Title: Sensation seeking, alcohol expectancies, and motivation for drinking among male and female college students
CITATION THUMBNAILS PAGE IMAGE ZOOMABLE
Full Citation
STANDARD VIEW MARC VIEW
Permanent Link: http://ufdc.ufl.edu/UF00102725/00001
 Material Information
Title: Sensation seeking, alcohol expectancies, and motivation for drinking among male and female college students
Physical Description: Book
Language: English
Creator: Henn, Linda
Copyright Date: 1992
 Record Information
Bibliographic ID: UF00102725
Volume ID: VID00001
Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
Resource Identifier: ltuf - AJG4531
oclc - 26529182

Full Text











SENSATION SEEKING, ALCOHOL EXPECTANCIES,
AND MOTIVATION FOR DRINKING AMONG
MALE AND FEMALE COLLEGE STUDENTS














BY

LINDA HENN


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA

1992


UNIVERSITY OFFLORIDtRI.I.s
















TABLE OF CONTENTS



page

ABSTRACT. ...... . . . . .. . . . . iv

CHAPTERS

1 INTRODUCTION . . . . . . .... .. . ..

Background and Overview . . . . ..... . 1
Theoretical Framework ..... . . . . . 6
Statement of the Problem . . . . . . . 9
Need for the Study . . . . . . . . . 9
Purpose of the Study . . . . . . . 11
Rationale for the Approach to the Study . . . 12
Research Questions . . . . . . . . 13
Definition of Terms . . . . . * * . 14
Overview of the Remainder of the Dissertation . .. 15

2 REVIEW OF RELATED LITERATURE ...... . . . 16

Review and Delineation of the Problem .. . . .. 16
Support for Theoretical Framework . .... . 31
Support for Need for the Study . . . ... . 34
Support for Approach to the Study . . .... . 37

3 METHODOLOGY ... . . . . . . . . 39

Overview of the Study . . . . . . . ... 39
Null Hypotheses . . . . . . . . . 39
Delineation of Relevant Variables . .... . 41
Description of the Assessment Instruments and
Resulting Data . . . . . . . 43
Description of Population and Sampling . . . 51
Description of Research Procedures . .... . 52
Description of Data Analyses . . . ... . 54
Description of Methodological Limitations . . 55

4 RESULTS OF THE STUDY . . . ..... . . 57

Introduction . . . . . . . . . 57
Description of the Sample . . . . ..... . .. 58
Analysis of the Data . . . . . .. .. ...... 58
Summary of the Findings . . . . . . . 70











5 DISCUSSION . . . . . . . . . . .

Summary of the Study . . . . . . . . .
Discussion of the Results . . . . . . . .
Conclusions and Implications . . . . . . .
Recommendations for Further Study . . . . . .

APPENDICES


DEMOGRAPHIC QUESTIONNAIRE . . .

KHAVARI ALCOHOL TEST . . . .

SENSATION SEEKING SCALE . . .

EXAMPLES OF ITEMS FROM THE ALCOHOL
EXPECTANCY QUESTIONNAIRE .

INFORMED CONSENT PROCESS . . .

DEBRIEFING SHEET . . . . .


REFERENCES . . . . . . . . .

BIOGRAPHICAL SKETCH . . . . . . .


. . . 85

. . . 86

. . . 90



. . . 96

. . . 97

. . . 99
















Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

SENSATION SEEKING, ALCOHOL EXPECTANCIES,
AND MOTIVATION FOR DRINKING AMONG
MALE AND FEMALE COLLEGE STUDENTS

By

Linda Henn

May 1992

Chairman: P. Joseph Wittmer
Major Department: Counselor Education

Sensation-seeking needs and alcohol expectancies have separately

been found to relate to collegiate alcohol use. In the current study,

these two motivational factors were investigated together in order to

explicate a disinhibition motive for drinking among male and female

college students. Two major research questions were posited: (a) the

influence of alcohol expectancies and gender on sensation-seeking needs

and (b) the joint influence of sensation-seeking needs, alcohol

expectancies, and gender on drinking level.

These questions were investigated utilizing survey instruments

including the Khavari Alcohol Test, the Sensation Seeking Scale Form V,

and the Alcohol Expectancy Questionnaire. The sample consisted of 464

undergraduates.

Examination of the results to the first question revealed that

alcohol expectancies of Social and Physical Pleasure and Social

Assertion were related to sensation-seeking needs. No gender

iv









differences existed in these relationships or in the level of sensation-

seeking needs.

Investigation of the second research question demonstrated that

higher scores in sensation-seeking needs, together with higher scores in

the alcohol expectancy of Social and Physical Pleasure, contributed to

heavier drinking among both male and female students. However, scores

in this alcohol expectancy increased at a slightly more rapid rate than

scores of sensation-seeking needs in relation to reported drinking

levels in the entire sample and in males. Although drinking level was

explained to some degree by individual consideration of sensation-

seeking needs, alcohol expectancies, and gender, a more accurate

prediction of drinking resulted from their joint consideration, which

explained 49% of variance in reported drinking level.

The motivational pattern of drinking delineated by these results

was similar for male and female students. The disinhibition motive for

drinking described by the relationships among sensation-seeking needs,

alcohol expectancies, and drinking level appeared to be the desire to

expand the usual limits on enjoyment of and stimulation from alcohol-

related social interactions and sensations. This disinhibition motive

depicted drinking as instrumental in enhancing positive affect and

situations. Suggestions were offered for the application of these

results to campus programs for the prevention of alcohol abuse.















CHAPTER 1
INTRODUCTION



Background and Overview

The use of alcohol in college has been the subject of considerable

research, beginning most notably in 1949 with a nationwide study of

17,000 college students (Straus & Bacon, 1953). Over the ensuing 30

years, some studies indicated an increasing incidence of drinking among

college students (e.g. Wechsler & McFadden, 1979). More recently

however, it appears that college use of alcohol has been declining. For

example, some researchers found that a lower percentage of students

reported daily use as compared to previous studies (Meilman, Stone,

Gaylor, & Turco, 1990). Nevertheless, a similar variety of alcohol-

related problems, including injuries, illegal activities, and academic

difficulties, have continued among students (O'Hare, 1990). Excessive

consumption of alcohol is a central feature of problem drinking by

students; heavier drinking is more often associated with negative

consequences (Berkowitz & Perkins, 1986). In spite of some declines in

alcohol use and problems, a disproportionate incidence of both have

persisted on college campuses (Gonzalez, 1986). It is expected that

results of the current study, which was designed to assess motivational

factors that contribute to student drinking, will be applicable to

campus prevention programs.








2

Motivation for alcohol use has traditionally been studied in terms

of personality correlates of drinkers in the general population (e.g.

Cahalan, Cisin, & Crossley, 1969) as well as in the college population

(e.g. Jaffe & Archer, 1987). More recently, a number of researchers

have investigated motivation for drinking in terms of expectancies of

reinforcing effects of alcohol use (e.g. Cooper, Russell, & George,

1988). Alcohol expectancies refer to beliefs about the effects that

alcohol has on behavior and mood (Leigh, 1989a). In the present study,

alcohol expectancies referred specifically to expected reinforcing or

positive effects of alcohol, or what have been called positive alcohol

expectancies by some researchers (e.g. Stacy, Widaman, & Marlatt, 1990).

In this study, only alcohol expectancies of positive effects were

considered because these should be most relevant to motivation for

drinking (Brown, Goldman, Inn, & Anderson, 1980).

In both general and college populations, it has consistently been

demonstrated that heavier drinking is associated with increased

expectancies of positive effects from drinking (Leigh, 1989c) and that

these expectancies are a major factor in motivation for drinking

(Connors & Maisto, 1988; Leigh, 1989a, 1989c; Maisto, Connors, & Sachs,

1981). The specific nature of alcohol expectancies has been

investigated through the development of several questionnaires (Brown et

al., 1980; Connors, O'Farrell, Cutter, & Thompson, 1987; Leigh, 1987;

Rohsenow, 1983; Southwick, Steele, Marlatt, & Lindell, 1981; Young &

Knight, 1989). According to Connors and Maisto (1988), the most widely

used measure has probably been the Alcohol Expectancy Questionnaire

(Brown et al., 1980). This questionnaire, which was designed to assess









3

only positive or reinforcing effects that individuals expect from

alcohol, measures expectancy factors of Global Positive Changes, Sexual

Enhancement, Social and Physical Pleasure, Social Assertion, Tension

Reduction, and Arousal and Aggression (Brown et al., 1980). Results

have been mixed regarding which of these positive expectancies

differentiate light- and heavy-drinking college students. For example,

although Brown et al. (1980) found that heavy-drinking students expected

more Sexual Enhancement and Arousal and Aggression, Brown (1985a) later

found that students whose drinking was heavy and most problematic

expected more Tension Reduction.

Several researchers have begun to investigate both personality and

alcohol expectancy variables in studies of motivation for drinking among

college students (Brown & Munson, 1987; Leonard & Blane, 1988; Mooney &

Corcoran, 1989). It has been suggested that further research is needed

to address the issue of personality variables or "individual differences

as moderators of expectancy-drinking associations" (Stacy et al., 1990,

p. 927). This researcher addressed this need by investigating whether

sensation-seeking needs of college students moderated the relationships

between their positive alcohol expectancies and their drinking levels.

This researcher's choice of the personality variable of sensation

seeking to study motivation for drinking reflected the fact it is a

central component of the personality dimension which has most often been

found to relate to student drinking (Brennan, Walfish, & AuBuchon,

1986a). In addition to sensation seeking, the variables comprising this

dimension include impulse expression, disinhibition, rebelliousness,

aggression, reduced conventionality, and extraversion (Brennan et al.,









4

1986a). This personality dimension of sensation seeking/impulse

expression has been found to be related to current heavy and problem

drinking by students as well as to future problem drinking (Brennan,

Walfish, & AuBuchon, 1986b). For example, college drinkers who were

more impulsive, nonconforming, or prone to deviant behavior were more

likely to become alcoholics or problem drinkers (Jones, 1968; Loper,

Kammeier, & Hoffmann, 1973; Donovan, Jessor, & Jessor, 1983).

Similarly, adolescents who scored higher on variables of Exhibition and

Impulsivity were later heavier users of alcohol and other drugs

(Labouvie & McGee, 1986).

The variable of sensation-seeking needs is expressed as behavior

which is aimed at obtaining a high degree of stimulation and which is

unconventional, impulsive, and extraverted (Zuckerman, Bone, Neary,

Mangelsdorff, & Brustman, 1972). Sensation-seeking needs, as measured

by the Sensation Seeking Scale (Zuckerman, 1979), have often been found

to be related to college drinking (e.g. Earleywine & Finn, 1991;

Schwarz, Burkhart, & Green, 1982; Segal, Huba, & Singer, 1980a, 1980b).

Among selected personality measures, the Sensation Seeking Scale has

been found to be the best predictor of student drinking (Schwarz,

Burkhart, & Green, 1978) and of both student use of alcohol and other

drugs (Jaffe & Archer, 1987). In another study, heavy-drinking students

were higher in sensation seeking but did not differ from light drinkers

in depression or self-esteem (Ratliff & Burkhart, 1984).

The Disinhibition factor of sensation seeking is most similar to

impulsivity (Zuckerman, 1983) and has been found to have a particularly

strong relationship with student drinking (Kohn & Coulas, 1985; Ratliff








5

& Burkhart, 1984; Schwarz et al., 1978). The researchers of one of

these studies concluded that students may primarily be motivated to

drink as "a culturally sanctioned 'time-out' from social control"

(Schwarz et al., 1978, p. 1145). Similarly, on the basis of the result

that the Disinhibition factor accounted for 44% of the variance between

light and heavy drinkers, Ratliff and Burkhart (1984) concluded that

"needs for varied experiences and disinhibition constitute a major

motivational pattern for alcohol use among both men and women college

students" (p. 31). On the basis of this result, together with the

findings that men scored higher on Disinhibition and reported more

alcohol-related problems with authorities, Ratliff and Burkhart (1984)

suggested that women may primarily expect drinking to enhance social

pleasure, whereas men may expect to become disinhibited in aggressive or

inappropriate behaviors as a result of drinking.

However, these interpretations may be questioned when compared with

the results of a study on college drinking in which researchers examined

the effect of extraversion on alcohol expectancies (Brown & Munson,

1987). Extraversion, which is correlated with sensation-seeking needs

(Zuckerman et al., 1972), was related to expectancies such that students

with high extraversion expected more Tension Reduction and Social and

Physical Pleasure, but not more Sexual Enhancement or Arousal and

Aggression (Brown & Munson, 1987). The relationships between alcohol

expectancies and sensation-seeking needs had not been investigated prior

to the current study.









6

Theoretical Framework

In order to investigate motivation for drinking in terms of

sensation seeking and alcohol expectancies, the current study was based

on a combination of personality and social learning theories. Current

personality theory of alcohol use has incorporated elements of trait

theory and drive theory (Cox, 1987). Social learning theory of alcohol

use is the product of many social learning theorists and lately has led

to an expectancy theory of alcohol use (Abrams & Niaura, 1987).

Trait theory posits that behavior is largely determined by traits,

such as sensation-seeking needs, that are fairly stable across various

times and situations (Mischel, 1981). According to personality theory

of alcohol use, motivation for drinking is the result of personality

traits in combination with the use of alcohol to manage affect, that is,

to enhance positive affect and/or reduce negative affect. The influence

of reinforcement principles came from drive theory and led most notably

to the development of the tension-reduction hypothesis of alcohol use

(Cox, 1987).

The tension-reduction hypothesis of alcohol use posits that alcohol

reduces negative affect, especially tension or anxiety, and that

therefore the motivation for drinking is escape or relief from this

negative affect (Cappell & Greeley, 1987). The results of research on

this hypothesis have been mixed; depending on a variety of factors,

alcohol may either enhance mood or increase anxiety and depression. It

is now accepted that such a general theory or unitary effect of alcohol

will continue to be insupportable (Cappell & Greeley, 1987). For the

current study, the relevance of the tension-reduction hypothesis is its








7

reflection in the common expectancy that drinking will produce an

outcome of relief of anxiety, stress, or tension (Goldman, Brown, &

Christiansen, 1987).

Personality theory of alcohol use has incorporated the research

findings on the relationship of sensation seeking to drinking (Cox,

1987). According to the theory, sensation seekers are motivated to

drink primarily for enhancement of stimulation and positive affect.

Based on research results of a relationship between sensation seeking

and normal alcohol use, it is suggested that initial drinking experience

and most normal use of alcohol are motivated primarily to enhance

positive affect (Cox, 1987).

In social learning theories of alcohol use, cognitive variables are

emphasized over personality variables (Abrams & Niaura, 1987).

Individual differences, pharmacological effects of alcohol, and

situational variables are construed as influencing behavior through the

cognitive processes of the person. On the basis of these various

influences, a person chooses whether to use alcohol to gain desired

outcomes. Central to this model are expectancies that individuals have

of the effects of alcohol and of their abilities to achieve, with and

without the use of alcohol, their desired goals. The latter type of

expectancy, known as a judgement of self-efficacy, is fundamental to

understanding how alcohol use is likely to become abuse when situational

demands exceed perceived coping ability, especially when individuals

have high expectancies of positive effects from drinking (Abrams &

Niaura, 1987).









8

An expectancy theory of alcohol use has recently been developed

from the broader social learning approach. Expectancy theory explains

motivation for drinking primarily in terms of individuals' expectations

of the reinforcing effects of drinking (Goldman et al., 1987).

Expectancy theory has roots in particular in the accumulation of

experimental results in which behavioral effects were caused by an

expectancy of being served alcohol, separate from actually receiving any

and therefore separate from any potential pharmacological effects of

alcohol (Marlatt & Rohsenow, 1980). In many of these experiments, the

relative effects of pharmacology and expectancy were tested in producing

either tension reduction or disinhibition of especially sexual and

aggressive behavior (Goldman et al., 1987).

Much of early alcohol research was dominated by a disinhibition

hypothesis which arose from beliefs about alcohol's effects on cortical

inhibitory functions (Ritchie, 1965) as well as from psychoanalytic

theory (Wilson, 1981). As with the case of the tension-reduction

hypothesis however, the relevance of the disinhibition hypothesis to

expectancy theory and to the current study is that alcohol-induced

disinhibition is so widely endorsed as fact in this culture (Critchlow,

1986).

According to social learning theory, behavior is best predicted by

consideration of both expectancies of the outcomes of the behavior and

the reinforcement value of those expectancies (Rotter, 1954). It has

been suggested that the reinforcement value of an alcohol expectancy may

be represented by the personality variable that is relevant to the

expectancy, as for example, to a nonassertive individual an alcohol









9

expectancy of increased social assertiveness would likely be highly

reinforcing (Mooney & Corcoran, 1989). In the current study, it was of

interest to similarly determine which alcohol expectancies predicted

drinking for students with varying levels of sensation-seeking needs.

Statement of the Problem

Motivation for drinking among college students has not been fully

investigated. The current study was designed to improve understanding

of this area by investigating motivation for student drinking among male

and female college students in terms of their sensation-seeking needs

and expectancies of reinforcing alcohol effects. In previous research,

individual relationships between college drinking and both sensation-

seeking needs and alcohol expectancies have been studied. Gender

differences in sensation seeking, alcohol expectancies, and the

relationships between alcohol expectancies and drinking have also been

studied. However, the relationships among gender, sensation-seeking

needs, alcohol expectancies, and drinking level had not been examined

together. As a result, it was unknown how sensation-seeking needs and

alcohol expectancies were related and how they together contributed to

student drinking.

Need for the Study

Results of this study should have implications for theory of

alcohol use in two major respects. First, if sensation seeking

predicted student drinking, then this result would support the inclusion

of personality variables in alcohol research in addition to demographic

and peer influences. A strong relationship between personality and

drinking in college students has, for example, been demonstrated when









10

the imperfections in the measures of both were taken into account

(Earleywine, Finn, & Martin, 1990). Also, the relevance of expectancy

theory to alcohol use would be supported if alcohol expectancies were

related to student drinking.

The current study would make a contribution to research by

investigating the relationships between sensation seeking and alcohol

expectancies and how they together contribute to drinking. If sensation

seeking was found to be most related to an alcohol expectancy of Sexual

Enhancement for example, this result would support the common

interpretation of a disinhibition motive for drinking (e.g. Brown &

Munson, 1987). This interpretation would be further supported if higher

scores on the expectancy of Sexual Enhancement, together with higher

scores on sensation seeking, contribute to heavier drinking. However,

if for example heavier drinkers scored higher in sensation-seeking needs

but expected more Tension Reduction as opposed to Sexual Enhancement or

Arousal and Aggression, then the concept of a disinhibition motive would

have been substantially altered. This study should also contribute to

the research literature on gender differences in sensation seeking, in

the relationships between sensation seeking and alcohol expectancies,

and in the relationships between drinking and both sensation seeking and

alcohol expectancies.

Results of this investigation would have several implications for

campus prevention efforts. If the predominant pattern of heavier

drinking was characterized by sensation seeking and expectancies related

to typical social situations, then alcohol education may best be

presented through students' primary social groups in order to adjust the









11

college drinking environment to limits of responsible use. In order to

satisfy sensation-seeking needs, alternate activities that provide

healthy stimulation might be encouraged. If cognitive expectancies of

disinhibition of sexual or aggressive behavior were involved, these

could be deprived of their excuse value by teaching that beliefs about

drinking, more than alcohol itself, can lead to disinhibited or

unacceptable behavior.

If heavier drinkers were found that appeared to be motivated to

drink to obtain relaxation or tension reduction, then these students

would benefit from an exploration of their expected reinforcement and

how it contrasts with any negative consequences they experience as well.

Coping skills training would be most appropriate for this group of

drinkers.

Purpose of the Study

In this study, motivation for drinking by college students was

examined in terms of their sensation-seeking needs and expectancies of

reinforcing alcohol effects. The relationships between drinking and

both sensation-seeking needs and alcohol expectancies have been studied

separately in previous research. It was the purpose of the current

study to investigate sensation-seeking needs, alcohol expectancies,

gender, and drinking level together to more fully explicate motivation

for student drinking. In particular, the relationships among these

variables were studied to improve understanding of the motivation for

drinking that has been inferred from the relationship often found

between sensation seeking and drinking. This relationship between

sensation seeking and drinking has usually been considered to reflect a









12

disinhibition motive or motivation to drink to escape social restraint

(e.g. Berkowitz & Perkins, 1986; Schwarz et al., 1978), and especially

to disinhibit sexual or aggressive behaviors in male students (Ratliff &

Burkhart, 1984). As a result of this study, it was hoped there would be

an improved understanding of motivation for student drinking,

particularly with regard to this purported disinhibition motive.

In order to study these relationships, two major questions were

asked. The first concerned the joint influence of gender and alcohol

expectancies on sensation-seeking needs. With this question, the

researcher sought to determine which alcohol expectancies were related

to sensation-seeking needs, whether gender differences existed in these

relationships, and whether a gender difference existed in sensation-

seeking needs. The second question concerned the multiple effects of

gender, sensation-seeking needs, and alcohol expectancies on drinking

level. This question included the investigation of relationships

between drinking level and both sensation-seeking needs and alcohol

expectancies, as well as gender differences in these relationships. In

addition, this question investigated which alcohol expectancies

contributed to drinking for varying levels of sensation-seeking needs

and whether gender differences existed in these relationships.

Rationale for the Approach to the Study

The current research problem was best addressed by a correlational

rather than experimental study. The primary reason for the choice of a

descriptive approach was the difficulty of effectively and ethically

manipulating and measuring the variables of interest. In particular,

ethical, safety, and liability issues would have needed to be resolved








13

before serving alcohol to experimental participants. If deception was

employed such that alcohol was actually not provided but was said to be,

the general expectancy of receiving alcohol must have been believable to

the subjects. The efficacy of manipulating conditions to elicit

specific expectancies would also have presented a difficulty. For

example, Schwarz et al. (1982) concluded that the manipulation to

produce an expectancy of a low-restraint situation may have been

inadequate to increase consumption in high sensation-seeking subjects.

The type of descriptive study recommended for the current research

problem was the survey method. Subjects' comfort and accuracy in

reporting drinking behavior is most probably facilitated by the

familiarity and anonymity of survey research. Measuring alcohol

consumption by self-report has generally been found to have adequate

reliability and validity (Polich, 1982; Williams, Aitken, & Malin,

1985). It has also been demonstrated within a college population that

the relationship between alcohol expectancies and alcohol use was not an

artifact of bias deriving from the self-report method of measuring both

constructs (Stacy et al., 1990). In addition, the survey method is an

efficient and practical means of acquiring data from the large sample

size needed for a descriptive study (Isaac & Michael, 1981).

Research Questions

In the current study, motivation for college student drinking was

investigated according to the following research questions:

1. Is there a difference in sensation-seeking needs between male

and female college students?









14

2. What are the relationships between the alcohol expectancies and

sensation-seeking needs of college students, and are there gender

differences in these relationships?

3. What is the relationship between college students' drinking

levels and their sensation-seeking needs, and is there a gender

difference in this relationship?

4. What are the relationships between college students' drinking

levels and their alcohol expectancies, and are there gender differences

in these relationships?

5. What are the relationships between college students' drinking

levels and their alcohol expectancies as a result of their levels of

sensation-seeking needs, and are there gender differences in these

relationships?

Definition of Terms

Variables that may require clarification are defined in the

following list in the manner in which they were used in this study:

1. Sensation-seeking needs describe behavior which is aimed at

achieving a high degree of stimulation and which is impulsive,

nonconforming, and extraverted. Sensation-seeking needs were measured

by the Sensation Seeking Scale (Zuckerman, 1979). This measure yields a

total score and scores for various subscales. The sensation-seeking

variable of this study was selected according to which of these three

scores was most correlated with drinking level in the sample: the total

scale, the Disinhibition subscale, and the Experience-Seeking subscale.

Thus sensation-seeking needs was operationally defined as the variable











among these three sensation-seeking scores that was most highly

correlated with drinking level in the study.

2. Alcohol expectancies were defined as expected, reinforcing

effects from drinking as measured by the Alcohol Expectancy

Questionnaire (Brown et al., 1980). Of the six factors or scales of

this questionnaire, the following five factors were the alcohol

expectancy variables in the current study: Sexual Enhancement, Social

and Physical Pleasure, Social Assertion, Tension Reduction, and Arousal

and Aggression.

3. Drinking level was defined as the ratio of the student's

reported alcohol consumption to his or her body weight. Consumption was

measured by the Khavari Alcohol Test (Khavari & Farber, 1978).

Overview of the Remainder of the Dissertation

The second chapter of this report contains a review of relevant

literature to the study. The third chapter focuses on methodology,

including descriptions of measurement instruments, research procedures,

and data analyses. In the fourth chapter, results of data analyses are

presented. The final chapter consists of a discussion of these results,

conclusions and implications, and recommendations for further study.















CHAPTER 2
REVIEW OF RELATED LITERATURE



This chapter consists of a review of research that is relevant to

the questions of the study. In the first section, research is presented

which describes variables similar to sensation-seeking needs and how

these variables and sensation seeking have been found to relate to

college student drinking in particular. Other research reviewed in the

first section covers areas of alcohol expectancies, their relationships

to student drinking, and the interrelationships among personality

variables, alcohol expectancies, and drinking behavior. In the

remaining sections of this chapter, literature is reviewed which

pertains more specifically to the theoretical framework, the need for

the study, and the approach to the study.

Review and Delineation of the Problem

Sensation-Seeking Needs and Similar Variables

Researchers have reported a fairly consistent pattern of

personality, behavioral, and demographic correlates of sensation seeking

(Zuckerman et al., 1972). Sensation seeking has been found to correlate

with extraversion (Farley & Farley, 1967; Zuckerman et al., 1972), the

MMPI scale of Hypomania (Thorne, 1971; Zuckerman & Link, 1968), the F

(deviant responding) and Psychopathic deviate scales of the MMPI

(Zuckerman et al., 1972), and Murray-type needs for Autonomy, Change,

and Exhibition (Zuckerman & Link, 1968). In some studies, a negative

16










correlation between sensation seeking and anxiety has been found (Segal,

1973; Zuckerman, Kolin, Price, & Zoob, 1964), whereas in others no

relationship was demonstrated (Zuckerman & Link, 1968). In terms of

behaviors, sensation seeking has been found to be related to gambling

(Anderson & Brown, 1984), sexual experience (Zuckerman et al., 1972),

smoking (Zuckerman, Ball, & Black, 1990), and both drug and alcohol use

in nonstudent populations (e.g. Marvel & Hartmann, 1986) and in the

college population (Jaffe & Archer, 1987; Kohn & Coulas, 1985; Segal et

al., 1980a, 1980b). With regard to major demographic factors, it has

been demonstrated that whites scored higher on sensation seeking than

blacks (Jaffe & Archer, 1987; Kaestner, Rosen, & Appel, 1977), males

scored higher than females (Jaffe & Archer, 1987; Ratliff & Burkhart,

1984; Zuckerman, Eysenck, & Eysenck, 1978), and scores declined with age

(Zuckerman et al., 1978).

In their review of college drinking, Brennan et al. (1986a)

described a variety of studies in which a relationship was found between

drinking and personality variables they grouped as a dimension of

sensation seeking/impulse expression. For example, Moos, Moos, and

Kulik (1977) found that students who increased or maintained heavy

consumption from the beginning of their freshman year to the end of it

were more impulsive and engaged in more deviant behavior. Results of

other longitudinal studies of adolescents or college students who became

problem drinkers include similar personality correlates. Jones (1968)

found that junior high students who were later tested in senior high and

as adults, and who became problem drinkers, were impulsive and

rebellious. Donovan et al. (1983) found that high school and college








18

students who became problem drinkers were more prone to a variety of

deviant or problem behaviors. Labouvie and McGee (1986) found that, in

adolescents aged 12, 15, and 18 who were tested three years later,

heavier drinking, smoking, and drug use were related to Autonomy,

Exhibition, and Impulsivity. Also, it was shown in a retrospective

study (Loper et al., 1973) that a sample of alcoholic patients who had

taken the MMPI as college students had at that time scored significantly

higher on F, Psychopathic deviate, and Hypomania scales.

According to Levenson, Oyama, and Meek (1987), the MacAndrew

Alcoholism scale and the Socialization scale of the California

Psychological Inventory measure a group of traits which is very similar

to sensation seeking. Using these two scales, Levenson et al. (1987)

found that subjects who scored according to the typical prealcoholic

profile showed a decreased physiological response to stress when given

alcohol. Another group of researchers who used these same two scales as

a measure of personality risk for drinking problems found that, in a

sample of male college students, scores indicating increased impulsive

and aggressive characteristics were correlated with quantity and

frequency of consumption (Earleywine et al., 1990).

Sensation Seeking and College Student Drinking

On the basis of the results of Earleywine et al. (1990) and other

alcohol research, Earleywine and Finn (1991) hypothesized that the

relationship between drinking and behavioral disinhibition as measured

by the MacAndrew and Socialization scales might be indirect and that

sensation seeking might be a third variable that, by being related both

to drinking and behavioral disinhibition, could account for their








19

correlation to each other. In this study, the factors of Experience

Seeking and Disinhibition from the Sensation Seeking Scale were included

in the researchers' measure of sensation seeking. Their analyses of

data on male undergraduates confirmed their hypothesis that sensation

seeking could account for the link between behavioral disinhibition and

drinking.

The relationship between college drinking and sensation seeking has

been found in a variety of studies. Noting the marked similarity

between premorbid personality correlates of problem drinkers and

sensation seeking, Schwarz et al. (1978) hypothesized that sensation

seeking might comprise a motive for college drinking. As suggested by

their title, "Turning On or Turning Off..." (p. 1144), Schwarz et al.

(1978) compared sensation seeking and tension reduction as motives for

college drinking. They found that all scales of Form IV of the

Sensation Seeking Scale (Zuckerman, 1979) were correlated with alcohol

use, whereas none of the scales of the S-R Inventory of General Trait

Anxiousness were correlated with alcohol use. Using multiple regression

analyses, they compared the relative contribution of anxiety and

sensation-seeking scores to alcohol use. The results of these analyses

were that anxiety contributed little (4% when entered first, 1% when

second), and sensation seeking accounted for most of the variance (30%

when entered second, 33% when first) in drinking scores. They concluded

that, because of the cultural expectancy that drinking disinhibits

behavior, individuals who have a high need for sensation seeking or

disinhibition will be motivated to use alcohol to obtain impunity for

less restrained behavior.








20

With the goal of establishing a causal relationship between

sensation seeking and drinking, Schwarz et al. (1982) conducted an

experimental study with 43 white male undergraduates. These students,

who had been told the study involved group processes and drinking and

that alcohol would be served, were randomly assigned to either of two

groups. One group, considered a stimulating, low-restraint condition,

was led to expect to be taken to a party at which alcohol would be

provided. The Sensation Seeking Scale (Zuckerman, 1979) and a measure

of mood were administered first. Following this, an excuse of having to

wait for late subjects was given in order to justify serving alcohol

while waiting for the party. Consumption of the punch (with 1 teaspoon

of alcohol added for each subject) was measured at three intervals. The

second group, a boring, high-restraint condition, was subject to the

same procedure except that it was led to expect a group discussion

rather than a party. The researchers' first hypothesis, that high

sensation seekers, defined by a median split of the Disinhibition scale,

would drink more, was not confirmed. This hypothesis was supported only

by using a median split of the General factor (Form IV) of sensation

seeking. The second hypothesis, that high sensation seekers would drink

more when expecting a party (stimulating, low-restraint condition), was

not supported using either factor of sensation seeking. Schwarz et al.

(1982) concluded that the laboratory setting and deception employed may

not have been sufficient to stimulate a disinhibition motive for

drinking.

In order to further define motivation for drinking among college

students, Ratliff and Burkhart (1984) conducted a study with a sample of









21

70 male and 70 female undergraduates. They compared men and women and

light and heavy drinkers on a variety of measures, including Form V of

the Sensation Seeking Scale, the Beck Depression Inventory, the Self-

Esteem Scale, the S-R Inventory of General Trait Anxiousness, and the

Reasons for Drinking Scale. Ratliff and Burkhart (1984) hypothesized

that heavy drinkers would be higher in sensation seeking and that women

who were heavy drinkers would exhibit more affective distress. The

results of primary interest here are that heavy drinkers scored higher

on all factors of sensation seeking, especially Disinhibition, and that

men scored higher than women on all but the Experience-Seeking factor.

Other motives based on personality variables did not appear to

differentiate drinkers, as there was no difference between light and

heavy drinkers in depression or self-esteem, and women who were heavy

drinkers did not show more affective disturbance. Ratliff and Burkhart

(1984) concluded that needs for sensation seeking and disinhibition

constitute a primary motive for student drinking. On the basis of these

and other findings, including the finding that men reported more

problems with authorities as a result of drinking, the researchers

suggested that women may expect drinking to produce social enhancement

and men may expect more disinhibition, particularly of aggressive

behavior.

Other researchers (Segal et al., 1980a, 1980b) have found a

relationship between sensation seeking and student use of both alcohol

and other drugs. Segal et al. (1980a) differentiated nonusers from

substance users among undergraduates on the basis of a variety of

personality measures, including all factors of the Sensation Seeking








22

Scale except Boredom Susceptibility (Zuckerman, 1979). The factors of

the Sensation Seeking Scale, and particularly the Experience-Seeking and

Disinhibition factors, as well as similar variables of Impulsivity and

Autonomy, contributed to the differentiation of users from nonusers.

Segal et al. (1980b) conducted a second study to assess various

reasons for drinking and to analyze them in conjunction with the

personality variables tested earlier (Segal et al., 1980a). Using a

series of canonical correlations, they found that the dimension on which

Experience Seeking and Disinhibition loaded was more typical of males

and was associated with several reasons for drinking, including both

social reasons (i.e. "get together more fun") and escapist reasons

(i.e. "feel better") (Segal et al., 1980b, p. 492). The researchers

suggested that this dimension probably reflected the normal pattern of

college drinking. As a result of these and other studies, Segal (1983)

concluded that a sensation-seeking motive is a common reason for alcohol

and drug use by college students and adolescents.

Other researchers have found a relationship between sensation

seeking and college student use of alcohol and other drugs. Kohn and

Coulas (1985) demonstrated that high scores on Disinhibition were

positively correlated with the use of alcohol, marijuana, and tobacco.

Students who scored high on Disinhibition also rated drug effects as

more desirable, whether those effects were stimulant or depressant in

nature. These researchers concluded that students who are motivated to

use alcohol and other drugs "seem attracted more by the general

consciousness alteration associated with their use than by specific








23

classes of attributed effects (e.g. stimulant vs. depressant)" (Kohn &

Coulas, 1985, p. 104).

In a study involving a wide range of drugs, Jaffe and Archer (1987)

compared the abilities of personality and alcohol and drug risk scales

to predict undergraduate use of substances. They found that the

Sensation Seeking Scale (total score) was the best predictor of use of

almost all substances, including alcohol. In addition, students

identified experience- and pleasure-seeking motives for alcohol and drug

use as more important than motives of self-medication and sociality.

The consensus of the literature on college drinking is that the

dimension of sensation seeking/impulse expression is the most

consistently supported personality correlate of student use of alcohol

(Brennan et al., 1986a). Prior to the current study, however, no

studies had been conducted to determine whether the relationship between

sensation seeking and drinking reflected a motive of becoming

disinhibited in terms of increased sexual, aggressive, and/or social

behaviors or in terms of being less inhibited by tension or anxiety.

Expectancy Effect of Disinhibition in Experimental Research

A number of experimental studies have been conducted to test a

general expectancy effect or instructional set of receiving alcohol.

Many of these studies employed a balanced-placebo design, where

conditions of receiving alcohol (yes/no) were crossed with conditions of

expecting alcohol (yes/no) via instructional set, in order to separate

pharmacological and general expectancy or placebo effects (Hull & Bond,

1986). The ability of instructional set, or expectancy as used in this

sense in experiments, to elicit behaviors in people who believed they









24

had been served an alcoholic beverage has been interpreted to be the

result of the specific outcome expectancies they hold with regard to

drinking (Marlatt & Rohsenow, 1980).

In a variety of these balanced-placebo design experiments, it has

been shown that the belief alone that alcohol was consumed has been

causally related to increased disinhibited behaviors of aggression,

sexual response, sociability, and laughter (Lang, 1983). Connors and

Maisto (1988) concluded that "expectancy seems to have its greatest

effects on the behaviors or emotions that society proscribes from free

expression" (p. 488). According to Hull and Bond (1986), an expectancy

effect of increased sexual arousal has been most consistently supported

in these experiments. In a review of the effects of alcohol and

expectancy on sexuality, Crowe and George (1989) concluded that

expectancy can disinhibit sexual arousal and thus allow the avoidance of

attribution of responsibility by the individual ("self-excusing") or to

the individual by society ("social excusing") (p. 382). It was

suggested by some researchers (Lang, Searles, Lauerman, & Adesso, 1980)

that such an evasion of an attribution of responsibility could account

for results in their experiment with male undergraduates. In this

experiment, students who expected alcohol, and particularly those with

high sex guilt, spent an increased amount of time viewing erotic slides.

Lang et al. (1980) concluded that individual differences could mediate

an expectancy effect in experimental research and that the

"disinhibiting action of an alcohol expectancy set is a function of its

potential reinforcing value for each individual" (p. 645).











Reinforcing Alcohol Expectancies

One researcher in the area of alcohol and sexuality has said that

"our societal definition of alcohol is that of a disinhibitor and

aphrodisiac" (Wilson, 1981, p. 34). In descriptive research on the

content of alcohol expectancies, disinhibition has been found to be a

common expectancy in a variety of populations (Critchlow, 1986; Goldman

et al., 1987). Southwick et al. (1981) found an expectancy factor of

Disinhibition in college students, as did Leigh (1987) in both college

and general population groups. More recently, a sample of Swedish

female undergraduates reported an expectancy of disinhibition for both

themselves and other drinkers (Gustafson, 1990).

In the study by Leigh (1987), the items describing alcohol effects

of becoming sexually aggressive, romantic, and friendly comprised a

factor she labeled Gregariousness. The factor of Disinhibition was

comprised of expected effects of becoming silly or loud, losing self-

control, or doing things not done when sober. The Disinhibition factor

on the questionnaire developed by Southwick et al. (1981) contained such

items as expecting to become more uninhibited, sexual, and talkative.

The discrepancies in these items that may constitute an alcohol

expectancy of disinhibition are in part a reflection of the difficulty

in defining the concept of disinhibition (Critchlow, 1986).

Nevertheless, the concept of alcohol-facilitated disinhibition continues

to figure prominently in both alcohol research and cultural expectancies

about drinking (Room & Collins, 1983).

The effects of Sexual Enhancement and Arousal and Aggression, which

are usually considered to be alcohol-induced disinhibition (Brown &








26

Munson, 1987), have been found to be expectancy factors of

questionnaires developed in the United States (Brown et al., 1980) and

in New Zealand (Young & Knight, 1989). Items describing expected

alcohol effects of increased sexual and aggressive behaviors have also

been found in the development of other questionnaires (Leigh, 1987;

Southwick et al., 1981).

The dimension of increased sociability is another major alcohol

expectancy, which has for example been described by alcohol effects of

becoming talkative and outgoing (Southwick et al., 1981) or friendly

(Leigh, 1987). Brown et al. (1980) found an expectancy factor of Social

and Physical Pleasure in the development of their questionnaire. A

similar factor of Social Enhancement was reported in a study of both

student and general populations in New Zealand (Young & Knight, 1989).

Both of these last two groups of researchers have also distinguished a

separate expectancy of Social Assertiveness.

An additional alcohol expectancy is reflected in the widespread

belief that drinking will reduce anxiety or promote relaxation (Goldman

et al., 1987). Southwick et al. (1981) found that the effect of

becoming relaxed comprised part of the factor of Pleasurable

Disinhibition on their questionnaire. A separate expectancy factor of

Tension Reduction was distinguished by Brown et al. (1980) in the

development of their questionnaire. Similarly, Young and Knight (1989)

derived a factor of Relaxation as an expected effect of drinking.

Alcohol Expectancies and Drinking Behavior

Researchers using alcohol expectancy questionnaires have

consistently found that expectancies are related to or predict drinking








27

behavior of college students, adolescents, the general population, and

alcoholics in treatment (Goldman et al., 1987; Leigh, 1989c). Similar

results were found by some researchers (McCarty, Morrison, & Mills,

1983) who, using a measure of beliefs and attitudes, found that heavier

drinking students believed that drinking results in more positive

experiences. Among college students and the general population, Leigh

(1987) found that heavier drinkers had increased alcohol expectancies of

both positive and negative effects, but they did not evaluate negative

effects as being as unpleasant as did lighter drinkers. In addition,

college students in that study reported becoming more disinhibited and

outgoing, as well as experiencing less of the negative effects of

drinking, than did the general population. Similarly, Rohsenow (1983)

reported that college students who were moderate and heavy drinkers

expected more positive effects but no more negative effects than light

drinkers. Again, similar results were obtained by Southwick et al.

(1981), who found that heavier drinking among college students was

associated with expectancy factors of greater Stimulation and

Pleasurable Disinhibition but not with more Behavioral Impairment as a

result of drinking a moderate dose of alcohol. In other studies of

college drinking, it has been demonstrated that alcohol expectancies of

positive effects were associated with a higher level of consumption

(Brown & Munson, 1987) and predicted subsequent alcohol use (Stacy et

al., 1990). Brown, Goldman, and Christiansen (1985) also found that

stronger expectancies were associated with excessive and abusive

drinking in college, general, and alcoholic populations. In addition, a

negative relationship has been found between alcohol expectancy scores









28

and the length of sobriety for alcoholics in recovery (Rather & Sherman,

1989) and with other measures of treatment success (Brown, 1985b).

Brown (1985a) reported that different patterns of college drinking

were differentially predicted by alcohol expectancies. She found that

nonproblematic drinking was best predicted by the expectancy of Social

and Physical Pleasure whereas an expectancy of Tension Reduction was

most predictive of problem drinking. However, Mooney, Fromme, Kivlahan,

and Marlatt (1987) found that alcohol expectancies differentially

predict quantity and frequency of college drinking. For example,

quantity of consumption was best predicted by expectancies of Social and

Physical Pleasure and Social Assertion.

In addition, Mooney et al. (1987) found that expectancies

differentially predicted frequency of consumption for male and female

students. Men who drank more frequently had stronger expectancies of

Social and Physical Pleasure, Global Positive Changes, and Sexual

Enhancement, whereas for women, frequent drinking was predicted by the

expectancy of Tension Reduction. Using a composite quantity-frequency

variable of consumption, Brown et al. (1980) found that men from

student and general populations expected more Arousal and Aggression and

women expected more Social and Physical Pleasure. However, Rohsenow

(1983) suggested that gender and drinking habit were confounded in the

study by Brown et al. (1980) because women were significantly lighter

drinkers. After statistically controlling for the gender difference in

consumption, Rohsenow (1983) found that female undergraduates expected

less Social and Physical Pleasure and Tension Reduction than men and did

not differ from men in expectancies of other positive effects. Thus








29

there is considerable variability in the research results on gender

differences in alcohol expectancies.

Personality. Expectancies. and Drinking Behaviors

Several studies have been conducted to determine how personality

may be related to individuals' outcome expectancies from drinking (Brown

& Munson, 1987; Leonard & Blane, 1988) and how, together, personality

and expectancies may relate to drinking patterns (Mann, Chassin, & Sher,

1987; Mooney & Corcoran, 1989). Brown and Munson (1987) tested the

hypotheses that college students with higher levels of extraversion and

trait anxiety would have stronger expectancies. They found that

extraverted students expected more Social and Physical Pleasure and

Tension Reduction than introverted students, whereas students with high

trait anxiety levels had increased scores of the other four expectancy

factors measured: Global Positive Changes, Sexual Enhancement, Social

Assertion, and Arousal and Aggression. In two samples of men, one from

a college population and the other from a general population, Leonard

and Blane (1988) studied the relationships between alcohol expectancies

and personality variables that appear relevant to those expectancies.

For both samples, a higher level of hostility was correlated with the

alcohol expectancy of Arousal and Aggression, and increased social

anxiety was correlated with alcohol expectancies of Social Assertion and

Global Positive Changes.

In the study by Mooney and Corcoran (1989), the hypothesis of

interest was whether the alcohol expectancy of Social Assertion would be

more predictive of heavy drinking in college students who were low in

assertiveness as compared to students high in assertiveness. Separate









30

multiple regression analyses were used to predict usual quantity, usual

frequency, and maximum quantity of consumption. For women who were low

in assertiveness, all three consumption variables were predicted only by

the alcohol expectancy of Social Assertion. This expectancy also

predicted the frequency of drinking for men who were low in

assertiveness. For highly assertive students, none of the consumption

variables was predicted by the alcohol expectancy of Social Assertion.

Mooney and Corcoran (1989) suggested that personality variables that are

relevant to particular alcohol expectancies may represent the

reinforcement value of those expectancies and therefore may improve

predictions of drinking behavior.

In a study of high school students, Mann et al. (1987) tested the

hypothesis that the relation between alcohol involvement (quantity,

frequency, and negative consequences of drinking) and alcohol

expectancies would vary with risk status, including personality risk for

problem drinking. Items from the Socialization scale of the California

Psychological Inventory and from the MacAndrew Alcoholism scale were

used as the measure of personality risk. Alcohol expectancies were

measured by the Alcohol Expectancy Questionnaire for Adolescents

(Christiansen, Goldman, & Inn, 1982). Results of this study included

the finding that expectancies of altered social behavior were more

predictive of alcohol involvement for adolescents who were low in

personality risk. For students high in personality risk, alcohol

involvement was predicted by alcohol expectancy factors of Tension

Reduction and Enhanced Cognitive and Motor Functioning. These results









31

are indicative of the importance of considering personality risk

variables in research on motives for drinking (Mann et al., 1987).

Support for Theoretical Framework

Some researchers have demonstrated that alcohol expectancies are

expressed by children and adolescents prior to drinking experience

(Christiansen et al., 1982; Miller, Smith, & Goldman, 1990) and that

their expectancies tend to become more homogeneous with drinking

experience (Christiansen et al., 1982). In a prospective study of

adolescent drinking, higher scores on some alcohol expectancy factors

predicted later drinking behavior, including the transition one year

later from normal to problem drinking (Christiansen, Smith, Roehling, &

Goldman, 1989). In another prospective study of adolescent drinking

(Bauman, Fisher, Bryan, & Chenoweth, 1985), it was shown that the

utility of expected consequences of drinking, determined by ratings of

probability and desirability of alcohol effects, was associated with

subsequent drinking level. In addition, it was found that earlier

drinking habits were related to later utility scores, or in other words,

that there was a reciprocal relation between expectancies and drinking.

This reciprocal relation may represent a process by which expectancies

both influence drinking and become modified through drinking experience

(Bauman & Bryan, 1980).

According to expectancy models of alcohol use, expectancy is "a

primary motivational factor for initiating drinking in a setting"

(Maisto et al., 1981, p. 13). In addition, drinkers' decisions "to

continue drinking at any time are mediated by what they anticipate the

effects of alcohol will be" (Connors & Maisto, 1988, p. 489). Drinkers'








32

perceptions of physiological, psychological, and behavioral changes that

result from drinking may be compared to desired and expected degrees of

change (Maisto et al., 1981). These assessments of alcohol effects then

influence decisions about further drinking (Goldman et al., 1987; Maisto

et al., 1981).

Pharmacological effects and social learning influences probably

interact in the development of alcohol expectancies at the individual

level (Brown et al., 1980) and at the cultural level (Critchlow, 1986).

Prior to drinking experience, alcohol expectancies develop as a result

of social learning (Miller et al., 1990). These expectancies may then

alternately influence and be modified by drinking experience (Bauman &

Bryan, 1980). In social learning and particularly expectancy theories

of drinking, it is recognized that both culture and individual

differences affect exposure to and perceptions of various learning

experiences about drinking (Abrams & Niaura, 1987; Goldman et al.,

1987).

It has been suggested that the relationship between drinking and

personality may be stronger than previously concluded in the literature

(Earleywine et al., 1990). Earleywine et al. (1990) hypothesized that

the correlation between drinking and personality could be underestimated

due to imperfect reliability of measures of both constructs. Using

confirmatory factor analysis to take into account measurement error,

these researchers found a correlation of .80 between alcohol consumption

and personality risk, as measured by the MacAndrew and Socialization

scales, in a sample of undergraduate men.









33

Personality variables may add to the explanations of motivation for

drinking which are based on alcohol expectancies due to a relationship

between personality and expectancies (Brown & Munson, 1987). For

example, personality variables may be a measure of the utility or

reinforcement value of individuals' alcohol expectancies (Lang et al.,

1980; Mooney & Corcoran, 1989). It has been demonstrated that the

relationship between drinking behavior and alcohol expectancies varies

with personality (Mann et al., 1987; Mooney & Corcoran, 1989).

The literature reviewed here has supported a consistent

relationship between drinking and especially the personality variable of

sensation seeking. It has been suggested that the relationship between

sensation seeking and drinking reflects a motive of disinhibition or

motivation to drink in order to express behaviors that are usually

subject to social restraint (Berkowitz & Perkins, 1986; Ratliff &

Burkhart, 1984; Schwarz et al., 1978). Schwarz et al. (1978)

hypothesized the link between sensation-seeking needs and drinking to be

the result of a cultural expectancy of disinhibition which allows

decreased social control over behavior during or following drinking.

They suggested that the "reinforcement potential provided by access to

such disinhibited states" (Schwarz et al., 1978, p. 1145) would be

particularly relevant for those with high sensation-seeking needs.

According to personality theory of alcohol use, sensation seeking

represents a motive to obtain stimulus enhancement or positive

reinforcement from drinking rather than a motive for reduction of

negative affect (Cox, 1987). This interpretation seems logical in terms

of the definition of sensation-seeking needs. Support for a predominant









34

positive reinforcement motive for student drinking has also come from

research in which it was found that most college students expressed a

greater desire to drink in pleasant settings and when in pleasant moods

rather than when in unpleasant settings or moods (Russell & Bond, 1979,

1980). However, in light of the finding by Segal et al. (1980b) of a

dimension of drinking that they interpreted to be the normal pattern of

college drinking, a pattern which was characterized by sensation seeking

and both social and escapist reasons for drinking, it is uncertain

whether a sensation-seeking motive for college drinking reflects solely

or primarily the enhancement of positive affect.

Support for Need for the Study

As has been suggested by Earleywine et al. (1990), the relationship

between personality and drinking may have been underestimated. In

another study (Mann et al., 1987), it was concluded that risk status for

problem drinking should include personality variables as well as family

risk factors. If in the current study sensation seeking contributed

significantly to drinking, this result would provide justification for

the inclusion of personality variables in research on alcohol use.

Similarly, expectancy theory would be supported if alcohol expectancies

were found to be related to drinking, a result which has been

consistently reported in the literature (Goldman et al., 1987; Leigh,

1989c).

This study would also provide an opportunity to replicate the

finding of higher sensation-seeking needs in male students (Jaffe &

Archer, 1987; Ratliff & Burkhart, 1984). In addition, results of this

study would help clarify gender differences in the relationships between









35

alcohol expectancies and drinking level. These issues have needed

clarification because the results of previous research have been

inconsistent (Brown et al., 1980; Rohsenow, 1983).

The current study represented an original test of relationships

between sensation seeking and alcohol expectancies. This question

addressed the need recognized by other researchers who have investigated

the relationship of personality to alcohol expectancies (e.g. Leonard &

Blane, 1988). In addition, the need to understand how personality

variables may moderate the relationship between drinking and alcohol

expectancies (Stacy et al., 1990) was addressed in the current study by

determining which alcohol expectancies contributed to drinking for

students with varying levels of sensation-seeking needs.

These questions about the relationships among drinking, sensation

seeking, and alcohol expectancies provided an opportunity to contribute

to theory and research on the nature of a disinhibition motive for

drinking. Despite the prevalence of a cultural expectancy of

disinhibition and the impact of disinhibition concepts on alcohol

research, it is unclear whether alcohol-induced disinhibition refers to

a deviation from social norms, loss of control, or simply increased

sexual, aggressive, or other social behaviors (Fingarette, 1983). If

sensation-seeking needs were found to be related to alcohol expectancies

of Sexual Enhancement and/or Arousal and Aggression, and if these

expectancies together with sensation-seeing needs were related to

drinking, then a disinhibition motive would be supported in the most

commonly understood sense (e.g. Brown & Munson, 1987).









36

The current study also provided information that is relevant to

alcohol prevention and treatment efforts. Motivation for college

drinking, including the influence of alcohol expectancies, needs to be

considered when designing campus abuse prevention programs (Berkowitz &

Perkins, 1986). Several researchers have suggested the clinical and

educational importance of exploring or assessing alcohol expectancies

(Berkowitz & Perkins, 1986; Connors & Maisto, 1988; Cooper et al.,

1988). In education and treatment programs, the exploration of alcohol

expectancies should illuminate instances where negative consequences

result from drinking in contrast to expectancies of desirable outcomes

(Connors & Maisto, 1988). An understanding of unrealistic expectations

of alcohol effects may be especially important in cases of alcohol abuse

(Maisto et al., 1981). Cognitive restructuring or alteration of

expectancies has been suggested as a valuable strategy in prevention and

treatment programs (Brown et al., 1985; Connors & Maisto, 1988; Mann et

al., 1987).

A knowledge of alcohol expectancies that motivate drinking may

assist in designing treatment programs for problem drinkers (Brown et

al., 1985; Rohsenow, 1983). In addition, an understanding of alcohol

expectancies and personality characteristics that are important

contributors to drinking may assist in tailoring treatment and

prevention programs to specific groups (Mooney & Corcoran, 1989).

Alternate ways of obtaining expected outcomes from drinking or

meeting sensation-seeking needs, for example, should also be promoted

(Connors & Maisto, 1988; Segal, 1983). For example, college drinkers

who are motivated to drink to enhance stimulation or to get "high" would









37

probably benefit most from the promotion of alternative activities

(Segal, 1983). Improved understanding of motivation for drinking should

also assist in the application of social and coping skills training in

prevention and treatment programs (Connors & Maisto, 1988).

Support for Approach to the Study

For this study on motivation for drinking, a descriptive approach

and survey method had several advantages. It is likely that alcohol

expectancies which have developed over an individual's lifetime would be

difficult to manipulate in a brief laboratory experience (Leigh, 1989c).

The difficulty in operationalizing specific expectancies was also noted

by Schwarz et al. (1982), who concluded that their experimental

condition of expecting a party may not have been sufficient to elicit a

disinhibition expectancy and hence increased alcohol consumption in high

sensation-seeking subjects. In addition, there are ethical issues

related to administering alcohol to experimental participants or

providing an effective deception or placebo (e.g. Schwarz et al., 1982).

Ethical and practical considerations often favor a self-report measure

of alcohol use (Fitzgerald, Mulford, & Mulford, 1987). One such

consideration is the efficiency of self-reported alcohol use for large

sample data collection (Isaac & Michael, 1981). Another advantage of

self-reported alcohol use is that anonymity can be assured, in contrast

to, for example, obtaining collateral reports of subjects' drinking

habits (see Polich, 1982).

A number of researchers have found the validity and reliability of

self-reports of alcohol consumption to be adequate for most research

purposes. In a large survey study of U.S. Air Force personnel, Polich









38

and Orvis (1979) found that, although self-reports represented up to 20%

less than alcohol sales, this degree of underreporting made an

insignificant contribution to their estimate of the prevalence of

alcohol problems. In a college population, researchers (Stacy, Widaman,

Hays, & DiMatteo, 1985) have demonstrated that self-reports and peer

reports among pairs of friends generally showed adequate convergent and

discriminative validity for use of alcohol, marijuana, and cigarettes.

In a household survey, Williams et al. (1985) tested both validity and

reliability of self-reported alcohol use. Their analyses of alternate-

form (14-day and 28-day recall periods) and test-retest reliabilities

yielded an average of .91. Concurrent validity between both self-report

forms and a drinking diary generally averaged .80 (Williams et al.,

1985).

In a recent study of an undergraduate population, Stacy et al.

(1990) investigated whether the self-report nature of both alcohol use

and expectancies might produce a bias that could account for the

relationship between drinking and alcohol expectancies. In order to

investigate this potential self-report bias, Stacy et al. (1990)

compared the factor correlation between self-reported consumption and

self-reported alcohol expectancies with the factor correlation between

peer-reported consumption and self-reported alcohol expectancies. The

correlation between self-reported drinking and expectancies was not

greater than that between peer-reported drinking and expectancies, which

indicated the absence of a self-report bias in the expectancy-drinking

association (Stacy et al., 1990).















CHAPTER 3
METHODOLOGY



Overview of the Study

Motivation for drinking among male and female college students was

examined in this study in terms of their sensation-seeking needs and

expectancies of reinforcing alcohol effects. In order to improve

understanding of motivation for drinking, two major questions were

investigated: (a) how alcohol expectancies and sensation-seeking needs

were related and (b) how sensation-seeking needs and alcohol

expectancies together contributed to drinking level. Gender differences

in these relationships were also examined.

Null Hypotheses

The null hypotheses of this study were:

1. Ho: There is no composite influence of alcohol expectancies,

gender, and their interactions on sensation-seeking needs of college

students.

a. Ho: There is no gender difference in sensation-seeking

needs of college students.

b. Ho: There are no relationships between alcohol

expectancies and sensation-seeking needs of college

students.









40

c. Ho: There are no gender differences in the relationships

between alcohol expectancies and sensation-seeking

needs of college students.

2. Ho: There is no composite influence of sensation-seeking needs,

alcohol expectancies, gender, and their interactions on drinking among

college students.

a. Ho: There is no relationship between sensation-seeking

needs and drinking among college students.

b. Ho: There is no gender difference in the relationship

between sensation-seeking needs and drinking among

college students.

c. Ho: There are no relationships between alcohol

expectancies and drinking among college students.

d. Ho: There are no gender differences in the relationships

between alcohol expectancies and drinking among

college students.

e. Ho: There are no differences in the relationships between

alcohol expectancies and drinking for differing

levels of sensation-seeking needs of college

students.

f. Ho: There are no gender differences in the relationships

between alcohol expectancies and drinking for

differing levels of sensation-seeking needs of

college students.

g. Ho: There is no gender difference in drinking among

college students.









41

Delineation of Relevant Variables

Demographic Variables

The students were requested to indicate their age, gender,

ethnicity, and body weight (see Appendix A). Gender was assessed in

order to determine differences in motivation for drinking among males

and females. Ethnicity choices were (a) White American, Caucasian

American; (b) Black American, African American; (c) Hispanic American;

(d) Asian American; and (e) other than above. Age, gender, and

ethnicity were used to describe the sample. Body weight was required to

standardize alcohol consumption as described in the next section.

Drinking Level

Alcohol consumption was measured by the Khavari Alcohol Test

(Khavari & Farber, 1978). For each type of alcoholic beverage, beer,

wine, and liquor, this questionnaire presents choices in frequency and

amount that correspond to a subject's usual and maximum consumption.

These variables are used to derive an aggregate index of consumption for

all types of alcoholic beverages. This aggregate index is designated

the annual absolute alcohol intake (AAAI).

Subjects' AAAI scores were then divided by their body weight in

order to adjust for the effect of alcohol depending on the weight of the

individual (K. A. Khavari, personal communication, May 31, 1991).

Standardization of consumption by body weight may be particularly useful

when comparing male and female drinkers to take into account the higher

average weight of males (Berkowitz & Perkins, 1987; Brennan et al.,

1986a; Ratliff & Burkhart, 1984). The weight-standardized consumption

was the variable denoting drinking level for subjects of this study.











Sensation-Seeking Needs

Sensation-seeking needs were measured by Form V of the Sensation

Seeking Scale (Zuckerman, 1979). This scale yields a total score as

well as scores derived from four factor subscales: Disinhibition,

Experience Seeking, Thrill and Adventure Seeking, and Boredom

Susceptibility. Because total sensation-seeking score and the subscales

of Disinhibition and Experience Seeking have most often been found to

relate to college drinking (e.g. Jaffe & Archer, 1987; Kohn & Coulas,

1985; Segal et al., 1980a, 1980b), these three scores were used in the

current study for the selection of the sensation-seeking variable.

Among the variables of total score, Disinhibition, and Experience

Seeking, the variable that was most highly correlated with drinking

level in the study sample was operationally defined as the variable of

sensation-seeking needs. Thus, the data analyses were simplified

through the reduction in number of sensation-seeking variables to the

one that was most related to drinking in the sample.

Alcohol Expectancies

Positive alcohol expectancies were assessed by the Adult Form III

of the Alcohol Expectancy Questionnaire (Brown et al., 1980). The

specific alcohol expectancies measured by this questionnaire are

represented by the following six factors or scales: Global Positive

Changes, Sexual Enhancement, Social and Physical Pleasure, Social

Assertion, Tension Reduction, and Arousal and Aggression. For the

current study, the entire questionnaire was administered, but the factor

of Global Positive Changes was not scored because it has previously been

found to be most highly intercorrelated with the other five factors









43

(Leigh, 1989b). Pearson correlations among the remaining five factors

were computed to determine if they were sufficiently distinct to be

included as five separate variables in the study. Any two factors with

an intercorrelation of .80 or higher were to be combined as a single

variable. If all intercorrelations among the five factors were less

than .80, the five alcohol expectancies would be retained as separate

variables in the study.

Description of the Assessment Instruments and Resulting Data

The Khavari Alcohol Test

In alcohol research, measures of consumption have traditionally

been based on some combination of variables denoting quantity and

frequency of consumption (e.g. Straus & Bacon, 1953). Using similar

variables, Cahalan et al. (1969) developed a measure called the volume-

variability index, which they reported as allowing greater

differentiation of types of drinkers, for example light, moderate, and

heavy drinkers. Khavari and Farber (1978) modified this index to

produce the Khavari Alcohol Test. This questionnaire was also designed

to include a variability or binge aspect of drinking behavior. However,

the Khavari Alcohol Test produces a continuous variable of consumption

without the necessity of classifying drinkers into types. If such a

classification is desired, the researcher may do so empirically rather

than by predetermined criteria. Khavari and Farber (1978) developed

standardized T scores for consumption variables of the instrument using

means and standard deviations of alcohol consumption reported by a

sample of 2237 college students and other adults.









44

Instructions for the Khavari Alcohol Test (see Appendix B) identify

the instrument as asking questions about alcohol use. Respondents are

instructed to check or write in their answers. The KAT consists of 12

questions, 4 questions being repeated for each alcoholic beverage type,

beer, wine, and liquor. For each beverage type, these questions assess

four aspects of the subjects' alcohol consumption: (a) usual frequency

(Fu), (b) usual quantity or volume (Vu), (c) maximum quantity on some

occasion (VM), and (d) the frequency of that maximum amount (FM). Nine

choices for frequency of both usual and maximum amounts consumed are

presented. These frequencies correspond to a range from daily use to

never having tried that type of alcoholic beverage. For the purposes of

computing an annual consumption score, these frequencies are converted

to their corresponding annual frequencies. For example, daily use is

converted to an annual frequency of 365, drinking once a week is

converted to an annual frequency of 52, and so on.

The variables of quantity consumed, both usual and maximum on some

occasion, are measured by a range of choices from never consuming that

beverage type to 15 cans, glasses, or bottles of that beverage. For any

quantity greater than 15 drinks, the subject is asked to write in that

amount. Usual and maximum quantities are then converted to ounces of

beverage consumed. The number of drinks is multiplied by the assumed

average number of ounces for that beverage type. These are 10 ounces

for a beer, 4 ounces for a glass of wine, and 1.5 ounces of liquor in a

drink.

Using the annual frequencies of both usual and maximum amounts

consumed and the number of ounces for both usual and maximum quantities,









45

an annual volume score (Va) is computed for each beverage type by the

formula: Va = (Fu Fm)Vu + Fm(VM). These computations result in

separate annual volume scores (Va) for beer, wine, and liquor.

In order to provide an aggregate consumption score, the differing

percentages of alcohol for each beverage type must be taken into

account. It is assumed that the average percentages of absolute alcohol

in the beverage types are 4.5% for beer, 15% for wine, and 45% for

liquor. Annual volumes (Va) for beer, wine, and liquor are converted to

absolute alcohol by multiplying each by their corresponding percentage

of absolute alcohol. The summation of these three converted volumes

results in the annual absolute alcohol intake (AAAI). These last two

procedures are represented by the formula: AAAI = Va(beer) x .045 +

Va(wine) x .15 + Va(liquor) x .45.

The annual absolute alcohol intake (AAAI) was divided by body

weight for each subject. This weight-standardized intake was the

variable of consumption or drinking level that was used in all data

analyses.

As a test of validity of the instrument, Khavari and Farber (1978)

compared the alcohol consumption of two alcoholic samples and three

nonalcoholic samples. The AAAI scores for the alcoholic samples were

two standard deviations above the mean of the normative sample used in

developing the test, whereas the AAAI scores for the three nonalcoholic

groups did not differ from the mean of the normative sample.

Khavari and Farber (1978) obtained data from the entire normative

sample on the Khavari Alcohol Test and on an alternate measure of

alcohol use. They reported a reliability coefficient of .80 between









46

these two measures of alcohol use. In addition, temporal stability of

responding was assessed by administering the Khavari Alcohol Test to a

sample of 17 college students and again administering it to the same

sample two weeks later. Khavari and Farber (1978) reported the

resulting test-retest reliability coefficient as averaging .92 for the

12 variables/questions of the instrument.

The Sensation Seeking Scale

The Sensation Seeking Scale was developed by Zuckerman et al.

(1964) to be a measure of the concept of optimal level of stimulation as

a motivating force (Leuba, 1955). According to Leuba (1955), organisms

are not only motivated by drive or tension reduction, but also by the

need to seek stimulation in order to regulate excitation around an

optimal level. This idea was expanded on the basis of evidence

suggesting an individual difference in threshold of stimulation required

to produce a response (Gray, 1964). Zuckerman et al. (1964) developed

the Sensation Seeking Scale to differentiate individuals low in an

optimal level of stimulation or sensation-seeking need from those with a

higher optimal level or need. This Sensation Seeking Scale (Zuckerman

et al., 1964) was found to have moderate correlations with other

measures of a tendency to increase stimulation, stimulus variability, or

change (Farley, 1971; Looft & Baranowski, 1971). Zuckerman (1979)

stated that "sensation seeking is a trait defined by the need for

varied, novel, and complex sensations and experiences and the

willingness to take physical and social risks for the sake of such

experience" (p. 10).











Factor analysis of the Sensation Seeking Scale has resulted in four

factors or subscales in addition to the total score: Thrill and

Adventure Seeking, Experience Seeking, Disinhibition, and Boredom

Susceptibility (Zuckerman, 1971). The Sensation Seeking Scale (Form V)

was later reconstructed on the basis of cross-cultural data from an

English sample, resulting in excellent reliability for all factors

except Boredom Susceptibility (Zuckerman et al., 1978).

An explanation of these factors was given by Zuckerman et al.

(1978). The first factor, Thrill and Adventure Seeking, describes

activities that involve speed or danger. The second factor, Experience

Seeking, consists of items that describe novel experiences and a

nonconforming lifestyle. The third factor, Disinhibition, which is most

similar to impulsivity (Zuckerman, 1983), primarily describes a desire

to drink alcohol and a preference for social and sexual disinhibition.

The fourth factor, Boredom Susceptibility, represents the dislike of

routine and the unexciting.

Zuckerman et al. (1978) constructed the 40-item Form V of the

Sensation Seeking Scale with 10 items for each factor, using the

criterion that items had a primary loading on the same factor for males

and females in both English and U.S. samples. Within these groups, the

internal reliabilities reported by Zuckerman et al. (1978) ranged from

.61 to .82 for factors of Thrill and Adventure Seeking, Experience

Seeking, and Disinhibition, and from .56 to .65 for the Boredom

Susceptibility factor. Internal reliabilities of the total scores of

the Sensation Seeking Scale for both U.S. and English samples ranged

from .83 to .86. Zuckerman (cited in Zuckerman, 1979) determined a








48

test-retest reliability of .94 for Form V which was administered at a 3-

week interval to a sample of 65 males and females.

The 40 questions found in the Sensation Seeking Scale are presented

as a choice between alternatives A and B (see Appendix C). The

instructions of the instrument ask respondents to choose the alternative

which best describes their likes or feelings rather than consider how

others may feel. In the event that subjects do not like either

alternative, they are instructed to choose the one they dislike least.

Subjects are encouraged to answer honestly and openly.

For each of the four factors, a score of 0 to 10 is possible, with

a value of 1 being assigned for each choice reflecting the sensation-

seeking alternative and a value of 0 for each nonsensation-seeking

alternative. The number of sensation-seeking choices out of the 10

items per factor scale are summed to give the scores for each factor.

Similarly, the total score, which may range from 0 to 40, is the

summation of the factor scores or total number of sensation-seeking

choices. As described above, scores for the total scale and the

subscales of Disinhibition and Experience Seeking were computed and

correlated with drinking level to select the one with which it was most

highly correlated.

The Alcohol Expectancy Questionnaire

The Alcohol Expectancy Questionnaire (Brown et al., 1980) was

developed to assess the nature of beliefs about the positive effects

that result from drinking. Brown et al. (1980) developed the

questionnaire using a three-step procedure: (a) interviewing a general

population sample of adults (N = 125) to obtain a list of positive









49

expectancies, (b) testing and refining the resulting questionnaire with

a similar sample (N = 400), and (c) testing and factor analyzing the

final questionnaire with data from a college sample (N = 440). This

Adult form of the questionnaire was later used in the development of an

Adolescent form (Christiansen et al., 1982).

The Adult form of the Alcohol Expectancy Questionnaire which was

used in this study consists of 90 items yielding six expectancy factors

or scales. Items are presented in an agree-disagree format (see

Appendix D for examples of items for each factor). The instructions

indicate that subjects should mark "agree" if the stated alcohol effect

is true, mostly true, or true for them some of the time. Similarly,

they are instructed to mark "disagree" if the effect is false or mostly

false for them. Subjects are directed to consider any type of alcoholic

beverage when the term alcohol is used within items. The instructions

encourage respondents to answer according to their current and personal

beliefs in contrast to what others may think about alcohol. Subjects

are asked to respond to the items even if they have no experience with

drinking alcohol. The instructions also ask subjects to answer honestly

and remind them that their answers are anonymous.

Agreement with an item, denoting belief in that particular positive

alcohol effect, is scored as 2, whereas disagreement is scored as 1. If

the subject has failed to respond to that item, it is scored as 0. The

number of items and range of scores (if items are answered) for each of

the six expectancy scales is as follows: (a) Global Positive Changes, 28

items, scores of 28-56; (b) Sexual Enhancement, 7 items, scores of 7-14;

(c) Social and Physical Pleasure, 9 items, scores of 9-18; (d) Social









50

Assertion, 11 items, scores of 11-22; (e) Tension Reduction, 9 items,

scores of 9-18; and (f) Arousal and Aggression, 5 items, scores of 5-10.

As described above, the Global Positive Changes factor was not utilized

as an expectancy variable in the current study due to its high

intercorrelations with the other five factors.

Reliability data for the Alcohol Expectancy Questionnaire were

reported by Brown, Christiansen, and Goldman (1987) for internal

consistency and test-retest reliability. In a sample of 176 normal

drinkers, coefficient alphas for the factor scales ranged from .72 to

.92 with a mean of .84. The mean test-retest reliability at an 8-week

interval was .64 in a sample of 465 college students.

The Alcohol Expectancy Questionnaire has been the most widely used

instrument among the currently available measures of alcohol

expectancies (Connors & Maisto, 1988; Leigh, 1989c; Young & Knight,

1989). Research with this instrument has consistently supported a

relationship between drinking and alcohol expectancies (e.g. Brown,

1985a; Brown et al., 1980) as has research using selected items from the

instrument (Rohsenow, 1983). Research using the Alcohol Expectancy

Questionnaire has been shown to distinguish alcoholics on the basis of

their greater endorsement of positive alcohol expectancies (Brown et

al., 1985; Connors, O'Farrell, Cutter, & Thompson, 1986; Zarantonello,

1986). Brown (1985a) also demonstrated that expectancies measured by

the instrument improved prediction of college drinking over the

inclusion of demographic or background variables.

Researchers have also shown a relationship between specific

expectancies measured by the questionnaire and relevant personality









51

characteristics (Leonard & Blane, 1988). In addition, drinking behavior

has been predicted by consideration of an alcohol expectancy and the

relevant personality variable, for example an expectancy of Social

Assertion for low assertive students (Mooney & Corcoran, 1989). Connors

and Maisto (1988) concluded that use of the questionnaire "can help to

clarify reasons for drinking" (p. 492).

Using alcohol-effect items similar to those of the Alcohol

Expectancy Questionnaire, Stacy et al. (1990) demonstrated the validity

of distinguishing positive alcohol expectancies or expectancies of

positive effects from drinking, from negative alcohol expectancies or

expectancies of negative effects from drinking. In their study of a

college sample, they found that only positive alcohol expectancies, such

as those measured by the Alcohol Expectancy Questionnaire, predicted

later alcohol use. In addition, Stacy et al. (1990) demonstrated the

validity of the expectancy construct as independent from attitudes about

drinking in predicting the intention to drink and subsequent alcohol

use.

Description of Population and Sampling

The population for this study was the undergraduate student body at

the University of Florida during the second summer term, 1991.

According to data obtained by the Drug and Alcohol Resource Center of

the University of Florida, approximately 85-90% of students on this

campus drink alcohol. This percentage of students is similar to that

found at other large, public universities (e.g. O'Hare, 1990). In order

to obtain a more homogeneous sample (see Mooney & Corcoran, 1989), only

students who currently drank alcohol were included in the study.









52

Based on demographic data from the Office of Academic Affairs at

the University of Florida, males and females comprised respectively 52%

and 48% of the undergraduate population of the spring term, 1991.

Ethnic composition of the undergraduate population of the spring term,

1991, was reported as approximately 80% white, 6% black, 5% Hispanic, 4%

Asian, and 5% other.

For the current study, data was to be collected on a minimum sample

of 400 students. Approximately 86% of the sample was obtained from three

psychology courses, and the remaining 14% of the sample was obtained

from courses in two other departments. The participating courses in

psychology and the approximate percentages of the sample comprised by

each were Personal Growth (34%), The Nature of Human Conflict (30%), and

General Psychology (22%). With respect to the General Psychology

course, students were recruited from the subject pool formed as a result

of course requirements for research participation. In addition, an

estimated 10% of the sample was comprised by students from the Social

Problems course in the Department of Sociology. The remaining

approximately 4% of the sample was obtained from Health Science

Education in Elementary School, a course in the Department of Health

Science Education.

Description of Research Procedures

Research procedures conformed to the Instructions to Researchers

Using the PSY 2013 Subject Pool. In accordance with these procedures,

this researcher was assigned a given number of subjects based on the

judgement of availability made by the Subject Pool Coordinator. The









53

maximum available number of participants in the subject pool for second

summer term, 300 students, was requested by the researcher.

In accordance with the required procedures, the researcher

completed and posted Experimental Information Sheets in the lobby of the

Psychology Building. Information that was provided to potential

participants included the researcher's name and phone number, the

purpose and procedures of the study, and the participation requirement

that subjects currently drank alcoholic beverages of some kind. A

choice of appointment times for research sessions was made available,

and participants were requested to select one. The information on the

location of research sessions was also provided.

At the beginning of each research session, the researcher

introduced the project as outlined in the protocol of the informed

consent process (see Appendix E). This protocol was prepared according

to the guidelines of the University of Florida Institutional Review

Board. Questionnaires were administered (in counterbalanced order)

according to the instructions described above. Following data

collection, participants were provided a more complete description of

the purpose and background of the study (see Appendix F). As required

of researchers using the psychology subject pool, this debriefing

process was designed to be an educational experience for research

participants.

Data collection occurred in a single session of approximately 30

minutes for each group of subjects. Groups of subjects from the PSY

2013 Subject Pool were limited to 50 students per data-gathering session

so as to attend to each participant's questions and learning experience.









54

Description of Data Analyses

Means and standard deviations were computed for the sample and

separately for males and females on drinking level, sensation-seeking

needs, and alcohol expectancies. These statistics were provided for

descriptive purposes.

In order to select the sensation-seeking variable of the study,

Pearson correlations were computed between drinking level and total

sensation seeking, Disinhibition, and Experience Seeking. The variable

among these three that was most correlated with drinking level in the

current sample was operationally defined as the variable of sensation-

seeking needs in this study.

Pearson correlations were also computed to investigate the

individual relationships among the variables of the study: sensation-

seeking needs, alcohol expectancies, and drinking level. If the

intercorrelations between any alcohol expectancies were .80 or higher,

these expectancy factors would be combined into a single variable. If

all intercorrelations among the five factors were less than .80, these

five would be retained as separate variables of alcohol expectancies in

this study: Sexual Enhancement, Social and Physical Pleasure, Social

Assertion, Tension Reduction, and Arousal and Aggression.

Multiple regression procedures were used to analyze all hypotheses

of the study. Two regression analyses were each conducted at a .05

alpha level.

Hypotheses for which sensation-seeking needs was the criterion or

dependent variable (hypotheses 1 through Ic) were tested by the first

regression model. The predictor or independent variables in this model









55

were gender, alcohol expectancies, and their interactions. This

regression equation tested the relationships between alcohol

expectancies and sensation-seeking needs and determined if there were

gender differences in these relationships. In addition, this model

tested a gender difference in level of sensation-seeking needs.

Hypotheses for which drinking level was the criterion or dependent

variable (hypotheses 2 through 2g) were tested in the second model. The

predictor or independent variables in this model were gender, sensation-

seeking needs, alcohol expectancies, and their interactions. This

regression model tested the relationships between drinking level and

gender, sensation-seeking needs, alcohol expectancies, and alcohol

expectancies as a result of sensation-seeking needs. In addition, the

model tested gender differences in these relationships. Thus, this

model tested all two-way interactions and the three-way interaction

among alcohol expectancies, sensation seeking, and gender. Although

this model provided information with respect to a gender difference in

drinking, this question was of secondary importance in the current

study.

Description of Methodological Limitations

Several methodological issues and limitations applied to this

research study. The results of this study, which was designed to assess

motivation for drinking in a college student sample, are less applicable

to other groups of drinkers (e.g. Crowley, 1991). Two methodological

limitations applied to the current study. The first limitation derived

from the imperfect factor structure of the Alcohol Expectancy

Questionnaire. Confirmatory factor analysis of the AEQ has resulted









56

particularly in concerns of low factor independence and discriminative

validity (Leigh, 1989b). There are substantial intercorrelations among

the subscales or expectancy factors. In particular, the Global Positive

Changes factor is highly correlated with other factors; therefore the

Global Positive Changes factor was not included as an expectancy

variable in this study. After finding similar methodological weaknesses

with two other measures of alcohol expectancies, Leigh (1989b) concluded

that alcohol expectancies are in reality overlapping. Therefore,

conclusions about differences in specific expectancies "should be made

with the knowledge that these concepts are intertwined" (Leigh, 1989b,

p. 276).

The second methodological limitation of this study resulted from

sampling procedures. The sample provided by the Psychology Subject Pool

and other undergraduate courses was not a random sample. Also, this

sample was not necessarily representative of the university population.

In order to increase the representativeness of the sample, students from

a variety of courses were requested to participate.















CHAPTER 4
RESULTS OF THE STUDY



Introduction

The purpose of the study was to examine motivation for drinking

among male and female undergraduate students in terms of the

relationships among their sensation-seeking needs, alcohol expectancies,

gender, and drinking level. In order to study these relationships, two

major questions were posited. The first question concerned the joint

influence of alcohol expectancies and gender on sensation-seeking needs

within the student sample; therefore, the researcher sought to determine

which alcohol expectancies were related to sensation-seeking needs and

whether gender differences contributed to these relationships. In

addition, it was of interest to determine whether a gender difference

existed in sensation-seeking needs. The second major question concerned

the influence of sensation-seeking needs, alcohol expectancies, and

gender together on drinking level of students. This question included

the investigation of relationships between drinking level and both

sensation-seeking needs and alcohol expectancies, as well as gender

differences in these relationships. In addition, the researcher sought

to determine which alcohol expectancies contributed to drinking for

varying levels of sensation-seeking needs and whether gender differences

existed in these relationships.









58

Description of the Sample

The sample consisted of undergraduates from the Departments of

Psychology, Sociology, and Health Science Education at the University of

Florida. After eliminating data which could not be scored due to

subjects' noncompliance with questionnaire instructions, a sample of 464

students was obtained. All questionnaires were retained that could be

scored despite some missing data, resulting in slightly different sample

sizes for various data analyses.

In the sample, the ages of students ranged from 17 to 50, with a

mean of 19.7 and a median of 19. Approximately two thirds of the sample

were females (n = 301) and one third were males (n = 163). Ethnic

groups were represented in the sample as 87% Caucasian American, 5%

Hispanic American, 3% African American, 2.8% Asian American, and 2.2%

other than the above.

Analysis of the Data

To determine whether the number of sensation-seeking variables

might be reduced, thus simplifying the analyses, the Pearson

correlations among total sensation-seeking needs, Experience Seeking,

Disinhibition, and drinking level were examined. This correlation

matrix is presented in Table 1. Disinhibition was selected as the

sensation-seeking variable of the study because it was most correlated

with drinking level.

Descriptive statistics for the study variables appear in Table 2.

The means and standard deviations of each variable are presented for the

total sample and for each gender.











Table 1
Pearson Correlations Among the Sensation-Seeking Variables and Drinking
Level


Variable 2 3 4


1. Disinhibition .28 .73 .48


2. Experience Seeking -- .66 .18


3. Total sensation seeking .39


4. Drinking level


Note: All correlations significant at p<.0001.
aDrinking level = annual absolute alcohol intake (ounces)/body weight
(pounds).



A preliminary test of individual relationships among sensation-

seeking needs, alcohol expectancies, and drinking level was performed to

determine the sizes of these relationships. The Pearson correlations

among the variables of the study are shown in Table 3. Because the

highest intercorrelation among the various alcohol expectancy factors

was .61, they were considered to have sufficient independent variance to

be included as separate variables in the multiple regression analyses.

All correlations among sensation-seeking needs, alcohol expectancies,

and drinking level were significant with the exception of that between

drinking level and the expectancy of Arousal and Aggression (VI).

Two multiple linear regressions were performed to test the two main

null hypotheses and their corresponding subhypotheses. The alpha level

was set at .05. Following are the data results as they relate to all

null hypotheses of the study.











Table 2
Descriptive Statistics for Sensation-Seeking Needs. Alcohol
Expectancies. and Alcohol Consumption for the Total Sample and by Gender


Variable Total sample Females Males

MSD SU M SU

SSNa 5.28 2.61 4.77 2.56 6.24 2.43


Alcohol expectancies

IIb 8.95 2.07 8.97 2.11 8.91 2.00


IIIc 15.48 2.04 15.30 2.10 15.83 1.89

IVd 17.91 3.67 17.80 3.76 18.13 3.50


Ve 14.12 2.56 13.94 2.58 14.47 2.49


Vl 7.38 1.20 7.39 1.16 7.36 1.28


AAAIg 321.99 382.21 226.53 267.39 498.25 486.75


Drinking level 2.28 2.68 1.83 2.26 3.09 3.15

aSSN = sensation-seeking needs. bl1 = Sexual Enhancement. CIII = Social
and Physical Pleasure. dIV = Social Assertion. eV = Tension Reduction.
fVI = Arousal and Aggression. gAAAI = annual absolute alcohol intake
(ounces). hDrinking level = AAAI (ounces)/body weight (pounds).


Ho 1: There is no composite influence of alcohol expectancies,

gender, and their interactions on sensation-seeking needs of

college students. This null hypothesis was rejected. In Table 4 are

presented the results of the regression of alcohol expectancies, gender,

and their interactions on the dependent variable of sensation-seeking









61

needs. The omnibus or overall test of the model was significant (E1,411

= 19.02, 9<.0001).

Ho la: There is no gender difference in sensation-seeking needs of

college students. The researcher failed to reject this null hypothesis

at the alpha level set. As shown in Table 4, there was no unique

contribution of gender to the explanation of sensation-seeking needs.

Ho Ib: There are no relationships between alcohol expectancies and

sensation-seeking needs of college students. This null hypothesis was


Table 3
Pearson Correlat ions Among


SPnntjnn- S~gCioki 1N3ooC. A1,nhnl


Expectancies. and Drinking Level


Variable 2 3 4 5 6 7


1. SSNa .33** .46** .45** .37** .17* .48**


2. IIb .45** .50** .46** .39** .23**


3. IIIc -- .57** .61** .32** .32**


4. IVd -- .57** .36** .27**


5. Ve -- .41** .29**


6. VI' .08


7. Drinking level8

aSSN = sensation-seeking needs. bII = alcohol expectancy of Sexual
Enhancement. CIII = alcohol expectancy of Social and Physical Pleasure.
dIV = alcohol expectancy of Social Assertion. eV = alcohol expectancy of
Tension Reduction. 'VI = alcohol expectancy of Arousal and Aggression.
9Drinking level = annual absolute alcohol intake (ounces)/body weight
(pounds).
*P<.001. **,<.0001.











Table 4
Multiple Linear Regression of Alcohol Expectancies and Gender on
Sensation-Seeking Needs


Source df SS E


Model 11 977.85 19.02*


Alcohol expectancies

IIb 1 15.29 3.27


IIIc 1 93.25 19.95*


IVd 1 104.66 22.39*


Ve 1 0.57 0.12


VIf 1 9.24 1.98


Gender 1 0.13 0.03


II*Gender 1 0.05 0.01


III*Gender 1 2.84 0.61


IV*Gender 1 3.27 0.70


V*Gender 1 1.36 0.29


VI*Gender 1 3.54 0.76


Error 411 1921.42


Total 422 2899.28

aV2 = .34. bII = Sexual Enhancement. CIII = Social and Physical
Pleasure. dIV = Social Assertion. eV = Tension Reduction. 'VI = Arousal
and Aggression.
*D<.0001.









63

rejected (see Table 4). The alcohol expectancy of Social and Physical

Pleasure (III) was significantly related to sensation-seeking needs

(Ei,411 = 19.95, p<.0001). In addition, the alcohol expectancy of Social

Assertion (IV) shared a unique portion of its variance with sensation-

seeking needs (E1,411 = 22.39, y<.0001). In Table 5 are presented the

estimates of the slopes and intercept of the lines denoting the

relationships of alcohol expectancies III and IV with sensation-seeking

needs. The values of both slopes are positive, indicating that

sensation-seeking needs in the student sample increased with increases

in scores of alcohol expectancies III and IV.



Table 5
Estimates of Slopes and Intercept for the Regression of Alcohol
Expectancies and Gender on Sensation-Seeking Needs


Parameter Estimate t


Intercept -3.63 -2.19*


Alcohol expectancies


IIIa 0.41 3.14**


IVb 0.23 3.35***


Note: E11,411 = 19.02, y<.0001.
all = Social and Physical Pleasure. bIV = Social Assertion.
*R<.05. **"<.01. ***R<.001.



Ho Ic: There are no gender differences in the relationships between

alcohol expectancies and sensation-seeking needs of college students.









64

The researcher failed to reject this null hypothesis at the

predetermined alpha level. As shown in Table 4, no significant

relationships were found between sensation-seeking needs and the

interactions of gender and alcohol expectancies within the student

sample.

Ho 2: There is no composite influence of sensation-seeking needs,

alcohol expectancies, gender, and their interactions on drinking among

college students. This null hypothesis was rejected at the .05 level of

significance. Appearing in Table 6 are the results of the regression of

sensation-seeking needs, alcohol expectancies, gender, and their

interactions on the dependent variable of drinking level. The omnibus

or overall test of the model was significant (F23,396 = 16.56, p<.0001).

The data for drinking level were logarithmically transformed because of

the violation of the assumption of homoscedasticity (equal variances of

drinking scores for all scores of each criterion variable).

Ho 2a: There is no relationship between sensation-seeking needs and

drinking among college students. This null hypothesis was rejected. As

indicated in Table 6, sensation-seeking needs made a significantly

unique contribution to the explanation of drinking level (E 1396 = 16.98,

P<.0001). The estimates of the slope and intercept of the line denoting

the relationship between sensation-seeking needs and drinking level are

given in Table 7. The positive value of the slope indicates that

drinking level increased with increases in scores of sensation-seeking

needs.

Ho 2b: There is no gender difference in the relationship between

sensation-seeking needs and drinking among college students. This null









65

hypothesis was rejected (see Table 6) at the .05 level of significance.

The interaction of sensation-seeking needs and gender was significantly

related to drinking level (E1,396 = 5.54, 9<.05) within the student

sample. Presented in Table 7 is the difference between the slopes of

the lines which represented the separate relationships for males and

females between sensation-seeking needs and drinking. The negative



Table 6
Multiple Linear Regression of Sensation-Seeking Needs. Alcohol
Expectancies, and Gender on Drinking Level


Source di S


Model


489.55


21.82


SSNb


16.56**-


16.98**-


Alcohol expectancies


0.01


24.36


1.54


0.37


1.99


8.75


7.13


0.09


Gender


SSN*Gender


II*Gender


III*Gender


1 10.22


0.01


18.95***


1.20


0.29


1.55


6.80**


5.54*


0.07


7.95**












Table 6--continued


Source f SS


IV*Gender 1 1.65 1.28


V*Gender 1 0.06 0.05


VI*Gender 1 0.10 0.08


SSN*II 1 0.00 0.00


SSN*III 1 6.81 5.30*


SSN*IV 1 1.48 1.15


SSN*V 1 0.03 0.03


SSN*VI 1 1.68 1.31


SSN*II*Gender 1 0.18 0.14


SSN*III*Gender 1 10.56 8.21**


SSN*IV*Gender 1 0.77 0.60


SSN*V*Gender 1 0.05 0.04


SSN*VI*Gender 1 1.64 1.27


Error 396 508.99


Total 419 998.55


Note: Data for drinking level were logarithmically transformed due to
the violation of homoscedasticity.
aR2 = .49. bSSN = sensation-seeking needs. CII = Sexual Enhancement.
dII = Social and Physical Pleasure. elV = Social Assertion. V =
Tension Reduction. 9VI = Arousal and Aggression.
*y<.05. **p<.01. ***-<.0001.











Table 7
Estimates of Slopes and Intercept for the Regression of Sensation-
Seeking Needs. Alcohol Expectancies. and Gender on Drinking Level


Parameter Estimate 1


Intercept -11.34 -5.60***


SSNa 1.35 3.99***

IIIb 0.74 3.98***


SSN*Genderc -0.98 -2.35*


III*Gender -0.58 -2.82**


SSN*III -0.09 -3.10**


SSN*III*Gender 0.10 2.87**


Note: E23,396 = 16.56, p<.0001. Data for drinking level were
logarithmically transformed due to the violation of homoscedasticity.
aSSN = sensation-seeking needs. bIII = alcohol expectancy of Social and
Physical Pleasure. cEstimates of slopes are given for females relative
to males wherever gender appears.
*H<.05. **y<.01. ***y<.0001.


value of this difference for females relative to males indicates a

steeper slope of the linear relationship for males. Thus, drinking

level increased at a more rapid rate for males for every unit increase

in sensation-seeking needs.

Ho 2c: There are no relationships between alcohol expectancies and

drinking among college students. This null hypothesis was rejected (see

Table 6) at the predetermined level of significance. The alcohol

expectancy of Social and Physical Pleasure (III) made a unique

contribution to the explanation of drinking level (Ez,396 = 18.95,








68

q<.0001). The slope and intercept of the line which represented this

relationship are found in Table 7. The positive value of the slope

indicates that drinking level increased with increases in scores of

alcohol expectancy III.

Ho 2d: There are no gender differences in the relationships between

alcohol expectancies and drinking among college students. This null

hypothesis was rejected. As shown in Table 6, a significant

relationship was found between drinking level and the interaction of the

alcohol expectancy of Social and Physical Pleasure (III) and gender

(E1,396 = 7.95, 9<.01). The negative difference, females relative to

males, between the slopes of the lines (see Table 7) indicates that for

males there was a steeper slope of the linear relationship between

drinking and the alcohol expectancy of Social and Physical Pleasure.

Therefore, drinking level increased at a more rapid rate for males for

every unit increase in scores of alcohol expectancy III.

Ho 2e: There are no differences in the relationships between

alcohol expectancies and drinking for differing levels of sensation-

seeking needs of college students. This null hypothesis was rejected

(see Table 6) as a significant relationship was found between drinking

level and the interaction of the alcohol expectancy of Social and

Physical Pleasure (III) and sensation-seeking needs (1,396 = 5.30,

y<.05). As indicated in Table 7, the slope of the plane describing the

relationships among sensation-seeking needs, alcohol expectancy III, and

drinking level was a relatively small negative value. This value

indicates that, in relation to drinking, an increase in sensation-









69

seeking needs was accompanied by a slightly greater increase in the

score of alcohol expectancy III.

Ho 2f: There are no gender differences in the relationships between

alcohol expectancies and drinking for differing levels of sensation-

seeking needs of college students. This null hypothesis was rejected at

the .05 level of significance. As revealed in Table 6, there was a

significant gender difference in the relationship between drinking level

and the interaction of the alcohol expectancy of Social and Physical

Pleasure (III) and sensation-seeking needs (E1,396 = 8.21, p<.01). That

is, the gender difference indicates that, in relation to drinking level,

an increase in sensation-seeking needs was accompanied by a slightly

greater increase in score of alcohol expectancy III for males than for

females in this study. The gender difference of the slopes of the

planes denoting the relationships of drinking level to the interaction

of sensation-seeking needs and alcohol expectancy III (.10 for females

relative to males) is given in Table 7. The positive difference

indicates that the slope of the plane for females was positive relative

to that of the males. To obtain actual estimates of the slopes of the

planes, separate regression analyses were performed for each gender.

The slopes of these planes were respectively .01 for females and -.09

for males, producing the .10 arithmetic difference for females relative

to males.

Ho 2g: There is no gender difference in drinking among college

students. This null hypothesis was rejected (see Table 6). That is,

gender made a significantly unique contribution to the explanation of









70

drinking level (E1,396 = 6.80, p<.01). As indicated in Table 2, the mean

drinking level was lower for females than for males.

Summary of the Findings

With regard to the first major question of the study, alcohol

expectancies and gender together were significantly related to

sensation-seeking needs, explaining 34% of sensation-seeking scores.

Alcohol expectancies of Social and Physical Pleasure (III) and Social

Assertion (IV) each made a unique contribution to the explanation of

sensation-seeking needs. No gender differences existed in these

relationships or in sensation-seeking needs.

With regard to the second major question of the study, sensation-

seeking needs, alcohol expectancies, and gender together were

significantly related to drinking level, explaining 49% of drinking

scores. When the complete model was specified, there was a gender

difference such that, in relation to drinking, males' scores of the

alcohol expectancy of Social and Physical Pleasure (III) increased at a

slightly more rapid rate than scores of sensation-seeking needs. When

gender was not accounted for, the same relationship appeared. Gender

differences also existed in the relationships between drinking level and

both sensation-seeking needs and the alcohol expectancy of Social and

Physical Pleasure (III). Males' drinking levels increased at a more

rapid rate than females' drinking levels with increases in scores of

sensation-seeking needs and the alcohol expectancy of Social and

Physical Pleasure. Males also had higher drinking levels than females.

Without taking gender into account, sensation-seeking needs and alcohol

expectancy III were both positively related to drinking level.















CHAPTER 5
DISCUSSION



Summary of the Study

Alcohol use and motivation for drinking among college students are

subjects of research due in part to a high incidence of heavy

consumption and alcohol-related problems (Gonzalez, 1986). A few

researchers have recently studied motivation for drinking by considering

how alcohol expectancies and personality factors were related (Brown &

Munson, 1987; Leonard & Blane, 1988) and how these variables together

contributed to student drinking (Mooney & Corcoran, 1989). The goal of

this study was to use both of these approaches to improve understanding

of motivation for student drinking, thereby assisting in campus

prevention efforts.

In particular, the purpose of the study was to examine the

relationships among sensation-seeking needs, alcohol expectancies,

gender, and student drinking. Among personality variables, sensation-

seeking needs have most often been found to relate to student drinking

(Brennan et al., 1986a). This relationship has been considered to

reflect a disinhibition motive for student drinking, interpreted as

drinking to facilitate the expression of normally restrained behavior or

to escape social control (Berkowitz & Perkins, 1986; Schwarz et al.,

1978), particularly with regard to sexual and aggressive behaviors

(Brown & Munson, 1987; Ratliff & Burkhart, 1984). Although alcohol-

71









72

induced disinhibition figures prominently in both cultural expectancies

and research constructs (Room & Collins, 1983), it is unclear exactly

what behaviors may become disinhibited (Fingarette, 1983), how these

effects hold particular reinforcement value for individuals with high

sensation-seeking needs, and whether these alcohol effects primarily

represent enhancement of positive affect or amelioration of negative

affect.

As a preliminary investigation of these issues relevant to a

disinhibition motive for student drinking, two major questions were

posited. The first concerned the joint influence of alcohol

expectancies and student gender on sensation-seeking needs. By means of

this question, the researcher examined the relationships between alcohol

expectancies and sensation-seeking needs, as well as gender differences

in these relationships and in sensation-seeking needs. The second major

question concerned the joint influence of sensation-seeking needs,

alcohol expectancies, and gender on drinking level of students.

Included in this question was the investigation of individual

relationships between drinking level and both sensation-seeking needs

and alcohol expectancies, as well as gender differences in these

relationships. Also examined was the joint contribution of sensation-

seeking needs and alcohol expectancies to drinking level and whether

gender differences existed in these relationships.

Survey instruments, consisting of a short demographic

questionnaire, the Khavari Alcohol Test (Khavari & Farber, 1978), the

Sensation Seeking Scale Form V (Zuckerman, 1979), and the Alcohol

Expectancy Questionnaire (Brown et al., 1980), were completed by a









73

sample of 464 undergraduates (301 females and 163 males) at the

University of Florida. The median age in the sample was 19.

The variables of the study consisted of drinking level, sensation-

seeking needs, five specific alcohol expectancies, and gender. Drinking

level, based on usual and maximum quantities of alcohol and their

corresponding frequencies of consumption, was standardized by body

weight. From the Sensation Seeking Scale, the Disinhibition subscale

was selected as the variable of sensation-seeking needs because it had

the strongest relationship with drinking level in this sample. The five

alcohol expectancies were represented by the following factors of the

Alcohol Expectancy Questionnaire: Sexual Enhancement, Social and

Physical Pleasure, Social Assertion, Tension Reduction, and Arousal and

Aggression.

With an alpha level set at .05, two multiple linear regressions

were performed to test these research questions. Regarding the first

question, alcohol expectancies and gender together were significantly

related to sensation-seeking needs, explaining 34% of the scores.

Alcohol expectancies of Social and Physical Pleasure (III) and Social

Assertion (IV) were each related to sensation seeking, and no gender

differences existed in these relationships or in sensation-seeking

needs.

With regard to the second research question, sensation-seeking

needs, alcohol expectancies, and gender together were significantly

related to drinking level, explaining 49% of the scores. A three-way

interaction was related to drinking such that males' scores of alcohol

expectancy III increased at a slightly more rapid rate than scores of









74

sensation-seeking needs. This relationship appeared as well when gender

was not taken into account. Two two-way interactions were related to

drinking such that males' drinking levels increased at a more rapid rate

with increases in, individually, sensation-seeking needs and alcohol

expectancy III. Again when gender was not considered, sensation-seeking

needs and alcohol expectancy III were each positively related to

drinking. Of less importance to the current study was the finding that

males had higher drinking levels than females.

Discussion of the Results

The first major research question of this study constituted an

indirect inquiry into motivation for drinking by relating sensation-

seeking needs to expected positive effects of drinking (with the

assumption that factors comprising motivation for drinking will

contribute to actual drinking behavior), whereas the second question was

a direct inquiry into how these two motivational factors together

contributed to drinking level itself. These approaches have been used

separately by previous researchers (e.g. Leonard & Blane, 1988; Mooney &

Corcoran, 1989). In the current study, both questions were posited in

order to compare the results and provide a more comprehensive view of a

disinhibition motive for drinking. With regard to the first question,

both the alcohol expectancies of Social and Physical Pleasure (III) and

Social Assertion (IV) were related to sensation-seeking needs. However,

the alcohol expectancy that best predicted drinking level was expectancy

III.

Examination of the results to the second research question reveals

that, although sensation-seeking needs, alcohol expectancy III, and









75

gender were individually related to drinking, each of these

relationships was influenced by the other variables. Higher scores in

sensation-seeking needs, together with higher scores in alcohol

expectancy III, contributed to heavier drinking for both male and female

students; however, scores in expectancy III increased at a slightly more

rapid rate than scores of sensation-seeking needs in relation to

drinking levels in the sample as a whole and in males. Also, males'

drinking levels increased at a more rapid rate with increases in,

individually, sensation-seeking needs and alcohol expectancy III.

Without consideration of gender, sensation-seeking needs and alcohol

expectancies were individually related to drinking level. Therefore,

although drinking level was explained to some degree by individual

consideration of sensation-seeking needs, alcohol expectancies, and

gender, a more accurate prediction of drinking resulted from their joint

consideration.

The dissimilarity between the two research questions exists not

only in that the second question was concerned with motivational

influences on actual drinking behavior but also in that these

motivational influences included the interactions among sensation-

seeking needs, alcohol expectancies, and gender. The second question

represented a more complete specification of motivational factors, both

personality and expectancies, as they jointly or interactively related

to drinking level. Indeed, the joint consideration of sensation-seeking

needs, alcohol expectancies, and gender explained almost half (R2 = .49,

Table 6) of drinking scores.









76

Thus previous research results of the relationship between

drinking behavior and sensation-seeking needs (e.g. Brennan et al.,

1986a) were supported here but were also amended by results of their

interactions with alcohol expectancies and gender in explanations of

drinking level. For example, by including gender as a variable, it was

seen that males' drinking levels increased at a more rapid rate with

increases in sensation-seeking needs.

In contrast to previous research (e.g. Ratliff & Burkhart, 1984),

males did not have higher sensation-seeking needs than females, as

assessed by the Disinhibition factor of the Sensation Seeking Scale.

The Disinhibition factor of sensation seeking describes drinking,

sexual, and social behaviors which are disinhibited or expressed beyond

the usual limits. In this study therefore, male and female students

tended to accept a similar degree of limit setting or social control

over these behaviors.

The finding that males had a higher average drinking level than

females is in agreement with most earlier research (e.g. Brennan et al.,

1986a). However, in contrast to the majority of this research,

consumption in this study was standardized by body weight to correct for

a higher average weight in males. In another study in which this same

method was used, no difference was found between consumption levels for

male and female heavy drinkers, but among light drinkers, males consumed

more than females (Ratliff & Burkhart, 1984). In the current study,

although the average drinking level of males was higher, one female had

the highest drinking score (16.7 ounces of absolute alcohol annually per









77

pound of body weight). A number of students, both male and female, had

very high drinking levels.

The finding here of a relationship between student drinking and

the alcohol expectancy of Social and Physical Pleasure (III) is

consistent with previous research results of positive relationships

between alcohol expectancies and drinking (Leigh, 1989c). By including

gender as a variable, it was also shown here that males' drinking levels

increased at a more rapid rate with increases in alcohol expectancy III.

Similar to the current study, Mooney et al. (1987) found that the amount

of drinking for both male and female students, as well as the frequency

of drinking for males, were best predicted by both the alcohol

expectancies of Social and Physical Pleasure (III) and Social Assertion

(IV). Alcohol expectancy III in particular was also found by Brown

(1985a) to predict one pattern of student drinking which was

characterized by heavy consumption and some negative physical effects

and which was considered the stereotypic pattern of collegiate alcohol

use.

Although in the current study only alcohol expectancy III

contributed to drinking, both expectancies III and IV were related to

sensation-seeking needs. No previous research has been conducted to

determine the relationships between sensation-seeking needs and alcohol

expectancies. Brown and Munson (1987) did, however, examine the

relationships between alcohol expectancies and extraversion, which is

correlated with sensation-seeking needs (Zuckerman et al., 1972). They

found that high extraversion was related to alcohol expectancies of

Tension Reduction and Social and Physical Pleasure. The results of the











current study converge to some degree with those of Brown and Munson

(1987).

The alcohol expectancy of Social and Physical Pleasure (III)

describes expected effects of appreciating the taste of alcoholic

beverages, feeling good from drinking, feeling better when "high", and

enjoying certain social situations that are enhanced by drinking. Thus

this expectancy factor measures expected alcohol effects of pleasurable

sensations and social interactions, or essentially expectations of

alcohol-induced enhancement of positive affect. The finding in the

present study that alcohol expectancy III contributed to student

drinking is consistent with other research in which students reported a

greater desire to drink when in pleasant settings and moods than when in

unpleasant ones (Russell & Bond, 1979, 1980).

The alcohol expectancy of Social Assertion (IV), which was also

related to sensation-seeking needs but which did not contribute to

student drinking here, describes alcohol effects such as being more

relaxed in social situations, decreasing shyness, and increasing

confidence. These items largely comprise alcohol effects of a reduction

in social anxiety or negative affect.

It appears therefore that sensation-seeking needs were related to

expectancies that alcohol will enhance some positive feelings and

pleasant social situations and will also reduce some negative affect or

social anxiety. In addition, it appears that actual drinking level was

best predicted by the expectancy of enhancement of positive affect.

However, the post hoc regression mentioned in Chapter 4 which was

performed separately for males and females resulted in the possibility









79

that expectancy IV rather than III might be significantly related to

drinking among female students. It is also possible that expectancies

III and IV, whose correlation was .57 (see Table 3), are sufficiently

similar so that variation from one sample to another may account for

which one is found to be related to drinking level. The difference in

results between the first and second research questions could be due to

a potential for III and IV to be somewhat interchangeable.

Because alcohol expectancy III and sensation-seeking needs jointly

contributed to drinking, it appears that students were primarily

motivated to enhance positive affect and positive situations through

drinking. However, it is important to consider that the alcohol

expectancy of Social Assertion (IV) is moderately correlated with the

alcohol expectancy of Social and Physical Pleasure (III), that both were

related to sensation-seeking needs, and that both expectancies III and

IV were most predictive of student drinking in a previous study (Mooney

et al., 1987). In addition, Segal et al. (1980b) found that the typical

pattern of collegiate alcohol use appeared to be characterized by high

sensation-seeking needs and both social and escapist reasons for

drinking. A more accurate interpretation, based on current and previous

research taken together, may be that student drinking is usually

motivated to enhance positive feelings and sociability but may also on

occasion be instrumental in ameliorating social anxiety or other

negative affect.

Because sensation-seeking needs and the alcohol expectancy of

Social and Physical Pleasure jointly contributed to student drinking,

these expected alcohol effects were most relevant to students as a











result of their sensation-seeking needs. Alcohol-related sensations and

social experiences appeared particularly reinforcing to students with

higher sensation-seeking needs or needs to engage in disinhibited,

impulsive behavior. Contrary to a common interpretation of a

disinhibition motive as drinking to facilitate the expression of

especially sexual and aggressive behaviors (Brown & Munson, 1987;

Ratliff & Burkhart, 1984), alcohol expectancies of Sexual Enhancement or

Aggression and Arousal did not contribute to drinking as a function of

students' sensation-seeking needs.

The motivational pattern of drinking described by these results

was very similar for male and female students. The relationships

between alcohol expectancies and sensation-seeking needs did not vary

with gender. More importantly, the alcohol expectancy of Social and

Physical Pleasure (III) contributed to drinking among both male and

female students as a function of their sensation-seeking needs. Despite

some gender differences in individual relationships among the variables

of the study, the overall motivational pattern described by drinking,

alcohol expectancy III, and sensation-seeking needs implied a similar

disinhibition motive for drinking among males and females. The results

of this study did not describe different motives for drinking for male

and female students as suggested by Ratliff and Burkhart (1984), who

concluded that female students may drink primarily to disinhibit or

enhance sociability, whereas male students may drink primarily to

disinhibit sexual or aggressive behaviors. Rather, a disinhibition

motive for drinking among both male and female students in this study









81

expressed sensation-seeking needs through drinking to enhance

sociability and pleasurable sensations.

Conclusions and Implications

In this study, a disinhibition motive for drinking appeared to be

the desire to expand the usual limits on enjoyment of and stimulation

from alcohol-related social interactions and sensations. The

disinhibition motive thus depicted drinking as instrumental in enhancing

positive affect and situations. Alcohol-facilitated pleasurable

sensations and social interactions appeared to be most reinforcing to

students with higher sensation-seeking needs as a result of their

dislike of restrained behavioral expression, or conversely, due to their

desire for high levels of new and exciting stimulation. However,

alcohol-facilitated disinhibition seemed to be particularly relevant to

a social context of drinking or as part of social activities. If, as

Schwarz et al. (1978) suggested, a cultural expectancy of disinhibition

allows high sensation seekers an opportunity to obtain impunity for

"exhibitionistic, hedonistic behavior" (p. 1145), then it appears that

the behaviors that these students wanted to express with impunity were

drinking-related social interactions and pleasurable sensations.

Some researchers (Connors & Maisto, 1988) have concluded that

alcohol expectancies may have the greatest influence on behaviors that

are subject to the most social control or restraint. However, the

expectancy that drinking enhances normal socializing and produces

pleasant sensations was most relevant to drinking among students with

higher sensation-seeking needs. Although sexual and aggressive

disinhibition often seem to increase with student drinking, these









82

effects were not directly related to the primary motivational pattern

for drinking in this study.

For both male and female students, drinking was best explained by

joint consideration of sensation-seeking needs and the alcohol

expectancy of Social and Physical Pleasure. Overall, a similar

motivational pattern or disinhibition motive for drinking existed for

male and female students. Some gender differences were found in the

rate of increase of scores of the motivational factors with respect to

each other and drinking level. It is unclear why males' drinking levels

increased at a more rapid rate than did females' drinking levels in

relation to increasing sensation-seeking needs and the expectancy of

Social and Physical Pleasure. One interpretation of these results is

that males' drinking levels were more sensitive to these motivational

factors.

Among the implications of the current results is that the primary

focus of alcohol prevention programs should be students' social groups

and drinking peer groups. Within the peer group, students are generally

reinforced for acting "wild and crazy", being silly or humorous, or

otherwise expressing largely positive affect beyond the usual limits

acceptable outside of drinking situations. It is within these groups

and the social context that acceptable limits may best be negotiated for

drinking and drinking-related behavior. Because sensation-seeking needs

contribute to the primary motivation for drinking, students with high

levels of these needs should benefit from these limits being a

structured part of their social groups and environment.











At the same time however, the positive aspects of drinking to

achieve stimulation and disinhibition should be facilitated. For

example, ways of meeting sensation-seeking needs, such as enabling

students to "let their hair down", should be encouraged within an

environment that helps to ensure safe and acceptable limits on drinking

and the behavioral effects of drinking. Recognition of both personality

needs and alcohol expectancies should be helpful in designing prevention

programs.

Recommendations for Further Study

Further delineation of a disinhibition motive for drinking should

result from employing several other statistical procedures in continued

research. It would be helpful to use hierarchical multiple regression

analyses, as did Mooney and Corcoran (1989) to determine which alcohol

expectancies best predicted drinking for students of each gender who

were low or high in assertiveness. In this manner, it could be

determined how much each expectancy contributes to the explanation of

drinking for students with low or high levels of sensation-seeking

needs. In addition, MANOVA procedures such as were used by Ratliff and

Burkhart (1984) could test differences in alcohol expectancies of

students according to gender, light/heavy drinking level, and low/high

sensation-seeking needs. These statistical procedures would be helpful

in determining whether the alcohol expectancy of Social Assertion or

that of Social and Physical Pleasure is more relevant to drinking among

female students.

Several other variables appear particularly relevant to

understanding the relationships among sensation-seeking needs, alcohol









84

expectancies, and student drinking. It would be helpful to include a

measure of reasons for drinking to determine how social and escapist

reasons relate to the variables of the current study. It also appears

important to consider how much of a disinhibition motive for student

drinking may be explained in terms of independence needs that are

characteristic of development in the undergraduate age group.

A consideration for further research on a disinhibition motive for

drinking is the fact that a number of students in the current study did

not understand the meanings of the terms, "swingers" and "jet set", on

the Sensation Seeking Scale. This questionnaire uses several terms and

items that appear to relate more to the 1960's era than to today.

Updating this scale or using another measure of sensation-seeking

tendencies are alternatives that researchers in this area may wish to


consider.















APPENDIX A
DEMOGRAPHIC QUESTIONNAIRE



The following questions ask your age, sex, ethnicity, and body weight.

Your answers are anonymous; therefore please do not place your name or

social security number anywhere on the questionnaire.

1. Your age is _

2. Your sex is: a) male b)female

3. Your primary ethnic background is:

a) White American, Caucasian American

b) Black American, African American

c) Hispanic American

d) Asian American

e) other than above.

4. Your weight is
















APPENDIX B
KHAVARI ALCOHOL TEST



This is a series of questions about the use of alcoholic

beverages--what beverages people drink, how much, and how often. Please

check the statement that best applies to you, or write in the answer.

Answers are anonymous; therefore please do not put a name or social

security number on the questionnaire.

1. How often do you usually drink beer?

daily 3 or 4 times a year

3 or 4 times a week _once a year

twice a week I have tried beer, but I don't

once a week drink beer now.

twice a month never had beer

once a month

2. Think of all the times you have had beer recently. When you drink

beer, how much beer do YOU USUALLY DRINK each time?

CANS OR GLASSES?

1 5 9 13

2 6 10 14

3 7 11 15

4 8 12 More?

I don't drink beer









87

3. Each time you drink beer, what is the MOST YOU DRINK?

CANS OR GLASSES?

1 5 9 13

2 6 10 14

3 7 11 15

4 8 12 More?

I don't drink beer

4. About HOW OFTEN do you drink this much beer?

daily 3 or 4 times a year

3 or 4 times a week _once a year

twice a week I don't drink beer

once a week

twice a month

once a month

5. How often do you usually have wine, or a punch containing wine?

daily 3 or 4 times a year

3 or 4 times a week once a year

twice a week I have tried wine, but I don't

once a week drink wine now.

twice a month never had wine

once a month

6. Think of all the times you have had wine recently. When you drink

wine, how much wine or a punch containing wine do YOU USUALLY DRINK each

time?


5 9 13


1











2

3

4



7. Each time you

YOU DRINK?


6


10

11


14


15


8 12 More?

I don't drink wine

drink wine or a punch containing wine, what is the MOST


1

2

3

4


5

6

7

8


8. About HOW OFTEN do you

daily

3 or 4 times a week

twice a week


9

10


13

14


11 1

12 More

I don't drink wine

drink this much wine?

3 or 4 times a year

once a year

I don't drink wine


once a week

twice a month

once a month

9. How often do you usually have drinks containing whiskey or liquor

(such as martinis, manhattans, highballs, or straight drinks)?

daily 3 or 4 times a year


3 or 4 times a week

twice a week

once a week


once a year

I have tried whiskey and liquor,

but I don't drink them now.


5

?









89

twice a month never had whiskey or liquor

once a month

10. Think of all the times you have had drinks containing whiskey or

iquor recently. When you drink whiskey or liquor, how much do YOU

USUALLY DRINK each time?


1

2

3

4



11. Each time you


1

2

3

4


5

6

7

8

r I

drink


5

6

7

8

_I


12. About HOW OFTEN do

daily

3 or 4 times a week

twice a week

once a week

twice a month

once a month


9 13

10 14

11 15

12 More?

don't drink whiskey or liquor

whiskey or liquor, what is the MOST YOU DRINK?

DRINKS?

9 13

10 14

11 15

12 More?

don't drink whiskey or liquor

you drink this much whiskey or liquor?

3 or 4 times a year

_once a year

I don't drink whiskey or liquor
















APPENDIX C
SENSATION SEEKING SCALE



Each of the items below contains two choices, A and B. Please

indicate which of the choices most describes your likes or the way you

feel. In some cases you may find items in which both choices describe

your likes or the way you feel. Please choose the one which better

describes your likes or feelings. In some cases you may find items in

which you do not like either choice. In these cases mark the choice you

dislike least. It is important you respond to all items with only one

choice, A or B. We are interested only in your likes or feelings, not

in how others feel about these things or how one is supposed to feel.

There are no right or wrong answers. Be frank and give your honest

appraisal of yourself. This survey is anonymous; therefore please do

not put a name or social security number on the questionnaire.

1. A. I like "wild" uninhibited parties.

B. I prefer quiet parties with good conversation.

2. A. There are some movies I enjoy seeing a second or even a third

time.

B. I can't stand watching a movie that I've seen before.

3. A. I often wish I could be a mountain climber.

B. I can't understand people who risk their necks climbing

mountains.











4. A. I dislike all body odors.

B. I like some of the earthy body smells.

5. A. I get bored seeing the same old faces.

B. I like the comfortable familiarity of everyday friends.

6. A. I like to explore a strange city or section of town by myself,

even if it means getting lost.

B. I prefer a guide when I am in a place I don't know well.

7. A. I dislike people who do or say things just to shock or upset

others.

B. When you can predict almost everything a person will do and say

he or she must be a bore.

8. A. I usually don't enjoy a movie or play where I can predict what

will happen in advance.

B. I don't mind watching a movie or play where I can predict what

will happen in advance.

9. A. I have tried marijuana or would like to.

B. I would never smoke marijuana.

10. A. I would not like to try any drug which might produce strange and

dangerous effects on me.

B. I would like to try some of the new drugs that produce

hallucinations.

11. A. A sensible person avoids activities that are dangerous.

B. I sometimes like to do things that are a little frightening.

12. A. I dislike "swingers".

B. I enjoy the company of real "swingers".











13.



14.






15.






16.



17.



18.


92

A. I find that stimulants make me uncomfortable.

B. I often like to get high (drinking liquor or smoking marijuana).

A. I like to try new foods that I have never tasted before.

B. I order the dishes with which I am familiar, so as to avoid

disappointment and unpleasantness.

A. I enjoy looking at home movies or travel slides.

B. Looking at someone's home movies or travel slides bores me

tremendously.

A. I would like to take up the sport of water-skiing.

B. I would not like to take up water-skiing.

A. I would like to try surf-board riding.

B. I would not like to try surf-board riding.

A. I would like to take off on a trip with no pre-planned or

definite routes, or timetable.

B. When I go on a trip I like to plan my route and timetable

fairly carefully.

A. I prefer the "down-to-earth" kinds of people as friends.

B. I would like to make friends in some of the "far-out" groups lik

artists or "hippies".

A. I would not like to learn to fly an airplane.

B. I would like to learn to fly an airplane.

A. I prefer the surface of the water to the depths.

B. I would like to go scuba diving.

A. I would like to meet some persons who are homosexual (men or

women).

B. I stay away from anyone I suspect of being "queer".


19.


20.



21.



22.


e









93

23. A. I would like to try parachute jumping.

B. I would never want to try jumping out of a plane with or

without a parachute.

24. A. I prefer friends who are excitingly unpredictable.

B. I prefer friends who are reliable and predictable.

25. A. I am not interested in experience for its own sake.

B. I like to have new and exciting experiences and sensations even

if they are a little frightening, unconventional or illegal.

26. A. The essence of good art is in its clarity, symmetry of form and

harmony of colors.

B. I often find beauty in the "clashing" colors and irregular

forms of modern painting.

27. A. I enjoy spending time in the familiar surroundings of home.

B. I get very restless if I have to stay around home for any length

of time.

28. A. I like to dive off the high board.

B. I don't like the feeling I get standing on the high board (or I

don't go near it at all).

29. A. I like to date members of the opposite sex who are physically

exciting.

B. I like to date members of the opposite sex who share my values.

30. A. Heavy drinking usually ruins a party because some people get loud

and boisterous.

B. Keeping the drinks full is the key to a good party.

31. A. The worst social sin is to be rude.

B. The worst social sin is to be a bore.









94

32. A. A person should have considerable sexual experience before

marriage.

B. It's better if two married persons begin their sexual experience

with each other.

33. A. Even if I had the money I would not care to associate with

flighty persons like those in the "jet set".

B. I could conceive of myself seeking pleasure around the world witi

the "jet set".

34. A. I like people who are sharp and witty even if they do sometimes

insult others.

B. I dislike people who have their fun at the expense of hurting th

feelings of others.

35. A. There is altogether too much portrayal of sex in movies.

B. I enjoy watching many of the "sexy" scenes in movies.

36. A. I feel best after taking a couple of drinks.

B. Something is wrong with people who need liquor to feel good.

37. A. People should dress according to some standards of taste,

neatness, and style.

B. People should dress in individual ways even if the effects are

sometimes strange.


38. A.

B.



39. A.

B.


h


e


Sailing long distances in small sailing crafts is foolhardy.

I would like to sail a long distance in a small but seaworthy

sailing craft.

I have no patience with dull or boring persons.

I find something interesting in almost every person I talk with.









95

40. A. Skiing fast down a high mountain slope is a good way to end up on

crutches.

B. I think I would enjoy the sensations of skiing very fast down a

high mountain slope.




University of Florida Home Page
© 2004 - 2010 University of Florida George A. Smathers Libraries.
All rights reserved.

Acceptable Use, Copyright, and Disclaimer Statement
Last updated October 10, 2010 - Version 2.9.7 - mvs