ArticlePDF Available

Dyadic Conflict, Drinking to Cope, and Alcohol-Related Problems: A Psychometric Study and Longitudinal Actor-Partner Interdependence Model

Authors:

Abstract and Figures

The motivational model of alcohol use posits that individuals may consume alcohol to cope with negative affect. Conflict with others is a strong predictor of coping motives, which in turn predict alcohol-related problems. Two studies examined links between conflict, coping motives, and alcohol-related problems in emerging adult romantic dyads. It was hypothesized that the association between conflict and alcohol-related problems would be mediated by coping-depression and coping-anxiety motives. It was also hypothesized that this would be true for actor (i.e., how individual factors influence individual behaviors) and partner effects (i.e., how partner factors influence individual behaviors) and at the between- (i.e., does not vary over the study period) and within-subjects (i.e., varies over the study period) levels. Both studies examined participants currently in a romantic relationship who consumed ≥12 alcoholic drinks in the past year. Study 1 was cross-sectional using university students (N = 130 students; 86.9% female; M = 21.02 years old, SD = 3.43). Study 2 used a 4-wave, 4-week longitudinal design with romantic dyads (N = 100 dyads; 89% heterosexual; M = 22.13 years old, SD = 5.67). In Study 2, coping-depression motives emerged as the strongest mediator of the conflict-alcohol-related problems association, and findings held for actor effects but not partner effects. Supplemental analyses revealed that this mediational pathway only held among women. Within any given week, alcohol-related problems changed systematically in the same direction between romantic partners. Interventions may wish to target coping-depression drinking motives within couples in response to conflict to reduce alcohol-related problems. (PsycINFO Database Record (c) 2015 APA, all rights reserved).
Content may be subject to copyright.
Journal of Family Psychology
Dyadic Conflict, Drinking to Cope, and Alcohol-Related
Problems: A Psychometric Study and Longitudinal
Actor–Partner Interdependence Model
Laura Lambe, Sean P. Mackinnon, and Sherry H. Stewart
Online First Publication, June 15, 2015. http://dx.doi.org/10.1037/fam0000098
CITATION
Lambe, L., Mackinnon, S. P., & Stewart, S. H. (2015, June 15). Dyadic Conflict, Drinking to
Cope, and Alcohol-Related Problems: A Psychometric Study and Longitudinal Actor–Partner
Interdependence Model. Journal of Family Psychology. Advance online publication.
http://dx.doi.org/10.1037/fam0000098
Dyadic Conflict, Drinking to Cope, and Alcohol-Related Problems:
A Psychometric Study and Longitudinal Actor–Partner
Interdependence Model
Laura Lambe, Sean P. Mackinnon, and Sherry H. Stewart
Dalhousie University
The motivational model of alcohol use posits that individuals may consume alcohol to cope with negative
affect. Conflict with others is a strong predictor of coping motives, which in turn predict alcohol-related
problems. Two studies examined links between conflict, coping motives, and alcohol-related problems
in emerging adult romantic dyads. It was hypothesized that the association between conflict and
alcohol-related problems would be mediated by coping-depression and coping-anxiety motives. It was
also hypothesized that this would be true for actor (i.e., how individual factors influence individual
behaviors) and partner effects (i.e., how partner factors influence individual behaviors) and at the
between- (i.e., does not vary over the study period) and within-subjects (i.e., varies over the study period)
levels. Both studies examined participants currently in a romantic relationship who consumed 12
alcoholic drinks in the past year. Study 1 was cross-sectional using university students (N130 students;
86.9% female; M21.02 years old, SD 3.43). Study 2 used a 4-wave, 4-week longitudinal design with
romantic dyads (N100 dyads; 89% heterosexual; M22.13 years old, SD 5.67). In Study 2,
coping-depression motives emerged as the strongest mediator of the conflict-alcohol-related problems
association, and findings held for actor effects but not partner effects. Supplemental analyses revealed
that this mediational pathway only held among women. Within any given week, alcohol-related problems
changed systematically in the same direction between romantic partners. Interventions may wish to target
coping-depression drinking motives within couples in response to conflict to reduce alcohol-related
problems.
Keywords: conflict, coping motives, alcohol-related problems, romantic, dyads
Supplemental materials: http://dx.doi.org/10.1037/fam0000098.supp
Emerging adulthood (approximately 18 to 25 years old) is the
period in which romantic relationships become more intimate and
serious (Arnett, 2000). In fact, emerging adults rate romantic
relationships high on their list of priorities, with only education
and achieving personal life goals ranked as more important (Ham-
mersla & Frease-McMahan, 1990). However, conflict (i.e., behav-
iors such as interruptions, disagreements, antagonism, negative
expressiveness, tension) is often an inherent part of these relation-
ships, and conflict is associated with reduced relationship satisfac-
tion, anxiety, depressive symptoms, and psychological distress
(Fincham & Beach, 1999). Emerging adulthood also constitutes
the developmental period where the prevalence of alcohol use is
the highest (Arnett, 2005). Indeed, 32.0% of Canadian college
students report hazardous levels of drinking (AUDIT scores 8),
with 43.9% reporting at least one alcohol-related problem (i.e.,
negative consequences associated with alcohol use such as mem-
ory loss or injuries; Adlaf, Demers, & Gliksman, 2004). Conflict
and alcohol-related problems tend to co-occur in marital and
dating relationships (Marshal, 2003; Shorey, Stuart, & Cornelius,
2011). We hypothesized one mechanism to explain this link:
Members of a couple may use alcohol to cope with negative affect
(Cooper, 1994) following conflict (Carrigan, Samoluk, & Stewart,
1998), which may in turn lead to alcohol-related problems.
Theory Review
Drinking Motives Theory
The motivational model of alcohol use (Cooper, 1994) postulates
two dimensions that explain why people drink: (a) Positive versus
negative reinforcement and (b) internal versus external motivation.
Laura Lambe, Department of Psychology and Neuroscience, Dalhousie
University; Sean P. Mackinnon and Sherry H. Stewart, Department of
Psychiatry and Department of Psychology and Neuroscience, Dalhousie
University.
Laura Lambe is now at the Department of Psychology at Queens
University.
This study was funded by the Dalhousie Psychiatry Research Fund. Sean
P. Mackinnon was supported by a postdoctoral fellowship from the Social
Sciences and Humanities Research Council. Laura Lambe was supported
by a Dalhousie Faculty of Science summer studentship.
We thank Marla Korecky, Matthew Orr, Dana Song, Madeline Park,
Lois Cartwright, Kevin Cox, Jennifer Swansburg, and Pam Collins for their
research assistance.
Correspondence concerning this article should be addressed to Sean P.
Mackinnon, Department of Psychology and Neuroscience, Dalhousie Uni-
versity, Life Sciences Centre, 1355 Oxford Street, PO Box 15000, Halifax,
Nova Scotia, Canada, B3H 4R2. E-mail: mackinnon.sean@gmail.com
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
Journal of Family Psychology © 2015 American Psychological Association
2015, Vol. 29, No. 4, 000 0893-3200/15/$12.00 http://dx.doi.org/10.1037/fam0000098
1
Thus, individuals may drink to obtain a desired outcome (positive
reinforcement) or to avoid a negative outcome (negative reinforce-
ment) and also for internal reasons (e.g., manipulating mood) or
external reasons (e.g., social approval). Crossing these dimensions
results in four distinct drinking motives: social (positive, external),
enhancement (positive, internal), conformity (negative, external), and
coping (negative, internal). Social motives involve drinking to affiliate
or celebrate with friends; in contrast, enhancement motives involve
drinking to increase positive moods, such as excitement. Conformity
motives, common in adolescents, involve drinking to fit in with peers.
Last, coping motives involve drinking to alleviate negative affect,
such as anxiety or depression (Cooper, 1994).
Each motive has distinct alcohol outcomes. Internal motiva-
tions (enhancement and coping motives) are significantly asso-
ciated with drinking quantity and frequency and alcohol-related
problems. Enhancement motives are predictive of alcohol-
related problems through increased consumption, whereas cop-
ing motives are directly related to alcohol-related problems
even after accounting for consumption. In contrast, social mo-
tives are related to drinking quantity and frequency but not to
alcohol-related problems. Last, conformity motives are nega-
tively associated with drinking quantity and frequency but are
positively associated with alcohol-related problems (Cooper,
Frone, Russell, & Mudar, 1995). Recently, coping motives have
been further subdivided into coping with anxiety motives
(CAM) and coping with depression motives (CDM), which both
uniquely predict alcohol-related problems (Grant, Stewart,
O’Connor, Blackwell, & Conrod, 2007). Drinking in response
to relationship conflict can be conceptualized as primarily a
negatively reinforcing, internal behavior (i.e., occurring in re-
sponse to an aversive emotional stimulus; Cooper et al., 1995).
Thus, this study focused largely on coping motives as an
outcome of conflict and as a predictor of alcohol-related prob-
lems.
Drinking Partnership Theory
Alcohol use by one partner often influences alcohol use by
the other (i.e., social influence) and congruence of drinking
patterns between partners is linked with relationship satisfac-
tion (Homish & Leonard, 2007). The similarity of alcohol use
between romantic partners may be related to both selection and
socialization (or social influence) processes. Selection occurs
through the influence of individual characteristics, which lead
an individual to seek out particular people. During emerging
adulthood, heavy episodic drinkers (five or more drinks in a
2-hr period) are more likely to create friendships with other
heavy episodic drinkers (McCabe et al., 2005). Conversely,
socialization refers to the influence of others on the individual.
In terms of alcohol use, compared with nonmembers, Greek
members show increased rates of heavy episodic drinking over
time, suggesting a socialization process (McCabe et al., 2005).
Social influence effects for alcohol are also found in married
(Leonard & Mudar, 2004) and dating (Mushquash et al., 2013)
couples. Thus, individuals are likely to initially seek out others
with similar patterns of alcohol use (selection) and may then
further escalate their drinking over time because of social
influence (socialization).
Previous Research
Conflict and Alcohol Use
Interpersonal conflict represents negative behavior occurring
between two individuals that may be hostile, critical, rejecting, or
inconsiderate. There are many different operationalizations and
names for this construct (e.g., social negativity, negative social
exchanges; Brooks & Schetter, 2011); we collectively operation-
alized this construct as conflict. Conflict is positively correlated
with alcohol-related problems among heavier drinkers (i.e., those
with a higher quantity–frequency of drinking; Carey, 1995), and it
represents a significant stressor that can lead to consumption
(Marlatt, 1996). In a daily diary study of emerging adult romantic
dyads, Levitt and Cooper (2010) found that conflict significantly
predicted alcohol use the following day (suggesting a directional
relationship). Because drinking motives are theorized to be the
most proximal predictor of alcohol outcomes, through which more
distal risk factors exert their influence (Cooper et al., 1995), we
propose that conflict leads to increased alcohol-related problems
indirectly through drinking to cope.
Conflict and Drinking to Cope
People commonly report drinking to cope with negative inter-
personal situations. Indeed, Carrigan et al. (1998) found that drink-
ing in situations of conflict significantly predicted over 36% of the
variance in coping motives scores among emerging adults. Al-
though there is little empirical research that specifically investi-
gates drinking to cope in response to romantic relationship conflict
among emerging adults, the cross-sectional association between
jealousy and alcohol-related problems is partially mediated by
coping motives in this population (DiBello, Neighbors, Rodriguez,
& Lindgren, 2014). Jealousy is a related construct as it can pro-
mote negative actions within a romantic relationship and is linked
with alcohol use (Foran & O’Leary, 2008). In general, these results
support the notion that individuals may drink to cope with roman-
tic relationship conflict.
Drinking to Cope and Alcohol-Related Problems
A direct link between CAM, CDM, and alcohol-related prob-
lems has been found both cross-sectionally (e.g., Cooper, 1994;
Grant et al., 2007) and longitudinally (Grant et al., 2007; Mack-
innon, Kehayes, Clark, Sherry, & Stewart, 2014). Overall, these
studies suggest that coping motives broadly defined, as well as
CDM and CAM more specifically, are significant predictors of
alcohol-related problems over time.
Partner Effects
In dyadic research, partner effects refer to the partner’s influ-
ence on the individual. In contrast, actor effects refer to how an
individual’s characteristics influence his or her own behavior
(Kenny & Cook, 1999). Partner effects are a significant contributor
to alcohol use within romantic dyads. Cross-sectionally, alcohol-
related problems are significantly related between emerging adult
romantic partners (Rodriguez, Knee, & Neighbors, 2014). Also
among emerging adult romantic dyads, Mushquash et al. (2013)
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
2LAMBE, MACKINNON, AND STEWART
found that a partner’s heavy episodic drinking predicted future
individual heavy episodic drinking 30 days later, even after ac-
counting for actor effects. Although research has yet to examine
partner effects of drinking motives within romantic dyads, adoles-
cent research indicates that peer coping motives significantly pre-
dict individual coping motives, which in turn predict individual
drinking behavior (Kuntsche & Stewart, 2009). Overall, this sug-
gests that both an individual’s drinking motives and alcohol use
may be influenced by the drinking motives and alcohol use of his
or her romantic partner.
Current Study
Given the above-reviewed associations, it is likely that drinking
to cope mediates the link between conflict and alcohol-related
problems in romantic relationships. In particular, little is known
about conflict and alcohol use among emerging adult couples, as
the majority of research in this area focuses only on older married
couples (e.g., Marshal, 2003), and to our knowledge, research
examining links between all three constructs in a single model has
not been conducted to date. Furthermore, as romantic partners
influence each other’s alcohol use (Mushquash et al., 2013), part-
ner influence effects must be considered when examining drinking
motives and alcohol-related problems within romantic relation-
ships. Thus, the current set of studies investigated the mediational
role of coping motives in the conflict to alcohol-related problems
relation, while also taking into account partner influence for alco-
hol use.
Study 1 tested the psychometric properties of the instruments to
be used, whereas Study 2 examined these associations in a four-
wave, 4-week longitudinal design with couples. This design re-
duced temporal confounding and recall bias, as participants were
asked to report on events closer to their actual occurrence. As there
is little research regarding romantic relationship conflict and drink-
ing motives, these studies fill a significant gap by integrating
drinking motives theory and theories of partner influence and by
using improved methods.
Study 1
In Study 1, we first validated the psychometric properties of a
variety of modified weekly measures. This was necessary as no
research to date has examined weekly drinking motives or alcohol-
related problems—a timeframe in which drinking to cope with
relationship conflict might reasonably occur. Hypotheses for Study
1 were as follows:
Hypothesis 1: Modified weekly measures of conflict, drinking
motives, and alcohol-related problems would demonstrate ac-
ceptable internal consistency (s.70).
That is, we predicted that all of our measures would be reliable.
Hypothesis 2. Modified weekly measures of conflict, drinking
motives, and alcohol-related problems would be positively
correlated (rs around .50) with the original measures, support-
ing criterion validity. That is, we predicted that all of our
modified measures would be similar to the original measures
on which they were based.
Method
Participants
Participants were 130 undergraduates (86.9% female, 79.9%
Caucasian) with an average age of 21.02 years (SD 3.43).
Participants were recruited using the university’s psychology par-
ticipant pool, flyers at local universities, and online ads. Partici-
pants heard about the study through the psychology subject pool
(58.5%), website ads (20.0%), e-mail ads (6.9%), flyers (3.1%),
class announcements (2.3%), or multiple sources (7.7%). To meet
eligibility requirements, participants were required to be currently
in a romantic relationship (of any duration–type), and must have
consumed 12 alcoholic beverages in the past year. All partici-
pants met these criteria when screened on arriving at the lab. The
second criterion excluded lifetime abstainers and former–
infrequent drinkers (Dawson, 2003). We focused on drinkers be-
cause participants would not have been able to respond to ques-
tions on drinking motives or alcohol-related problems if they did
not drink alcohol.
The majority of the sample was dating one person (83.7%),
whereas others were cohabiting (14.7%) or married (1.6%). On
average, participants rated their romantic relationships as “serious”
(M3.22, SD 0.76, on a 4-point scale) and had face-to-face
contact with their romantic partner an average of 4.33 days/week
(SD 2.52). In the past week, participants also reported consum-
ing an average quantity of 6.93 standard drinks per week (SD
7.84, range 0 –53 drinks/week).
1
A standard drink was defined as
one bottle of beer (12 oz, 5%), one glass of wine (5 oz, 10%–12%),
or one shot of hard liquor (1.5 oz, 43%–50%; Collins, Kashdan,
Koutsky, Morsheimer, & Vetter, 2008). Additionally, 33.1% of
participants reported clinically significant alcohol-related prob-
lems (3-year Rutgers Alcohol Problem Index [RAPI] summed raw
scores 15; Thombs & Beck, 1994).
Materials
Conflict. Relationship conflict was analyzed as a composite
variable consisting of the Social Conflict Scale (Abbey & An-
drews, 1985), the Partner-Specific Rejecting Behaviors Scale
(Murray, Griffin, Rose, & Bellavia, 2003), and the Interpersonal
Qualities Scale (Oishi & Sullivan, 2006). All scales measured
relationship conflict in the past week. The Social Conflict Scale
consisted of five items (e.g., “Argued with your partner about
something”) rated ona1(not at all)to5(a great deal) scale. The
Partner-Specific Rejecting Behaviors Scale consisted of seven
questions (e.g., “I snapped or yelled at my partner”) rated on a 1
(strongly disagree)to9(strongly agree) scale. The Interpersonal
Qualities Scale consisted of five questions pertaining to the par-
ticipant’s interpersonal characteristics when they were with their
romantic partner (e.g., “Critical/judgmental”) rated ona1(not at
all characteristic)to9(completely characteristic) scale. Averaged
subscale totals were used for descriptive statistics and correlations.
For hypothesis tests, factor scores were created for each item,
1
One outlier value of 53 drinks occurred for one participant who
reported on the week of St. Patrick’s Day. The next highest value was 38
drinks. As number of drinks was not analyzed in results below, this outlier
was retained.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
3
CONFLICT, DRINKING TO COPE, ALCOHOL-RELATED ISSUES
which were then summed to create a single composite conflict
index (DiStefano, Zhu, & Mîndrilaˇ 2009). These subscales have
been shown to possess excellent psychometric properties in pre-
vious work (Abbey & Andrews, 1985; Mackinnon et al., 2012).
Because the measures were altered from their original past week
timeframe and because no prior research has combined these three
measures together into a single composite, in Study 1 we tested
whether this composite measure of conflict had good factorial
validity, criterion validity, and internal consistency.
Modified Drinking Motives Questionnaire—Revised. The
Modified Drinking Motives Questionnaire—Revised (Modified
DMQ-R; Grant et al., 2007), a modified version of Cooper’s
(1994) four-factor measure was used to measure motives. The
Modified DMQ-R is a reliable and valid 28-item measure of five
drinking motives: five-item enhancement (“because it is excit-
ing”), five-item social (“to be sociable”), five-item conformity (“so
I will not feel left out”), nine-item CDM (“to numb my pain”), and
four-item CAM (“to reduce my anxiety”). Participants were asked
to indicate how often they drank for each reason ona1(almost
never/never)to5(almost always) scale. Two versions of this
questionnaire were fielded: The original version, which asks par-
ticipants to consider their drinking motives over the past 3 years,
and a newly developed version which asked participants to con-
sider the past 7 days. Items were averaged for each subscale to
create a total score for analyses. Prior research has shown that the
past 3 years version of the Modified DMQ-R has excellent test–
retest reliability, internal consistency, and factorial validity (Grant
et al., 2007). The psychometric properties of the 7-day version
were tested in Study 1.
Rutgers Alcohol Problem Index. The RAPI (White &
Labouvie, 1989) was used to measure alcohol-related problems
(i.e., negative consequences from alcohol use). The RAPI is a
23-item measure asking about the occurrence of specific alcohol-
related problems (e.g., “not able to do your homework or study for
a test”) with each question scored from 0 (never)to4(4 or more
times). The RAPI can be used over various timeframes (e.g., past
year, past month, etc.) while maintaining favorable psychometric
properties, with test–retest reliabilities ranging from r.83 (one
month) to r.88 (one year; Miller et al., 2002) and good alpha
reliability (␣⫽.92; Grant et al., 2007). Both 3-year and 7-day
versions of the RAPI were administered in the present study. They
were both scored by dichotomizing each item (i.e., 1 presence
and 0 absence of the problem measured by that item) and then
summing the items into a single value for analysis (Martens,
Neighbors, Dams-O’Connor, Lee, & Larimer, 2007).
Procedure
Eligible participants completed the above pen-and-paper ques-
tionnaires in the lab as part of a larger questionnaire battery.
Participants were compensated with 1.0 bonus credits or $5.
Results
Missing Data and Normality
Overall, 19 of the 130 participants abstained from alcohol over
the past 7 days. As these participants could not have drinking
motives or alcohol-related problems within the past week time-
frame, they were excluded. Item-level missing data was minimal
(0.01%) and was imputed using an expectation-maximization
method in SPSS. When missing data are minimal (5%), a single
imputation using an expectation-maximization approach is rela-
tively unbiased and improves statistical power (Scheffer, 2002).
Data were screened for univariate normality and outliers using
kurtosis–skewness values, histograms, normal P-P plots, and box
plots. On the basis of these diagnostics, the past 7-days and past
3-years version of CDM, CAM, conformity motives, and the 7-day
RAPI were log
10
-transformed to reduce the impact of nonnormal-
ity and outliers.
Descriptive Statistics and Bivariate Correlations
Descriptive statistics appear in Table 1. The past 3 years drink-
ing motive scores were comparable to past samples of undergrad-
uate students (i.e., within 1 standard SD of the mean; Grant et al.,
2007), as were scores for the RAPI using the dichotomous scoring
(Martens et al., 2007),
2
and scores on the Interpersonal Qualities
scale and the Partner-Specific Rejecting Behaviors scale (Mack-
innon et al., 2012). All measures exceeded acceptable levels of
internal consistency (all s were .72; see Table 1).
Bivariate correlations revealed that all five modified weekly
drinking motives were strongly positively correlated with their
original past 3-years versions (rs from .69 to .77), demonstrating
criterion validity. Moreover, the 7-day version of the RAPI was
strongly correlated with the 3-year version of the RAPI (r.50).
As expected, all five 3-year drinking motives were significantly
and positively correlated with alcohol-related problems over the
past 3 years (rs from .21 to .43). When looking at 7-day measures,
all drinking motives remained significantly and positively corre-
lated with alcohol-related problems except for CAM. In general,
the conflict measures were positively correlated with alcohol-
related problems, CDM, and CAM; the other drinking motives
were unrelated to conflict.
Principal-Components and Factor Analyses
Principal-components analysis (PCA) of the 17 conflict items,
using parallel analysis to determine the number of factors
(O’connor, 2000), was used to assess the unidimensionality of the
conflict measures. Comparison of obtained eigenvalues with 95th
percentile eigenvalues indicated that the conflict measures were
best represented by a one-component solution, which explained
44.9% of the item variance. Factor loadings ranged from .47 to .81,
suggesting it was appropriate to combine these measures into a
single factor.
The unidimensionality of each of the five subscales of the 7-day
version of the DMQ-R was also tested before hypothesis tests
using a similar procedure. The items for each of the five subscales
were tested in five separate PCAs. Each of the five subscales had
a unidimensional structure when assessed with parallel analysis,
with percentage of variance accounted for by the factor as follows:
CDM (60.0%), CAM (60.2%), enhancement (56.0%), social
(58.8%), and conformity motives (65.4%). Factor loadings ranged
from .43 to .92 across subscales.
2
Summed scale totals using raw (rather than dichotomized) scores for
the 3-year Rutgers Alcohol Problem Index were also comparable to past
samples of young drinkers (M12.11, SD 10.45; Grant et al., 2007).
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
4LAMBE, MACKINNON, AND STEWART
Because component items for the 7-day version of the RAPI
were dichotomously scored (see Martens et al., 2007), an alterna-
tive to PCA was needed to assess the unidimensionality of the
RAPI. We used confirmatory factor analysis with all items
3
load-
ing on a single factor, categorical indicators, and a weighted least
squares (WLSMV) estimator in Mplus 7.11 (Yu, 2002). This
analysis suggested that a one-factor model was a reasonable fit for
the data,
2
(N130) 216.08, p.09;
2
/df 1.14; compar-
ative fit index (CFI) 0.91; Tucker–Lewis index (TLI) 0.90;
root-mean-square error of approximation (RMSEA) .03 (90%
CI [.00, .05]), weighted root mean square residual (WRMR)
4
0.87, with standardized factor loadings ranging from .48 to .92.
Discussion
Using a cross-sectional design, in Study 1 we validated the
psychometric properties of modified weekly measures of conflict,
drinking motives, and alcohol-related problems. As hypothesized,
all measures demonstrated acceptable levels of internal consis-
tency and provided some support for criterion validity by converg-
ing with the original measures. We also found general support for
the unidimensional structure of each measure. Thus, our modified
measures are reliable and valid when administered over a weekly
timeframe. Additionally, conflict was uniquely associated with
both CDM and CAM, demonstrating that only coping motives are
likely to mediate the conflict-alcohol-related problems link. Al-
though CDM were positively associated with alcohol-related prob-
lems in a weekly context, CAM were not. This was unexpected, as
past research looking at CAM over a 3-year period has found that
they are associated with alcohol-related problems (Grant et al.,
2007). Study 1 is limited by its cross-sectional design. Further-
more, correlating the modified measures with their original is a
limited measure of concurrent criterion validity; future studies may
wish to examine the predictive validity of these measures and
concurrent associations with a wider range of criterion variables.
As data was collected from individuals instead of romantic dyads,
Study 1 can neither examine partner effects that may occur for
alcohol use in romantic dyads (e.g., Mushquash et al., 2013) nor
conceptualize conflict as a property of the dyad (e.g., Mackinnon
et al., 2012). Study 2 addresses these limitations directly by col-
lecting data from romantic dyads using a longitudinal design.
Study 2
Study 2 examined both CDM and CAM as potential mediators
of the association between dyadic conflict and alcohol-related
problems using a rigorous four-wave, 4-week longitudinal meth-
odology. Longitudinal research allows examination of change and
stability over time and permits stronger causal inferences than
cross-sectional research. Additionally, past research (e.g., Dibello
et al., 2014) gathered data from only one member of the romantic
dyad. Study 2 gathered data from couples, allowing investigation
of both actor and partner effects (Kenny & Ledermann, 2010).
Partner effects may occur directly (i.e., the partner’s motive im-
pacts the individual’s alcohol use) or indirectly (e.g., partner
motives predict partner alcohol-related problems, which in turn
predict individual alcohol-related problems). Moreover, the dyadic
nature of the data allowed us to operationalize conflict as a dyadic
variable that includes elements of social negativity from both
3
Two items were omitted from the coping with anxiety motives (CFA;
#20 “felt you were going crazy” and #22 “felt physically or physiologically
dependent on alcohol”) because they were each endorsed by only one
person. The lack of variance produced a nonpositive definite matrix when
they were included in the CFA. These items were retained when calculating
the total summed score for the Rutgers Alcohol Problem Index, however,
for consistency with the prior literature using this measure.
4
WRMR values less than 1.0 are generally considered to fit the data well
(Yu, 2002).
Table 1
Descriptive Statistics, Bivariate Correlations, and Alpha Reliabilities for Study 1
123456789101112131415
1. Social conflict
2. Rejecting behaviors .76
ⴱⴱ
3. Interpersonal qualities .62
ⴱⴱ
.76
ⴱⴱ
4. CDM (7 days) .26
ⴱⴱ
.24
.21
5. CDM (3 years) .22
.18
.22
.77
ⴱⴱ
6. CAM (7 days) .26
ⴱⴱ
.13 .10 .51
ⴱⴱ
.54
ⴱⴱ
7. CAM (3 years) .29
ⴱⴱ
.25
.20
.57
ⴱⴱ
.63
ⴱⴱ
.70
ⴱⴱ
8. EM (7 days) .19 .10 .10 .34
ⴱⴱ
.33
ⴱⴱ
.41
ⴱⴱ
.36
ⴱⴱ
9. EM (3 years) .14 .12 .05 .34
ⴱⴱ
.38
ⴱⴱ
.46
ⴱⴱ
.51
ⴱⴱ
.76
ⴱⴱ
10. SM (7 days) .06 .13 .17 .18 .28
ⴱⴱ
.21
ⴱⴱ
.19
.56
ⴱⴱ
.48
ⴱⴱ
11. SM (3 years) .01 .04 .03 .24
.31
ⴱⴱ
.34
ⴱⴱ
.38
ⴱⴱ
.51
ⴱⴱ
.62
ⴱⴱ
.73
ⴱⴱ
12. CM (7 days) .17 .03 .10 .37
ⴱⴱ
.29
ⴱⴱ
.34
ⴱⴱ
.28
ⴱⴱ
.39
ⴱⴱ
.33
ⴱⴱ
.41
ⴱⴱ
.34
ⴱⴱ
13. CM (3 years) .13 .07 .08 .35
ⴱⴱ
.44
ⴱⴱ
.33
ⴱⴱ
.34
ⴱⴱ
.31
ⴱⴱ
.42
ⴱⴱ
.41
ⴱⴱ
.42
ⴱⴱ
.69
ⴱⴱ
14. RAPI (7 days) .12 .15 .18
.26
ⴱⴱ
.26
ⴱⴱ
.07 .04 .35
ⴱⴱ
.24
ⴱⴱ
.48
ⴱⴱ
.34
ⴱⴱ
.35
ⴱⴱ
.23
ⴱⴱ
15. RAPI (3 years) .18
.24
.17 .21
.39
ⴱⴱ
.25
ⴱⴱ
.29
ⴱⴱ
.29
ⴱⴱ
.43
ⴱⴱ
.29
ⴱⴱ
.39
ⴱⴱ
.29
ⴱⴱ
.39
ⴱⴱ
.51
ⴱⴱ
M(SD) 1.90
(0.73)
2.22
(1.25)
2.85
(1.48)
1.26
(0.54)
1.73
(0.79)
2.45
(1.00)
2.98
(1.18)
1.22
(0.56)
1.37
(0.55)
1.82
(0.79)
2.65
(0.83)
3.21
(0.86)
1.33
(0.63)
1.55
(2.54)
7.28
(4.55)
Alpha reliability .84 .84 .75 .91 .72 .80 .82 .86 .89 .79 .76 .76 .91 .88 .88
Note. N 111. Coping with depression motives (CDM), coping with anxiety motives (CAM), Coping motives (CM, both versions), and Rutgers Alcohol
Problem Index (RAPI; 7-day) were log transformed to reduce nonnormality. Ms and SDs are shown for averaged subscale totals for motives and conflict
variables. Items were dichotomized then summed for the presented RAPI means. EM enhancement motives; SM social motives.
p.05.
ⴱⴱ
p.01.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
5
CONFLICT, DRINKING TO COPE, ALCOHOL-RELATED ISSUES
partners (see Mackinnon et al., 2012). Finally, data were analyzed
using multilevel structural equation modeling (Preacher, Zyphur,
& Zhang, 2010) to take advantage of their longitudinal nature. This
approach partitions variance into between-subjects (i.e., the por-
tion of variance that stays constant across 4 weeks) and within-
subjects components (i.e., the portion of variance that changes
from week to week). We proposed the following hypotheses:
Hypothesis 1: Conflict will have an indirect effect on prob-
lematic alcohol use through coping with depression motives—
but not coping with anxiety motives. That is, dyadic conflict
will lead to drinking to cope with depression, which in turn
will lead to alcohol-related problems.
Hypothesis 2: The indirect effect proposed in Hypothesis 1
will be true for both actor and partner effects. That is, coping
with depression motives in each partner will uniquely contrib-
ute to more alcohol-related problems in both partners.
Hypothesis 3: The indirect effect proposed in Hypothesis 1
will be true for both the within-subjects and between-subjects
components. That is, results will hold when considering av-
erage rates of each variable over the entire 4 weeks (between
subjects) and when considering co-occurring changes within
any given week (within subjects).
Method
Participants
Participants were 100 romantic dyads (11 same-sex female
dyads; 89 heterosexual dyads) gathered from Dalhousie University
and the surrounding community. Participants heard about the study
through flyers (31.5%), word of mouth (29.0%), website ads
(15.0%), psychology subject pool (5.5%), class announcements
(3.5%), multiple sources (9.5%), or did not specify (0.5%). Eligi-
bility criteria were the same as Study 1. Participants (55.5%
female, 83.5% Caucasian) were 22.13 years old (SD 5.67) on
average. At Wave 1, the majority (89.0%) reported that their
relationship was “serious” (M3.39, SD 0.69, on a 4-point
scale), 41.5% reported cohabiting (5% married),
5
and had face-to-
face contact with their partner an average of 6.15 days/week (SD
1.37). At Wave 1, participants reported consuming 8.44 drinks in
the past week (SD 10.64, range 0 –105 drinks/week),
6
with
38.8% reporting at least one alcohol-related problem in the past
week (M1.35, SD 2.29, range 0 –19). As well, 40.5% of
participants reported clinically significant alcohol-related prob-
lems (3-year RAPI summed raw scores 15; Thombs & Beck,
1994).
Materials
Conflict was analyzed as a composite variable created from
factor scores in the same manner as Study 1. Consistent with
theory and the findings of Study 1, the CAM and CDM subscales
from the 7-day version of the Modified DMQ-R (Grant et al.,
2007) were used to measure drinking motives. Finally, the 7-day
version of the RAPI (White & Labouvie, 1989) was used to
measure alcohol-related problems. Questionnaires were identical
to those in Study 1.
Procedure
Participants were recruited using flyers, e-mails, online ads, and
the psychology subject pool. Couples always arrived at the lab
together and completed the study at the same times. On arriving at
the lab, dyads were asked two questions (“Are you currently in a
romantic relationship? Have you had at least 12 alcoholic bever-
ages in the past year?”) to ensure they met the eligibility require-
ments. Eligible participants were given time to review the consent
form and gave informed consent. Participants then scheduled
follow-up sessions and completed the questionnaire battery. They
returned to the lab once a week for the following 3 weeks (4 weeks
total). There could only be 7–13 days in between appointments.
7
If
participants were unable to attend one of the follow-ups, data was
counted as missing for that wave. During follow-ups, participants
completed the same questionnaires. Participants were compensated
$5.00 or one credit point for each wave they completed, and they
were debriefed at completion.
Data Analytic Strategy
Data from Study 2 were analyzed using an Actor-Partner Inter-
dependence Model (APIM; Kenny & Ledermann, 2010) in a
multilevel structural equation modeling framework (Preacher et
al., 2010) using Mplus 7.0 software. We examined internal con-
sistency at the between- and within-subjects levels using a multi-
level adaptation of Cronbach’s alpha (Geldhof, Preacher, & Zy-
phur, 2014). The APIM used data from both members of each
couple and incorporated all four waves of data. At the within-
subjects level, significant effects demonstrate that changes in X
co-occur with changes in Y within any given week. The between-
subjects level effects indicate that, when averaged across 4 weeks,
the trait-like component of X is correlated with the trait-like
component of Y. As the sample consisted of several same-sex
romantic dyads, an APIM with indistinguishable dyads was chosen
(Peugh, DiLillo, & Panuzio, 2013), which allowed us to include all
participants.
8
Members of each couple were randomly assigned to
be either Partner 1 or Partner 2, and paths between partners were
constrained to be equal. As this method precludes a test of sex
differences, a supplementary analysis using only the 89 heterosex-
5
A series of one-way analyses of variance revealed that married, co-
habiting, and dating couples did not significantly differ in terms of alcohol-
related problems, F(2, 190) 1.70, p.19; conflict, F(2, 192) 2.56,
p.08; coping with anxiety motives, F(2, 140) 0.12, p.89); or
coping with depression motives, F(2, 141) 0.73, p.49. Thus, analyses
were not broken down based on these demographic features.
6
There was one outlier recruited from the community who reported 105
drinks/week at Wave 1 and subsequently dropped out of the study. This
participant was retained in presented analyses.
7
The schedule for assessments was weekly, but participants were able to
reschedule within an additional 6-day window (a week 6) to maximize
retention rates. If this occurred, the following waves would also be re-
scheduled to occur 1 week from the rescheduled session. This procedure
allowed us to ensure that there would always be a minimum of 7 days in
between appointments (necessary as our measures referred specifically to
the past week). We chose a 6-day window as this was the maximum time
between sessions before participants would be entering the next week of
data collection (i.e., a 7-day window would be the next week).
8
This approach was chosen both to maximize power, and because
ethical guidelines suggest that no individuals should be excluded from
research based on sexual orientation (Canadian Institutes of Health Re-
search, 2010).
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
6LAMBE, MACKINNON, AND STEWART
ual couples was conducted to explore potential sex differences in
the magnitude of paths. For the APIM, a RMSEA around .05, a
standardized root-mean-square residual (SRMR) around .08, and a
CFI and TLI around .95 indicate excellent fit (Kline, 2011).
Significance of indirect effects was calculated using a Monte Carlo
approach (Selig & Preacher, 2008). To account for nonnormality,
we used a robust estimator of fit indices and standard errors (MLR
estimator in Mplus).
Results
On average, couples completed 3.55 (SD 0.90) out of four
waves, with 72.5% completing all four waves. More specifically,
81.5% completed Wave 2, 86.0% completed Wave 3, and 85.0%
completed Wave 4. During the study, two romantic dyads reported
ending the relationship; data these couples provided before break-
ing up was included in the analysis, and the rest was coded as
missing. On average, there were 7.89 days (SD 1.30) between
study appointments, demonstrating good compliance with the re-
quested weekly scheduling. Overall, 12.1% of data were missing,
with covariance coverage ranging from .74 to 1.00. A significant
Little’s MCAR test,
2
(882) 988.22 p.007, revealed the data
were not missing completely at random. Closer examination of
separate variance ttests using revealed that missing data could be
significantly predicted by other variables in the model. Specifi-
cally, people with higher levels of coping-depression and coping-
anxiety motives at Wave 1 were more likely have missing data at
future waves. Thus, data were missing at random (i.e., can be
predicted by variables in the model). Full information maximum
likelihood estimation was used to account for this missing data in
Mplus 7.0, which provides relatively unbiased parameter estimates
when data is missing at random and produces less bias compared
to listwise deletion (Enders & Bandalos, 2001).
Means and standard deviations are shown in Table 2. Drinking
motives scores were comparable to (i.e., within 1 SD of the mean)
those observed in Study 1. Within- and between-subjects correla-
tions, intraclass correlations, and reliabilities appear in Table 3.
Intraclass correlations show the percentage of the variance avail-
able to be explained at the between-subjects level; the majority of
the variance was at the between-subjects level (i.e., trait like
variance) for all variables. At both the between- and within-
subjects levels, all variables were significantly and positively
correlated with one another, with the magnitudes of the correla-
tions being greater at the between-subjects level. All measures
demonstrated good reliability at both the between- and within-
subjects levels, with the one exception being the RAPI at the
within-subjects level. Within-subjects reliabilities may be under-
estimated when cluster sizes are small, and cutoff scores for
reliability are not as well-established in multilevel models (Geld-
hof et al., 2014). Thus, we proceeded with the planned analyses
despite one low reliability value.
Mediation
The hypothesized model fit well,
2
(20) 27.31 p.13;
CFI .97; TLI .94; RMSEA .03; SRMR (within) .08;
SRMR (between) .07. Unstandardized path coefficients and R
2
values are presented in Figure 1 and indirect effects in Table 4.
At the within-subjects level, conflict significantly predicted both
CDM and CAM. A significant actor effect was found for CDM
only, as they significantly predicted alcohol-related problems in
the same individual. Partner effects were not significant for CDM.
Neither actor nor partner effects were observed for CAM. At the
within-subjects level, the correlated error terms between the ac-
tor’s and partner’s alcohol-related problems were significant. Al-
though the correlated error terms between the actor’s and partner’s
CDM (B0.52, p.43) and CAM (B0.10, p.53) were not
significant, CDM and CAM were strongly related within the
same individual (B1.95, p.001). Tests of indirect effects
revealed that CDM significantly mediated the link between
conflict and alcohol-related problems and that the actor’s
alcohol-related problems significantly mediated the link be-
tween the actor’s CDM and the partner’s alcohol-related prob-
lems (an indirect partner effect). No other significant indirect
effects were found.
At the between-subjects level, conflict remained a significant
predictor of both CDM and CAM. However, no significant actor or
partner effects were observed. The correlated error terms between
alcohol-related problems were not significant. Although the cor-
related error terms between CDM (B1.84, p.23) and CAM
(B⫽⫺.08, p.90) remained unrelated, CDM and CAM main-
tained their strong relationship within the same individual (B
Table 2
Means and Standard Deviations for Study 2
Variable
Wave 1 Wave 2 Wave 3 Wave 4
MSDMSDMSDMSD
Social conflict 2.19 0.72 2.02 0.83 1.95 0.86 1.90 0.86
Rejecting behaviors 2.57 1.40 2.28 1.43 2.18 1.35 2.12 1.35
Interpersonal qualities 2.91 1.44 2.59 1.43 2.60 1.56 2.60 1.45
CDM 1.33 0.57 1.22 0.52 1.21 0.56 1.22 0.58
CAM 1.87 0.82 1.69 0.69 1.65 0.66 1.72 0.72
RAPI 1.63 2.51 1.27 1.82 1.32 2.29 1.15 2.42
Note. N 200. Ms and SDs are shown for averaged subscale totals for
motives and conflict variables. Items were dichotomized then summed for
the presented Rutgers Alcohol Problem Index (RAPI) means. CDM
coping with depression motives; CAM coping with anxiety motives.
Table 3
Correlation Matrix, Intraclass Correlations, and Reliability
Analyses at Between- and Within-Subjects Levels for Study 2
Variable 1 2 3 4
1. Conflict .15
.28
ⴱⴱ
.30
ⴱⴱ
2. CDM .58
ⴱⴱⴱ
— .45
ⴱⴱⴱ
.21
ⴱⴱⴱ
3. CAM .53
ⴱⴱⴱ
.73
ⴱⴱⴱ
— .12
4. RAPI .58
ⴱⴱⴱ
.51
ⴱⴱⴱ
.55
ⴱⴱⴱ
ICC .65 .67 .69 .59
Alpha reliability (within-subjects) .88 .98 .94 .64
Alpha reliability (between-subjects) .96 .99 .95 .92
Note. N 200. Between-subjects correlations are below the diagonal;
within-subjects correlations are above the diagonal. CDM coping with
depression motives; CAM coping with anxiety motives; RAPI Rut-
gers Alcohol Problem Index; ICC intraclass correlation.
p.05.
ⴱⴱ
p.01.
ⴱⴱⴱ
p.00.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
7
CONFLICT, DRINKING TO COPE, ALCOHOL-RELATED ISSUES
6.92, p.001). No significant indirect effects were found at the
between-subjects level.
9
Supplementary Analysis
A supplementary analysis was conducted on the heterosexual
couples only, to explore potential sex differences. These analyses
revealed some sex differences. Specifically, at both levels of the
model, dyadic conflict was a significant predictor of CDM and
CAM for women only. Thus, significant mediation was only
observed for women.
10
Discussion
Study 2 aimed to overcome the limitations of Study 1 by using
a longitudinal design with couples. As hypothesized, only CDM
mediated the link between conflict and alcohol-related problems.
This relationship was significant at the within-subjects level:
Within any given week, increases in conflict predicted increases in
CDM, which in turn predicted increases in alcohol-related prob-
lems. This also held at the between-subjects level after accounting
for collinearity (see footnote 8). Although CAM emerged as a
significant correlate of alcohol-related problems, it did not add
much beyond CDM in the prediction of these problems. Thus,
romantic partners are turning to alcohol as a maladaptive coping
mechanism for dealing specifically with depressive affect. Con-
trary to hypotheses, this only occurred at the individual level (i.e.,
no partner effects were found). Although coping motives did not
9
Significant actor effects for coping with depression motives (CDM)
and coping with anxiety motives (CAM) were not observed because of
strong collinearity of the coping motives at the between-subjects level (see
Table 3). When analyzed in separate models, both CDM (B0.19, p
.05) and CAM (B0.39, p.001) predicted alcohol problems.
10
Detailed results are available in the online supplementary materials.
Partner 1
.22
-.09
.02
.26
.02
.22
-.09
.12
.26
.10
Coping-Anxiety
Motives
Coping-Depression
Motives
Coping-Anxiety
Motives
Coping-Depression
Motives
Alcohol-Related
Problems
Alcohol-Related
Problems
.07**
.07**
.03**
.03**
.07
.04
.08
.51
.34
.31
Partner 2
Between-Subjects Model
Partner 1
.02
.004
-.04
.004
-.01
.02
.07*
.01
Coping-Anxiety
Motives
Coping-Depression
Motives
Coping-Anxiety
Motives
Coping-Depression
Motives
Alcohol-Related
Problems
Alcohol-Re lated
Problems
.02**
.02**
.01**
.01**
.01
.01
.01
.40**
Partner 2
Within-Subjects Model
Dyadic
Conflict
Dyadic
Conflict
Figure 1. Multilevel APIM (Study 2). Black lines indicate significant paths, gray dashed lines indicate
nonsignificant paths, and gray solid lines indicate paths that become significant when coping with depression
motives and coping with anxiety motives are entered separately (which accounts for the strong collinearity
caused by the strong between-subjects correlation between motives). Rectangles indicate manifest variables.
Single-headed arrows indicate paths. Double-headed arrows indicate covariances. Coefficients are unstandard-
ized and paths were constrained to equality across both partners. R
2
values are indicated in bold italics in the
upper right hand corner of endogenous variables. Readers should note that R
2
values from multilevel models may
not be directly comparable in interpretation to R
2
values from conventional regression. Moreover, R
2
values may
vary between partners despite equality constraints because each partner group has slightly different variances
(Kline, 2011).
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
8LAMBE, MACKINNON, AND STEWART
vary systematically between romantic partners, partners did show
similar increases in alcohol-related problems within any given
week, creating an indirect partner effect. Thus, even without the
direct influence of a partner’s CDM on the individual, the presence
of these motives led to alcohol-related problems for both partners.
General Discussion
Across both studies, dyadic conflict was a consistent predictor
of both types of coping motives. This is consistent with drinking
motives theory (Cooper et al., 1995) and suggests that emerging
adult romantic dyads may use alcohol to cope with negative affect
following conflict with their partner. This is maladaptive, as it
leads to alcohol-related problems. In particular, this appears to
occur through depressive affect (not anxiety), as couples who
experienced relationship conflict tended to drink to cope with
feelings of depressed affect, which in turn predicted increases in
alcohol-related problems—suggesting this is the main mechanism
through which dyadic conflict leads to alcohol-related problems in
emerging adult romantic dyads.
Partner Effects
Significant partner effects were not found, suggesting coping mo-
tives may only influence alcohol use patterns at the individual level.
Furthermore, within any given week, changes in coping motives in
one partner were unrelated to changes in coping motives for the other
partner, and over time, coping motives were unrelated between ro-
mantic partners. This was unexpected, given that individual drinking
motives (including coping motives) are related to peer drinking mo-
tives in adolescents (Kuntsche & Stewart, 2009). It is possible that the
lack of partner effect for the coping motives can be attributed to the
contextual factors of drinking. For example, those endorsing coping
motives are more likely to drink alone (Cooper, 1994). Thus, perhaps
partner effects were not found for coping motives because they are
solitary processes that are not readily found in social situations where
partner effects may occur. It is also possible that other drinking
motives (e.g., enhancement) are more visible to others (and thus more
influential). In fact, among adolescents, enhancement motives are
transmitted among peers earlier than coping motives (Stewart,
Castellanos-Ryan, Vitaro, & Conrod, 2014). The null findings for
partner effects might also be because of the short 1-week time lag, the
focus on emerging adults, or the focus on dating couples rather than
peers or married couples (see limitations).
However, a significant indirect partner effect was found, such that
the actor’s alcohol-related problems mediated the link between the
actor’s CDM and the partner’s alcohol-related problems. Thus, within
any given week, individuals who drink to cope with depression tended
to have more problems associated with their drinking (e.g., black outs,
missing work, neglecting responsibilities), which in turn encouraged
that individual’s partner to consume alcohol in similarly problematic
ways. This is consistent with previous work demonstrating that ro-
mantic partners exhibit similar longitudinal escalations in problematic
drinking (Mushquash et al., 2013). Indeed, it is likely that simply
being exposed to a context of alcohol-related problems impacts the
problematic drinking of the romantic partner. This may occur through
modeling, as experimental research demonstrates that participants
tend to model their drinking after that of a confederate; for instance,
participants drink more when the confederates also drink more (Bor-
sari & Carey, 2001). Thus, perhaps when the individual is drinking to
cope and experiencing alcohol-related problems, their partner is
drinking in a similar fashion but is motivated to drink for a different
reason. For example, perhaps a couple might go to a party, drink
excessively, and both miss work the next day due to a hangover—
even though one partner is motivated to “get drunk” (enhancement),
whereas the other is drinking to cope with a hard day at work
(coping). In this way, the drinking behavior might be emulated
through social learning, whereas motives—which, in this example,
are internal and unobservable—affect one’s own behavior but not the
partner’s.
Clinical Implications
The current results have several clinical implications. In particular,
future intervention efforts may wish to develop strategies to enable
emerging adult romantic dyads to cope more adaptively with conflict,
and/or to use targeted strategies to reduce CDM, as reducing these
factors might reduce alcohol-related problems. This may be particu-
larly important for women. Previous work has shown that
personality-matched strategies significantly reduce coping motives
and alcohol-related problems in samples of adolescents (Conrod,
Castellanos-Ryan, & Mackie, 2011), suggesting that similar interven-
Table 4
Tests of Indirect Effects for the Multiple Mediation Model With Coping With Depression Motives
(CDM) and Coping With Anxiety Motives (CAM; Study 2)
Predictor (X) Mediator (M) Outcome (Y)
95% CI
within-subjects
95% CI
between-subjects
Dyadic conflict Actor’s CDM Actor’s RAPI [.0005, .003]
[.007, .03]
a
Dyadic conflict Actor’s CDM Partner’s RAPI [.0003, .001] [.02, .003]
Actor’s CDM Actor’s RAPI Partner’s RAPI [.001, .07]
[.06, .20]
Actor’s CDM Partner’s RAPI Actor’s RAPI [.007, .03] [.13, .03]
Dyadic conflict Actor’s CAM Actor’s RAPI [.002, .001] [.0003, .02]
a
Dyadic conflict Actor’s CAM Partner’s RAPI [.002, .0007] [.0001, .02]
Actor’s CAM Actor’s RAPI Partner’s RAPI [.05, .04] [.07, .28]
Actor’s CAM Partner’s RAPI Actor’s RAPI [.07, .02] [.07, .32]
Note. Confidence intervals (CI) were derived using a Monte Carlo method with 20,000 resamples using
unstandardized coefficients. RAPI Rutgers Alcohol Problem Index.
a
Indirect effect becomes significant when collinearity is accounted for (i.e., model with single mediator).
p.05.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
9
CONFLICT, DRINKING TO COPE, ALCOHOL-RELATED ISSUES
tions may be advantageous in other samples. Additionally, interven-
tions should target the couple rather than the individual. Although
individual level interventions can be effective in reducing levels of
alcohol-related problems, these interventions have been shown to be
less successful when individuals remain in a social network of heavier
drinkers (Carey, Scott-Sheldon, Carey, & DeMartini, 2007). Indeed,
results showed that alcohol-related problems tend to change system-
atically in the same direction between partners, suggesting dyadic-
level interventions may be advantageous.
Limitations and Future Directions
Although the current study makes important contributions, it has
several limitations. Although our longitudinal design can make stron-
ger causal inferences than cross-sectional designs, no longitudinal
data can unambiguously disentangle cause and effect. Future research
might use experimental designs, which induce mild conflict in the lab
by asking couples to discuss contentious issues and examining part-
ners’ resultant alcohol use. Moreover, it is unknown if a time lag of
one week is appropriate for examining this relationship. Future studies
might examine these variables over a different time lag (e.g., daily
diary or event contingent), as drinking to cope may occur relatively
quickly following conflict. Future research might also examine neg-
ative affect as a potential mediating process between conflict and
coping motives. Furthermore, because the same measures were ad-
ministered during each wave, participants may have been influenced
by demand characteristics or reactance, which may have inadvertently
influenced their responses over the follow-up period. Our sample
consisted mainly of young, dating couples, so findings may not
generalize to other samples (e.g., older married couples, peers, or
clinical samples with more serious alcohol-related problems). Finally,
though theory suggests they are unrelated to dyadic conflict, future
research may wish to examine other drinking motives to see if partner
effects exist for these.
Conclusions
This study strengthened and integrated both drinking motives the-
ory and theories of romantic relationship influence by demonstrating
that CDM mediated the association between dyadic conflict and
alcohol-related problems. Furthermore, the partner’s CDM indirectly
influenced the individual’s alcohol-related problems by way of the
partner’s alcohol-related problems, suggesting alcohol-related prob-
lems co-occur within emerging adult romantic dyads. Supplementary
analyses of heterosexual dyads suggested women are more likely to
drink to cope with negative emotions in response to relationship
conflict. These sex differences are consistent with prior work (Levitt
& Cooper, 2010) and suggest this is a particularly important route to
alcohol-related problems for women. Such knowledge is important
for understanding the interpersonal factors associated with alcohol-
related problems and may inform future intervention efforts such as
couples’ treatment for alcohol-related problems.
References
Abbey, A., & Andrews, F. M. (1985). Modeling the psychological deter-
minants of life quality. Social Indicators Research, 16, 1–34. http://dx
.doi.org/10.1007/BF00317657
Adlaf, E. M., Demers, A., & Gliksman, L. (2004). Canadian Campus
Survey: Highlights. Retrieved from http://www.camh.ca/en/research/
research_areas/community_and_population_health/Documents/
CCS_2004_report.pdf
Arnett, J. J. (2000). Emerging adulthood. A theory of development from
the late teens through the twenties. American Psychologist, 55, 469 –
480. http://dx.doi.org/10.1037/0003-066X.55.5.469
Arnett, J. J. (2005). The developmental context of substance use in emerg-
ing adulthood. Journal of Drug Issues, 35, 235–254. http://dx.doi.org/
10.1177/002204260503500202
Borsari, B., & Carey, K. B. (2001). Peer influences on college drinking: A
review of the research. Journal of Substance Abuse, 13, 391– 424.
http://dx.doi.org/10.1016/S0899-3289(01)00098-0
Brooks, K. P., & Dunkel Schetter, C. (2011). Social negativity and health:
Conceptual and measurement issues. Social and Personality Psychology
Compass, 5, 904 –918. http://dx.doi.org/10.1111/j.1751-9004.2011
.00395.x
Canadian Institutes of Health Research, Natural Sciences and Engineering
Research Council of Canada, and Social Sciences and Humanities Re-
search Council of Canada. (2010, December). Tri-Council Policy State-
ment: Ethical Conduct for Research Involving Humans. Retrieved from
http://www.pre.ethics.gc.ca/pdf/eng/tcps2/TCPS_2_FINAL_Web.pdf
Carey, K. B. (1995). Heavy drinking contexts and indices of problem
drinking among college students. Journal of Studies on Alcohol, 56,
287–292. http://dx.doi.org/10.15288/jsa.1995.56.287
Carey, K. B., Scott-Sheldon, L. A., Carey, M. P., & DeMartini, K. S.
(2007). Individual-level interventions to reduce college student drinking:
A meta-analytic review. Addictive Behaviors, 32, 2469 –2494. http://dx
.doi.org/10.1016/j.addbeh.2007.05.004
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. http://dx.doi.org/10.1016/S0005-7967(98)00024-2
Collins, R. L., Kashdan, T. B., Koutsky, J. R., Morsheimer, E. T., & Vetter,
C. J. (2008). A self-administered timeline followback to measure vari-
ations in underage drinkers’ alcohol intake and binge drinking. Addictive
Behaviors, 33, 196 –200. http://dx.doi.org/10.1016/j.addbeh.2007.07
.001
Conrod, P. J., Castellanos-Ryan, N., & Mackie, C. (2011). Long-term
effects of a personality-targeted intervention to reduce alcohol use in
adolescents. Journal of Consulting and Clinical Psychology, 79, 296 –
306. http://dx.doi.org/10.1037/a0022997
Cooper, M. L. (1994). Motivations for alcohol use among adolescents:
Development and validation of a four-factor model. Psychological As-
sessment, 6, 117–128. http://dx.doi.org/10.1037/1040-3590.6.2.117
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.
http://dx.doi.org/10.1037/0022-3514.69.5.990
Dawson, D. A. (2003). Methodological issues in measuring alcohol use.
Alcohol Research & Health, 27, 18 –29.
DiBello, A. M., Neighbors, C., Rodriguez, L. M., & Lindgren, K. (2014).
Coping with jealousy: The association between maladaptive aspects of
jealousy and drinking problems is mediated by drinking to cope. Addic-
tive Behaviors, 39, 94 –100. http://dx.doi.org/10.1016/j.addbeh.2013.08
.032
DiStefano, C., Zhu, M., & Mîndrilaˇ, D. (2009). Understanding and using
factor scores: Considerations for the applied researcher. Practical As-
sessment, Research & Evaluation, 14, 1–11. Available online http://
pareonline.net/getvn.asp?v14&n20.
Enders, C. K., & Bandalos, D. L. (2001). The relative performance of full
information maximum likelihood estimation for missing data in struc-
tural equation models. Structural Equation Modeling, 8, 430 – 457.
http://dx.doi.org/10.1207/S15328007SEM0803_5
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
10 LAMBE, MACKINNON, AND STEWART
Fincham, F. D., & Beach, S. R. (1999). Conflict in marriage: Implications
for working with couples. Annual Review of Psychology, 50, 47–77.
http://dx.doi.org/10.1146/annurev.psych.50.1.47
Foran, H. M., & O’Leary, K. D. (2008). Problem drinking, jealousy, and
anger control: Variables predicting physical aggression against a partner.
Journal of Family Violence, 23, 141–148. http://dx.doi.org/10.1007/
s10896-007-9136-5
Geldhof, G. J., Preacher, K. J., & Zyphur, M. J. (2014). Reliability
estimation in a multilevel confirmatory factor analysis framework. Psy-
chological Methods, 19, 72–91. http://dx.doi.org/10.1037/a0032138
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 Be-
haviors, 32, 2611–2632. http://dx.doi.org/10.1016/j.addbeh.2007.07.004
Hammersla, J. F., & Frease-McMahan, L. (1990). University students’
priorities: Life goals vs. relationships. Sex Roles, 23, 1–14. http://dx.doi
.org/10.1007/BF00289874
Homish, G. G., & Leonard, K. E. (2007). The drinking partnership and
marital satisfaction: The longitudinal influence of discrepant drinking.
Journal of Consulting and Clinical Psychology, 75, 43–51.
Kenny, D. A., & Cook, W. (1999). Partner effects in relationship research:
Conceptual issues, analytic difficulties, and illustrations. Personal Re-
lationships, 6, 433– 448. http://dx.doi.org/10.1111/j.1475-6811.1999
.tb00202.x
Kenny, D. A., & Ledermann, T. (2010). Detecting, measuring, and testing
dyadic patterns in the actor-partner interdependence model. Journal of
Family Psychology, 24, 359 –366. http://dx.doi.org/10.1037/a0019651
Kline, R. B. (2011). Principles and practice of structural equation mod-
eling (3rd ed.). New York, NY: Guilford Press.
Kuntsche, E., & Stewart, S. H. (2009). Why my classmates drink:
Drinking motives of classroom peers as predictors of individual
drinking motives and alcohol use in adolescence—A mediational
model. Journal of Health Psychology, 14, 536 –546. http://dx.doi.org/
10.1177/1359105309103573
Leonard, K. E., & Mudar, P. (2004). Husbands’ influence on wives’
drinking: Testing a relationship motivation model in the early years of
marriage. Psychology of Addictive Behaviors, 18, 340 –349. http://dx.doi
.org/10.1037/0893-164X.18.4.340
Levitt, A., & Cooper, M. L. (2010). Daily alcohol use and romantic
relationship functioning: Evidence of bidirectional, gender-, and
context-specific effects. Personality and Social Psychology Bulletin, 36,
1706 –1722. http://dx.doi.org/10.1177/0146167210388420
Mackinnon, S. P., Kehayes, I. L., Clark, R., Sherry, S. B., & Stewart, S. H.
(2014). Testing the 4-factor model of personality vulnerability to alcohol
misuse: A three-wave, one-year longitudinal study. Psychology of Ad-
dictive Behaviors, 28, 1000 –1012. http://dx.doi.org/10.1037/a0037244
Mackinnon, S. P., Sherry, S. B., Antony, M. M., Stewart, S. H., Sherry,
D. L., & Hartling, N. (2012). Caught in a bad romance: Perfectionism,
conflict, and depression in romantic relationships. Journal of Family
Psychology, 26, 215–225. http://dx.doi.org/10.1037/a0027402
Marlatt, G. A. (1996). Taxonomy of high-risk situations for alcohol re-
lapse: Evolution and development of a cognitive-behavioral model.
Addiction, 91 (Suppl.), S37–S49. http://dx.doi.org/10.1111/j.1360-0443
.1996.tb02326.x
Marshal, M. P. (2003). For better or for worse? The effects of alcohol use
on marital functioning. Clinical Psychology Review, 23, 959 –997. http://
dx.doi.org/10.1016/j.cpr.2003.09.002
Martens, M. P., Neighbors, C., Dams-O’Connor, K., Lee, C. M., & La-
rimer, M. E. (2007). The factor structure of a dichotomously scored
Rutgers Alcohol Problem Index. Journal of Studies on Alcohol and
Drugs, 68, 597– 606. http://dx.doi.org/10.15288/jsad.2007.68.597
McCabe, S. E., Schulenberg, J. E., Johnston, L. D., O’Malley, P. M.,
Bachman, J. G., & Kloska, D. D. (2005). Selection and socialization
effects of fraternities and sororities on US college student substance use:
A multi-cohort national longitudinal study. Addiction, 100, 512–524.
http://dx.doi.org/10.1111/j.1360-0443.2005.01038.x
Miller, E. T., Neal, D. J., Roberts, L. J., Baer, J. S., Cressler, S. O., Metrik,
J., & Marlatt, G. A. (2002). Test-retest reliability of alcohol measures: Is
there a difference between internet-based assessment and traditional
methods? Psychology of Addictive Behaviors, 16, 56 – 63. http://dx.doi
.org/10.1037/0893-164X.16.1.56
Murray, S. L., Griffin, D. W., Rose, P., & Bellavia, G. M. (2003).
Calibrating the sociometer: The relational contingencies of self-esteem.
Journal of Personality and Social Psychology, 85, 63– 84. http://dx.doi
.org/10.1037/0022-3514.85.1.63
Mushquash, A. R., Stewart, S. H., Sherry, S. B., Mackinnon, S. P., Antony,
M. M., & Sherry, D. L. (2013). Heavy episodic drinking among dating
partners: A longitudinal actor-partner interdependence model. Psychol-
ogy of Addictive Behaviors, 27, 178 –183. http://dx.doi.org/10.1037/
a0026653
O’connor, B. P. (2000). SPSS and SAS programs for determining the
number of components using parallel analysis and velicer’s MAP test.
Behavior Research Methods, Instruments, & Computers, 32, 396 – 402.
http://dx.doi.org/10.3758/BF03200807
Oishi, S., & Sullivan, H. W. (2006). The predictive value of daily vs.
retrospective well-being judgments in relationship quality. Journal of
Experimental Social Psychology, 42, 460 – 470. http://dx.doi.org/
10.1016/j.jesp.2005.07.001
Peugh, J. L., DiLillo, D., & Panuzio, J. (2013). Analyzing mixed-dyadic
data using structural equation models. Structural Equation Modeling,
20, 314 –337. http://dx.doi.org/10.1080/10705511.2013.769395
Preacher, K. J., Zyphur, M. J., & Zhang, Z. (2010). A general multilevel
SEM framework for assessing multilevel mediation. Psychological
Methods, 15, 209 –233. http://dx.doi.org/10.1037/a0020141
Rodriguez, L. M., Knee, C. R., & Neighbors, C. (2014). Relationships can
drive some to drink: Relationship-contingent self-esteem and drinking
problems. Journal of Social and Personal Relationships, 31, 270 –290.
http://dx.doi.org/10.1177/0265407513494037
Scheffer, J. (2002). Dealing with missing data. Research Letters in the
Information and Mathematical Sciences, 3, 153–160.
Selig, J. P., & Preacher, K. J. (2008, June). Monte Carlo method for
assessing mediation: An interactive tool for creating confidence intervals
for indirect effects [Computer software]. Available from http://quantpsy
.org/
Shorey, R. C., Stuart, G. L., & Cornelius, T. L. (2011). Dating violence and
substance use in college students: A review of the literature. Aggression
and Violent Behavior, 16, 541–550. http://dx.doi.org/10.1016/j.avb.2011
.08.003
Stewart, S. H., Castellanos-Ryan, N., Vitaro, F., & Conrod, P. J. (2014).
Why my friends drink is why I drink. Alcoholism, Clinical and Exper-
imental Research, 38 (Suppl. s1), 187a.
Thombs, D. L., & Beck, K. H. (1994). The social context of four adolescent
drinking patterns. Health Education Research, 9, 13–22. http://dx.doi
.org/10.1093/her/9.1.13
White, H. R., & Labouvie, E. W. (1989). Towards the assessment of
adolescent problem drinking. Journal of Studies on Alcohol, 50, 30 –37.
http://dx.doi.org/10.15288/jsa.1989.50.30
Yu, C. Y. (2002). Evaluating cutoff criteria of model fit indices for latent
variable models with binary and continuous outcomes (Doctoral disser-
tation, University of California Los Angeles).
Received August 21, 2014
Revision received December 2, 2014
Accepted April 29, 2015
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
11
CONFLICT, DRINKING TO COPE, ALCOHOL-RELATED ISSUES
... Scores from both parents were summed into a single couple score to capture collective parent depression. This creation of a composite score across parents is similar to the approach we used in past studies with couples to capture a dyadic level variable [55,56]. Depression scores between Parent A and Parent B were highly correlated (r = 0.49, p ≤ 0.001), providing psychometric support for this approach. ...
... Possible scores ranged from 1 to 27, with greater scores indicating higher levels of anxiety. Parent scores were summed into one collective anxiety score, which is similar to the approach we used in past studies with couples to capture a dyadic level variable [55,56]. The high correlation (r = 0.45, p = <0.001) between Parent A and Parent B provides psychometric support for this approach. ...
... Both parents' scores were summed into a single couple score, representing their collective stress. In past studies with couples, we used a similar approach to capture a dyadic level variable [55,56]. Parent A and Parent B stress scores were highly correlated (r = 0.5, p ≤ 0.001), which provides psychometric support for this approach. ...
Article
Full-text available
Background/Objectives: Most studies have linked mandatory homeschooling during COVID-19 to mental health harm in parents and children, while a minority have found non-significant or beneficial effects. Past studies have not measured mandatory homeschooling continuously over an extended period; consequently, they could not capture compounding mental health effects, which may explain conflicting results. We asked whether children’s cumulative time spent homeschooled during COVID-19 school closure mandates caused compounding harm for parent and child mental health, and whether parent employment, child internet access and educational support from schools (live and pre-recorded online classes, home learning packs) impacted this relationship. We aimed to identify the families at greatest risk of mental health harm during mandatory homeschooling and the educational support that may have mitigated this risk. Methods: Couples completed retrospective, cross-sectional survey questionnaires assessing parent depression, anxiety and stress, child internalizing and externalizing symptoms, and the family’s homeschooling experience. Data were analyzed using mediation analysis total effects, ordinary least squares regression and simple slopes analysis. Results: Both parents and children experienced compounding mental health harm during mandatory homeschooling. Live online classes protected parents and children, while home learning packs protected children. Unexpectedly, reliable internet access and the employment of both parents placed children at greater risk. Conclusions: Findings suggest that long-term mandatory homeschooling during COVID-19 placed families at greater risk of mental health harm. To protect family mental health during homeschooling mandates, schools should provide children with evidence-based educational support.
... Conflict behaviors, which refer to critical, hostile, or rejecting behaviors directed at one's partner (Kim, Smith et al., 2022), have been found to be positively correlated with depressive symptoms, dissatisfaction with the relationship, drinking and gambling to cope with depression, and decreased life satisfaction (Hagen et al., 2023;Lambe et al., 2015;Mackinnon et al., 2012Mackinnon et al., , 2017. Research on the stress generation hypothesis within couples has focused primarily on relationship discord, such as disagreements on household tasks, money, and sex, and its reciprocal associations with depressive symptoms (e.g., Choi & Marks, 2008). ...
... The total scores for each conflict measure were standardized and combined into a composite. 1 Previous research using these measures has also reported good internal consistency and test-retest reliability, as well as criterion validity and coherence on a single underlying factor (Kim, Smith et al., 2022;Lambe et al., 2015;Mackinnon et al., 2012Mackinnon et al., , 2017. ...
... Table 1 provides descriptive statistics and correlations among variables. The scores for depressive symptoms and conflict enactment measures were within the expected range based on means derived from normative samples of emerging adults (Lambe et al., 2015;Mackinnon et al., 2017;Radloff, 1977Radloff, , 1991. All measures showed good to excellent internal consistency in the present sample (α = .86-.93). ...
Article
Introduction: In this study we investigated the stress generation hypothesis in romantic relationships, testing how individuals’ depressive symptoms can lead to interpersonal stress that worsens their depressive symptoms. Interpersonal stress was operationalized as conflict enactment (critical, hostile, and rejecting behaviors toward one's romantic partner) and breakup rumination (persistent and intense thoughts of ending the relationship). Methods: Participants included 226 heterosexual adult romantic couples, with a mean age of 21.48 years for women and 22.35 years for men. A short-term longitudinal design was employed, involving three waves of data collection, including a daily diary portion. Depressive symptoms were measured in the lab at the beginning of the study (Wave 1) and again one month later (Wave 3). Following Wave 1, conflict enactment and breakup rumination were measured online over a 14-day period (Wave 2) through daily diary entries, which were then averaged across the 14 days into a single score for each construct. Results: Using actor-partner interdependence mediation modelling, we found women's and men's depressive symptoms at Wave 1 positively predicted their own conflict enactment and breakup rumination at Wave 2, as well as their own depressive symptoms at Wave 3 (actor effects). Women's depressive symptoms at Wave 1 positively predicted men's conflict enactment at Wave 2 (partner effect). Additionally, women's conflict enactment at Wave 2 positively predicted their own depressive symptoms at Wave 3 (actor effect) and mediated the increase in their own depressive symptoms from Wave 1 to 3 (indirect effect). Discussion: Findings illustrate how depressive symptoms are embedded in an interpersonal context and exacerbated by conflictual behaviors within romantic relationships over a period of one month.
... The link between conflict and addictive behavior via coping motives has been demonstrated in the alcohol literature, where coping drinking motives were shown to mediate the association between dyadic conflict and alcohol-related problems, both crosssectionally and longitudinally (Lambe et al., 2015). However, the prior study did not examine gambling problems, and was conducted in a young (M age = 22.13 years old, SD = 5.67) and age-restricted sample which limits generalizability to other stages of adulthood. ...
... The present study sought to expand our understanding of coping gambling motives in response to romantic conflict. This would both extend this body of knowledge from the alcohol literature (i.e., Lambe et al., 2015) to the gambling literature and additionally directly test the mediating role of negative affect in a sample of adult partnered gamblers. Specifically, we examined if romantic conflict was positively associated with gambling problems, and, whether negative affect and coping motives sequentially mediated this relationship. ...
... The original versions were included to validate the seven-day versions. A seven-day version of romantic conflict has already been validated (Lambe et al., 2015), and coping (Murray et al., 2003). b Negative affect measured by combining the subscales from the Depression, Anxiety, and Stress Scale (DASS-21; Lovibond and Lovibond, 1995). ...
Article
Full-text available
Introduction: While individuals have many motives to gamble, one particularly risky motive for gambling is to cope with negative affect. Conflict with one's romantic partner is a strong predictor of negative affect, which may elicit coping motives for gambling and, in turn, gambling-related problems. Support for this mediational model was demonstrated in relation to drinking-related problems. We extended this model to gambling. Method: Using a cross-sectional design, we examined links between romantic conflict (Partner-Specific Rejecting Behaviors Scale), negative affect (Depression, Anxiety, and Stress Scales-21), coping gambling motives (Gambling Motives Questionnaire, coping subscale), and gambling-related problems [Problem Gambling Severity Index (PGSI)] in 206 regular gamblers (64% men; mean age = 44.7 years; mean PGSI = 8.7) who were in a romantic relationship and recruited through Qualtrics Panels in July 2021. Results: Results supported our hypothesis that the association between romantic conflict and gambling-related problems would be sequentially mediated through negative affect and coping gambling motives, β = 0.38, 95% CI [0.27, 0.39], and also showed a strong single mediation pathway through negative affect alone, β = 0.27, 95% CI [0.17, 0.38]. Discussion: Negative affect and coping gambling motives partially explain the link between romantic conflict and gambling-related problems. Interventions should target both negative affect and coping gambling motives in response to romantic conflict to reduce gambling-related problems in partnered gamblers.
... Another pre-pandemic study from Lambe et al. examined the links between couples' conflict and alcohol outcomes in a sample of 100 emerging adult couples [38]. They examined the individual's own coping motives and the partner's drinking motives as possible mediators of the conflict-drinking outcomes link [38]. ...
... Another pre-pandemic study from Lambe et al. examined the links between couples' conflict and alcohol outcomes in a sample of 100 emerging adult couples [38]. They examined the individual's own coping motives and the partner's drinking motives as possible mediators of the conflict-drinking outcomes link [38]. In indistinguishable dyad analyses (i.e., not distinguished by gender), this study showed that the individual's coping-with-depression motives mediated the relation between dyadic conflict (combined reporting by both partners) and the individual's alcohol-related problems [38]. ...
... They examined the individual's own coping motives and the partner's drinking motives as possible mediators of the conflict-drinking outcomes link [38]. In indistinguishable dyad analyses (i.e., not distinguished by gender), this study showed that the individual's coping-with-depression motives mediated the relation between dyadic conflict (combined reporting by both partners) and the individual's alcohol-related problems [38]. Within the mixed-gender couples, this mediation held true for women but not men. ...
Article
Full-text available
The COVID-19 pandemic spurred public health measures to reduce viral spread. Concurrently, increases in alcohol consumption and conflict in romantic partnerships were observed. Pre-pandemic research demonstrated a bidirectional association between couples’ conflict and drinking. Recent research shows one’s drinking motives (proximal predictors of drinking behavior) can influence another person’s drinking in close relationships. It is possible that individuals are drinking to cope with distress following romantic conflict. The current study examined 348 cohabitating couples during the first lockdown in the spring of 2020. Our analyses examined coping motives as a mediator between dyadic conflict and drinking behavior using actor–partner interdependence models. Results showed that conflict was associated with greater reports of own drinking in gendered (distinguishable) and nongendered (indistinguishable) analyses through coping motives. Further, in mixed-gender couples, men partners’ coping motives predicted less drinking in women, while women partners’ coping motives predicted marginally more drinking in men. Partner effects may have been observed due to the increased romantic partner influence during the COVID-19 lockdown. While these results suggest that men’s coping motives may be protective against women’s drinking, more concerning possibilities are discussed. The importance of considering dyadic influences on drinking is highlighted; clinical and policy implications are identified.
... Item ratings were made on a scale from 1 (strongly disagree) to 9 (strongly agree). Research has demonstrated adequate internal reliability, test-retest reliability, and criterion validity for this measure when used with romantic partners (Lambe et al., 2015;Mackinnon et al., 2012). ...
... As recommended by Martens et al. (2007), the individual RAPI items were dichotomized (i.e., 1 = presence, 0 = absence of problematic drinking) and then summed into a single value (possible range 0-23) for analysis. The dichotomized RAPI possesses a unidimensional factor structure and has high internal reliability (Lambe et al., 2015). ...
Article
Full-text available
Friendships are important for the mental well-being of emerging adults. Socially prescribed perfectionism, where individuals feel pressured to be perfect by others, can be destructive, leading to conflict with others, depressive symptoms, and problematic drinking. However, its impact on friendships is not well-explored. This study examined 174 emerging adult friendship dyads using a 4-wave, 4-month dyadic design. Data were analyzed using longitudinal actor-partner interdependence models. Using a novel friend-specific measure of socially prescribed perfectionism, we found that an individual's perceived expectation to be perfect from a friend was positively associated with increased conflict between friends, as well as with higher levels of depressive symptoms and problematic drinking in the individual. Findings lend credence to longstanding theoretical accounts and case histories suggesting socially prescribed perfectionism leads to harmful individual and relational outcomes and extends them to the specific context of friendships.
... For example, in a sample comprised of nearly all (>90%) dating couples, partners demonstrated similarity in their drinking behaviors and drinking motivations. 4,5 Emerging research shows that many aspects of adults' health behaviors have been altered in mostly negative directions since the onset of COVID-19, including sleep, physical activity, dietary habits, and relationship aggression. 6 Despite longstanding recognition that romantic relationships play an important role in how individuals' health may change over time in response to the pandemic onset, 7 few studies have been able to test dyadic-informed hypotheses about connections between partners' COVID-19 experiences (eg, coping behaviors, perceived adjustment). ...
Article
Background: Research has established associations between romantic partners' health-related behaviors, although links between partners' experiences of the COVID-19 pandemic at the time of an important medical event remain untested. Methods: The sample was drawn from an existing study of patients who received a new opioid prescription at an emergency department visit for acute pain. We assessed COVID-19 experiences of 97 patients and their romantic partners from April 2021 through June 2022. Results: Romantic partners reported similar ratings of COVID-19 impact and were likely to agree on their coping with the pandemic by engaging in more time on activities like puzzles or books, using marijuana, and drinking alcohol. Partners also demonstrated high concordance in their COVID-19 vaccination statuses. Conclusions: These findings extend a robust literature showing romantic partners' concordance in a host of health-relevant behaviors to their COVID-19 experiences.
Article
Committed romantic relationships between two individuals with Substance Use Disorder (or dual-SUD couples) are prevalent. Dual-SUD couples have poor treatment engagement and outcomes. Research has established a reciprocal link between relationship dynamics (e.g., conflict, intimacy) and substance use. Thus, the couple’s relationship presents a distinct social context for both partner’s substance use. Dual-SUD couples face unique challenges due to substance use being a shared behavior that may serve as a rewarding source of compatibility, closeness, and short-term relationship satisfaction despite it being at the cost of other alternative sources of substance-free reinforcement. Yet, treatment options for these couples are scarce. Dual-maladaptive health behaviors (e.g., dual-substance use) are challenging to treat; however, theory and preliminary research suggests that transformation of couple’s joint motivation toward adaptive health behavior change may result in a more satisfying relationship and improved treatment outcomes for both individuals. The current paper reviews the extant literature on dual-SUD couples from theoretical, empirical, and treatment research and proposes an expanded paradigm regarding how we understand dual-SUD couples with the aim of informing basic research and treatment development.
Chapter
The traditional—and still prevalent—model of family violence primarily focuses on individuals as either perpetrators or victims. However, research demonstrates that violence within couples does not occur within a vacuum, is dynamic, and typically involves behaviors from both partners. In an attempt to incorporate a dyadic perspective to family violence, this chapter demonstrates why a dyadic perspective is important, outlines obstacles to obtaining dyadic data and methods for overcoming those obstacles, and presents two primary methodological and analytical techniques for analyzing dyadic data: the actor–partner interdependence model (APIM) and concordance analysis (CA). We highlight that these methods are complementary and have different strengths. Whereas APIM models are best used by researchers, CA may be used by researchers, clinicians, and practitioners who are interested in quickly categorizing couples by their dyadic concordance type.
Article
Background Although alcohol consumption is a well-known risk factor for intimate partner violence (IPV) perpetration, few studies have been conducted among young males in low- and middle-income countries. Alcohol consumption and IPV are both complex phenomena, whose association requires more in-depth exploration regarding drinking patterns and the alcohol-related manifestation of five different forms of IPV. Objective In this study, we sought to explore the relationship between alcohol use and IPV in young Tanzanian men and to identify differences in the magnitude of past-year IPV perpetration among alcohol drinkers and abstainers. Furthermore, we aimed to assess the association between various drinking patterns with the perpetration of different forms of IPV. Methods A cross-sectional survey of 1002 young males residing in Mwanza, Tanzania, was conducted in 2021–2022. Data on alcohol consumption were collected using the alcohol use disorder identification test. IPV perpetration was assessed using an index total of 19 items on acts of physical, sexual, economic, emotional abuse, and controlling behaviour. Logistic regression models were conducted to estimate the relationship between alcohol use and the perpetration of each form of IPV. Results Among partnered respondents currently consuming alcohol (n = 189, 18.8%), the most and the least prevalent IPV forms in the past 12 months were controlling behaviour (84.1%) and physical IPV (25.4%), respectively. Those reporting recent alcohol consumption reported higher rates of all forms of past-year IPV perpetration compared to abstainers. While no form of IPV was associated with low-risk consumption versus abstention, all forms of IPV were associated with hazardous drinking. Conclusion Young men who drink alcohol, especially those drinking hazardously, are also more likely to report perpetrating IPV. An understanding of the different drinking patterns and manifestations of forms of IPV can contribute to better-tailored alcohol-related interventions and has the potential to improve young adults’ health and reduce IPV perpetration.
Article
Full-text available
The 4-factor model of personality vulnerability identifies 4 personality risk factors for alcohol misuse: hopelessness, anxiety sensitivity, impulsivity, and sensation seeking. These personality traits are associated with distinct mechanisms and motivations for alcohol misuse. Individuals high in hopelessness drink to regulate dysphoric affect, while those high in anxiety sensitivity drink to reduce anxiety and to conform to peer expectations. Individuals high in sensation seeking are highly sensitive to the rewarding properties of alcohol, and misuse alcohol to maximize enjoyment. Impulsivity is a broad risk factor contributing to all drinking motives. We hypothesized that personality vulnerabilities would indirectly predict alcohol quantity and problems through specific drinking motives theorized by the 4-factor model. The present study tested hypotheses using a 3-wave, 1-year longitudinal study of undergraduate drinkers (N 302). Data were analyzed using multilevel path analysis. Hopelessness and impulsivity were positively related to drinking motives in the expected fashion. Anxiety sensitivity was related to coping-anxiety and conformity motives only in the between-subjects model (partially supporting hypotheses), while sensation seeking was generally unrelated to all drinking motives and alcohol outcomes (failing to support hypotheses). Enhancement motives predicted alcohol quantity and problems at both levels, coping-depression motives predicted alcohol problems at the between-subjects level only, and coping-anxiety, conformity, and social motives failed to predict alcohol outcomes beyond other motives. Overall, this study partially supports the 4-factor model, with the strongest support emerging for impulsivity and hopelessness. This study suggests that personality traits such as impulsivity and hopelessness may be important targets in prevention and treatment with undergraduate drinkers.
Article
Full-text available
Heavy episodic drinking (HED) is a major health problem for young adults. Rates of HED have remained consistently high among young adults for the past two decades. Though research has identified various intrapersonal, interpersonal, and environmental contributors to HED, the majority of research focuses on intrapersonal factors. As such, more research is needed to test the role that specific interpersonal relationships play in perpetuating HED. This study tests the partner influence hypothesis that suggests partners in romantic relationships influence one another's HED over time. A sample of 208 young, nonmarried, heterosexual dating couples completed HED measures at baseline and again 28 days later. Actor-partner interdependence modeling revealed significant actor effects, demonstrating stability in HED within each partner over time. Results also showed significant partner effects where HED in both young men and women in dating relationships positively influenced their partners' future HED over a relatively short time frame. Patterns in the results suggest both women and men are more affected by their own than by their partner's prior level of HED. Nonetheless, small partner effects were present for both women and men. Results support the partner influence hypothesis and suggest HED is a self-propagating behavior sustained, in part, by a pattern of interpersonal influence. These results highlight the importance of considering both intrapersonal and interpersonal factors when implementing prevention and intervention programs for young adults' HED.
Article
Full-text available
According to the social disconnection model, perfectionistic concerns (i.e., harsh self-scrutiny, extreme concern over mistakes and others' evaluations, and excessive reactions to perceived failures) confer vulnerability to depressive symptoms indirectly through interpersonal problems. This study tested the social disconnection model in 226 heterosexual romantic dyads using a mixed longitudinal and experience sampling design. Perfectionistic concerns were measured using three partner-specific self-report questionnaires. Conflict was measured as a dyadic variable, incorporating reports from both partners. Depressive symptoms were measured using a self-report questionnaire. Perfectionistic concerns and depressive symptoms were measured at Day 1 and Day 28. Aggregated dyadic conflict was measured with daily online questionnaires from Days 2 to 15. Data were analyzed using structural equation modeling. There were four primary findings: (a) Dyadic conflict mediated the link between perfectionistic concerns and depressive symptoms, even when controlling for baseline depressive symptoms; (b) depressive symptoms were both an antecedent and a consequence of dyadic conflict; (c) perfectionistic concerns incrementally predicted dyadic conflict and depressive symptoms beyond neuroticism (i.e., a tendency to experience negative emotions) and other-oriented perfectionism (i.e., rigidly demanding perfection from one's partner); and (d) the relationships among variables did not differ based on gender. As the most rigorous test of the social disconnection model to date, this study provides strong support for this emerging model. Results also clarify the characterological and the interpersonal context within which depressive symptoms are likely to occur.
Article
Full-text available
How much and why individuals in romantic relationships drink alcohol may be a function of both personal and relational influences. The current research examined factors that predict vulnerability to health-related risky behavior (i.e., drinking to cope and drinking problems) in response to relationship difficulties. We consider the possibility that for individuals whose self-worth is contingently tied to the fluctuations of their relationship, feeling less satisfied may predict increased drinking problems; moreover, this may be mediated by drinking to cope. This study evaluated relationship-contingent self-esteem (RCSE) as a moderator of the association between relationship satisfaction and coping motives, which was expected to predict alcohol problems in couples. Both members in committed relationships (N = 78 dyads) reported relationship satisfaction, RCSE, drinking to cope motives, and alcohol problems in a cross-sectional survey. Actor–Partner Interdependence Model analyses revealed significant mediated moderation among men, such that higher RCSE interacted with lower relationship satisfaction to predict stronger coping motives, in turn predicting increased drinking problems. Implications and future directions are discussed.
Article
Full-text available
Following an exploratory factor analysis, factor scores may be computed and used in subsequent analyses. Factor scores are composite variables which provide information about an individual's placement on the factor(s). This article discusses popular methods to create factor scores under two different classes: refined and non-refined. Strengths and considerations of the various methods, and for using factor scores in general, are discussed.
Article
Full-text available
Mixed-dyadic data, collected from distinguishable (nonexchangeable) or indistinguishable (exchangeable) dyads, require statistical analysis techniques that model the variation within dyads and between dyads appropriately. The purpose of this article is to provide a tutorial for performing structural equation modeling analyses of cross-sectional and longitudinal models for mixed independent variable dyadic data, and to clarify questions regarding various dyadic data analysis specifications that have not been addressed elsewhere. Artificially generated data similar to the Newlywed Project and the Swedish Adoption Twin Study on Aging were used to illustrate analysis models for distinguishable and indistinguishable dyads, respectively. Due to their widespread use among applied researchers, the AMOS and Mplus statistical analysis software packages were used to analyze the dyadic data structural equation models illustrated here. These analysis models are presented in sufficient detail to allow researchers to perform these analyses using their preferred statistical analysis software package.
Article
Full-text available
The theory of emerging adulthood has been proposed as a way of conceptualizing the developmental characteristics of young people between the ages of 18 and 25. Here, the theory is applied to explaining the high rates of substance use in this age group. Specifically, five developmentally distinctive features of emerging adulthood are proposed: the age of identity explorations, the age of instability, the age of self-focus, the age of feeling in-between, and the age of possibilities. Then, each of these features is applied to an explanation of drug use in emerging adulthood.
Article
Previous research has shown that both alcohol use and jealousy are related to negative relationship outcomes. Little work, however, has examined direct associations between alcohol use and jealousy. The current study was aimed to build upon existing research examining alcohol use and jealousy. More specifically, findings from current jealousy literature indicate that jealousy is a multifaceted construct with both maladaptive and adaptive aspects. The current study examined the association between maladaptive and adaptive feelings of jealousy and alcohol-related problems in the context of drinking to cope. Given the relationship between coping motives and alcohol-related problems, our primary interest was in predicting alcohol-related problems, but alcohol consumption was also investigated. Undergraduate students at a large Northwestern university (N=657) in the US participated in the study. They completed measures of jealousy, drinking to cope, alcohol use, and alcohol-related problems. Analyses examined associations between jealousy subscales, alcohol use, drinking to cope, and drinking problems. Results indicated that drinking to cope mediated the association between some, but not all, aspects of jealousy and problems with alcohol use. In particular, the more negative or maladaptive aspects of jealousy were related to drinking to cope and drinking problems, while the more adaptive aspects were not, suggesting a more complex view of jealousy than previously understood.
Article
The social context of adolescent drinking involves the combined influence of motivational and situational factors. This investigation assessed the usefulness of the Social Context of Drinking Scales in discriminating among four adolescent drinker types. Light and moderate drinkers were classified by relatively low scores on alcohol consumption measures and the Rutgers Alcohol Problem Index (RAPI), a measure of drinking consequences. Heavy drinkers were identified by high consumption scores, but a low RAPI score. High-consequence drinkers were those with high RAPI scores. A discriminant analysis of the drinker groups yielded three statistically significant functions. The first one most clearly distinguished light from high-consequence drinkers, and was strongly correlated with the variable ‘Social Facilitation’. The second function, which best separated heavy from high-consequence drinkers, was dominated by the variable ‘Stress Control’. Alcohol use intensity was not important to the discrimination between these two types of alcohol abusers. Moderate and heavy drinkers were distinguished from one another by gender on a third function. The findings support the discriminant validity of the Social Context of Drinking Scales and point to social psychological differences among types of adolescent drinkers.