Content uploaded by Cynthia D Mohr
Author content
All content in this area was uploaded by Cynthia D Mohr
Content may be subject to copyright.
Coping-Anxiety and Coping-Depression Motives Predict Different Daily
Mood-Drinking Relationships
Valerie V. Grant and Sherry H. Stewart
Dalhousie University
Cynthia D. Mohr
Portland State University
Individuals with different drinking motives show distinctive patterns of alcohol use and problems.
Drinking to cope, or endorsing strong coping motives for alcohol use, has been shown to be particularly
hazardous. It is important to determine the unique triggers associated with coping drinking. One
limitation of past research has been the failure to contend with the complexities inherent in coping
motives. Using the Modified Drinking Motives Questionnaire–Revised (Grant, Stewart, O’Connor,
Blackwell, & Conrod, 2007), which separates coping-anxiety and coping-depression motives, we
investigated whether these motives moderated relationships between daily mood and subsequent drinking
(statistically controlling for sex, baseline anxious and depressive symptomatology, initial alcohol prob-
lems, and additional drinking motives). College students (N⫽146) provided daily reports of mood and
alcohol consumption online for 3 weeks. Multilevel modeling analyses revealed that, as hypothesized,
stronger initial coping-depression motives predicted higher daily depressed mood-alcohol consumption
slopes. Also consistent with expectation, stronger initial coping-anxiety motives predicted higher anxious
mood-alcohol consumption slopes. We discuss how this identification of the unique mood triggers
associated with each type of coping drinking motive can provide the basis for targeted interventions.
Keywords: alcohol, drinking motives, coping, depression, anxiety
Introduction
Despite increased efforts to reduce heavy episodic alcohol use
(i.e., binge drinking) among college students, the proportion of
frequent heavy episodic drinkers in this population increased be-
tween 1993 and 2001 (Wechsler et al., 2002). Furthermore, rates of
alcohol-related problems among college students either remained
the same or grew over this time period (Wechsler et al.). Perhaps
most concerning is the finding of a 1999 survey that over a third
of college students met criteria for an alcohol use disorder (abuse
or dependence) in the previous year (Knight et al., 2002).
Motivational Model of Alcohol Use
Clearly, given the high rates of heavy drinking and related
problems in the college student population, it is important to
determine which individuals are at relatively higher risk of prob-
lematic alcohol use. To this end, much individual differences
research has focused on alcohol use motives, or reasons for drink-
ing. In fact, drinking motives have been posited to comprise the
“final common pathway to alcohol use” through which other
influences on drinking behaviors operate (Cox & Klinger, 1988, p.
168).
1
In Cooper’s (1994) motivational model of alcohol use,
drinking motives can be distinguished along two dimensions: the
type of reinforcement sought (positive or negative) and the source
of the desired effects of drinking (internal or external). As com-
pared with the external drinking motives (positive-reinforcement
“social,” i.e., drinking for social/celebratory reasons, and negative-
reinforcement “conformity,” i.e., drinking to avoid social disap-
proval), the internal, affect-regulation motives (positive-
reinforcement “enhancement,” i.e., drinking to enhance positive
feelings, and negative-reinforcement “coping,” i.e., drinking to
avoid or alleviate negative emotional states, such as sadness or
anxiety), as measured by the Drinking Motives Questionnaire–
Revised (DMQ-R; Cooper), have been found to be concurrently
associated with riskier alcohol use. Cooper determined that en-
hancement and coping motives were both positively related to
typical frequency and quantity of alcohol consumption, heavy
drinking, and alcohol problems. Only coping motives were posi-
1
However, others have argued that alcohol outcome expectancies are a
more essential psychological determinant of drinking behavior (Goldman,
Del Boca, & Darkes, 1999).
Valerie V. Grant and Sherry H. Stewart, Department of Psychology,
Life Sciences Centre, Dalhousie University, Halifax, Nova Scotia, Canada;
Cynthia D. Mohr, Department of Psychology, Portland State University.
This study was supported by a grant from the Social Sciences and
Humanities Research Council of Canada (SSHRC) awarded to Sherry H.
Stewart and by a dissertation grant award from the Society for a Science of
Clinical Psychology awarded to Valerie V. Grant. This study was con-
ducted as a component of a doctoral dissertation by Valerie V. Grant under
the supervision of Sherry H. Stewart. A condensed version of this manu-
script was presented in a poster at the 69th Annual Convention of the
Canadian Psychological Association in June 2008 in Halifax, Nova Scotia,
Canada. Valerie V. Grant has been funded by two Nova Scotia Health
Research Foundation Student Awards, a SSHRC Doctoral Fellowship, and
Honourary Killam Predoctoral Scholarships over the course of the com-
pletion of this research. Sherry H. Stewart is currently supported by a
Killam Research Professorship from the Faculty of Science at Dalhousie
University; she was supported by an Investigator Award from the Canadian
Institutes of Health Research at the time this research was conducted.
Cynthia Mohr was supported by NIAAA grant R03-AA014598.
Correspondence concerning this article should be addressed to Valerie
V. Grant, Department of Psychology, Life Sciences Centre, Dalhousie
University, Halifax, Nova Scotia, Canada, B3H 4J1.
Psychology of Addictive Behaviors © 2009 American Psychological Association
2009, Vol. 23, No. 2, 226–237 0893-164X/09/$12.00 DOI: 10.1037/a0015006
226
tively related to alcohol problems when usual levels of alcohol use
were statistically controlled.
Furthermore, Cooper, Frone, Russell, and Mudar (1995) estab-
lished that coping motives were more proximal determinants of
alcohol consumption and related problems in adolescents and
adults from the general community than the negative mood states
thought to underlie coping drinking. Nonetheless, in a comparable
investigation with college students, Read, Wood, Kahler, Mad-
dock, and Palfai (2003) found weaker evidence of the importance
of coping motives. Global negative affect was concurrently related
to coping motives, which were related to alcohol problems (but not
to alcohol use). Prospectively though, coping motives did not
predict alcohol problems when baseline alcohol use and problems
were statistically controlled. Other studies investigating the asso-
ciations among negative mood, coping motives, alcohol use,
and/or alcohol problems have yielded varied results (Bradizza,
Reifman, & Barnes, 1999; Kassel, Jackson, & Unrod, 2000; Park
& Levenson, 2002; Perkins, 1999).
Decomposition of Coping Motives Into Anxiety- and
Depression-Relevant Subtypes
Evidently, the relations among negative affect, coping motives,
alcohol use, and alcohol problems are complex, especially in
college students, as reflected in the mixed research findings de-
scribed above. This complexity has prompted investigators to more
closely examine the current notion of coping motives. The coping
motives construct tapped by Cooper’s (1994) DMQ-R is generic,
with depression- and anxiety-related items loading onto one sub-
scale. However, some evidence suggests that depression and anx-
iety are associated with disparate patterns of alcohol use. For
instance, a U.S. national survey of college students in 2002 showed
that both alcohol use and heavy episodic drinking were more likely
in those students with higher (vs. lower) levels of self-reported
depression (The National Center on Addiction and Substance
Abuse [CASA], as cited in CASA, 2003). Conversely, some evi-
dence suggests a negative relationship between anxiety (especially
social anxiety) and alcohol-consumption variables (e.g., mean
number of binges per week) in college students (see review by
Morris, Stewart, & Ham, 2005). Nevertheless, despite their differ-
ent relations to alcohol use, both depression and anxiety are
associated with alcohol use disorders in college students (Dawson,
Grant, Stinson, & Chou, 2005).
Since depression and anxiety may each be linked to different
patterns of alcohol use, separate measures of depression-related
and anxiety-related coping motives for drinking have been devel-
oped. In an effort to capture the complexities of coping-motivated
drinking, Blackwell and Conrod (2003; also see Blackwell, Con-
rod, & Hansen, 2002) modified and extended earlier versions of
Cooper’s drinking motives measures (Cooper, 1994; Cooper, Rus-
sell, Skinner, & Windle, 1992) to create the Modified DMQ-R,
which distinguishes between drinking to cope with depression
(coping-depression) and drinking to cope with anxiety (coping-
anxiety). Using confirmatory factor analysis to examine the Mod-
ified DMQ-R scores of college students, Grant, Stewart,
O’Connor, Blackwell, and Conrod (2007) found support for a
correlated five-factor structure (with separate social, conformity,
enhancement, coping-depression, and coping-anxiety factors).
Moreover, the correlated five-factor structure fit the data better
than a correlated four-factor structure theoretically equivalent to
that proposed by Cooper (i.e., with all of the coping items con-
strained to one generic coping factor). In addition, Grant, Stewart,
O’Connor, and colleagues (2007) found that coping-depression
and coping-anxiety motives were related to different patterns of
alcohol use and related problems. For instance, coping-depression
motives, but not coping-anxiety motives, concurrently predicted
greater typical quantity of drinks consumed per occasion, over and
above basic demographic effects and controlling for the effects of
the other drinking motives. Over and above the effects of demo-
graphics and initial alcohol use, and controlling for the other
drinking motives, only coping-depression motives prospectively
predicted alcohol problems. Nevertheless, when alcohol use
(drinks per week) at Time 2 was also accounted for, only coping-
anxiety motives prospectively predicted alcohol problems.
Between- Versus Within-Person Effects
Grant, Stewart, O’Connor, and colleagues’ (2007) findings are
important because they suggest that coping-depression and coping-
anxiety motives are not redundant in their relationships to alcohol
use and related problems, and that both are important in predicting
risk for alcohol misuse. Nevertheless, the methodology used by
Grant, Stewart, O’Connor, et al. and the other studies reviewed
thus far is somewhat limited in that it only allows for clarification
of between-person research questions. For example, a between-
person research question is: do people with higher (vs. lower)
coping-depression motives consume more drinks per week, on
average? (Grant, Stewart, O’Connor, et al.). However, as pointed
out by Mohr et al. (2005), the predictions made by theoretical
models of drinking motives are inherently process-oriented. For
instance, it is expected that individuals who report that they are
motivated to drink to cope with depression would increase their
alcohol consumption at those times when they feel depressed.
Thus, to adequately test the key predictions of motivational models
of alcohol use, within-person analyses, which assess the process
aspect of drinking behavior, are required. In a daily process (e.g.,
daily diary or experience sampling) study, participants record
information about daily occurrences (e.g., about daily moods and
daily drinking behavior) close to the time that they actually hap-
pen. Accordingly, a major strength of the daily process design is
that recall biases are minimized.
Drinking Motive Moderation of Within-Person
Mood-Drinking Relations
Few investigations have examined the effects of coping drinking
motives on within-person mood-alcohol consumption relation-
ships. In their first study of community adults (moderate-to-heavy
drinkers), in which structured paper-and-pencil nightly diaries
were used, Todd, Armeli, Tennen, Carney, and Affleck (2003)
found that same-day associations between nervous mood and
drinking and between sad mood and drinking were not moderated
by measures of generic coping motives. In their second study of
moderate-to-heavy drinkers, Todd et al. (2003) had participants
use hand-held electronic computers for diary entries, ensuring data
collection in near-real time. In this second study, generic coping
motives for drinking positively moderated the relationships be-
tween weekday nervous mood and subsequent weekend drinking
227
MOTIVES PREDICT DAILY MOOD–ALCOHOL RELATIONSHIPS
and between weekday sad mood and subsequent weekend drink-
ing. In addition, generic coping motives positively moderated the
relationship between evening nervous mood and concurrent
evening alcohol consumption. In a sample of relatively frequent
drinkers, Todd and colleagues (2005) found that generic coping
motives positively moderated nervous mood-drinking associa-
tions, but not sad mood-drinking associations.
A small number of studies have focused on the moderating
effects of coping motives for alcohol use on mood-drinking asso-
ciations in college students. Park, Armeli, and Tennen (2004)
found that generic coping drinking motives did not moderate
within-person associations between global negative affect and
drinking. However, Hussong, Galloway, and Feagans (2005)
found that generic coping motives did moderate daily sadness-
drinking relations, with individuals displaying relatively weak
coping motives showing no linear relation between sadness and
alcohol consumption and those with stronger coping motives un-
expectedly drinking less on days with greater sadness. Generic
coping motives did not moderate linear daily fear-drinking asso-
ciations, but in quadratic models they did moderate the effects of
daily fear on daily drinking in theoretically expected ways. Indi-
viduals with relatively high coping motives had an increased
likelihood of drinking on days with moderate or high levels of fear
(vs. days with low levels of fear), while those with weaker coping
motives displayed essentially no relationship between fear and
daily drinking. Mohr and colleagues (2005) found that generic
coping motives positively moderated the relationship between
daily generic negative mood and daily drinking at home.
An additional two studies have examined situational coping
motives (i.e., elicited daily reports of the extent to which alcohol
was used to cope). Flynn (2000) determined that mean daily
depressed mood (over a 45-day reporting period) and mean daily
(situational) alcohol coping (i.e., the extent to which alcohol was
used to cope with the most stressful event of that day) were
positively correlated in a sample of college students. Conversely,
Galloway (2007) found that daily (situational) generic coping
motives did not interact with daily generic negative affect in
predicting daily alcohol consumption. Likewise, dispositional ge-
neric coping motives did not moderate daily generic negative
affect-drinking associations and dispositional sadness-specific
coping motives did not moderate daily sad affect-alcohol con-
sumption relations. However, Galloway did find a significant inter-
action of daily sadness-related coping motives and daily sad mood,
such that there was a positive relationship between sad mood and
drinking when greater daily sadness-related coping motives were
endorsed and a negative relationship between sad mood and alco-
hol consumption when daily sadness-related coping motives were
weaker. This result, coupled with the finding that daily generic
coping motives did not interact with daily generic negative affect,
highlights the potential importance of decomposing generic neg-
ative affect and generic coping motives into specific subtypes
when testing potential moderation of daily mood-alcohol con-
sumption relationships by coping drinking motives.
The Current Study
Given the mixed findings described earlier, it appears that
further research is necessary to clarify the possible moderating
effects of alcohol use motives on daily mood-alcohol consumption
relations. It is possible that accounting for the complexities inher-
ent in coping motives and the negative affect theoretically posited
to underlie coping motives (see Grant, Stewart, O’Connor, et al.,
2007) might help to elucidate the associations among motives,
daily moods, and daily drinking. The current investigation is an
extension of the Mohr et al. (2005) daily process study that
examined the moderation of within-person daily mood-alcohol use
associations by drinking motives among college students. How-
ever, instead of using just one measure of daily global negative
mood and the four-factor DMQ-R (Cooper, 1994) with a single
generic coping motives subscale, we used separate measures of
daily depressed and daily anxious affect along with the five-factor
Modified DMQ-R (Grant, Stewart, O’Connor, et al.), which dif-
ferentiates coping-depression from coping-anxiety motives. We
were most interested in determining whether coping-depression
and coping-anxiety motives moderated daily depressed mood-
drinking and daily anxious mood-drinking associations, respec-
tively; thus daily positive affect was excluded from the model
tested in the current study. Each afternoon for 3 weeks, partici-
pants logged onto a secure Web site to rate their moods for that day
and to report the number of alcoholic beverages they had con-
sumed the previous evening. The data had a hierarchical structure,
with 3 weeks of daily assessments of mood and alcohol consump-
tion nested within each person.
The hypotheses were: (1) that students with stronger (vs.
weaker) coping-depression (but not coping-anxiety) drinking mo-
tives would consume more alcohol in the evenings following days
with higher depressed mood, and (2) that those with stronger (vs.
weaker) coping-anxiety (but not coping-depression) motives
would drink more in the evenings following days with greater
anxious mood. The analyses were focused on coping motives for
drinking, because of their demonstrated relationship to problematic
outcomes (Cooper, 1994; Grant, Stewart, O’Connor, et al., 2007).
To rule out the possibility that any potential coping-motive
moderation of anxious mood- or depressed mood-drinking slopes
might be entirely attributable to differences in anxious or depres-
sive symptomatology, we controlled for baseline symptoms of
anxiety and depression in the analyses. The other (non-coping)
drinking motives were also statistically controlled. Moreover, to
ensure that any evidence of motive moderation of the mood-
drinking slopes was not simply the result of relationships among
motives and alcohol problems, we controlled for initial alcohol
problems in the analyses. Sex was also controlled in the analyses,
given the possibility that the relationships among mood and drink-
ing for individuals with different drinking motives might vary as a
function of sex (e.g., Birch, Stewart, Girling, & Berish, 2006).
Method
Participants
College students enrolled in psychology courses at Dalhousie
University were recruited from mass screenings conducted in 2005
(from a paper-and-pencil screening battery administered in intro-
ductory psychology classes) and 2006 –2007 (from online screen-
ing open to students taking undergraduate psychology courses at
all levels), for a study described as “An Examination of Daily
Health and Daily Activities” (see detailed description of Sample 3
in Grant, Stewart, O’Connor, et al., 2007, which is the same as the
228 GRANT, STEWART, AND MOHR
current sample).
2
To be eligible, students had to indicate that they
had consumed alcohol at least 4 or 5 times in the 30 days prior to
the screening (on the Lifestyles Questionnaire). One hundred and
eighty-four undergraduates agreed to participate in the study, but 6
never actually commenced the study, resulting in 178 participants,
of which 5 indicated that they wished to have their data be
excluded from the analyses (an option given to all participants for
ethical reasons pertaining to avoidance of coercion of students to
participate in research). Remaining participants were excluded
because their screening data indicated that they had consumed
alcohol only 2 to 3 times in the 30 days prior to the screening (n⫽
1), because they did not complete all of the questionnaires in the
initial assessment (n⫽2), because they were much older than the
majority of the undergraduate participants (n⫽1), or because of
a technical problem with the daily survey (n⫽1),
3
leaving 168
participants.
As the relevant daily alcohol consumption items pertained to
drinking done in the previous evening, a lead variable for alcohol
consumption was created so that daily mood could predict subse-
quent evening drinking. Thus, each missed survey resulted in loss
of 2 reporting days and the upper limit of reporting days was
reduced to 20 per participant. Only those reporting days for which
there were both a daily mood report and a subsequent alcohol
consumption report were retained. Thirteen (8%) of the remaining
168 remaining participants were excluded from the remaining
analyses because they did not have any full daily entries. An
additional 9 (6%) of the remaining 155 participants were elimi-
nated from the rest of the analyses because they abstained from
drinking during the entirety of the daily reporting component of
the study, leaving a final sample of 146 participants (with 1,488
full daily entries).
An analysis of level-1 residuals (see Raudenbush, Bryk,
Cheong, Congdon, & du Toit, 2004) revealed 29 outlying daily
entries (i.e., ⬎[M⫹3.29 SD]; Tabachnick & Fidell, 2001) for 28
separate participants. The outlying daily entries were removed,
leaving 1,459 daily reports (i.e., 50% of a possible 20 ⫻146 ⫽
2,920), corresponding to a mean number of reporting days of 9.99
per participant (SD ⫽4.72). The participants included in the final
sample were predominantly female (n⫽113; 77%), Caucasian
(n⫽137; 94%), and in the first (n⫽82; 56%) or second (n⫽35;
24%) year of undergraduate studies.
Materials
Screening
Lifestyles Questionnaire. This author-compiled measure as-
sesses an array of lifestyle behaviors performed in the past 30 days.
It was administered during the mass screenings and again in the
initial assessment battery, but for the purposes of the current study,
just one item administered at screening was analyzed. Specifically,
one alcohol-related item that asked about the frequency of alcohol
consumption in the past 30 days was used to determine eligibility
for the current study (i.e., participants had to have consumed
alcohol at least 4 or 5 times in the 30 days before the screening).
Scale anchors were 0 (Not Applicable [Only if you did NOT drink
alcohol in the past 30 days]) and 4 (6 or more times). Of the final
sample of participants included in the current study, 75 individuals
(51%) reported having consumed alcohol 4 or 5 times in the 30
days prior to the screening, and the rest (n⫽71; 49%) reported
having consumed alcohol 6 or more times in the 30 days before the
screening.
Initial Assessment Battery
A variety of measures were administered to the current sample,
but only those pertinent to the present study will be described
herein. The initial assessment battery took about 25 minutes to
complete (including the measures not described below).
Demographics Questionnaire. In order to minimize the sa-
lience of the alcohol-related questions (Babor, Brown, & Del
Boca, 1990), this 29-item questionnaire presented open-ended
items about usual alcohol consumption (quantity and frequency)
amongst items about other behaviors (e.g., exercise, caffeine in-
take, and cigarette smoking) and typical demographics items.
Participants reported their usual number of drinking occasions per
week (or per month, if less often than weekly; or per year, if less
often than monthly). In addition, participants indicated their typi-
cal number of alcoholic beverages (one alcoholic beverage ⫽one
bottle of beer, one cooler, one small [4-ounce] glass of wine, or
one shot/mixed drink containing an ounce of hard liquor) per
occasion. At the outset of the study, participants reported that, on
average, they drank 1.61 times per week (SD ⫽1.00) and had a
mean of 6.03 alcoholic beverages per occasion (SD ⫽2.84).
Modified Drinking Motives Questionnaire–Revised. The 28-
item Modified DMQ-R (Blackwell & Conrod, 2003; Grant, Stew-
art, O’Connor, et al., 2007) is a modified version of Cooper’s
(1994; Cooper et al., 1992) earlier drinking motives measures
which differentiates between drinking alcohol to cope with de-
pressed feelings (coping-depression) and drinking to cope with
anxious emotions (coping-anxiety). It has items representing five
motives for drinking: 5 social items (alphas presented are for final
sample in current study; ␣⫽.68), 5 conformity items (␣⫽.90),
5 enhancement items (␣⫽.82), 9 coping-depression items (␣⫽
.94), and 4 coping-anxiety items (␣⫽.68). Participants indicated
how often they are motivated to drink for the reason provided in
each item on a 5-point Likert scale ranging from 1 (almost never/
never)to5(almost always/always). Each subscale is scored as the
mean of its component items. See Grant, Stewart, O’Connor, et al.
for evidence of the good to excellent psychometric properties of
this measure. See Table 1 for means and SDs of the Modified
DMQ-R subscales.
2
One participant who did not complete the entire initial questionnaire
battery was inadvertently excluded from the sample description in Grant,
Stewart, O’Connor, et al. (2007), but is included in the current sample
description.
3
Because of a technical problem, some participants were able to access
the daily survey more than once per day. If a participant provided multiple
daily reports on the same day, all of that participant’s entries for that day
were deleted, because it is impossible to determine which report was
intended for that actual day. In total, only 131 daily mood surveys for 57
unique reporting days (i.e., just 2% of 2,289 unique reporting days) for 31
participants were deleted, leaving 2,232 unique mood reporting days.
Similarly, 96 daily alcohol surveys for 40 unique reporting days (i.e., 2%
of 2,277 unique reporting days) for 16 participants were deleted, leaving
2,237 unique daily alcohol reports. One participant had only duplicate
entries and was thus excluded from the remainder of the analyses.
229
MOTIVES PREDICT DAILY MOOD–ALCOHOL RELATIONSHIPS
Rutgers Alcohol Problem Index (RAPI). The RAPI (White &
Labouvie, 1989) has 23 items that measure alcohol-related prob-
lems among adolescents and young adults. On a scale of 0 (never)
to4(more than 10 times), respondents indicated how frequently
they had experienced each consequence of alcohol use in the past
3 years. As indicated in Table 1, the mean RAPI total (summed)
score was 16.25, which is above the score of 15 that indicates
significant alcohol problems (Thombs & Beck, 1994). In the
current study, Cronbach’s ␣for this subscale was .93. RAPI scores
will be referred to as “alcohol problems” throughout the Results
and Discussion sections.
Mood and Anxiety Symptom Questionnaire (MASQ). The
MASQ (Watson & Clark, 1991; Watson, Clark, et al., 1995)
consists of 90 items that were derived from anxiety and mood
disorder symptom criteria. As part of the initial assessment, par-
ticipants were asked to indicate the extent to which they had
experienced each symptom in the past week, including the day that
they completed the MASQ, on a scale from 1 (not at all)to5
(extremely). For the current study, two subscales were analyzed
(see Watson, Weber, et al., 1995): anxious arousal (sum of 17
items; current study ␣⫽.86), which includes anxiety-specific
symptoms of body tension and physical hyperarousal, and anhe-
donic depression
4
(sum of 21 items; current study ␣⫽.91), which
is comprised of depression-specific items reflecting anhedonia and
loss of interest, as well as reverse-scored items reflecting positive
affect (see Table 1 for descriptive statistics).
Daily Questionnaire Battery
Daily affect. Participants were instructed to indicate the degree
to which each of 13 words described their mood “today.” To
conform to the procedure of Mohr et al. (2005), participants
selected (clicked) a response for each item from a 5-point scale
ranging from 0 (not at all)to4(extremely). We used the same
items from the self-report mood scales used in prior research
(Grant & Stewart, 2007; Grant, Stewart, & Birch, 2007) to mea-
sure depressed (sad,depressed, and blue) and anxious affect
(nervous,anxious, and tense, from Shacham [1983]). The remain-
ing 7 items measured angry and positive affect, but were not
employed in the current analyses. Mood subscale scores were
means of component items. Internal consistencies for the mood
subscales were computed for 3 days representing the beginning
(day 2; n⫽63), middle (day 9; n⫽62), and end (day 16; n⫽78)
of the daily reporting component of the study, which resulted,
respectively, in ␣⫽.92, .87, and .86 for depressed mood and ␣⫽
.88, .87, and .89 for anxious mood.
Daily alcohol use. Each day, participants were asked to report
the number of alcoholic beverages they had consumed the previous
evening (i.e., subsequent to completing the previous survey or after
6:00 p.m. on the eve of the first day they completed the daily
reporting; see Mohr et al., 2005) at home and away from home.
Consistent with the Mohr et al. investigation, participants in the
current study indicated on a scale from 0 to 13 (with 13 indicating
greater than 12 drinks) the number of drinks they consumed at
each location. Operational definitions of an alcoholic drink (i.e.,
one alcoholic beverage ⫽one bottle of beer, one cooler, one small
[4-ounce] glass of wine, or one shot/mixed drink containing an
ounce of hard liquor) were provided. Predictions concerning loca-
tion of alcohol consumption were not central to the current inves-
tigation, so drinks at home and drinks away from home were
summed for the analyses reported herein. In an effort to improve
the distribution of the alcohol outcome variable, which was ini-
tially positively skewed and kurtotic (S⫽3.40, s
s
⫽0.06; K⫽
14.10, s
k
⫽0.13), the cap on alcoholic drinks per day was reduced
from 26 (13 home ⫹13 away from home) to 6, resulting in
reduced skew and kurtosis (S⫽2.05, s
s
⫽0.06; K⫽2.70, s
k
⫽
0.13).
Given that reactivity of self-monitoring may be reduced by
reporting on multiple types of behavior (Hayes & Cavior, 1977),
4
Because this questionnaire was completed by participants on a secure
Web site in the current study, we did not have the ability to immediately
respond appropriately to a participant who threatened to harm themselves.
Therefore, for ethical reasons, we removed the one MASQ item on the
anhedonic depression scale that pertains to suicidal ideation.
Table 1
Descriptive Statistics and Bivariate Correlations
Variable MSD 1 2 3 4 56789 101112
1. Sex
a
—— —
2. Social 3.04 0.78 ⫺.11 —
3. Conformity 1.42 0.72 ⫺.06 .30
ⴱⴱⴱ
—
4. Enhancement 2.99 0.98 ⫺.02 .55
ⴱⴱⴱ
.15 —
5. Coping-depression 1.49 0.77 ⫺.12 .07 .61
ⴱⴱⴱ
.10 —
6. Coping-anxiety 1.94 0.74 ⫺.24
ⴱⴱ
.19
ⴱ
.46
ⴱⴱⴱ
.30
ⴱⴱⴱ
.66
ⴱⴱⴱ
—
7. Anhedonic depression 54.03 13.38 .07 ⫺.11 .21
ⴱ
⫺.12 .47
ⴱⴱⴱ
.34
ⴱⴱⴱ
—
8. Anxious arousal 27.64 9.03 .01 .11 .15 .12 .22
ⴱⴱ
.25
ⴱⴱ
.27
ⴱⴱ
—
9. Alcohol problems
b
16.25 14.29 ⫺.01 .23
ⴱⴱ
.28
ⴱⴱ
.35
ⴱⴱⴱ
.37
ⴱⴱⴱ
.44
ⴱⴱⴱ
.20
ⴱ
.36
ⴱⴱⴱ
—
10. Depressed mood
c
0.55 0.60 .03 .01 .34
ⴱⴱⴱ
.03 .41
ⴱⴱⴱ
.32
ⴱⴱⴱ
.45
ⴱⴱⴱ
.39
ⴱⴱⴱ
.25
ⴱⴱ
—
11. Anxious mood
c
0.88 0.76 .07 ⫺.02 .28
ⴱⴱ
⫺.01 .27
ⴱⴱ
.28
ⴱⴱ
.26
ⴱⴱ
.38
ⴱⴱⴱ
.18
ⴱ
.65
ⴱⴱⴱ
—
12. Drinks
c,d
0.86 0.93 ⫺.17
ⴱ
.12 ⫺.02 .27
ⴱⴱ
.09 .17
ⴱ
.00 .23
ⴱⴱ
.32
ⴱⴱⴱ
⫺.07 ⫺.03 —
Note.N⫽146.
a
Sex was coded so that women ⫽⫹1 and men ⫽⫺1.
b
Alcohol problems ⫽initial Rutgers Alcohol Problem Index total scores.
c
Person-level means
were used in the analyses.
d
Mean score of each participant’s mean daily sum of alcoholic beverages consumed at home and those consumed away from
home, capped at 6 drinks.
ⴱ
p⬍.05.
ⴱⴱ
p⬍.01.
ⴱⴱⴱ
p⬍.001.
230 GRANT, STEWART, AND MOHR
the daily alcohol use items were embedded amongst several addi-
tional questions pertaining to caffeine use, cigarette smoking,
Internet usage, exercise, studying, and social interactions. A daily
events checklist was also administered, but not included in the
current analyses. The entire daily questionnaire battery took ap-
proximately 5 minutes to complete. No studies have explicitly
tested the validity of daily Internet self-reports of alcohol con-
sumption and mood states. However, daily self-reports of drinking
using interactive voice response technology have been validated
against objective measures of alcohol consumption (i.e., blood
alcohol concentrations measured from breath samples; Perrine,
Mundt, Searles, & Lester, 1995). Furthermore, Knapp and Kirk
(2003) found that mode of questionnaire administration (pencil-
and-paper, touch-tone telephone, or Internet, as in the current
study) did not affect participants’ responses to a survey covering a
broad range of topics (e.g., sexual behavior, alcohol use, and
suicidal ideation) when anonymity was guaranteed.
Procedure
Eligible screened students who had indicated preliminary inter-
est in participating in “An Examination of Daily Health and Daily
Activities” were sent an e-mail providing them with access to a
secure, password-protected Web site. At the outset of the study,
participants were presented with an online informed consent form
and were given the option to not participate, to participate as a
research “observer” only (i.e., data were not to be used in the
analyses), or to participate as a typical research participant (i.e.,
data were to be used in the analyses).
After students provided consent and commenced the question-
naires, they were given a maximum of 1 week to complete the
initial assessment battery. Subsequently, participants had approx-
imately 2 weeks in which to complete their first daily online
interview. To help ensure consistent reporting, automatic reminder
e-mails (containing a link to the secure study Web site) were sent
to participants each morning for the 21 consecutive days of the
study.
5
The online questionnaires were only available between
3:30 p.m. and 7:00 p.m., consistent with Mohr and colleagues
(2005). Missed daily surveys could not be completed at later date
(if they were, they were not included in the data analyses
6
).
Participants were debriefed in person and were awarded credit
points towards their eligible psychology course grade. Those who
completed at least 18 of the 21 possible daily surveys were
rewarded with a bonus of $20 Canadian.
Results
Descriptive Statistics and Correlations
Of the 1,459 reporting days in the final sample, 24% (n⫽354)
were drinking days. On drinking days (with the alcohol variable
capped at 6 drinks used for the calculations), participants con-
sumed a mean of 3.56 alcoholic beverages (SD ⫽2.04). See Table
1 for descriptive statistics for the daily drinks variable averaged
across all reporting days (i.e., not just across drinking days). Daily
reporting was generally consistent across days of the week, rang-
ing from 11% (n⫽161) of reporting occurring on Fridays to 16%
(n⫽235) of reporting occurring on Wednesdays. Drinking levels
varied by day of the week, with the highest alcohol consumption
occurring on Friday nights (M⫽2.27 drinks, SD ⫽2.57) and the
lowest consumption occurring on Monday nights (M⫽0.09
drinks, SD ⫽0.52).
Table 1 displays descriptive statistics and zero-order correla-
tions among the variables included in the multilevel regression
analyses. It should be noted that the daily mood and daily drinks
variables presented in this table are aggregated (i.e., they are
derived from person-level summary scores) and do not, therefore,
address the within-person associations of interest. All of the drink-
ing motives intercorrelations were positive and significant, except
the social— coping-depression, enhancement— coping-
depression, and enhancement— conformity correlations. The high-
est correlation was between coping-anxiety and coping-depression,
but these subscales fell well short of multicollinearity (i.e., rⱖ.90;
Tabachnick & Fidell, 2001). In general, men and women endorsed
similar levels of each type of drinking motive, with the exception
of coping-anxiety, which was more heavily endorsed by men (M⫽
2.27, SD ⫽0.83) than by women (M⫽1.85, SD ⫽0.69), t(144) ⫽
2.94, p⫽.004. Each of the drinking motives, as well as the initial
measures of depressive symptomatology and anxious symptom-
atology, were significantly positively correlated with initial alco-
hol problems. Despite their respective significant correlations with
baseline alcohol problems, social motives, conformity motives,
coping-depression motives, and baseline depressive symptomatol-
ogy were not related to mean drinks per day consumed during the
daily reporting portion of the study. However, the other drinking
motives (enhancement and coping-anxiety) and baseline anxious
symptomatology were positively correlated with drinking during
the daily reporting study component.
Mean daily depressed mood was positively correlated with both
types of coping motives, but more strongly so to coping-depression
than to coping-anxiety motives, and was also positively correlated
with conformity motives and baseline depressive and anxious
symptomatology, as well as with mean daily anxious mood. Fur-
thermore, though mean daily depressed mood was unrelated to
mean drinking during the study, it was positively related to base-
line alcohol problems. Mean daily anxious mood was positively
correlated to coping-anxiety motives, coping-depression motives,
conformity motives, baseline depressive and (more so) to anxious
symptomatology, as well as to baseline alcohol problems, but not
to mean daily drinking.
5
Because of some technical problems, several participants were not sent
daily reminder e-mails for a couple of days.
6
Because of a technical problem, some participants were able to access
the daily survey more than once per day. If a participant provided multiple
daily reports on the same day, all of that participant’s entries for that day
were deleted, because it is impossible to determine which report was
intended for that actual day. In total, only 131 daily mood surveys for 57
unique reporting days (i.e., just 2% of 2,289 unique reporting days) for 31
participants were deleted, leaving 2,232 unique mood reporting days.
Similarly, 96 daily alcohol surveys for 40 unique reporting days (i.e., 2%
of 2,277 unique reporting days) for 16 participants were deleted, leaving
2,237 unique daily alcohol reports. One participant had only duplicate
entries and was thus excluded from the remainder of the analyses.
231
MOTIVES PREDICT DAILY MOOD–ALCOHOL RELATIONSHIPS
Multilevel Analyses Testing the Central Hypotheses
Data Analytic Plan for Multilevel Modeling
To test the multilevel (hierarchical) hypotheses, we used HLM
software (Version 6.06; Raudenbush, Bryk, & Congdon, 2005).
This program permits estimation of within-person (daily, level-1)
and between-person (level-2) effects in diary data with missing
level-1 observations. Each individual’s (level-2) final regression
coefficient estimates are weighted to account for their number of
level-1 observations; thus individuals with fewer (vs. more) level-1
entries have less influence on the ultimate results (Raudenbush &
Bryk, 2002). Given that the level-1 outcome variable (daily alco-
holic drinks consumed) was a count with many zeros in the
distribution, a Poisson sampling model (with constant exposure)
with a log-link function was employed in the analyses (Mohr et al.,
2005; Raudenbush et al., 2004). The population-average model
with robust standard errors was interpreted and an over-dispersion
parameter was included. First, in the level-1 analyses, we tested the
within-person main effects of daily mood on subsequent evening
alcohol consumption, without considering possible moderation of
these associations by individual difference variables (e.g., drinking
motives, sex). Next, in the level-2 analyses, we tested our central
hypotheses by examining whether drinking motives moderated the
within-person main effects of daily mood on subsequent evening
drinking.
Level-1 Analyses: Mood-Drinking Associations
As displayed in Equation 1, nightly alcohol consumption each
day (capped at 6 drinks) was modeled as a function of depressed
and anxious mood ratings provided earlier in the day:
Drinksti ⫽b0i⫹b1i共depressed moodti兲
⫹b2i共anxious moodti兲⫹eti (1)
Although not demonstrated in Equation 1, the level-1 model
controlled for possible day-of-week effects using six day-of-week
dummy variables for Tuesday through Sunday, inclusive, which
were modeled as fixed effects (as in Mohr et al., 2005). Each of the
level-1 predictors was person-centered (e.g., an individual’s mean
daily depressed mood was subtracted from each of his/her daily
depressed mood subscale scores), with the exception of the day-
of-week variables. Accordingly, the level-1 intercept, b
0i
, can be
interpreted as the predicted number of drinks at day tfor individual
iat mean levels of the individual’s predictor variables. The level-1
slopes, b
1i
and b
2i
, are the partial within-person regression coef-
ficients, or the respective expected changes in daily drinks asso-
ciated with a unit increase in daily depressed and anxious moods,
controlling for the other mood and for the day-of-week effects, for
person i. The overdispersion parameter is represented by e
ti
.
Consistent with Mohr et al. (2005), the random components of
the level-1 slopes (but not of the level-1 intercept) were set to 0.
Without considering the impact of the level-2 individual difference
predictors, on average, the depressed mood-drinking partial slope
was not significantly different from 0, b
1
⫽⫺0.11, SE ⫽0.09, p⫽
.24. The average anxious mood-drinking partial slope was negative
and significant, b
2
⫽⫺0.16, SE ⫽0.08, p⫽.04. So, in general,
across the entire sample, daily depressed mood did not trigger
subsequent evening alcohol consumption and daily anxious mood
was protective against subsequent evening drinking. In multilevel
models using the Poisson sampling model with a log-link function,
exponentiation of partial slope coefficients yields odds ratios and
thus can aid in the interpretation of the findings by offering an
effect size index (see Mohr et al., 2005). The exponentiated anx-
ious mood-drinking partial slope, exp(b
2
⫽⫺0.16) ⫽0.85, indi-
cates that for each unit increase in daily anxious mood (controlling
for the other level-1 predictors), subsequent evening drinking
decreased by 15%. As expected, each of the day-of-week fixed
effects was positive and significant (indicating higher alcohol
consumption relative to Monday): for Tuesday, b⫽1.16, SE ⫽
0.38, p⫽.003; for Wednesday, b⫽1.77, SE ⫽0.41,p⬍.001;
for Thursday, b⫽2.36, SE ⫽0.42,p⬍.001; for Friday, b⫽3.25,
SE ⫽0.41,p⬍.001; for Saturday, b⫽3.12, SE ⫽0.41,p⬍.001;
and for Sunday, b⫽1.31, SE ⫽0.38,p⫽.001.
Level-2 Analyses: Drinking Motive Moderation of
Mood-Drinking Associations
The central hypotheses are concerned with testing individual differ-
ences in (specifically, drinking motive moderation of) daily mood-
drinking relations. Statistically, individual differences in daily mood-
drinking relations are expressed as cross-level interactions (e.g., level-2
coping-depression motives by level-1 depressed mood). To determine the
effects of the between-person (level-2) variables on the within-person
(level-1) mood-drinking relations, the level-1 regression coefficients were
each regressed on all of the level-2 predictors (including sex, drinking
motives, initial depression and anxiety symptoms, and initial alcohol
problems). The intercept model (i.e., average daily alcohol consumption,
b
0i
, at person-centered levels of daily depressed and anxious moods,
controlling for the day-of-week effects, predicted by sex, motives, de-
pression and anxiety symptoms, and alcohol problems) is displayed in
Equation 2 (see results in Table 2). Equations 3 and 4, which test the key
hypotheses concerning drinking motive moderation of daily mood-
drinking associations, show the level-2 regression models predicting the
within-person, level-1 mood-drinking partial slopes for depressed (b
1i
)
and anxious (b
2i
) moods, respectively.
b0i⫽␥00 ⫹␥01共sex兲⫹␥02共social 兲⫹␥03共conformity兲
⫹␥04共enhancement兲⫹␥05共coping-depression兲
⫹␥06共coping-anxiety兲⫹␥07共anhedonic depression兲
⫹␥08共anxious arousal 兲⫹␥09共alcohol problems兲⫹u0i
(2)
b1i⫽␥10 ⫹␥11共sex兲⫹␥12共social 兲⫹␥13共conformity兲
⫹␥14共enhancement兲⫹␥15共coping-depression兲
⫹␥16共coping-anxiety兲⫹␥17共anhedonic depression兲
⫹␥18共anxious arousal 兲⫹␥19共alcohol problems兲(3)
b2i⫽␥20 ⫹␥21共sex兲⫹␥22共social 兲⫹␥23共conformity兲
⫹␥24共enhancement兲⫹␥25共coping-depression兲
⫹␥26共coping-anxiety兲⫹␥27共anhedonic depression兲
⫹␥28共anxious arousal 兲⫹␥29共alcohol problems兲(4)
232 GRANT, STEWART, AND MOHR
All of the level-2 predictors (with the exception of contrast-
coded sex) were grand-mean centered, so ␥
10
and ␥
20
are inter-
preted as the average within-person partial regression coefficients
for the daily mood-alcohol consumption slopes (for depressed and
anxious mood, respectively) across the participants. The sex coef-
ficient, ␥
11,
for instance, is interpreted as the difference between
women (coded as ⫹1) and men (coded as ⫺1) in the depressed
mood-alcohol consumption partial relation, controlling for the
other level-2 predictors (i.e., drinking motives, etc.). The social
coefficient, ␥
12
, for example, is interpreted as the partial relation-
ship between social motives and the daily depressed mood-
drinking partial slopes, or the cross-level interaction between
level-2 social motives and level-1 depressed mood.
Level-2 depressed mood-drinking slope model. As predicted,
coping-depression motives positively moderated the partial asso-
ciation between daily depressed mood and daily alcohol consump-
tion (controlling for the levels of the other level-2 predictors; see
Table 2). We conducted follow-up simple slope tests for the
average level-1 mood-drinking associations, which involved in-
specting the level-2 intercepts (i.e., average) of the level-1 slopes
when the relevant level-2 predictor (coping-depression in this
case) was recentered so that 0 represented the level or group of
interest (Aiken & West, 1991). Using this procedure, we found that
when coping-depression motives were re-centered so that 0 repre-
sented relatively strong coping-depression motives, operationally
defined as 1 SD above the mean, the average level-1 depressed
mood-alcohol consumption partial slope was not statistically dif-
ferent from 0, ␥
10
⫽0.02, SE ⫽0.12, p⫽.88, exp(0.02) ⫽1.02.
Conversely, when weaker coping-depression motives, defined as 1
SD below the mean on coping-depression, were entered in the
model, the average level-1 depressed mood-drinking partial slope
was significantly negative, ␥
10
⫽⫺0.49, SE ⫽0.21, p⫽.02,
exp(⫺0.49) ⫽0.62. Specifically, for each unit increase in de-
pressed mood, controlling for the levels of the other level-2 pre-
dictors, for day-of-week effects and for the level of daily anxious
mood, subsequent alcohol consumption decreased by 38% among
those with relatively weak coping-depression motives. See Figure
1 for a graphical representation of the Coping-depression ⫻De-
pressed mood cross-level interaction.
Surprisingly, social motives also positively moderated the par-
tial association between daily depressed mood and drinking (see
Table 2). When relatively strong social motives (M⫹1SD) were
entered in the model, the average level-1 depressed mood-drinking
partial slope was not statistically different from 0, ␥
10
⫽0.05,
SE ⫽0.19, p⫽.78, exp(0.05) ⫽1.05. However, when relatively
weak social motives (M–1SD) were entered in the model, the
average level-1 depressed mood-drinking partial slope was signif-
icantly negative, ␥
10
⫽⫺0.52, SE ⫽0.18, p⫽.004,
exp(⫺0.52) ⫽0.60. So, for each unit increase in depressed mood,
controlling for the levels of the other level-2 predictors, for day-
of-week effects and for the level of daily anxious mood, conse-
Table 2
Prediction of Daily Mood-Alcohol Use Intercept and Slopes by Sex, Drinking Motives, Initial Symptoms of Depression and Anxiety,
and Alcohol-Related Problems
Predictor
Intercept model
(average daily drinking)
Daily mood-alcohol consumption slope models
Depressed mood-drinking Anxious mood-drinking
Coefficient SE p Coefficient SE p Coefficient SE p
Intercept ⴚ2.40 0.42 ⬍.001 ⫺0.23 0.12 .06 ⴚ0.19 0.08 .02
Sex
a
ⴚ0.22 0.07 .003 0.04 0.12 .74 ⫺0.02 0.08 .80
Social 0.14 0.10 .18 0.37 0.17 .04 ⴚ0.31 0.11 .007
Conformity ⴚ0.35 0.17 .04 ⴚ0.38 0.14 .008 0.45 0.11 ⬍.001
Enhancement 0.07 0.09 .41 ⴚ0.30 0.13 .02 0.07 0.10 .47
Coping-depression 0.20 0.10 .05 0.33 0.16 .04 ⫺0.22 0.14 .12
Coping-anxiety 0.08 0.13 .55 0.20 0.16 .22 0.23 0.11 .04
Anhedonic depression 0.00 0.01 .62 ⫺0.01 0.01 .07 0.01 0.01 .22
Anxious arousal 0.02 0.01 .10 0.01 0.01 .22 0.00 0.01 .87
Alcohol problems
b
0.01 0.00 .12 ⫺0.01 0.01 .09 0.00 0.01 .73
Note.N⫽146. Outcome variable is daily sum of alcoholic beverages consumed at home and those consumed away from home, capped at 6 drinks. These
analyses controlled for day-of-week effects. The population-average model with robust standard errors is reported. Significant partial regression coefficients
(at p⬍.05) are bolded. Coefficients are unstandardized and thus cannot be directly compared with one another.
a
Sex was coded so that women ⫽⫹1 and men ⫽⫺1.
b
Alcohol problems ⫽initial Rutgers Alcohol Problem Index total scores.
0.00
0.05
0.10
0.15
0.20
hgiHwoL
Depressed Mood
Number of Drinks
High coping-depression
(M+ 1 SD)
Low coping-depression
(M-1 SD)
0.00
0.05
0.10
0.15
0.20
hgiHwoL
Depressed Mood
Number of Drinks
High coping-depression
(M+ 1 SD)
Low coping-depression
(M-1 SD)
Figure 1. The cross-level interaction of coping-depression and daily
depressed mood on daily alcohol consumption.
233
MOTIVES PREDICT DAILY MOOD–ALCOHOL RELATIONSHIPS
quent drinking decreased by 40% among those with relatively
weak social motives.
In addition, as seen in Table 2, conformity and enhancement
motives both negatively moderated the partial association between
daily depressed mood and alcohol consumption. At relatively high
levels of conformity and enhancement motives (M⫹1SD), the
average level-1 depressed mood-drinking partial slopes were sig-
nificantly negative (for conformity, ␥
10
⫽⫺0.51, SE ⫽0.17, p⫽
.003, exp[⫺0.51] ⫽0.60, and for enhancement, ␥
10
⫽⫺0.53,
SE ⫽0.18, p⫽.004, exp[⫺0.53] ⫽0.59). However, at relatively
low levels of conformity and enhancement motives (M–1SD), the
average level-1 depressed mood-drinking partial slopes were flat
(for conformity, ␥
10
⫽0.04, SE ⫽0.15, p⫽.79, exp[0.04] ⫽
1.04, and for enhancement, ␥
10
⫽0.07, SE ⫽0.17, p⫽.70,
exp[0.07] ⫽1.07).
Level-2 anxious mood-drinking slope model. As anticipated,
coping-anxiety motives positively moderated the association be-
tween daily anxious mood and daily alcohol consumption (see
Table 2). A simple slope test revealed that when relatively strong
coping-anxiety motives (M⫹1SD) were used in the model, the
average anxious mood-drinking partial slope was not significantly
different from 0, ␥
20
⫽⫺0.02, SE ⫽0.10, p⫽.83, exp(⫺0.02) ⫽
0.98. When the coping-anxiety variable was re-centered to repre-
sent relatively weak coping-anxiety motives (M–1SD), the
average level-1 anxious mood-alcohol consumption partial slope
was significantly negative ␥
20
⫽⫺0.36, SE ⫽0.13, p⫽.006,
exp(⫺0.36) ⫽0.70. See Figure 2 for a graphical representation of
the Coping-anxiety ⫻Anxious mood cross-level interaction.
As shown in Table 2, conformity motives also positively mod-
erated the daily anxious mood-drinking relationship. Using stron-
ger conformity motives (M⫹1SD) in the model, the average
anxious mood-drinking partial slope was not significantly different
from 0, ␥
20
⫽0.14, SE ⫽0.11, p⫽.20, exp(0.14) ⫽1.15. When
weaker conformity motives (M–1SD) were entered in the model,
the average anxious mood-drinking partial slope was significantly
negative, ␥
20
⫽⫺0.52, SE ⫽0.12, p⬍.001, exp(⫺0.52) ⫽0.60.
Finally, social motives negatively moderated the daily anxious
mood-alcohol consumption relationship (see Table 2). Simple
slope tests revealed that at the level of relatively strong social
motives (M⫹1SD), the average anxious mood-drinking partial
slope was significantly negative, ␥
20
⫽⫺0.43, SE ⫽0.13, p⫽
.001, exp(⫺0.43) ⫽0.65, whereas at the level of weaker social
motives (M–1SD), the average anxious mood-drinking partial
slope was flat, ␥
20
⫽0.05, SE ⫽0.10, p⫽.64, exp(0.05) ⫽1.05.
7
Discussion
Motivational models of alcohol use (e.g., Cooper, 1994) predict
that individuals who endorse high (vs. low) levels of coping
motives would have more positive associations between negative
emotion and alcohol consumption. However, the results of
process-oriented studies testing this prediction in college students
have been fairly mixed (e.g., Mohr et al., 2005; Park et al., 2004).
It is possible that the findings to date have been equivocal because
researchers have not accounted for the complexities inherent in the
coping motives construct (i.e., they have typically used generic
coping motives measures). The current study extends prior re-
search by examining the impact of two specific subtypes of coping
motives, coping-depression and coping-anxiety, on daily mood-
drinking relations.
Drinking Motive Moderation of Within-Person
Mood-Drinking Relations
There was support for both of the hypotheses of this investiga-
tion, which were related to coping drinking motive moderation of
daily mood-drinking relationships. As predicted, we found that
coping-depression alcohol use motives positively moderated the
daily depressed mood-alcohol consumption association. Weaker
coping-depression motives (controlling for sex differences, the
levels of the other drinking motives, baseline depressive and
anxious symptomatology, and baseline alcohol problems) were
associated with a negative average daily depressed mood-drinking
partial slope, whereas stronger coping-depression motives were
associated with a flat average depressed mood-alcohol consump-
tion partial slope. In addition, as hypothesized, coping-anxiety
motives positively moderated the daily anxious mood-alcohol con-
sumption association. Relatively weak coping-anxiety motives
(controlling for sex differences, the levels of the other drinking
motives, baseline depressive and anxious symptomatology, and
7
We re-estimated the multilevel models with the purpose of trimming
effects that did not achieve marginal significance ( p⬍.10) in either the
intercept model or in the slope models (i.e., anxious arousal). The pattern
of significant results was largely the same as in the original models, with
two exceptions: alcohol problems became a significant positive predictor in
the intercept model, ␥
09
⫽0.01, SE ⫽0.00, p⫽.02, and the Coping-
depression ⫻Depressed mood cross-level interaction no longer reached
significance, ␥
15
⫽0.29, SE ⫽0.18, p⫽.10. Nevertheless, the form of the
Coping-depression ⫻Depressed mood cross-level interaction was highly
similar to that in the original analysis, with a basically flat slope at the level
of high coping-depression motives, ␥
10
⫽0.01, SE ⫽0.13, p⫽.97,
exp(0.01) ⫽1.01, and a (marginally) significant negative slope at the level
of low coping-depression motives, ␥
10
⫽⫺0.44, SE ⫽0.23, p⫽.05,
exp(-0.44) ⫽0.64. Thus, following Snijders and Bosker (1999), we re-
tained the original analysis, because even nonsignificant covariates can
have a modest influence on parameter estimates and should be included if
theory dictates.
0.00
0.05
0.10
0.15
0.20
hgiHwoL
Anxious Mood
Number of Drinks
High coping-a nxiety
(M+ 1 SD)
Low coping-anxiety
(M-1 SD)
0.00
0.05
0.10
0.15
0.20
hgiHwoL
Anxious Mood
Number of Drinks
High coping-a nxiety
(M+ 1 SD)
Low coping-anxiety
(M-1 SD)
Figure 2. The cross-level interaction of coping-anxiety and daily anxious
mood on daily alcohol consumption.
234 GRANT, STEWART, AND MOHR
baseline alcohol problems) were associated with a negative aver-
age daily anxious mood-alcohol consumption partial slope. Con-
versely, stronger coping-anxiety motives were associated with an
effectively flat daily anxious mood-drinking partial slope. The
current investigation is the first to demonstrate that two specific
subtypes of coping motives for drinking moderate the relationships
between corresponding specific subtypes of daily negative affect
and alcohol consumption. That is, coping-depression, but not
coping-anxiety, motives positively moderated the daily depressed
mood-drinking relationship and coping-anxiety, but not coping-
depression, motives positively moderated the daily anxious mood-
drinking association.
The present coping motives findings are consistent with certain
aspects of findings from prior daily process investigations. For
instance, the current results are congruent with the Mohr et al.
(2005) finding that generic coping motives moderated the negative
affect-alcohol consumption (at home) relation in college students.
Other researchers (e.g., Hussong et al., 2005; Park et al., 2004;
Todd et al., 2003, 2005) might have found more reliable evidence
of the expected coping motive moderation of negative affect-
drinking relations had they examined specific types of negative
affect (i.e., depressed and anxious mood, respectively) and corre-
sponding specific coping motives subtypes (i.e., coping-depression
and coping-anxiety, respectively).
Galloway (2007) was the only prior investigator to attempt to
match a specific subtype of coping motives (sadness-specific) to a
specific type of negative affect (sadness) in a study of drinking
motive moderation of daily mood-drinking relations. His finding
that daily sadness-specific coping motives positively moderated
daily sad mood-alcohol consumption slopes is consistent with the
current results. However, his finding that dispositional sadness-
specific coping motives did not moderate daily sadness-drinking
relations is at odds with the present findings. Perhaps the discrep-
ancy between the current results and Galloway’s findings is be-
cause of a difference in psychometric characteristics. The current
coping-depression motives subscale has been psychometrically
validated (see Grant, Stewart, O’Connor, et al., 2007), whereas
Galloway’s measure of dispositional sadness-specific coping mo-
tives has not undergone rigorous psychometric evaluation. The
above explanation is speculative of course, as the absence of a
comprehensive psychometric evaluation of Galloway’s disposi-
tional sadness-specific coping motives measures precludes a com-
parison of its psychometric properties with those of the current
coping-depression scale. The psychometric properties of Gallo-
way’s measure of daily sadness-specific coping motives have not
been tested either, and are not necessarily better than those of his
measure of dispositional sadness-specific coping motives. How-
ever, tests of daily (vs. dispositional) motive moderation of affect-
drinking relations may be relatively less stringent because daily
motives are measured concurrently with daily alcohol use, whereas
tests of dispositional motive moderation of affect-drinking associ-
ations are prospective.
In addition to the effects discussed earlier, there were several
unhypothesized moderation effects. First, conformity motives pos-
itively moderated the daily anxious mood-drinking association.
This finding is perhaps not surprising in light of some evidence
that measures of social anxiety are positively correlated with
conformity motives for drinking in undergraduates (Stewart, Mor-
ris, Mellings, & Komar, 2006). On the other hand, conformity
motives negatively moderated the daily depressed mood-drinking
association. Taken with Mohr and colleagues’ (2005) finding that
conformity motives negatively moderated the global negative
mood-drinking at home association, the current findings (that
conformity motives positively moderated the anxious mood-
drinking association and negatively moderated the depressed
mood-drinking relation) highlight our earlier conclusion that it is
crucial to differentiate between subtypes of negative affect in
examining daily mood-drinking relationships.
Unexpectedly, social motives positively moderated the daily
depressed mood-drinking relation. Thus, it seems that depressed
mood is more likely to precede drinking among individuals en-
dorsing high levels of social motives than among those endorsing
low levels of social motives. It is possible that information about
drinking context (i.e., solitary vs. with others; see Cooper et al.,
1992) might help us to understand why social motives and coping-
depression motives showed surprisingly similar cross-level inter-
actions with daily depressed mood in the prediction of daily
drinking. Perhaps among those with higher (vs. lower) social
motives, depressed mood is more likely to trigger drinking in
social contexts, with the aim of allowing for enjoyment of good
times with friends. On the other hand, the depressed mood-
triggered drinking that is more likely among those with higher (vs.
lower) coping-depression motives might be more likely to occur in
solitary contexts, with the aim of eliminating the depressed affect.
Alternatively, it may be that, as Mohr and colleagues (2005)
suggested, endorsement of social motives represents a socially
desirable justification for drinking behavior instead of representing
one’s true reasons for drinking. In contrast to social motives’ effect
on the daily depressed mood-drinking association, social motives
negatively moderated the daily anxious mood-drinking associa-
tion, once more underscoring the importance of differentiating
between subtypes of negative affect and between subtypes of
coping motives. The simple slopes analyses revealed that individ-
uals with high social motives showed the normative pattern (seen
in the level-1 model) of inhibiting drinking when anxiety levels
increased, which may explain why social motives are “safer” in
terms of relations with heavy drinking and alcohol-related prob-
lems (Cooper, 1994). It is unclear, though, why there was a
difference in the direction of social motives’ moderating influence
on depressed mood-drinking versus anxious mood-drinking rela-
tions. It may be that low social motives are a proxy for high social
anxiety, which might explain why drinking is not inhibited by
anxious mood at the level of low social motives, similar to the
pattern seen at relatively high levels of coping-anxiety and con-
formity motives. Finally, enhancement motives negatively moder-
ated depressed-mood drinking slopes. If depressed mood is as-
sumed to be negatively correlated with positive mood, this finding
is perhaps not surprising given Mohr and colleagues’ (2005)
finding that enhancement motives positively moderated positive
affect-alcohol consumption (at home) slopes.
Strengths, Limitations, and Clinical Implications
The main strength of this research is that it is more ecologically
valid than laboratory investigations which have examined the
impact of drinking motives on the mood triggers of alcohol con-
sumption (e.g., Birch et al., 2006) and that it employs daily process
methodology to evaluate the inherently process-oriented research
235
MOTIVES PREDICT DAILY MOOD–ALCOHOL RELATIONSHIPS
questions posed by motivational models of alcohol use (Tennen &
Affleck, 1996). The current study does, however, have several
limitations. First, for a variety of reasons (i.e., participant noncom-
pliance, technical problems, outliers) only 50% of possible daily
reports were included in the analyses. Our sample, though it was
reflective of the sex composition of undergraduates taking psy-
chology courses, did not permit a comprehensive test of sex
differences given the relatively small number of male participants.
Also, our sample was predominantly Caucasian, which is consis-
tent with previous studies at this university, but which may limit
the generalizability of the results. There is some evidence to
suggest that coping motives are more predictive of drinking be-
havior and/or consequences for females (vs. males, e.g., Birch et
al., 2006) and for Black adolescents (vs. White adolescents, e.g.,
Cooper et al., 2008). Thus, future studies should endeavor to
recruit a higher proportion of males and individuals from racially
diverse populations. Furthermore, it would be interesting to mea-
sure daily motives for alcohol use (as per Galloway, 2007) to
determine whether the current conclusions would hold. Nonethe-
less, there are good reasons for studying dispositional drinking
motives. First, drinking motives have been shown to be relatively
stable across time (Grant, Stewart, O’Connor, et al., 2007) and
second, the clinical implications are much stronger if dispositional
motives are shown to prospectively predict drinking outcomes (vs.
being concurrently associated with drinking behavior, like in stud-
ies of daily alcohol use motives).
The present findings, especially those pertaining to the coping-
depression and coping-anxiety motives for drinking that have been
found to be particularly risky (Grant, Stewart, O’Connor, et al.,
2007), can inform clinical intervention approaches. According to
the results of a recent meta-analytic review of individual-level
interventions designed to reduce drinking among college students,
interventions providing feedback on alcohol expectancies and/or
drinking motives were more successful at diminishing alcohol-
related problems than those that did not (Carey, Scott-Sheldon,
Carey, & DeMartini, 2007). It is conceivable that incorporating
personalized feedback about undergraduates’ distinct subtypes of
coping motives for drinking might enhance the success of extant
interventions. As the current study is the first to examine the
differential impact of coping-anxiety and coping-depression mo-
tives on daily negative affect-drinking relations, it is important that
additional research is undertaken before the potential clinical util-
ity of the Modified DMQ-R is evaluated. Nonetheless, the current
findings help to elucidate the intricate associations among coping
drinking motives, negative affect, and alcohol use.
References
Aiken, L. S., & West, S. G. (1991). Multiple regression: Testing and
interpreting interactions. Thousand Oaks, CA: Sage.
Babor, T. F., Brown, J., & Del Boca, F. K. (1990). Validity of self-reports
in applied research on addictive behaviors: Fact or fiction? Behavioral
Assessment, 12, 5–31.
Birch, C. D., Stewart, S. H., Girling, A. N., & Berish, M. J. (2006). The
impact of mood, drinking motives, and gender on laboratory alcohol
consumption. In A. V. Clark (Ed.), Psychology of moods: New research
(pp. 1–36). Hauppauge, NY: Nova Science Publishers.
Blackwell, E., & Conrod, P. J. (2003). A five-dimensional measure of
drinking motives. Unpublished manuscript, Department of Psychology,
University of British Columbia.
Blackwell, E., Conrod, P. J., & Hansen, N. (2002, June). Negative cogni-
tions, hopelessness, and depression related drinking motives. Poster
presented at the Scientific Meeting of the Research Society on Alcohol-
ism, San Francisco, CA. Abstract in Alcoholism: Clinical and Experi-
mental Research, 26(Suppl.), 27A.
Bradizza, C. M., Reifman, A., & Barnes, G. M. (1999). Social and coping
reasons for drinking: Predicting alcohol misuse in adolescents. Journal
of Studies on Alcohol, 60, 491– 499.
Carey, K. B., Scott-Sheldon, L. A. J., Carey, M. P., & DeMartini, K. S.
(2007). Individual-level interventions to reduce college student drinking:
A meta-analytic review. Addictive Behaviors, 32, 2469 –2494.
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., Krull, J. L., Agocha, V. B., Flanagan, M. E., Orcutt, H. K.,
Grabe, S., et al. (2008). Motivational pathways to alcohol use and abuse
among Black and White adolescents. Journal of Abnormal Psychology,
117, 485–501.
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.
Cox, W. M., & Klinger, E. (1988). A motivational model of alcohol use.
Journal of Abnormal Psychology, 97, 168 –180.
Dawson, D. A., Grant, B. F., Stinson, F. S., & Chou, P. S. (2005).
Psychopathology associated with drinking and alcohol use disorders in
the college and general adult populations. Drug and Alcohol Depen-
dence, 77, 139 –150.
Flynn, H. A. (2000). Comparison of cross-sectional and daily reports in
studying the relationship between depression and use of alcohol in
response to stress in college students. Alcoholism: Clinical and Exper-
imental Research, 24, 48 –52.
Galloway, C. A. (2007). Do coping motives moderate daily mood-drinking
covariation: Disentangling a paradox (Doctoral dissertation, University
of North Carolina at Chapel Hill, 2006). Dissertation Abstracts Inter-
national: Section B: The Sciences and Engineering, 67, 6053.
Goldman, M. S., Del Boca, F. K., & 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.
Grant, V. V., & Stewart, S. H. (2007). Impact of experimentally induced
positive and anxious mood on alcohol expectancy strength in internally
motivated drinkers. Cognitive Behaviour Therapy, 36, 102–111.
Grant, V. V., Stewart, S. H., & Birch, C. D. (2007). Impact of positive and
anxious mood on implicit alcohol-related cognitions in internally moti-
vated undergraduate drinkers. Addictive Behaviors, 32, 2226 –2237.
Grant, V. V., Stewart, S. H., O’Connor, R. M., Blackwell, E., & Conrod,
P. J. (2007). Psychometric evaluation of the five-factor Modified Drink-
ing Motives Questionnaire-Revised in undergraduates. Addictive Behav-
iors, 32, 2611–2632.
Hayes, S. C., & Cavior, N. (1977). Multiple tracking and the reactivity of
self-monitoring. Behavior Therapy, 8, 819 – 831.
Hussong, A. M., Galloway, C. A., & Feagans, L. A. (2005). Coping
motives as a moderator of daily mood-drinking covariation. Journal of
Studies on Alcohol, 66, 344 –353.
Kassel, J. D., Jackson, S. I., & Unrod, M. (2000). Generalized expectancies
for negative mood regulation and problem drinking among college
students. Journal of Studies on Alcohol, 61, 332–340.
Knapp, H., & Kirk, S. A. (2003). Using pencil and paper, Internet and
touch-tone phones for self- administered surveys: Does methodology
matter? Computers in Human Behavior, 19, 117–134.
Knight, J. R., Wechsler, H., Kuo, M., Seibring, M., Weitzman, E. R., &
236 GRANT, STEWART, AND MOHR
Schuckit, M. A. (2002). Alcohol abuse and dependence among U.S.
college students. Journal of Studies on Alcohol, 63, 263–270.
Mohr, C. D., Armeli, S., Tennen, H., Temple, M., Todd, M., Clark, J., et
al. (2005). Moving beyond the keg party: A daily process study of
college student drinking motivations. Psychology of Addictive Behav-
iors, 19, 392– 403.
Morris, E. P., Stewart, S. H., & Ham, L. S. (2005). The relationship
between social anxiety disorder and alcohol use disorders: A critical
review. Clinical Psychology Review, 25, 734 –760.
Park, C. L., Armeli, S., & Tennen, H. (2004). The daily stress and coping
process and alcohol use among college students. Journal of Studies on
Alcohol, 65, 126 –135.
Park, C. L., & Levenson, M. R. (2002). Drinking to cope among college
students: Prevalence, problems and coping processes. Journal of Studies
on Alcohol, 63, 486 – 497.
Perkins, H. W. (1999). Stress-motivated drinking in collegiate and post-
collegiate young adulthood: Life course and gender patterns. Journal of
Studies on Alcohol, 60, 219 –227.
Perrine, M. W., Mundt, J. C., Searles, J. S., & Lester, L. S. (1995).
Validation of daily self-reported alcohol consumption using interactive
voice response (IVR) technology. Journal of Studies on Alcohol, 56,
487– 490.
Raudenbush, S. W., & Bryk, A. S. (2002). Hierarchical linear models:
Applications and data analysis methods (2nd ed.). Thousand Oaks, CA:
Sage.
Raudenbush, S. W., Bryk, A. S., Cheong, Y. F., Congdon, R. T., & du Toit,
M. (2004). HLM 6: Hierarchical linear and nonlinear modeling. Lin-
colnwood, IL: Scientific Software International.
Raudenbush, S. W., Bryk, A. S., & Congdon, R. T. (2005). HLM 6:
Hierarchical linear and nonlinear modeling (Version 6.06) [Computer
software]. Lincolnwood, IL: Scientific Software International.
Read, J. P., Wood, M. D., Kahler, C. W., Maddock, J. E., & Palfai, T. P.
(2003). Examining the role of drinking motives in college student
alcohol use and problems. Psychology of Addictive Behaviors, 17, 13–
23.
Shacham, S. (1983). A shortened version of the Profile of Mood States.
Journal of Personality Assessment, 47, 305–306.
Snijders, T. A. B., & Bosker, R. J. (1999). Multilevel analysis: An intro-
duction to basic and advanced multilevel modeling. Thousand Oaks,
CA: Sage.
Stewart, S. H., Morris, E., Mellings, T., & Komar, J. (2006). Relations of
social anxiety variables to drinking motives, drinking quantity and
frequency, and alcohol-related problems in undergraduates. Journal of
Mental Health, 15, 671– 682.
Tabachnick, B. G., & Fidell, L. S. (2001). Using multivariate statistics (4th
ed.). Needham Heights, MA: Allyn & Bacon.
Tennen, H., & Affleck, G. (1996). Daily processes in coping with chronic
pain: Methods and analytic strategies. In M. Zeidner & N. S. Endler
(Eds.), Handbook of coping: Theory, research, applications (pp. 151–
177). New York: John Wiley & Sons.
The National Center on Addiction and Substance Abuse (CASA) at
Columbia University. (2003). Depression, substance abuse and col-
lege student engagement: A review of the research literature. New
York: The National Center on Addiction and Substance Abuse
(CASA) at Columbia University. Retrieved June 25, 2008, from
http://www.bringingtheorytopractice.org/pdfs/LitReviewDec03.pdf
Thombs, D. L., & Beck, K. H. (1994). The social context of four adolescent
drinking patterns. Health Education Research, 9, 13–22.
Todd, M., Armeli, S., Tennen, H., Carney, M. A., & Affleck, G. (2003). A
daily diary validity test of drinking to cope measures. Psychology of
Addictive Behaviors, 17, 303–311.
Todd, M., Armeli, S., Tennen, H., Carney, M. A., Ball, S. A., Kranzler,
H. R., et al. (2005). Drinking to cope: A comparison of questionnaire and
electronic diary reports. Journal of Studies on Alcohol, 66, 121–129.
Watson, D., & Clark, L. A. (1991). The Mood and Anxiety Symptom
Questionnaire. Unpublished manuscript, University of Iowa, Depart-
ment of Psychology, Iowa City.
Watson, D., Clark, L. A., Weber, K., Assenheimer, J. S., Strauss, M. E., &
McCormick, R. A. (1995). Testing a tripartite model: II. Exploring the
symptom structure of anxiety and depression in student, adult, and
patient samples. Journal of Abnormal Psychology, 104, 15–25.
Watson, D., Weber, K., Assenheimer, J. S., Clark, L. A., Strauss, M. E., &
McCormick, R. A. (1995). Testing a tripartite model: I. Evaluating the
convergent and discriminant validity of anxiety and depression symptom
scales. Journal of Abnormal Psychology, 104, 3–14.
Wechsler, H., Lee, J. E., Kuo, M., Seibring, M., Nelson, T. F., & Lee, H.
(2002). Trends in college binge drinking during a period of increased
prevention efforts. Journal of American College Health, 50, 203–217.
White, H. R., & Labouvie, E. W. (1989). Towards the assessment of
adolescent problem drinking. Journal of Studies on Alcohol, 50, 30 –37.
Received June 29, 2008
Revision received December 9, 2008
Accepted December 9, 2008 䡲
237
MOTIVES PREDICT DAILY MOOD–ALCOHOL RELATIONSHIPS
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.