cma Team 93 93 ~93...» _.omn _.owm _.o-Z
58898: 30%;
Um?” 9; 9mm 93 :3 9w; PM: -93 9A”. 95 993 9me No.3
wz: -98 98 :w 905 9.3; _.o-Z -98 9cm 93 93M 9%: _._ 3
>mo 93 98 um— rmec 996 N330 -9 E 9 _ 0 93 93c 9an _.Nma
mom» manna—d. 68 93 93 ~93: _.oao .bwa _.omq 93 93 3.3-_._... remq _bua _.oqo
UmH-W x mom“ magma :8 o on 98 Now _.omu 903 fame -98 98 um: 90.: 993 roo—

 

 

 

 

 

 

 

 

 

 

 

 

 

 

._ me A .8. 23.828. 2.... u A .3. ”so-8:3.

 

 

251

 

H35 3

rommmao ”$2890: 32:38 9338 an: m8 33:2: 9508 S55 Zoagamoa I 00> manna

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

_ 2E8 _ moan—om
_ w _ mm F «ii ox _ 8.x. 9 _ w _ mm # é»; _ ow _ 8.x. 9
mama Zona—
Umrw -93 9mm 93 933 9%: run; -98 98 9: 9me 93A raw
E": -92 98 98 90mm 93.). _.8m -93 98 PM: 930 9.3x 98a
>mo 98 9 fl -\ _.g _.NAo 9ch _.38 98 9 _ m 93 I: m 93mm _.wmo
>355 Kong
DmH-W -98 93 9% 933 930 _.uum -98 98 93 933 9A.; _.omm
ES” -98 98 9.: 93m 9me #28 -9 H u o.oo AB... 930 93m 903
g 98 93 #3 _.Nfi 9mm.» .38 98 93 9S rem 933 rum—
OO> -93 98 9d 930 9 _ ma _.93 -98 93 o.oo 9on 93¢ 8.38
383030: 30%
Umrw -98 93 9mm 9%; 93¢ ~03 -93 93 _.u-w 92$. 9H: flag
9": -92 9cm 9.5 903 933 fomw -93. 93 93... 93a 9.3m 9o8
>mo 98 93 #3 _.Na 9w? ~33 9S 9; 93 _.oz 933 5.3
00> -93 92 93 933 93w r80 -93 9% 92 90% 93» N8;
Umrw X 00> 98 98 9B rm: 983 SEC #3 9% .98 A33 9w$ 3.88
*5 A .3. 25-839 In A .3. arc-8:3.

 

 

 

252

significant individual variable. Students that had a one unit higher level of expectation
were 1.163 (95% C1 = 1.088, 1.244) times more likely to be a Frequent Binge Drinker
versus a Nonbinging Drinker (see Table 82, p. 255). Students that had a greater
perception equal to one standard deviation (8.73) were 3.74 times more likely to be Binge
Drinkers. Additionally, age was a significant factor, as students that were 1 year older
were 1.594 (95% CI = 1.069, 2.376) times more likely to be Frequent Binge Drinkers.
Students that were two years older were 2.54 times more likely to be Frequent Binge
Drinkers.

The Interaction Model also was significant overall 08(5) = 29.83, p < .001, RZCO,‘
& 5...," = .21, from“, = .32) and had a successful prediction rate of 78.4%; however, it
did not show a block improvement over the Additive Model (380) = 1.51, p = .219). Age
(OR = 1.603, 95% CI = 1.074, 2.392) and expectations of tension reduction (OR = 1.176,
95% CI = 1.095, 1.262) remained the only significant individual predictors (see Table 82,
p. 255).

Females. The Additive Model with tension reduction expectations was significant
as an overall model 08(4) = 55.84, p < .001, R20,“ 5...... = .33, mm = .44). The
model had a successful prediction rate of 73.9%, an increase of 22.5% over not having a
model. Expectations of alcohol serving as a tension reducer changed the odds of being a
Frequent Binge Drinker versus a Nonbinging Drinker by 1.229 (95% CI = 1.139, 1.325)
for a one unit difference in expectations (see Table 82, p. 255). By changing the
difference in expectations to a unit equivalent to one standard deviation for expectations

(8.81) there is a 6.14 times greater likelihood of being a Frequent Binge Drinker.

253

Additionally, BMI was a significant contributor changing the likelihood of being a
Frequent Binge Drinker (OR = .765, 95% CI = .624, .937).

The Interaction Model also was significant overall (38(5) = 57.9l,p < 001,122“,
& gm“ = .34, from = .46) and had a successful prediction rate of 75.4%; however, it
did not show a block improvement over the Additive Model (12(1) = 2.08, p = .150).
Expectations of tension reduction (OR = 1.239, 95% CI = 1.143, 1.343) and BMI (OR =
.776, 95% CI = .633, .953) were the only significant individual predictors (see Table 82,
p. 255).
Dijferentiation & Expectations of Alcohol Serving as a Social Lubricant

Males. The Additive Model with social lubrication expectations was significant as
an overall model (36(4) = 25.63, p < 001,18“, g 3,.“ = .19, 1(2an = .28). The model
had a successful prediction rate of 80%, an increase of 4.8% over not having a model.
Expectations of alcohol serving as a social lubricant (OR = 1.182, 95% CI = 1.093,
1.279) and age (OR = 1.601, 95% CI = 1.063, 2.411) were significant individual
predictors of Binge Drinking category (see Table 83, p. 258). Students that had a greater
expectation of social lubrication equal to one standard deviation (7.89) were 3.74 times
more likely to be Frequent Binge Drinkers. Additionally, students that were two years
older were 2.57 times more likely to be Frequent Binge Drinkers.

The Interaction Model also was significant overall (38(5) = 26.52, p < .001, RIC...
& 5.," = .19, 1?sz = .28) and had a successful prediction rate of 79.2%; however, it

did not show a block improvement over the Additive Model (x2(1)= .89, p = .345).

254

 

 

Haze mm

rommmno Wannammmos wnoamomam £333 93 man 33:2: 95on 388 Zoacmamoa l mxvoonmmoam om H038: Won—5:0:

 

 

 

 

 

 

 

_ K38 _ moBEnm
_emmm_€aa_ow~e§9_e_mm_€sa_ow_ ease
wean Zona—
cm; -93 9mm 9d 38 out: 3% -95 8M 9: 33 on: _.am
we: -92 98 can 8% Sam _.So -9: o.oo an: case 3; o.oea
2a SN 9: _.ma 5.5 38 ES 98 9; 92 _.o; 3% 5%
>355 Kean—

 

UmTW 9mm 93 93 _.wna 9m: 920 9.: 93 PM: PEN 9ch era:

 

W?= -98 90m 93 90.: 93a _._3 -93 95 92¢... 9.5m 9a: 903

 

>mo 93 93 MB... _.moa _.omo No.3 -95 98 9c— 9ww_ 9me _.N:

 

Hoammoawacoaos 93 98 3.8.... _.au romm H93 95 90A N931. _.Nno #30 _.wNm

 

588030: zona—

 

Uwzw 93 93 _.Ne _bwm 93A FM: 9S 9Au Na Nome 9ch 9qu

 

W74: -99 9cm 98 90.3 9mm~ _.coA -9~m 93 9mm... 9.3a 99am 98w

 

>mo 93 98 mm»... _.gw _.o: when -93 98 93 9mm~ 9mg _.NAo

 

H253: ”3:030: 93 92 ~98: _.Sm _.oem HRS 93 9o... 3.3.1 _.Nwo _._Aw #qu

 

 

 

 

 

 

 

 

 

 

 

 

 

Umzwx H958: Waggon 93 com #3 fog 90am _._wm -98 98 fee 9on 93¢ To”:

 

 

_.. A .3. 25-3:59 _._... no A .3. 25-859

 

 

255

Expectations for social lubrication (OR = 1.192, 95% CI = 1.099, 1.292) and age
(OR = 1.634, 95% CI = 1.080, 2.472) remained the only significant individual variables
in predicting Binge Drinkers versus Nonbinging Drinkers (see Table 83, p. 258).

Females. The Additive Model with social lubrication expectations was significant
as an overall model (38(4) = 27.49, p < .001, 3%,, .e s...“ = .18, mm“... = .24). The
model had a successful prediction rate of 68.1%, an increase of 16.7% over the not
having a model. Expectations of alcohol serving as a social lubricant changed the odds of
being a Frequent Binge Drinker versus a Nonbinging Drinker by 1.144 (95% CI = 1.074,
1.218) for every one unit change in perceptions (see Table 83, p. 258). Students with a
difference in expectations equal to one standard deviation (7.74) were 2.82 times more
likely to be a Frequent Binge Drinker. Additionally, BMI was a significant contributor in
predicting Frequent Binge Drinkers versus a Nonbinging Drinkers (OR = .855, 95% CI =
.731, .999).

The Interaction Model also was significant overall (£6) = 28.19, p < .001, R20”,
& 5...." = .19, RZNW, = .25) and had a successful prediction rate of 68.8%; however, it
did not show a block improvement over the Additive Model (x20) = .70, p = .404). BMI
was no longer a significant individual variable with the addition of the interaction term
(see Table 83, p. 258). Expectations for social lubrication was the only significant
individual predictor (OR = 1.145, 95% C1 = 1.074, 1.221) in predicting Frequent Binge
Drinkers versus a Nonbinging Drinkers.
The Comprehensive Model

Males. The Comprehensive Model was significant according to the model chi-

square (x2( 13) = 64.24, p < .001, RIC“ & 3m" = .40, Rznmmm = .60). The model had an

256

88.8% success rate in predicting the correct binge drinking classifications, which was a
13.6% improvement over having no model. Age (OR = 2.431) and perceptions of best
friend use (OR = 1.040) were the only individual variables to significantly improve the
odds of being classified as a Frequent Binge Drinker versus a Nonbinging Drinker (see
Table 84, p. 259).

Females. The Comprehensive Model was significant according to the model chi-
square (x203) = 100.54, p < .001, RIC... & 5,... = .52, hm...“ = .69) and showed a
significant improvement over the Basic Model (3800) = 94.73, p < .001). The model .
illustrated an 85.5% success rate in predicting the correct drinking classifications, an
improvement of 26.8% over the Basic Model. Perceptions of best fiiend use (OR =
1.050) and expectations of tension reduction (OR = 1.251) were the only individual
variables to significantly improve the odds of being classified as a Frequent Binge
Drinker versus a Nonbinging Drinker (see Table 84, p. 259).

Frequent Binge Drinkers versus Occasional Binge Drinkers

Differentiation and binge drinking

Males. The Basic Model was not significant according to the model chi-square
(78(3) = 6.83, p = .08, see Table 85, p. 261).

Females. The Basic Model was not significant according to the model chi-square
(£6) = 4.588, p = .205). The model had a 52.7% success rate in predicting the correct
drinking classifications which was 1.3% worse than having no model. Additionally, there
were no significant individual predictors for the sample, although BMI approached

significance (p = .054) (see Table 85, p. 261).

257

 

Haze. 8

homage Womaommmo: 32:38 mean-:5 9m: :3 3383 agmflmwgm 283808 I 9:88:03 om moan: r3185:

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

_ :38 _ moafiow
_ e :3 sea _ow = «use P e I: 55 _9: 8.89
momma Zona—
UmH-W -98 98 9.8 93.. flew: _.MNA -95 98 9: 9me 93A _.omm
wz: -98 98 9mm 98“ meg _.o-a -98 98 PB... 9on 9.3x 9o:
.98 9mm 93 :3 _.NS :88 _QNN 98 93 98 _.SM 9.3m 8%
>833 Zena
Um?” 93 9.: 93 :3 :88 Na: 93 9mm _.8 _.m8 9qu 8.33
mg: -98 98 9 3 934 = 88 :8 -9 _ a 98 9mg. 98m 98. 980
.98 93 9B 93... :2: __=8 NA: -92 93 93 98a 93a 83
meow». reggae: 9: 98 3.8.1. :8 = :8 _.Nqo 98 98 35.... It. :8; _.NE
58303:: Kong
Ume 98 93 9mm :8. 1.88 98m 93 98 :5 88 9:8 988
ES: -98 98 98 90.8 :88 :80 -9 _ m 98 8.9 980 98m 298
>8 9% 92 9.2... _.85 _n :86 PBS -93 93 98 98w 9:0 8%
mega reggae: 9; 98 3.3.... :8 _M :8 _.Nou 93 98 3.8: :3 _.ofi _.NB
Ume x mega reggae: 98 98 o.oo :89 388 :8 -98 98 98 980 980 :80

 

 

 

 

 

 

 

 

 

 

 

 

 

.3: A .8. 3.5-8:3. _._... .u A .3. "So-8:3.

 

 

258

 

 

885 8

rommmmo wnmammmo: 32:38 mama—:8 3m» :8 33:03 9:83 «638 20:28.5 I noavnoroamE Zona—

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Kanm «2:28
w mm «$8 0% 88 Q 3 mm $35 0w 8.8 Q
Om?» 9.8 98 98 _.83 988 o.ooo -9 _ m 98 98 98c 983 w. _ 8
ES— -9 _ o 9 3 98 980 933 _._8 -98 9 I 8.3 9m _ o 93— m _.83
>mo 98 98 3.8.... N8. _.83 P83 -9 _ 3 98 98 988 938 #88
Km: 88888 93 93 93 #8” 908 _.98 -93 93 98 908 933 _.Sa
Umrw x ch 880308 98 98 98 .98 98~ _.8m -98 98 N3 98c 98.. _.o8
8% 1:33 98 9S _ _.No: _.8c 2: 3 8.88 98 93 3.8.. _._ _.8o :88 _.83
Umzw x mam: $.53 98 98 _.8 _.88 90% _.8o -98 98 98 98c 98¢ _.88
00> L .3 93 8.8 98m 98m #58 -98 98 98 93.3 9 _ 8 93¢
Um?” x 00> _.8 _.8 _.3 8.80 988 8.98 8.8 _.8 _.8 .88 988 88.83
.358: ”@988: 98 98 8.3 _.o8 98m _.8o 98 98 8.8.1. #8. _._ _ _ r88
Um?” x .338: 833:: -98 9: 98 930 938 _.NNA -98 9: _.8 98m 933 :8
won?— rccnomao: 9: 98 8.8 _._ E 988 _.N8 -98 98 98 908 983 _.omm
Um?” x moan. r5188: 93 93 93 _._8 98~ _.88 93 98 98 _._8 983 .88

 

 

 

 

_._ A .8. 2535. _._. u A .2. Eoéi.

 

 

259

F our- factors of the DSI-R with binge drinking

Males. The overall four-factor model was significant (x2(6) = 15.82, p = .015,
RICOx a 3m" = .10, Rznmm = .13). The model had the same successful prediction rate of
63%, a 2% improvement over having no model. The “1” Position subscale was a
significant predictor of binge drinking categories. Students that had a one unit higher “I”
Position score were .488 (95% CI = .268, .889) times more likely to be Frequent Binge
Drinkers versus Occasional Binge Drinkers (see Table 86, p. 261). Additionally,
Emotional Cut-off subscale approached significance (p = .056).

Females. As was the case in the Basic Model, the overall four-factor model failed
to reach a level of significance (12(6) = 9.05, p = .171, see Table 86, p. 261).
Dflrentiation & perception of average MSU student’s alcohol use

Males. The Additive Model was significant as an overall model (38(4) = 11.031, p
= .026, RZCOX a 5m“ = .069, Rzumm = .094). The model had a successful prediction rate
of 62.3%, an improvement of 1.3% over having no model. Perceptions of MSU student
use was the only significant individual predictor. Although perception of MSU student
use approached significance (p = .053), none of the individual predictors significantly
predicted being more likely to be Frequent Binge Drinkers than Nonbinging Drinkers

(see Table 87, p. 263).

260

 

 

 

 

Hugo mu

Famine Womaommmos vanamozsm mama—.8 9»: m8 man—:63 952$ 82% 0888:». Eamon“ I

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

93.23225:
3E8 mmaaom
w mm (Sam OW 3.x. 9 w mm gig OW 3.x. Q
02% -93 93 mum out obmu fog o.oo 9.3 9.3 room oboe f2:
ESH .93 93 PS 902 cud _.omm -9: o.om P: chow 9qu _.oow
>mo PNA 93 New _.Nuu 903 _.Sm .93 0.3 93 ob: 93o rug
.6 A .3. 23-6339 .1. E A .3. go-§_oa.

 

 

 

Hugo mo

rename Womaowmmon mammomnm wagon—Hm 33 m8 «Bacon. wmamonm <o~m=m 0883:»— wmamoa I moan 3883

 

 

 

 

 

 

 

 

 

3&8 moan—am

w mm (SE 0% 8.x. 9 v mm gin OW 8.x. Q
5330:». Womoaia. PAN o8 r3 _.MNA 930 N30 -98 98 Pow ohm.» chem rm:
:1 womaos bus 9”: 93%. ohm» obam 9mg 93 ohm Pd _.mqa o.oo. Pum—
mBomoaa 0:8? boo ohm w .3 odom ooaw _.o _ N -o. G PNA ohm 9me out :3 m
mama: <19 09»; -9: ox: 95 Pmom Puma _.ouu -ouso Puo _.3 Pad oowu fwd
wz: .98 93 93 o.omu chow _.ofl -o; _ o.oo 93 chew A53 _.oca
>mo obs o. ~ u #3 _.Nmu o.ouo _.omw .98 o. H A 93 935 obs _.Nwo

 

 

 

 

 

 

 

 

 

 

 

 

 

_._ h A .3. go-§_oa. I. .u A .3. 25.659

 

 

261

The Interaction Model was significant as an overall model (x-2(5) = 14.96, p =
.011, R2“, & s..." = .09, mm“... = .13), had a successful prediction rate of 64.9%, and
showed a significant block improvement over the Additive Model (360) = 3.93, p =
.047). None of the individual predictors significantly predicted binge drinking category,
although the DSI-R (p = .051), perceptions of MSU use (p = .054), and the interaction
variable (p=.061) each approached significance (see Table 87, p. 263).

Females. The Additive Model with perceptions of the average MSU student’s
alcohol consumption was not significant as an overall model (12(4) = 4.61 , p = .330, see
Table 87, p. 263). The Interaction Model also was not significant as an overall model
(38(5) = 6.280, p = .280).

Difi'erentiation & Perception of Best Friend at MSU ’3 Alcohol Use

Males. The Additive Model with perceptions of their best friend at MSU’s alcohol
consumption was significant (752(4) = 33.27, p < .001, Ric... a 3,... = .19, alum“, = .26)

I and had a successful prediction rate of 71.4%, which was 10.4% greater than having no
model. Students that had one unit higher in perceptions of best friend use were 1.018
(95% CI = 1.010, 1.027) times more likely to be a Frequent Binge Drinker versus an
Occasional Binge Drinker (see Table 88, p. 265). Students that had a higher level of
perception of best friend use equivalent to one standard deviation (72.43) were 3.68 times

more likely to be a Frequent Binge Drinker.

262

 

flew—n mu

homage Womuommmoa 32:35 maaoam 2:: m8 125:2: Eamon. <32... 0885:». :3:QO I 338295 0».
Km: manna Ema

 

_ Kaom _ moan—om
_6 _mm _ 2% _ 9: 8:9 _ 6 _mm_€sa_ 9: 3x9

 

momma Zeno:

 

63.” -98 93 whom 9mg 9N3 _.oao 98 93 93 _.ocm oboe _.03

 

ES: -92 93 93 903 okuu _.omw -9: 9g 93 9mg 9qu room

 

>mo 9?: 9: New _.Nuu 93a _.Sm .98 93 93 993 92o _.Nfi

 

>95?” Zona—

 

Umzw -9% 9.3 no.3 9am 926 _.ooa 9S 9uo 9: 5:3 930 No3

 

wz: -93 93 _.5 93m 933 50$ -95 90a um: 93» 9.3a room

 

>mo 9~A 93 who _.N-N-N ohm-o. _.mow -93 9: 9: 993 93¢ _.wa

 

ZmC Cma 93 93 9.3 _.3o _.ooo _.omo o.oo o.oo 98 :93 98M _.ooo

 

38330: gono—

 

UmZN -odw 93 93 930 9wa _.ooN 9: 9: 98 3.30 93m NZ:

 

ES: -93 93 _.qo 98m 993 _.owa -9: 9cm 93 9mg 9.3m _.oom

 

>mo ohm 9G 98 :boo 93m 3.20 -98 93 93 90% 9w: 3ng

 

ZmCCmo 93 93 9.: 3.3: _.ooo _.oNN o.oo o.oo 98 _.ooo 98: _.ooc

 

 

 

 

 

 

 

 

 

 

 

 

 

UmTWX ch Cmo -98 93 9mm 93c 90% _.o3 -93 93 _.3 90% 93M _.ooa

 

 

.5 A .8. 2.0-8:09 uuwA .3. ”so-fizoa.

 

 

263

The Interaction Model also was significant (38(5) = 33.87, p < .001, R20,” 3.,“ =
.20, [gum-km = .27) and had a successful prediction rate of 70.1%, but did not show a
block improvement over the Additive Model (380) = .60, p = .438). The only significant
individual predictor in the model was perception of best friend use (OR = 1.019, 95% C1
= 1.010, 1.028, see Table 88, p. 265).

Females. The Additive Model with perceptions of their best fi'iend at MSU’s
alcohol consumption was significant (x2(4) = 22.36, p < .001, R2“,- at 5....“ = .14, RZNaaelkerke
= .19) and had a successful prediction rate of 64.4%, an increase of 11.7% over the Basic
Model. Students that had one unit higher in perceptions of best friend use were 1.018
(95% CI = 1.009, 1.027) times more likely to be Frequent versus Occasional Binge
Drinkers (see Table 88, p. 265). Students that had a higher level of perception of best
fi'iend use equivalent to one standard deviation (41.48) were 2.11 times more likely to be
a Frequent versus Occasional Binge Drinkers. Additionally, BMI was a significant
contributor in predicting drinking categories, as subjects BMI values increased they were
less likely to be Frequent Drinkers (OR = .866, 95% CI = .762, .978).

The Interaction Model also was significant (x2(5) = 22.69, p < .001, R20,“ 5m“ =
.14, [(21-1.8.“,ch = .19) and had a successful prediction rate of 63.7%, but did not show a
block improvement over the Additive Model (380) = .33, p = .565). The only significant
individual predictors in the model remained perception of best friend use (OR = 1.019,

95% CI = 1.009, 1.029) and BMI (.866, 95% CI = .762, .983, see Table 88, p. 265).

264

 

Hugo mm

rommmao ”3838: 32:33 mama-=8 :5» >3 33:2: 95mg 538 0885:». Esme-m I $088303 om mam, magma Cmo

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Zaam # moan—om
s _ mm _ 55 _ ow _ 8.x. 9 _ s _ mm _ é»: _ ow a 8: Q
mum? Zena
Umzw o.oo 98 9mm 9%.» 923 _.oao o.oo o.oo 9 H o room o.ooo _.ow:
ES: -93 93 93 902 o.od _.omm -9: o.oo 9.: o.oon 93A _.ooN
>mo 9?: 93 Now _.Nd o.ooo _.oqo o.oo 93 93 998 92¢ _.Nfi
25:76 Kean:
Ume 9% 93 93 ohom o.ooo :oo obo 93 93 Ewen o.ooo 93m
:35 o.oo o.oo Poo 92m 93o :93 -93 o.oo who. 93: 9.3» 93m
>mo ohm o. _ a mom flown o.ooo _.m _ A -93 o. _ m 98 o.ooo o.ooo _.Nao
mom” magma Cmo o.om o.oo Sou: _.o; _.o_o rowq 98 93 3:33.. _.o_ m #80 _.owu
583020: Zoom.
Ume -9: 9.3 _.Ao 99% onto _.u; 9: 9.3 9.5 _.Nwm 93o Noon
wz: o.oo o.oo _.oo o.omw 92 m 9ko -93 o.oo are: .. o.ooo th o.ooo
>mo 98 o. _ o 93 _.u :o o.oo: _.m _ o -9ow o. _ m obo 93o o.ooo _.NAo
wow" magma can 98 9o: 3.“: 3.. _.oE _.o_o _.owm o.om 9o: _93: _.o_o _.ooo 980
Um?” X wow" magma can 93 93 o.oo _.oou o.ooo rowu o.oo 9o: 93 :.ooA o.ooo _.omo

 

 

 

 

 

 

 

 

 

 

 

 

 

 

_._ me A .8. 25.8%? 1.5 A .or Rio-8:09

 

 

265

Dzfllerentiation & COA Status

Males. The Additive Model with COA status failed to reach a level of
significance (36(4) = 7.30, p = .121, see Table 89, p. 268). The Interaction Model also
failed to achieve a significance as an overall model (£6) = 7.309, p = .199).

Females. The Additive Model with COA status failed to reach a level of
significance (x2(4) = 5.65, p = .227, see Table 89, p. 268). The Interaction Model was
significant as an overall model 08(5) = 11.19, p = .05, RZCM 3,... = .07, Rzmm =
.10). The model had a successful prediction rate of 52.7%, which was an improvement of
1.3% over having no model. Students with a one unit higher interaction value were 6.773
(95% CI = 1.209, 37.949) times more likely to be a Frequent Binge Drinker than an
Occasional Binge Drinker (see Table 89, p. 268, see Figure 7, p. 74 for interaction
effect).

Dijfirentiation & Expectations of Alcohol Serving as a Tension Reducer

Males. The Additive Model with tension reduction expectations was significant as
an overall model 08(4) = 15.09, p = .005, R2“,- & 3....- = .09, RIM-m = .13). The model
had a successful prediction rate of 62.3%, which was a 1.3% improvement over having
no model. Expectation of alcohol serving as a tension reducer was the only significant
individual variable. Students that had a one unit higher level of expectation were 1.073
(95% CI = 1.021, 1.127) times more likely to be a Frequent versus an Occasional Binge
Drinkers (see Table 90, p. 269). Students that had a greater perception equal to one
standard deviation (8.73) were 1.84 times more likely to be Frequent Binge Drinkers.

Additionally, age was a significant factor, as students that were 1 year older were 1.338

266

(95% CI = 1.001, 1.787) times more likely to be Frequent Binge Drinkers. Students that
were two years older were 1.79 times more likely to be Frequent Binge Drinkers. '

The Interaction Model also was significant overall (x2(5) = 16.62, p = .005, R20,,(
& 30¢" = .10, Rznmnmkc = .14) and had a successfill prediction rate of 61 .7%; however, it
did not show a block improvement over the Additive Model (x20) = 1.54, p = .215). Age
(OR = 1.377, 95% CI = 1.024, 1.852) and expectations of tension reduction (OR = 1.074,
95% CI = 1.022, 1.128) remained the only significant individual predictors (see Table 90,
p. 269).

Females. The Additive Model with tension reduction expectations was significant
as an overall model (36(4) = 15.89, p = .003, R20,“ 5,... = .10, £2an = .14). The
model had a successful prediction rate of 57.5%, an increase of 6. 1% over not having a
model. Expectations of alcohol serving as a tension reducer changed the odds of being a
Frequent Binge Drinker versus an Occasional Binge Drinker by 1.096 (95% CI = 1.035,
1.160) for a one unit difference in expectations (see Table 90, p. 269). By changing the
difference in expectations to a unit equivalent to one standard deviation for expectations
(8.81) there is a 2.23 times greater likelihood of being a Frequent Binge Drinker.

The Interaction Model also was significant overall (36(5) = 15.97, p = .007, RICO.
& 3n..." = .19, Rznmm = .27) and had a successful prediction rate of 59.6%; however, it
did not show a block improvement over the Additive Model (x20) = .08, p = .778).
Expectations of tension reduction remained the only significant individual predictor (OR

= 1.095, 95% CI = 1.035, 1.160, see Table 90, p. 269).

267

 

o.oo—a we

rommmoo ”£838: 3.8335 maaoam 5m» >8 33:2: Eamon 825 008925— wmsmoa l 00> mBEm

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

_ 3E8 _ 123—om
L a :m fléea _ ow _ 8.x. 2 _ e _ mm _ 2% _ ow _ 3.x. 9
momma Zona—
UmTW o.oo o.oo o.oo out ohwu #vo o.oo o.oo o. _ o o.ooo o.ooo o.oo—
Eé -o.oA o.om o.oo o.ooo o.ooo o.omm -o._ _ o.oo Pd o.oom o.ooo _.oou
>mo ob... o. H a New _.Nd o.ooo _.ouo o.oo o. _ A o. o o 903 o..2o ENS
>353 Zona—
UmEN o.oo o.oo 93 ohm— o.Nco _.og o.oo o.oo o.oo _.omu o.oo_ _.ooo
wzo o.oo o.om o.oo o.oo.“ o.od o.oow -o.: o.oo who o.ooo. o.ooo _.ooc
.9? ob.“ o. 3 NS _.Nqu o.oom _.oqm -o.oq 9: ob.“ o.ooo ohom _.NK
00> -o.w_ o.oo o.oo o.ooo o.ooo TEN ohm ohm o.om _.mao o.ooo Pin
588020: gono—
Umzw -93 o.oo NE om: obom _.o _N -obu o.oo 93 o.ooo. o.oom _.Aoo
w?= -o.oA 93 ohm o.oo; 93m _.oou -o._~ o.oo who o.oo.» o.ooo foo—
>m@ on» o. _ A New _.Nd o.oou _.oqo o.oo o._ m ohm o.ono o.ooo _.Nwo
00> oh: ohm 93 9qu o.ooo _.qmm ohm o.oo ohm _.A _ w ouqo 9A3
szw x 00> o.oq o.oo o.oo _.oom oboo 95$ _.3 o.oo Pd... 99d _.Noo 3.vo

 

 

 

 

 

 

 

 

 

 

 

 

 

 

n to A .8. 23.8%? 3h A or 2.5-8:.”9

 

 

268

 

Hugo co

rommmno Wowammmo: 38323 £888 5:: >3 man—8:» mgmoa <29; Geomagm— 259m l mxuooflmaoa ow H038:

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

W888:
_ 3&8 _ maBEom
_ a o mm _ éoa _ o: _ 8&9 _ o 4 mm _ fins 8 o: _ 8:9
wmmmo :81

Uwzw o.oo o.oo o.oo out ohoq :vo o.oo o.oo o. A o _.oon o.ooo :3:
BE -o.oA o.om o.oo. o.ofl o.ooo _.omo -o.: o.oo on: o.oow o.ooA _.ooN
>mo o.~A o. A New _.Nd o.ooo o.ooo o.oo o. _ A o. _ o o.oAm o..:o A.NA.\

>835 382

Ume -obm ohm ohm ohwm o.ooo A.Awo o.oo ohm who _.dq o.omu Pmoo
wz: -o.oA o.ou o.AA o.ooo o.od _.ooo -o._ _ o.oo Non o.ooo o.ooo _.S u
>mo obo o.Am o.oo... o.ooo _.oo_ _.qmq .98 93 93 o.oum on“; _.Nd

#38: W885: o.oq o.oo $3.1. o.ooo _.oB :3 o.oo 98 com: _.ooo room :3
5838:: 38¢.

Ume .93 o.om o.oo o.ooo .wmA _.A_o o.oo o.oo NuA ooNA o.oun o.ooo
wz: -o.oA 93 9% o.ooo o.ooo _.oon -o. _ o o.oo ~.oA o.ooo o.ooo _.oB
>mo ohm o._u A.Ao... $3 _.oNA _.omn -o.oA o._m o.oo 99% oh: _.NE

Adamo: ”8:80: o.oq o.ow Soon... _.oQA _.oun _._No o.oo 98 who: _.oou _.ouu Zoo
Ume x H038: W883: o.oo 93 _.Au 9vo o.ooo _.ouA .93 93 o.oo o.ooo o.co_ _.omn

 

 

 

 

 

 

 

 

 

 

 

 

 

 

. n A .8. 398:8. _._. a A .3. 253.8.

 

 

269

Dzfikrentiation & Expectations of A lcohol Serving as a Social Lubricant

Males. The Additive Model with social lubrication expectations was significant as
an overall model (38(4) = 9.53, p = .049, 82¢... a 5,... = .06, 82%“, = .08). The model
had a successful prediction rate of 60.4%, an increase of .6% over not having a model.
Although the model was significant, none of the individual variables were significant in
the prediction of binge drinking category (see Table 91, p. 271). The Interaction Model
failed to reach a level of significance (38(5) = 9.54, p = .089, see Table 91, p. 271).

Females. The Additive Model with social lubrication expectations was significant
as an overall model (36(4) = 12.18, p = .016, 13¢,” 5,... = .08, 112an = .11). The
model had a successful prediction rate of 59.6%, an increase of 8.2% over not having a
model. Expectations of alcohol serving as a social lubricant changed the odds of being a
Frequent Binge Drinker versus an Occasional Binge Drinker by 1.078 (95% CI = 1.020,
1.140) for every one unit change in perceptions (see Table 91, p. 271). Students with a
difference in expectations equal to one standard deviation (7.74) were 1.79 times more
likely to be a Frequent Binge Drinker.

The Interaction Model also was significant overall (36(5) = 12.18, p = .032, 18%,,
a 5...." = .08, thi..g.m.ke = .11) and had a successful prediction rate of 71.8%; however, it
did not show a block improvement over the Additive Model (x20) = .00, p = .982).
Expectations for social lubrication remained the only significant individual predictor (OR
= 1.078, 95% CI = 1.020, 1.140) in predicting Frequent versus Occasional Binge

Drinkers (see Table 91, p. 271).

270

 

 

Hugo 3

homage Woman-ammo: 32:38 930:8 92 >8 mama—ma Ego; 58% 088mg:— wm:mo~m I mxnooBmoPa ow mega

 

 

 

 

 

 

 

rccnommo:
— 3&8 _ mafia—am
_ e _ mm _éa: o: _ moan: _ a _ mm. as; _ ox _ 392:
mean Kong
Um?” -98 98 8.8 9qu 9mm: _.vo 98 93 98 room 926 be“:
male: -98 98 93 903 938 98a -9: 98 8.: 98» 93A _.oou
go 9N8. 9: 98 _.N8 98a _.a-R -98 9: 93 903 9.2m _.NS
>aaE<o zona—

 

UmH-W -9ao 9mm _.wA 9.3a 9qu “bu-A 93 98A _.8 _.os 98— 8.30

 

w?= -98 98 93 93m 98o _.oqu -95 98 8.3 9x£ 980 _.ooo

 

>mo 93 93 93 _.uow 98m _.qwm -98 93 98 93M 98¢ _.Nom

 

meow»: rccnomao: 99A 98 N65 _.oAu 903 _._8 98 98 98...... _.o-Nm _.ouc _._Ao

 

38898: Zona—

 

UmH-W -93 98 _.wh 93g 98a fuNA 9% 9w» _.3 _.3o 93w “:8

 

W73 -98 98 93 93m 933 _.o-Nw -95 o.oo 8.3 9%: 9qu _.cow

 

>Mn 93 9; 8.3 _.woo 93A 9qu -98 9G 98 93M 9.3g _.Nom

 

moan—“15183:: 98 98 N8 2.3 993 ram 98 98 $81.. _.oqm _.owc _.Eo

 

Umfi-WX mega rcvaowmo: coo 98 9o— ?ook— 90: _._oo 98 98 o.oo _.oS 9on Tom:

 

 

 

 

 

 

 

 

 

 

 

 

 

 

e h A .8. 25.838. _._... .u A .2. go-Szoa.

 

 

27l

The Comprehensive Model

Males. The Comprehensive Model was significant according to the model chi-
square (38(13) = 46.1 1, p < .001, R20,” 3...,- = .26, mm...“ = .35). The model
illustrated a 73.4% success rate in predicting the correct drinking classifications, a 12.4%
improvement over having no model. Age (OR = 1.472), perceptions of best friend use
(OR = 1.018), and the differentiation X expectations of tension reduction interaction (OR
= .826, see Figure 8, p. 77, for interaction) were the only individual variables to
significantly improve the odds of being classified as a Frequent Binge Drinker versus an
Occasional Binge Drinker (see Table 92, p. 273).

Females. The Comprehensive Model was significant according to the model chi- .
square (x203) = 39.46, p < .001, R2c°x ,- 3...- = .24, 12sz = .32) and showed a
significant improvement over the Basic Model (x2(10)= 34.88, p < .001). The model had
a 69.2% success rate in predicting the correct drinking classifications, an improvement of
16.5% over the Basic Model. Perceptions of best friend use (OR = 1.019) and
expectations of tension reduction (OR = 1.081) were the only individual variables to
significantly improve the odds of being classified as a Frequent versus an Occasional

Binge Drinker (see Table 92, p. 273).

272

 

Had—o 0N

homage ”om—6&8: 3.35:8 magma 9R Ed 3.00:2: Esau; 388 08333. 9803 I 02:96:03?» gono—

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Zaom man—Ema
8 mm £35 0% 8.x. 9 8 mm 83m OW 8.x. Q
69% -98 98 _.8 98w 98¢ :88 9mm 93 _.8 .88 9mm.“ 8.88
wz: -98 98 Hum 908 9a _ m _.98 -9 8 98 u .8 988 9.88 _.o . o
>m 980 9; 8.8... _.SN _.8. ~88 -98 93 93 908 9.88 _.80
Km: 02.80.83 98 93 98— ro8 900. :03 98 93 95 900m 908 .88
Um?” X ch 0280308 -98 93 N00 908 908 :08 -92 93 98 900m 980 for.
mam" manna 98 9S 5&0: _.o; _.8m _.8m 98 93 3.3.1. 2:0 _.8m #8—
Umzw X mam. 3.5:. 98 92 8.3. _.c; 900m _.80 98 9o— 9: :88 908 _.ouu
00> -93 98 9d 998 980 _.88 -9 5 9m. 98 98— 988 N88
62% X 00> 9N0 _.8 98 #88 93c 2.38 _.8 _.8 N8 986 988 8.88
dogma: X8230: 98 98 93 78¢ 908 :3 98 98 8.3... _.2: _.8m :8
0m?” X H955: ”8:30: .93 98 .30... 98¢ 988 908 -98 98 98 908 98w _._ _N
moomfi 551832. -98 98 98 98m 908 :08 98 98 _._u _.80 908 _._ _m
09% X mafia r5183: 93 9C0 Pam _._m0 908 _.8m 98 98 9G _.80 9m0_ :8

 

 

 

 

*5 A .8. "so-3.0m. _._... E A .3. 25-889

 

 

273