Content uploaded by Sherry H Stewart
Author content
All content in this area was uploaded by Sherry H Stewart
Content may be subject to copyright.
Mood-Induced Increases in Alcohol Expectancy Strength
in Internally Motivated Drinkers
Cheryl D. Birch and Sherry H. Stewart
Dalhousie University
Anne-Marie Wall
York University
Sherry A. McKee
Yale University
Shondalee J. Eisnor and Jennifer A. Theakston
Dalhousie University
This study investigated whether exposure to musical mood induction procedures (MMIP) differentially
increases the strength of specific alcohol expectancies for coping motivated (CM) versus enhancement
motivated (EM) drinkers. Participants were 86 undergraduates who had elevated scores on either the CM
or EM subscale of the Drinking Motives Questionnaire (M. L. Cooper, 1994). Participants were randomly
assigned to either a positive or negative mood condition. The Alcohol Craving Questionnaire (E. G.
Singleton, S. T. Tiffany, & J. E. Henningfield, 1994) was administered at baseline and after MMIP to
assess phasic changes in alcohol expectancy strength. Consistent with hypotheses, only CM drinkers in
the negative mood condition reported increased relief expectancies, and only EM drinkers in the positive
mood condition reported increased reward expectancies. Theoretical and clinical implications are
discussed.
People who report a tendency to drink to regulate their emotions
(positive or negative) are often heavy or problem drinkers (e.g.,
Cooper, 1994; Cooper, Frone, Russell, & Mudar, 1995). Cooper’s
(1994) motivational model of drinking posits that people drink to
obtain either positive or negative reinforcement that is meant to
improve either their emotional (internal) or social (external) well-
being. There are two types of internal motives for drinking: (a)
coping motives (CM, drinking to alleviate negative emotions) and
(b) enhancement motives (EM, drinking to enhance positive emo-
tions). These internal motives for drinking are thought to be
phenomenologically distinct, in part, because they are associated
with unique drinking outcomes. Cross-sectional research has indi-
cated that although both CM and EM are associated with frequent
heavy drinking and with problem drinking, only CM predicts
problem drinking directly after controlling for usual consumption
levels (cf. Cooper, 1994). In any case, drinking for affect regula-
tion is more problematic than drinking for affiliative reasons
(Cooper, 1994).
To date, there has been very little research investigating the
relationship between drinking motives and the affective anteced-
ents of alcohol consumption. Self-reported frequent drinking in
situations involving unpleasant emotions and conflict with others
has been found to predict high CM scores on the Drinking Motives
Questionnaire (DMQ; Cooper, Russell, Skinner, & Windle, 1992),
and self-reported frequent drinking in situations involving pleasant
emotions and pleasant times with others has been found to predict
high EM scores on the DMQ (Carrigan, Samoluk, & Stewart,
1998). In an effort to overcome the limitations of this retrospec-
tive, self-report research, Stewart, Hall, Wilkie, and Birch (2002)
conducted a cognitive experiment involving a primed Stroop task
(Segal, Gemar, Truchon, Guirguis, & Horowitz, 1995). On this
task, alcohol schema were considered activated when participants
had longer color-naming latencies for alcohol target words versus
nonalcohol target words. Targets were displayed following the
presentation of neutral, positive affect, and negative affect prime
words. Results indicated that, as expected, negative affect (but not
neutral) primes activated alcohol schema for CM drinkers, and
positive (but not negative) affect primes activated alcohol schema
for EM drinkers. An unexpected finding, however, was that pos-
itive affect primes also activated alcohol schema for CM drinkers
and neutral primes also activated alcohol schema for EM drinkers.
Further research was thus needed to investigate the complex ques-
tion of whether there is indeed drinking motive specificity in the
affective triggers of alcohol cognitions and to identify a mecha-
nism to account for the observed priming effects. For example,
Cheryl D. Birch, Sherry H. Stewart, Shondalee J. Eisnor, and Jennifer A.
Theakston, Department of Psychology, Dalhousie University; Anne-Marie
Wall, Department of Psychology, York University; Sherry A. McKee,
Department of Psychiatry, Yale University.
This study was supported by a grant from the Social Sciences and
Humanities Research Council of Canada awarded to Sherry H. Stewart.
This study was conducted as a component of a doctoral dissertation by
Cheryl D. Birch under the supervision of Sherry H. Stewart. Cheryl D.
Birch was supported by a Social Sciences and Humanities Research Coun-
cil doctoral fellowship and by an Honorary Izaak Walton Killam Memorial
Scholarship. Sherry H. Stewart was supported by an Investigator Award
from the Canadian Institutes of Health Research. We gratefully acknowl-
edge the assistance of Michael Birch, Peter Hoaken, and Tarek Loubani.
Correspondence concerning this article should be addressed to Cheryl
D. Birch, Department of Psychology, Dalhousie University, Life Sciences
Centre, 1355 Oxford Street, Halifax, Nova Scotia B3H 4J1, Canada.
E-mail: cdbirch@dal.ca
Psychology of Addictive Behaviors Copyright 2004 by the Educational Publishing Foundation
2004, Vol. 18, No. 3, 231–238 0893-164X/04/$12.00 DOI: 10.1037/0893-164X.18.3.231
231
could the priming effects resulting from exposure to affectively
valenced prime words be due to transient changes in mood states?
One promising method for studying the relationship between
affective cues and alcohol cognition is by investigating the impact
of mood states on alcohol expectancies. Although alcohol expec-
tancies have been defined in many different ways, we define them
as learned beliefs about the mental, emotional, and behavioral
consequences of drinking (cf. Leigh, 1989; Marlatt & Gordon,
1985). These beliefs have been reliably found to operate as potent
incentives for alcohol consumption (cf. Goldman, Del Boca, &
Darkes, 1999). It should be noted that alcohol expectancies are
conceptually and empirically distinct from motives, even though
expectancies are logically implied by drinking motives (i.e., people
are not likely motivated to drink for tension-reduction reasons
unless they believe that alcohol can effectively ameliorate their
tension; cf. Cooper et al., 1995; Leigh, 1990). Expectancies are
necessary but not sufficient for motives (Leigh, 1990), and the fact
that they are often endorsed at widely different rates attests to their
discriminant validity. For example, people widely endorse tension-
reduction expectancies (see Leigh, 1989, for a review), but only a
small minority of drinkers report frequent drinking for coping-
related reasons (Stewart, Karp, Pihl, & Peterson, 1997). Motives
and expectancies can also be distinguished because motives are
considered to be trait-like individual-difference variables, whereas
current thinking conceptualizes expectancies as more state-like (cf.
Cooney, Gillespie, Baker, & Kaplan, 1987; MacLatchy-Gaudet &
Stewart, 2001; Schulze & Jones, 2000; Wall, McKee, & Hinson,
2000; Wall, McKee, Hinson, & Goldstein, 2001).
There has been very limited prior research investigating the
effects of both positive and negative mood states on alcohol
expectancies. Hufford (2001) administered an abbreviated Alcohol
Expectancy Questionnaire (AEQ; Brown, Christiansen, & Gold-
man, 1987) to undergraduates after inducing positive and negative
mood states with both musical and visual (slides) mood manipu-
lation procedures. Participants randomly assigned to the negative
mood condition endorsed a significantly greater number of posi-
tive alcohol expectancies (including more tension-reduction ex-
pectancies) than participants randomly assigned to the positive
mood condition. This research did not, however, take individual
differences in drinking motive status into account, and it did not
measure the impact of mood states on alcohol expectancy strength
(i.e., the strength with which individuals hold specific alcohol
expectancies; see Collins, Lapp, Emmons, & Isaac, 1990, for a
discussion of the importance of measuring expectancy strength).
To our knowledge, only one study to date has directly examined
the impact of both mood and drinking motives (CM and EM) on
alcohol expectancies (Goldstein, Wall, McKee, & Hinson, in
press). In this study, mood states were induced with musical mood
induction procedures (MMIP; Mongrain & Trambakoulos, 1997),
and alcohol expectancies were assessed by asking participants to
self-generate them in an open-ended format. Negative mood did
not predict increased self-generation of any type of expectancy and
positive mood predicted increased self-generation of expectancies
coded as social/situational enhancement outcomes. The discrep-
ancy of results reported by Hufford (2001) versus Goldstein et al.
(in press) could be due to the fact that Hufford measured explicit
expectancies (i.e., consciously accessed beliefs about which a
person is fully aware), whereas Goldstein et al. measured implicit
expectancies (i.e., beliefs that can influence behavior without
conscious recall or introspection; cf. Wiers et al., 2002). It should
also be noted that Goldstein et al. did not find that CM or EM
moderated the relationship between mood and alcohol expectan-
cies but this could be due to the fact that they performed a median
split on drinking motives scores to identify groups relatively high
and low in CM and EM rather than specifically selecting partici-
pants on the basis of extreme scores on the CM and EM subscales
of the DMQ (cf. Stewart et al., 2002).
In the present investigation we examined whether explicit alco-
hol expectancies vary as a function of mood state and drinking
motives. Positive and negative mood states were induced with
MMIPs (Mongrain & Trambakoulos, 1997). The dependent vari-
able was the strength of two types of positive, affect-regulation
alcohol expectancies (i.e., expectancies for emotional relief and
reward). The design was a 2 ⫻ 2 ⫻ 2 mixed factorial. The
between-subjects variables were drinking motives (CM or EM)
and mood (positive or negative), and the within-subject variable
was time (pre- and postinduction). We anticipated that CM (but not
EM) drinkers would show increases in emotional relief expectan-
cies after exposure to the negative MMIP. In contrast, we antici-
pated that EM (but not CM) drinkers would show increases in
emotional reward expectancies after exposure to the positive
MMIP. We tested the hypotheses using a priori planned compar-
isons (cf. Segal et al., 1995; Tabachnick & Fidell, 2001).
Method
Participants
The Drinking Motives Questionnaire—Revised (DMQ–R; Cooper,
1994) was included in a mass screening survey that was administered to
809 psychology undergraduate volunteers (70% female) from Dalhousie
University at the beginning of the academic year. The participants in this
study (n ⫽ 86; 62 women) were randomly selected from the larger group
of students in the screening sample who were identified as either CM or
EM drinkers. The male–female ratio in the resultant sample was represen-
tative of the imbalanced gender ratio in the screening sample. Half of the
participants in this study were selected according to their status as CM
drinkers (n ⫽ 42; 32 women) and the other half as EM drinkers (n ⫽ 44;
30 women). An individual was identified as a CM or an EM drinker if his
or her highest z score was either a CM or an EM subscale score and if his
or her high CM or EM score fell at least one standard deviation above the
student norm for that subscale (cf. Stewart et al., 2002). On average,
participants in this study were 19.3 years old (SD ⫽ 1.7), and the majority
(80.2%) were in their first year of university. The sample was 94.2%
Caucasian, and they were largely from middle- to upper-class family
backgrounds (e.g., 61.6% of participants reported an annual family salary
greater than $51,000 CDN [$38,500 U.S. dollars]).
As anticipated, the present sample of CM and EM drinkers reported
heavy drinking patterns and high rates of drinking-related problems (cf.
Cooper, 1994). Their average number of drinks per week was significantly
higher than the average reported by a different sample of students (Stewart,
Loughlin, & Rhyno, 2001) who were not expressly selected according to
their status as CM or EM drinkers (M ⫽ 10.6, SD ⫽ 12.2, vs. M ⫽ 6.4,
SD ⫽ 5.3 drinks/week, respectively), t(238) ⫽ 3.0, p ⬍ .01. On the Rutgers
Alcohol Problem Index (RAPI; White & Labouvie, 1989) the average score
for the present sample was 17.2 (SD ⫽ 12.2), which is above the cutoff
score of 15, indicating substantial alcohol problems (cf. Thombs & Beck,
1994).
We computed a series of 2 ⫻ 2 analyses of variance (ANOVAs) with
drinking motive (CM vs. EM) and mood condition (positive vs. negative)
as between-subjects factors to determine whether there were motive or
232
BIRCH ET AL.
mood group differences in alcohol use patterns and problems. CM and EM
drinkers did not differ in their reported number of alcoholic drinks per
week (M ⫽ 11.2, SD ⫽ 13.3, vs. M ⫽ 10.0, SD ⫽ 11.2, respectively). CM
drinkers did, however, have significantly higher RAPI scores than EM
drinkers (M ⫽ 20.8, SD ⫽ 13.7, vs. M ⫽ 13.8, SD ⫽ 9.4, respectively),
F(1, 82) ⫽ 7.87, p ⬍ .01. Only the CM drinkers’ average fell above the
cutoff score of 15 thought to indicate substantial drinking problems (cf.
Thombs & Beck, 1994). A series of 2 ⫻ 2 analyses of covariance
(ANCOVAs), controlling for drinking motive group differences in RAPI
scores, indicated that CM and EM drinkers did not differ on any of the
demographic variables (i.e., age, gender, year of university, or annual
family income range; all ps ⬎ .19). No effects involving mood condition
emerged in any of these analyses.
Materials
Demographics questionnaire. This author-compiled questionnaire as-
sessed demographic characteristics of participants, such as age, year of
university, gender, ethnicity, and family-of-origin annual income range.
Typical alcohol use was assessed with quantity–frequency methods (cf.
Stewart, Peterson, & Pihl, 1995).
DMQ–R. The DMQ–R (Cooper, 1994) is a 20-item self-report instru-
ment that yields scores on four subscales representing the four motives for
drinking (including CM and EM) identified in Cooper’s (1994) model. Five
items comprise each subscale, and the average of these items yields the
subscale score. Respondents estimate how often they are motivated to drink
for the reason specified in each item on a 5-point Likert scale. It has been
well-established that the DMQ–R has excellent psychometric properties
(Cooper, 1994; MacLean & Lecci, 2000). The internal consistencies of the
CM and EM subscales in the present experiment were .84 and .83,
respectively.
RAPI. The RAPI (White & Labouvie, 1989) is a 23-item well-
validated, self-report measure of drinking problems commonly experienced
by both clinical and community samples of adolescents and young adults
(Leccese & Waldron, 1994; White & Labouvie, 1989, 2000; Winters,
1999). Respondents indicate on a 5-point scale how many times (from
never to more than 10 times) during the last 3 years they have experienced
specific negative consequences due to their alcohol use (e.g., went to
school or work drunk). Responses were summed, as recommended by the
authors of the RAPI, to yield a problem frequency composite score (cf.
Winters, 1999). Cronbach’s alpha for the RAPI in this study was .91.
Alcohol Craving Questionnaire. Alcohol expectancies were assessed
with a craving questionnaire, the Alcohol Craving Questionnaire (ACQ–
Now; Singleton et al., 1994), because recently developed multidimensional
craving instruments have been found to reliably measure levels of different
alcohol-related cognitions, including positive alcohol expectancies, that are
thought to give rise to craving states (Drobes & Thomas, 1999; Love,
James, & Willner, 1998) and because alcohol expectancies are often
conceptualized as an important dimension of craving (cf. Drobes &
Thomas, 1999; Marlatt & Gordon, 1985). A multidimensional craving
instrument was also selected because they are thought to be highly sensi-
tive to phasic changes in alcohol-related cognitions (Love et al., 1998;
Sinha & O’Malley, 1999). The 47-item ACQ–Now was administered
because it measures two types of affect regulation alcohol expectancies in
separate 9-item subscales: (a) Anticipation of Positive Outcome (emotional
reward expectancies; e.g., “Drinking would make things seem just per-
fect”) and (b) Anticipation of Relief From Negative Outcome (emotional
relief expectancies; e.g., “If I used alcohol right now, I would feel less
tense”). ACQ–Now instructions state that “We are interested in how you
are thinking or feeling right now as you are filling out the questionnaire.”
Respondents rate their level of agreement with each item on a 7-point
Likert scale ranging from 1 (strongly disagree)to7(strongly agree). Mean
scores on each expectancy subscale were calculated. Expectancy subscales
showed good to excellent levels of internal consistency across the two
administrations of the ACQ–Now, with alpha values ranging from .76 to
.88.
Visual analogue scales. Mood induction research has often used visual
analogue scales (VAS) to measure experimentally induced fluctuations in
mood intensity (Martin, 1990). Participants rated the characteristics of their
current mood state on four positive affect (cheerful, happy, glad, and
pleased) and three negative affect (sad, depressed, and blue) continuum
scales (Mongrain & Trambakoulos, 1997). The scale endpoints very and
not at all were connected by 100-mm horizontal lines. Participants drew a
vertical line to intersect each horizontal continuum at the position that best
reflected their current mood state.
MMIP. The musical stimuli used in this experiment were developed by
Mongrain and Trambakoulos (1997) and are a modified version of MMIPs
originally developed by Pignatiello, Camp, and Rasar (1986). These
MMIPs have been found to reliably induce target positive and negative
mood states in university students (Goldstein et al., in press; McKee, Wall,
Hinson, Goldstein, & Bissonnette, 2003; Mongrain & Trambakoulos,
1997). The positive and negative MMIP are both 10-min compilations of
nonlyrical classical and popular musical pieces.
Procedure
Individual appointments were made for eligible participants. During the
initial telephone contact the experiment was described as an “examination
of lifestyle and musical preferences” (Goldstein et al., in press). To
minimize the possibility of eliciting demand characteristics, participants
were not informed prior to the experiment that they would be exposed to
MMIP (cf. Westermann, Spies, Stahl, & Hesse, 1996). The experimenter
remained blind to each participant’s motive group status and mood group
assignment.
On arriving at the laboratory, and after providing informed consent,
participants completed the first set of questionnaires, including the demo-
graphics questionnaire, the RAPI, and a test day DMQ–R, as well as the pre
mood induction ACQ–Now and VAS. After they completed these instru-
ments, the experimenter gave each participant a set of headphones. Prere-
corded instructions informed participants that they would hear a 10-min
arrangement of musical pieces and that they should focus on the music.
Participants were left alone during the MMIP (cf. Martin, 1990). After the
10-min MMIP, prerecorded instructions informed participants that they
should complete the second set of questionnaires, including the postmood
induction VAS and ACQ–Now. Participants listened to mood-condition-
appropriate background music, also developed by Mongrain and Tramba-
koulos (1997), while they completed the postmood induction question-
naire. At study completion, participants were fully debriefed, and those
who had received the negative MMIP were given the positive MMIP (cf.
Frost & Green, 1982).
Results
DMQ–R Scores
We computed a 2 ⫻ 2 ⫻ 2 ⫻ 2 mixed factorial ANOVA on the
standardized average DMQ–R CM and EM scores. Between-
subjects factors were motive group (CM drinkers vs. EM drinkers)
and mood condition (positive vs. negative). Within-subject factors
were test time (screen day vs. test day) and DMQ–R subscale (CM
subscale vs. EM subscale). There was a significant main effect of
motive group, F(1, 82) ⫽ 21.5, p ⬍ .0005, that was qualified by
significant Motive Group ⫻ Subscale, F(1, 82) ⫽ 161.9, p ⬍
.0005, and Motive Group ⫻ Subscale ⫻ Test Time, F(1, 82) ⫽
11.8, p ⬍ .001, interactions.
To further explore the three-way interaction, we conducted
simple effects analyses. There was only one significant simple test
233
MOOD, DRINKING MOTIVES, AND ALCOHOL EXPECTANCIES
time effect that emerged when we examined the effects of test time
for each subscale and motive group separately, collapsed across
mood condition. Initially extreme standardized EM scores reported
by EM drinkers decreased significantly from screen day to test day
(M ⫽ 1.4, SD ⫽ 0.4, vs. M ⫽ 1.2, SD ⫽ 0.6, respectively), F(1,
43) ⫽ 6.0, p ⬍ .05, but still remained substantially elevated at test
day relative to EM subscale norms. In contrast, standardized CM
scores reported by EM drinkers, and the EM and CM scores
reported by CM drinkers, remained stable from screen day to test day.
We then examined the Motive Group ⫻ Subscale interaction by
analyzing the simple effects of motives group for each subscale,
collapsed across mood condition and test times. EM drinkers
scored significantly higher than CM drinkers on the EM subscale,
F(1, 84) ⫽ 24.1, p ⬍ .0005 (M ⫽ 1.3, SD ⫽ 0.4, vs. M ⫽ 0.8,
SD ⫽ 0.6, respectively). Conversely, CM drinkers scored signifi-
cantly higher than EM drinkers on the CM subscale, F(1, 84) ⫽
116.8, p ⬍ .0005 (M ⫽ 1.7, SD ⫽ 0.7, vs. M ⫽ 0.3, SD ⫽ 0.5,
respectively). We also analyzed the simple effects of subscale for
each motive group separately, collapsed across mood condition
and test times. The EM drinkers scored significantly higher on the
EM subscale than the CM subscale, F(1, 43) ⫽ 158.3, p ⬍ .0005.
Conversely, CM drinkers scored significantly higher on the CM
subscale than the EM subscale, F(1, 41) ⫽ 45.7, p ⬍ .0005. Thus,
drinking motives remained sufficiently stable over time, and it is
reasonable to conclude that the participants who were identified as
CM or EM drinkers at screening were still CM or EM drinkers,
respectively, at the time of their study participation.
Manipulation Check
We first computed 2 (mood condition) ⫻ 2 (motive group)
between-subjects ANOVAs on mean baseline positive and nega-
tive affect scores from the VAS separately and found significant
motive group effects. As Table 1 indicates, CM drinkers reported
significantly less positive affect and more negative affect than EM
drinkers at baseline, Fs(1, 82) ⫽ 4.6 and 6.5, both ps ⬍ .05. There
were no mood group differences in baseline mood state.
We then computed a 2 (mood condition) ⫻ 2 (motive group)
ANCOVA on mean postinduction positive affect scores on the
VAS using mean baseline positive affect scores as the covariate.
As expected, the covariate-adjusted mean postinduction positive
affect score for participants in the positive mood group was sig-
nificantly higher than that for participants in the negative mood
group, F(1, 81) ⫽ 37.4, p ⬍ .05 (see Table 2). A parallel 2 ⫻ 2
ANCOVA on mean postinduction negative affect scores indicated
that participants who had been exposed to the negative MMIP had
a significantly higher baseline-adjusted mean negative affect score
than those exposed to the positive MMIP, F(1, 81) ⫽ 107.3, p ⬍
.001; see Table 2). No effects of motive group were observed in
either ANCOVA.
ACQ–Now Scores
We computed a 2 (mood condition) ⫻ 2 (motive group)
between-subjects ANOVA on the mean baseline expectancy
scores measured by the two expectancy subscales of the ACQ.
There were no mood or motive group differences in mean pre
mood induction expectancies; CM and EM drinkers within both
mood groups did not differ in their baseline endorsement of
emotional relief or emotional reward expectancies.
We examined mood-induced changes in these expectancies with
relation to the initial hypotheses by decomposing the full 2 ⫻ 2 ⫻
2 (Mood Condition ⫻ Motive Group ⫻ Test Time) table of means
into a series of planned comparisons (cf. Stewart et al., 2002). We
chose planned comparisons to analyze the most interesting com-
parisons first, using conventional alpha levels, as recommended by
Tabachnick and Fidell (2001). We specifically examined the de-
gree of change in the strength of each of the two expectancies from
pre to post mood induction separately for each mood condition and
motive group.
Consistent with hypotheses, only CM drinkers in the negative
mood condition showed a significant increase in emotional relief
expectancy strength, F(1, 20) ⫽ 9.3, p ⬍ .01,
2
⫽ .318. EM
drinkers in the negative mood group did not show a significant
change in relief expectancies, F(1, 18) ⫽ 0.2, ns. Similarly, neither
CM nor EM drinkers in the positive mood condition showed
significant changes in emotional relief alcohol expectancies from
the pre to post mood induction time: F(1, 20) ⫽ 0.5, ns, and F(1,
24) ⫽ 1.3, ns, respectively (see Figure 1).
Also consistent with hypotheses, only EM drinkers in the pos-
itive mood condition showed a significant increase in emotional
reward expectancy strength, F(1, 24) ⫽ 4.7, p ⬍ .05,
2
⫽ .165.
CM drinkers in the positive mood group did not show a significant
change in reward expectancies, F(1, 20) ⫽ 0.7, ns. Similarly,
neither CM nor EM drinkers in the negative mood condition
Table 1
Motive Group Differences in Mean Positive and Negative Affect
at Baseline
Group
Premood induction visual
analogue scale scores
Positive affect Negative affect
MSDMSD
Coping motivated 55.5 18.3
a
26.6 22.4
a
Enhancement motivated 64.2 21.2
b
15.8 18.4
b
Note. Comparisons are between means in the same column. Different
subscripts indicate a significant difference between means (p ⬍ .05).
Table 2
Mood Group Differences in Covariate-Adjusted Mean Positive
and Negative Affect at Postmood Induction Time
Group
Postmood induction visual
analogue scale scores
Positive affect Negative affect
MSDMSD
Positive mood 69.0 16.6
a
16.3 12.7
a
Negative mood 60.9 16.6
b
26.4 12.7
b
Note. Postmood induction means were covariate adjusted to account for
motive group differences in baseline mood state. Comparisons are between
means in the same column. Different subscripts indicate a significant
difference between means (p ⬍ .05).
234
BIRCH ET AL.
showed significant changes in emotional reward alcohol expect-
ancies from pre to post mood induction time: F(1, 20) ⫽ 0.0, ns,
and F(1, 18) ⫽ 0.1, ns, respectively (see Figure 2).
1
Although
absolute levels of reward expectancy strength ratings at post mood
induction appear somewhat comparable across groups of both CM
and EM drinkers who received the positive mood induction, only
the EM drinkers in the positive mood group showed a significant
increase in the strength of their reward expectancies from their pre
mood induction baseline ratings.
Discussion
These findings support the notion that alcohol expectancies vary
as a function of drinking motives and mood states. Drinking
motives appear to moderate the relationship between mood states
and alcohol expectancy strength. As anticipated, only CM drinkers
displayed an increase in positive expectancy strength after expo-
sure to the negative mood induction, and only EM drinkers dis-
played an increase in positive expectancy strength after exposure
to the positive mood induction. To the extent that alcohol expec-
tancies are proximal predictors of drinking (Goldman et al., 1999),
the increase in positive expectancy strength observed in this ex-
periment should predict a reciprocal increase in the inclination to
drink (Collins et al., 1990; Darkes & Goldman, 1998). Although
further research is warranted, the results of this study are consistent
with prior self-report research concerning the unique affective
antecedents of drinking behavior among CM versus EM drinkers
(Carrigan et al., 1998), and together these results suggest that
negative emotions may be antecedents of actual drinking behavior
for CM drinkers and that positive emotions may be antecedents of
actual drinking for EM drinkers. Overall, the findings from this
study also help to provide converging evidence that any analysis of
the relationship among mood, alcohol cognition, and actual drink-
ing behavior should take individual differences, such as drinking
motives, into account.
The pattern of findings reported here is similar to that reported
by Stewart and colleagues (2002). In both studies, negative affect
cues increased alcohol cognition for CM drinkers, and positive
affect cues increased alcohol cognition for EM drinkers. Confi-
dence in these findings is justifiable because results were repli-
cated despite the use of different methodologies. It remains unclear
why Stewart et al. (2002) had less specificity in their study,
however, as we previously found that both negative and positive
affect (but not neutral) cues activated alcohol cognition for CM
drinkers and that both positive affect and neutral (but not negative
affect) cues activated alcohol cognition for EM drinkers. It is
1
Although there is a moderate effect size associated with the finding that
negative mood activates relief expectancies among CM drinkers, there is a
somewhat smaller effect size associated with the finding that positive mood
activates reward expectancies among EM drinkers.
Figure 1. Mood-induced changes in emotional relief alcohol expectancies. The asterisk indicates a significant
increase from pre- to postmood induction. CM ⫽ coping motivated; EM ⫽ enhancement motivated.
235
MOOD, DRINKING MOTIVES, AND ALCOHOL EXPECTANCIES
possible that the affective cues for implicit alcohol cognitions are
different than the affective cues for explicit alcohol cognitions.
The Stroop task used in Stewart et al.’s (2002) study is thought to
tap implicit cognitions (e.g., Roehrich & Goldman, 1995), whereas
for questionnaire responding in the present study, explicit alcohol
expectancies were measured (Goldman et al., 1999). Future re-
search should investigate whether implicit and explicit alcohol
expectancies vary comparably as a function of mood states and
drinking motives.
The results of this study are inconsistent with Goldstein et al.’s
(in press) finding that drinking motives do not moderate mood–
expectancy relations. Although the discrepancy could be due to
differences in expectancy assessment across the two studies, it
could also be due to differences in the method of identifying CM
and EM drinkers. It is possible that motives exert a moderating
influence only when individuals have extreme CM and EM scores.
This study is of particular interest because it demonstrates that
the strength of specific types of emotion regulation alcohol ex-
pectancies can vary as a function of both mood states and drinking
motives. Negative mood specifically strengthened only emotional
relief (vs. reward) expectancies for CM drinkers, and positive
mood specifically strengthened only emotional reward (vs. relief)
expectancies for EM drinkers. There are several possible explana-
tions for this mood- and motive-dependent specificity in the types
of expectancies that were strengthened. It could be due to individ-
ual differences in sensitivity to the reinforcing effects of alcohol.
Research has shown for example, that some individuals (possibly
CM drinkers) are more likely to achieve tension reduction from
alcohol, whereas others (possibly EM drinkers) are highly sensi-
tive to the euphoric effects of alcohol (for reviews, see Pihl &
Peterson, 1995; Verheul, van den Brink, & Geerlings, 1999).
Hence, some individuals may be “hard wired” to learn that alcohol
can regulate the quality of their emotional experiences in very
specific ways. Motive group differences in personality could also
explain how drinking motives moderate the relationship between
mood and specific alcohol expectancies (Cooper et al., 1995;
Stewart & Devine, 2000). Cooper et al. (1995) speculated that CM
drinkers are highly responsive to punishment cues and strongly
compelled to regulate aversive motivation, whereas EM drinkers
are highly sensitive to reward cues and strongly compelled to
regulate appetitive motivation. Further research is needed to in-
vestigate whether there are reliable drinking motive differences in
personality and/or in sensitivity to alcohol’s various effects that
could account for the observed mood- and motive-specific pattern
of alcohol expectancy strengthening.
Some research has indicated that specific mood states are asso-
ciated with specific expectancies, but this research to date has not
also considered the role of drinking motives. Hufford (2001) found
negative mood to result in reports of more tension-reduction alco-
hol expectancies than positive mood. Within the smoking litera-
Figure 2. Mood-induced changes in emotional reward alcohol expectancies. The asterisk indicates a significant
increase from pre- to postmood induction. CM ⫽ coping motivated; EM ⫽ enhancement motivated.
236
BIRCH ET AL.
ture, McKee et al. (2003) found that participants who received a
positive mood induction were more likely to self-generate positive
reinforcement smoking expectancies than those in the negative or
neutral mood group and, conversely, those in the negative or
neutral mood group were more likely to self-generate negative
reinforcement expectancies than participants in the positive mood
group. Further research is needed to investigate whether motives
for substance use moderate the relationship between mood states
and specific mood-regulation expectancies.
An interpretive difficulty of this study pertains to the fact that
we did not control for reports of higher levels of alcohol problems
among CM versus EM drinkers. Thus, results may be due to group
differences in problem drinking versus group differences in drink-
ing motives. Cooney, Litt, Morse, Bauer, and Gaupp (1997) found,
however, that problem drinking severity is unrelated to the extent
of negative mood-induced craving. Only anxious and depressed
problem drinkers who had a history of drinking in response to
unpleasant emotions, as measured by the Inventory of Drinking
Situations (IDS; Annis, Graham, & Davis, 1987), reported in-
creased craving after their sad mood manipulation. Similarly,
Zack, Toneatto, and MacLeod (1999) found that, among problem
drinkers, only those who reported high psychiatric distress showed
strong cognitive associations between negative affect and alcohol
concepts. This suggests that individual differences, like drinking
motives, are more likely to account for negative mood reactivity
than problem drinking severity per se.
Some further limitations of this research include a lack of a
neutral mood control group and the fact that we did not examine
gender differences. It should also be acknowledged that in absolute
terms, positive affect remained higher than negative affect in the
negative mood group. Although this finding is consistent with
other studies that have used a mood induction (cf. McKee et al.,
2003), further similar research may yield a different pattern of
results if a stronger affect manipulation is used. It is especially
important to replicate the finding of a positive-mood reward ex-
pectancy strengthening among EM drinkers, because this effect
was less pronounced than the negative-mood relief expectancy
strengthening among CM drinkers (see footnote 1). Future re-
search might also investigate drinking motive group differences in
the mood-induced strengthening of specific positive expectancies
other than affect regulation alcohol expectancies (cf. Hufford,
2001). It is perhaps most important for future research to investi-
gate whether this mood-induced increase in expectancy strength
among CM and EM drinkers will correspond to a mood-induced
increase in their actual drinking behavior. Prior research identify-
ing expectancies as an important proximal cause of drinking be-
havior suggests there should be a correspondence (Darkes &
Goldman, 1998; Goldman et al., 1999).
Our major finding that there are drinking motive differences in
the affective triggers for positive alcohol expectancies has impor-
tant clinical implications for the prevention and treatment of heavy
and problem drinking. Clients could be assessed with the DMQ–R
(Cooper, 1994) to determine whether they endorse “risky” internal
motives for drinking, and treatment should be tailored appropri-
ately to client needs. Clients who endorse internal motives on the
DMQ–R need to be trained to recognize the affective triggers
(negative for CM clients and positive for EM clients) that may
increase their inclination to drink and the expectancy processes
through which this might occur. Each of these types of alcohol
expectancies could be challenged, perhaps through modifications
to existing expectancy challenge protocols developed to challenge
positive reinforcement expectancies (e.g., Darkes & Goldman,
1998). In addition, CM drinkers may need interventions to lower
or manage their relatively high levels of negative affect.
References
Annis, H. M., Graham, J. M., & Davis, C. S. (1987). Inventory of Drinking
Situations (IDS) user’s guide. Toronto, Ontario, Canada: Addiction
Research Foundation.
Brown, S. A., Christiansen, B. A., & Goldman, M. S. (1987). The Alcohol
Expectancy Questionnaire: An instrument for the assessment of adoles-
cent and adult alcohol expectancies. Journal of Studies on Alcohol, 48,
483–491.
Carrigan, G., Samoluk, S. B., & Stewart, S. H. (1998). Examination of the
short form of the Inventory of Drinking Situations (IDS–42) in a young
adult university student sample. Behaviour Research and Therapy, 36,
789–807.
Collins, R. L., Lapp, W. M., Emmons, K. M., & Isaac, L. M. (1990).
Endorsement and strength of alcohol expectancies. Journal of Studies on
Alcohol, 51, 336–342.
Cooney, N. L., Gillespie, R. A., Baker, L. H., & Kaplan, R. F. (1987).
Cognitive changes after alcohol cue exposure. Journal of Consulting and
Clinical Psychology, 55, 150–155.
Cooney, N. L., Litt, M. D., Morse, P. A., Bauer, L. O., & Gaupp, L. (1997).
Alcohol cue reactivity, negative mood reactivity, and relapse in treated
alcoholic men. Journal of Abnormal Psychology, 106, 243–250.
Cooper, M. L. (1994). Motivations for alcohol use among adolescents:
Development and validation of a four-factor model. Psychological As-
sessment, 6, 117–128.
Cooper, M. L., Frone, M. R., Russell, M., & Mudar, P. (1995). Drinking to
regulate positive and negative emotions: A motivational model of alco-
hol use. Journal of Personality and Social Psychology, 69, 990–1005.
Cooper, M. L., Russell, M., Skinner, J. B., & Windle, M. (1992). Devel-
opment and validation of a three-dimensional measure of drinking
motives. Psychological Assessment, 4, 123–132.
Darkes, J., & Goldman, M. S. (1998). Expectancy challenge and drinking
reduction: Process and structure in the alcohol expectancy network.
Experimental and Clinical Psychopharmacology, 6, 64–76.
Drobes, D. J., & Thomas, S. E. (1999). Assessing craving for alcohol.
Alcohol Research and Health, 23, 179–192.
Frost, R. O., & Green, M. L. (1982). Velten mood induction procedure
effects: Duration and postexperimental removal. Personality and Social
Psychology, 8, 341–347.
Goldman, M., Del Boca, F., & Darkes, J. (1999). Alcohol expectancy
theory: The application of cognitive neuroscience. In K. E. Leonard &
H. T. Blane (Eds.), Psychological theories of drinking and alcoholism
(2nd ed., pp. 203–246). New York: Guilford Press.
Goldstein, A., Wall, A. M., McKee, S., & Hinson, R. (2004). Accessibility
of alcohol expectancies from memory: Impact of mood and motives in
college student drinkers. Journal of Studies on Alcohol, 65, 95–104.
Hufford, M. R. (2001). An examination of mood effects on positive alcohol
expectancies among undergraduate drinkers. Cognition and Emotion,
15, 593–613.
Leccese, M., & Waldron, H. B. (1994). Assessing adolescent substance
use: A critique of current measurement instruments. Journal of Sub-
stance Abuse Treatment, 11, 553–563.
Leigh, B. C. (1989). In search of the seven dwarves: Issues of measurement
and meaning in alcohol expectancy research. Psychological Bulletin,
105, 361–373.
Leigh, B. C. (1990). Alcohol expectancies and reasons for drinking:
Comments from a study of sexuality. Psychology of Addictive Behav-
iors, 4, 91–96.
237
MOOD, DRINKING MOTIVES, AND ALCOHOL EXPECTANCIES
Love, A., James, D., & Willner, P. (1998). A comparison of two alcohol
craving questionnaires. Addiction, 93, 1091–1102.
MacLatchy-Gaudet, H. A., & Stewart, S. H. (2001). The context-specific
positive alcohol outcome expectancies of university women. Addictive
Behaviors, 26, 31–49.
MacLean, M. G., & Lecci, L. (2000, March). A comparison of models of
drinking motives in a university sample. Psychology of Addictive Be-
haviors, 14, 83–87.
Marlatt, G. A., & Gordon, J. R. (1985). Relapse prevention: Maintenance
strategies in the treatment of addictive behaviors. New York: Guilford
Press.
Martin, M. (1990). On the induction of mood. Clinical Psychology Review,
10, 669–697.
McKee, S. A., Wall, A. M., Hinson, R. E., Goldstein, A., & Bissonnette, M.
(2003). Effects of an implicit mood prime on the accessibility of smok-
ing expectancies in college women. Psychology of Addictive Behaviors,
17, 219–225.
Mongrain, M., & Trambakoulos, J. (1997, August). Vulnerability and
changes in dysfunctional attitudes following a mood induction. Poster
presented at the 105th Annual Convention of the American Psycholog-
ical Association, Chicago.
Pignatiello, M. F., Camp, C. J., & Rasar, L. A. (1986). Musical mood
induction: An alternative to the Velten technique. Journal of Abnormal
Psychology, 95, 295–297.
Pihl, R. O., & Peterson, J. B. (1995). Alcoholism: The role of different
motivational systems. Journal of Psychiatry and Neuroscience, 20,
372–396.
Roehrich, L., & Goldman, M. S. (1995). Implicit priming of alcohol
expectancy memory processes and subsequent drinking behavior. Ex-
perimental and Clinical Psychopharmacology, 3, 402–410.
Schulze, D., & Jones, B. T. (2000). Desire for alcohol and outcome
expectancies as measures of alcohol cue-reactivity in social drinkers.
Addiction, 95, 1015–1020.
Segal, Z. V., Gemar, M., Truchon, C., Guirguis, M., & Horowitz, L. M.
(1995). A priming methodology for studying self-representation in major
depressive disorder. Journal of Abnormal Psychology, 104, 205–213.
Singleton, E. G., Tiffany, S. T., & Henningfield, J. E. (1994). The multi-
dimensional aspects of craving for alcohol. Unpublished research, In-
tramural Research Program, National Institute on Drug Abuse, National
Institutes of Health, Baltimore.
Sinha, R., & O’Malley, S. S. (1999). Craving for alcohol: Findings from
the clinic and the laboratory. Alcohol & Alcoholism, 34, 223–230.
Stewart, S. H., & Devine, H. (2000). Relations between personality and
drinking motives in young adults. Personality and Individual Differ-
ences, 29, 495–511.
Stewart, S. H., Hall, E., Wilkie, H., & Birch, C. (2002). Affective priming
of alcohol schema in coping and enhancement motivated drinkers.
Cognitive Behaviour Therapy, 31, 68–80.
Stewart, S. H., Karp, J., Pihl, R. O., & Peterson, R. A. (1997). Anxiety
sensitivity and self-reported reasons for drug use. Journal of Substance
Abuse, 9, 223–240.
Stewart, S. H., Loughlin, H. L., & Rhyno, E. (2001). Internal drinking
motives mediate personality domain–drinking relations in young adults.
Personality and Individual Differences, 30, 271–286.
Stewart, S. H., Peterson, J. B., & Pihl, R. O. (1995). Anxiety sensitivity and
self-reported alcohol consumption rates in university women. Journal of
Anxiety Disorders, 9, 283–292.
Tabachnick, B. G., & Fidell, L. S. (2001). Using multivariate statistics (4th
ed.). Needham Heights, MA: Allyn & Bacon.
Thombs, D. L., & Beck, K. H. (1994). The social context of four adolescent
drinking patterns. Health Education Research: Theory and Practice, 9,
13–22.
Verheul, R., van den Brink, W., & Geerlings, P. (1999). A three-way
pathway psychobiological model of craving for alcohol. Alcohol and
Alcoholism, 34, 197–222.
Wall, A. M., McKee, S., & Hinson, R. E. (2000). Assessing variation in
alcohol outcome expectancies across environmental context: An exam-
ination of the situational-specificity hypothesis. Psychology of Addictive
Behaviors, 14, 367–375.
Wall, A. M., McKee, S. A., Hinson, R. E., & Goldstein, A. (2001).
Examining alcohol outcome expectancies in laboratory and naturalistic
bar settings: A within-subject experimental analysis. Psychology of
Addictive Behaviors, 15, 219–226.
Westermann, R., Spies, K., Stahl, G., & Hesse, F. W. (1996). Relative
effectiveness and validity of mood induction procedures: A meta-
analysis. European Journal of Social Psychology, 26, 557–580.
White, H. R., & Labouvie, E. W. (1989). Towards the assessment of
adolescent problem drinking. Journal of Studies on Alcohol, 50, 30–37.
White, H. R., & Labouvie, E. W. (2000). Longitudinal trends in problem
drinking as measured by the Rutgers Alcohol Problem Index [Abstract].
Alcoholism: Clinical and Experimental Research, 24, 76A.
Wiers, R. W., Stacy, A. W., Ames, S. L., Noll, J. A., Sayette, M. A., Zack,
M., & Krank, M. (2002). Implicit and explicit alcohol-related cognitions.
Alcoholism: Clinical and Experimental Research, 26, 129–137.
Winters, K. C. (1999). Screening and assessing adolescents for substance
use disorders (DHHS Publication No. SMA-99-3282). Retrieved June 6,
2002, from http://www.health.org/govpubs/BKD306/
Zack, M., Toneatto, T., & MacLeod, C. M. (1999). Implicit activation of
alcohol concepts by negative affective cues distinguishes between prob-
lem drinkers with high and low psychiatric distress. Journal of Abnormal
Psychology, 108, 518–531.
Received December 20, 2002
Revision received November 20, 2003
Accepted November 20, 2003 䡲
238
BIRCH ET AL.
A preview of this full-text is provided by American Psychological Association.
Content available from Psychology of Addictive Behaviors
This content is subject to copyright. Terms and conditions apply.