Hazardous Alcohol Use and Intimate Partner Violence in the Military:
Understanding Protective Factors
Heather M. Foran
University of Braunschweig
Richard E. Heyman and Amy M. Smith Slep
New York University
Jeffery D. Snarr
The College at Brockport, SUNY
United States Air Force Family Advocacy
Lackland-Kelly Air Force Base
Hazardous alcohol use is a well-established risk factor for men’s intimate partner violence (IPV), with
dozens of studies demonstrating the association. The current study extends understanding of the
hazardous alcohol use-IPV link by examining what factors moderate this association in a more systematic
and broader way that has been done in past studies. Individual, family, workplace, community, and
developmental factors were tested as moderators of the hazardous alcohol use and IPV link in a large,
representative sample of active duty service members (the 2006 Community Assessment), and the results
were tested for replicability in a hold-out sample. Two family variables (relationship satisfaction and
parent–child satisfaction), 1 community variable (community safety), and 3 developmental variables
(years in the military, marital length, and family income/pay grade) cross-validated as significant
moderators of the association between men’s hazardous alcohol use and IPV. Across the significant
moderators, the association between hazardous alcohol use and men’s IPV was weakened by maturation/
development, improved community safety, and better relationship functioning. No individual or work-
place variables were significant moderators for men, and there were no significant moderators found for
women. The results support the importance of a developmental and relational perspective to understand-
ing the hazardous alcohol use-IPV link, rather than solely an individual coping perspective.
Keywords: intimate partner violence, alcohol abuse, problem drinking, family violence, military
Hazardous alcohol use1is a well-established risk factor for
men’s perpetration of intimate partner violence (IPV). Three pre-
vious meta-analyses with different study inclusion criteria have all
provided support for a small-to-moderate association between
men’s drinking and IPV (average effect size rs for the three
studies ? .22, .23, 24; Ferrer, Bosch, Garcia, Manassero, & Gili,
2004; Foran & O’Leary, 2008; Stith, Smith, Penn, Ward, & Tritt,
2004). Further, multivariate analyses have demonstrated that the
association between alcohol and men’s IPV is significant after
controlling for other relevant predictors such as age, socioeconomic
status, ethnicity (Leonard, Bromet, Parkinson, Day, & Ryan, 1985;
Pan, Neidig, & O’Leary, 1994), hostility (Leonard & Senchak, 1993),
normative views of IPV (Kaufman-Kantor & Straus, 1990), drug
problems (Pan et al., 1994), and marital satisfaction (Leonard &
Senchak, 1993). In addition, there is support for the association
between women’s problem drinking and their perpetration of IPV
(Foran & O’Leary, 2008), although there are substantially fewer
studies of women and the effect size tends to be smaller than is
found in studies with men.
It is also clear that alcohol use/abuse does not lead to IPV for all
men or women. Theoretically, an array of factors across ecological
levels could interfere with individuals’ regulation abilities, thus
making them more vulnerable to the effects of heavy drinking on
1There are many operational definitions of alcohol use problems (e.g.,
hazardous drinking, problem drinking, alcohol abuse, alcohol dependence,
binge drinking). The focus of the current study is on hazardous alcohol use
as defined by the Alcohol Use Disorders Identification Test (Saunders et
al., 1993). Many of the studies reviewed measured alcohol use problems in
different ways and, thus, the applicable term is used throughout the review
when referring to these studies.
This article was published Online First March 26, 2012.
Heather M. Foran, Institute for Psychology, University of Braun-
schweig; Richard E. Heyman and Amy M. Smith Slep, Department of
Cardiology and Comprehensive Care, New York University; Jeffery D.
Snarr, Department of Psychology, The College at Brockport, SUNY;
United States Air Force Family Advocacy Program, Lackland-Kelly Air
Force Base, San Antonio, TX.
Contributors from the United States Air Force Family Advocacy Pro-
gram were, in alphabetical order, Major Rachel E. Foster, Lt. Col. David J.
Linkh, and Lt. Col. James D. Whitworth.
This research was supported by U.S. Department of Defense grant
W81XWH0710328. The opinions expressed in this article are solely those
of the authors and do not necessarily represent the official views of the U.S.
Government, the Department of Defense, or the Department of the Air
Force. We would like the thank the personnel at the local Air Force site
who were responsible for promoting the survey, to Caliber Associates
(especially Dr. Chris Spera) for administering the web survey and to the
active duty members and their spouses who took time to complete it.
Correspondence concerning this article should be addressed to Heather M.
Foran, University of Braunschweig, Institute of Psychology, 33 Humboldtstr.,
Braunschweig, Germany 38106. E-mail: email@example.com
Psychology of Addictive Behaviors
2012, Vol. 26, No. 3, 471–483
© 2012 American Psychological Association
0893-164X/12/$12.00 DOI: 10.1037/a0027688
aggressive behaviors (Schumacher, Homish, Leonard, Quigley, &
Kearns, 2008; Wills, Sandy, & Yaeger, 2002). Individuals with
poor personal coping, relationship distress, and/or elevated depres-
sive symptoms, for example, may be below the threshold for IPV
when not drinking, but be more likely than people without such
problems to cross over the threshold when they drink heavily and
their impulse control is lowered further. This conceptualization is
consistent with the multiple threshold model of IPV (Fals-Stewart,
Leonard, & Birchler, 2005), self-regulation theories that predict
risk for alcohol-related negative consequences (Wills et al., 2002),
and theories that suggest alcohol leads to impairment in impulse
control (e.g., alcohol myopia; Steele & Josephs, 1990; Pernanen,
Over the past two decades, several studies, primarily with men,
have found support for moderation effects of alcohol-aggression
association with experimental designs assessing laboratory aggres-
sion and intoxication (Denson, White, & Warburton, 2009; Eck-
hardt, 2007; 2008; Giancola, 2002a, 2002b, 2003; Giancola et al.,
2002; Giancola, Saucier, & Gussler-Burkhardt, 2003; Parrott &
Giancola, 2004), and with field studies assessing IPV and problem
or heavy drinking behaviors (Foran & O’Leary, 2007; Grekin,
Sher, & Larkins, 2004; Heyman, O’Leary, & Jouriles, 1995; Leon-
ard & Blane, 1992; Leonard & Senchak, 1993; Margolin, John, &
Foo, 1998; Schumacher et al., 2008). These studies have identified
high hostility, low dispositional empathy, poor coping, jealousy,
poor anger control, trait anger, marital dissatisfaction, ethnicity,
antisocial personality disorder, negative life events, trait aggres-
sivity, and alcohol expectancies regarding aggressive behaviors as
moderators of alcohol-IPV relations.
Although there has been considerable progress both theoreti-
cally and empirically in understanding the alcohol-IPV link (Leon-
ard, 2005), there are a number of limitations in the current litera-
ture that hamper a fuller understanding of the alcohol-IPV link,
and the purpose of the present study is to examine the alcohol-IPV
link while addressing some of these limitations. First, most studies
have examined only one or two moderator relationships at a time,
and little is known about which type of factors are more likely to
play a moderating role. Theoretical models suggest that modera-
tors increase/decrease threshold levels for IPV, but questions about
whether there are classes of variables that are most important have
yet to be addressed (e.g., family factors vs. workplace factors).
Second, past research on IPV and alcohol problems has primarily
focused on individual and family level risk factors (e.g., Eckhardt,
2007; Schumacher et al., 2008) and little research has examined
community and workplace factors as moderators despite growing
evidence that community and workplace level factors also play a
role in hazardous alcohol use and IPV (e.g., Cunradi, 2007; Cun-
radi, Todd, Duke, & Ames, 2009; Foran, Slep, Heyman, & U.S.
Air Force Family Advocacy Program, 2011a; Raghavan, Menner-
ich, Sexton, & James, 2006; Slep, Foran, Heyman, & Snar, 2010).
Third, consideration of the developmental context is also
needed. Both hazardous alcohol use and IPV prevalences vary
across developmental levels (i.e., higher among younger and un-
married individuals; O’Leary & Woodin, 2005). Developmental
factors are fundamental components of contemporary etiological
theories on alcohol (Sher, Grekin, & Williams, 2005; Zucker,
Fitzgerald, & Moses, 1995) and have also been incorporated into
models of IPV (O’Leary, 1999; Vickerman & Margolin, 2008;
Williams, Connolly, Pepler, Craig, & Laporte, 2008). However,
the impact of adult development on the alcohol-IPV connection
remains largely unexplored. Factors such as age, marital status,
and family income may moderate the risk for alcohol-related
aggression through enhanced self-regulation (i.e., impulse control)
that develops through maturation and changes in exposure to
environmental risk factors (e.g., transition to marriage leading to
different expectations for social behavior; Leonard & Mudar,
2003). An older individual may be able to better regulate their
emotions and behaviors while under the influence of alcohol,
leading to less risk for IPV. It is also possible that the association
between alcohol and IPV may become stronger with age. Given
that there is a general trend for older individuals to drink less
heavily and aggress less, those who continue to engage in those
behaviors and not “age-out” may be reflective of individuals with
more stable, lifetime persistent trajectories toward poor impulse
control and regulation (Chassin, Pitts, & Prost, 2002; Littlefield,
Sher, & Wood, 2009; Moffitt, Caspi, Harrington, & Milne, 2002).
Fourth, research has largely focused on testing stable trait-like
characteristics as moderators (e.g., Grekin, Sher, & Larkins, 2004)
rather than factors that are potentially modifiable and could be
targeted with prevention and treatment programs. A better under-
standing of which modifiable factors may moderate the alcohol-
IPV association at different ecological levels has clear implications
for interventions within a particular level (e.g., community-wide
prevention, workplace initiatives, and family therapy).
Fifth, as mentioned earlier, most of the studies examining
moderators of alcohol and IPV perpetration have been con-
ducted with men (e.g., Foran & O’Leary, 2007). The role of
alcohol in women’s IPV is less clear and studies that have
assessed moderators have found mixed results (Foran &
O’Leary, 2008; Leonard & Eiden, 2007; Simmons, Lehmann,
Cobb, 2008). In a recent longitudinal study that did examine
moderators of both men’s and women’s drinking-IPV relations,
a significant moderation effect was only found for men (Schu-
macher et al., 2008). The lack of findings for women across
studies may be due to insufficient power to detect the effect due
to a small number of heavy drinking women (Leonard & Eiden,
2007; Schumacher et al., 2008). Further, both meta-analytical
results of the association between alcohol and IPV and
laboratory-based research on alcohol intoxication and aggres-
sion suggest that the strength of the bivariate association be-
tween alcohol and IPV perpetration is smaller among women
(Foran & O’Leary, 2008; Giancola et al., 2009).
Consistent with this, in an earlier study with the sample used
in the current study, the correlation between alcohol problems
and IPV was twice as large for men (r ? .18, p ? .001) as for
women (r ? .09, p ? .001; Slep et al., 2010) and this difference
is statistically significant (z ? 3.41, p ? .001). Although the
association may be weaker for women, the Slep et al., 2010
study examined more than 20 risk factors of IPV and alcohol
problems significantly predicted both men’s and women’s per-
petration of IPV, even when controlling for other significant
The goal of the current study was to build on this previous
study and address limitations in the alcohol-IPV literature by
testing an array of risk factors across individual, family, work-
place, developmental, and community levels that could plausi-
bly exert additive or buffering effects on the alcohol problems
and IPV association. Factors within each ecological level were
FORAN ET AL.
selected for inclusion in the study if they were potentially
modifiable (e.g., depressive symptoms, relationship satisfac-
tion, personal coping) or could be used to identify high-risk
groups for targeted prevention efforts (e.g., number of years in
the military). Earlier studies with this sample identified the
above listed factors as related to both IPV and alcohol problems
(as main effects; Foran et al., 2011a; Slep et al., 2010).The
current study takes the step by testing which factors specifically
moderate the alcohol problems—IPV association.
Further, by examining interaction effects across an array of risk
factors, patterns of interaction effects across variables may be
identified. Are there certain “protective” factors that dampen the
alcohol-IPV association when elevated? Do the self-regulation and
threshold models of the alcohol-IPV link hold across different
types of factors (e.g., family or community level)? As stated
above, previous research has generally examined only a few mod-
eration effects in a single study, limiting inferences that could be
drawn to inform theoretical development of alcohol and IPV. The
current exploratory study therefore has potential to inform both
theoretical understanding of the alcohol-IPV link as well as pre-
vention and intervention efforts geared at different ecological
levels (e.g., individual vs. community).
In the current study, we sought to test moderators of the alcohol-
aggression relationship in a large, representative sample of active
duty Air Force members. Examination of hazardous alcohol use
and IPV in active duty members is particularly timely given the
increased strain placed on military personnel and their families
with the two ongoing wars in Iraq and Afghanistan. Rates of IPV
among active duty members are similar to rates found in civilian
samples (Foran, Slep, Heyman, & U.S. Air Force Family Advo-
cacy Program, 2011b; Heyman & Neidig, 1999), but rates of heavy
drinking are higher (Ames & Cunradi, 2004–2005). Further, in the
context of two wars, current active duty members may be at risk
for relationship problems (e.g., McLeland, Sutton, & Schumm,
2008), which is an important risk factor and predictor of IPV in
both military (Slep et al., 2010) and civilian (Stith, Green, Smith,
& Ward, 2008) samples.
All analyses were conducted separately for men and women
given the limited literature and mixed findings on the role of
alcohol in women’s IPV (Foran & O’Leary, 2008; Leonard &
Eiden, 2007; Simmons, Lehmann, Cobb, 2008). The current study
addresses this limitation in the extant literature by examining
women’s alcohol problems and IPV in a representative sample of
over 8,000 women. Few past studies of moderation effects have
had sufficient power to appropriately test for moderation for men
or women (Whisman & McClelland, 2005).
Given the exploratory nature of the current study, all significant
moderator effects identified were then tested in a separate sample
to determine whether the results cross-validate. Moderation effects
are notoriously difficult to detect and are often sample-specific,
and thus, cross-validation is particularly important (Whisman &
McClelland, 2005). Further, generalizability of the moderator ef-
fects was also tested across different regions of the U.S., rural/
urban locations, parental status, and marital status. The large
representative sample is a significant strength of the current study,
providing the ability to test for cross-validation and generalizabil-
ity of the results.
Active duty (AD) members (N ? 128,950) of the United States
Air Force (AF) were invited to complete the 2006 Community
Assessment (CA), a biennial, anonymous survey conducted at 82
AF sites worldwide (a copy of the survey is available by contacting
the commander of the Air Force Family Advocacy Program: Major
Travis at firstname.lastname@example.org). Sampling for the CA occurred
concurrently with sampling for another large survey of AF mem-
bers; linear programming was used to minimize members’ selec-
tion into both surveys (Bigelow, 2007). The sample was represen-
tative at the base and AF levels. The response rate was 45% (N ?
54,543). Only those that were currently in relationships (n ?
42,744) were eligible for the current study.
The CA was administered by Caliber Associates between April
27 and June 23, 2006. Each AD member selected was e-mailed an
invitation containing the Web survey link and an access code at the
time of launch. Weekly e-mails were sent reminding members to
participate; each base also conducted community-wide campaigns
to encourage participation (e.g., posters and flyers on base an-
nouncing the study, postings in base newspapers). This study was
approved by the institutional review board; participation was vol-
untary, there was no compensation for participation, and data were
anonymous. The survey took approximately 45 min to 1 hr to
complete and could be completed across multiple sessions.
The CA included demographic questions, quantitative items
(e.g., typical number of hours worked each week), and scales
measuring individual, family, workplace, and community con-
structs. Measures were based on existing scales or developed based
on a working group with the help of outside consultants for the CA
in 2003. Following the 2003 CA, the structural validity, internal
consistency, response distributions, and convergent validity of the
scales were evaluated and minor modifications (e.g., items
dropped that performed poorly and new items added to enhance
scale properties) were made for the 2006 CA. The scales are
described briefly below. Scales were scored by averaging across
items with higher scores indicating higher levels of the construct
Partner physical aggression perpetration.
completed 15 items assessing the frequency of IPV acts against
their partners in the previous year from never ? 0 to more than 10
times ? 5 (slapping, slamming against a wall, burning or scalding
a partner, choking, biting a partner, scratching a partner, using a
weapon, pushing or shoving, twisting hair, hitting with an object
that could hurt, kicking, grabbing, throwing something that could
hurt, punching or hitting; Heyman & Slep, 2003). Acts were
generally similar to those in the Physical Assault subscale of the
revised Conflict Tactics Scale (CTS2; Straus, Hamby, Boney-
McCoy, & Sugarman, 1996) and have demonstrated content, cri-
terion, and convergent validity in military samples (Heyman,
Snarr, Slep, & U.S. Air Force Family Advocacy Program, 2011).
MODERATING RISK FOR INTIMATE PARTNER VIOLENCE
Respondents were also able to indicate “other” and write in a
particular act not listed and the frequency in which it occurred in
the past year. The write in responses were reviewed by two of the
authors and included as acts of IPV if they fit with IPV definitions
(Heyman & Slep, 2006). The majority of write-in responses (87%)
did not qualify as IPV and were very clear to differentiate (e.g.,
“argued” vs. “physically struck me”). Scores were computed by
averaging across the 15 items with higher scores indicating more
aggression perpetrated in the past year (? ? .82).
Hazardous alcohol use.
The 10-item Alcohol Use Disorders
Identification Test (AUDIT; Saunders, Aasland, Babor, de la Fu-
ente, & Grant, 1993) was administered. The AUDIT has been
validated against other clinical assessments and scoring criteria for
identifying hazardous alcohol use have been developed (Rumpf,
Hampke, Meyer, & John, 2002) (? ? .85). In addition to the
standard 10-item measure of hazardous alcohol use, we computed
an 8-item measure, removing two items that could potentially yield
a spurious relationship between hazardous alcohol use and IPV
(“During the last year, have you or someone else been injured as
a result of your drinking?” and “How often during the last year
have you had feelings of guilt or remorse because of your drink-
ing?”). We compared the correlation coefficients between the
original AUDIT and the 8-item AUDIT measure with IPV perpe-
tration for men and women. The correlation was slightly higher
with the 8-item AUDIT measure and IPV perpetration for men and
women, and thus, we opted to err on the conservative side of
reporting the association between hazardous drinking and IPV
perpetration and used the validated 10-item AUDIT measure that
had a slightly lower correlation with IPV. Although previous
studies have demonstrated cut-off criteria using the AUDIT (Ba-
bor, Biddle-Higgins, Saunders, & Monteiro, 2001), recommenda-
tions vary from study to study with different criteria being recom-
mended for men and women. To maximize variability and not
sacrifice potentially important information provided by the full
range of scores, the continuous severity measure of the AUDIT
(scores ranging from 0 to 40 with higher scores indicating more
hazardous alcohol use) was used for all analyses.
Demographic Maturation Factors
Years in the military.
long they had served in the military to date; responses ranged from
less than 1 year to 40 years or more.
Length of marriage.
Survey participants were asked how
long they had been married; responses ranged from less than 1
year to 40 years or more.
Family income/pay grade.
estimated from pay grade, years of military experience, and time
reportedly spent working at a second job. Spouse income was
estimated from employment status, demographic information, and
census figures (U.S. Census Bureau, 2006). These were combined
to yield family income.
Survey participants were asked how
The AD member’s income was
1992) of the widely used Center for Epidemiological Studies
Depression Scale (Radloff, 1977) were included (? ? .87) to
assess how many days participants experienced symptoms of de-
Seven items (Mirowsky & Ross,
pression over the past week on a 4-point scale from 1 ? none to
4 ? 5 to 7 days.
Perceived financial stress.
Social Change in Canada Survey (Krause & Baker, 1992), a
financial strain scale (Vinokur, Price, & Caplan, 1996), and a
measure of family economic pressure (Conger et al., 1993) (? ?
.90). Respondents were asked to rate their financial stress on a
Likert scale from 1 to 5 with higher scores indicating more stress.
Items assessed difficulty paying bills, living on total household
income, going without things really needed, and stress related to
finances over the past 12 months.
Nine items assessed participants’ ability to
cope with stress and to manage work and family demands. Six
items were drawn from the General Self-Efficacy Scale (Scholz,
Gutierrez, Sud, & Schwarzer, 2002) and were rated on a 4-point
scale from not at all true to exactly true, and three were developed
and tested in the 2003 CA for military samples (e.g., “how well do
manage your Air Force responsibilities and demands” rated on a
7-point scale from extremely poorly to extremely well). Internal
consistency of the 9-item scale was good (? ? .84).
Six items from the Short Form-8 Health
Survey (Ware, Kosinski, Dewey, & Gandek, 2001) inquired about
overall current health, pain, energy levels, sleep, diet, and exercise
patterns (? ? .65). Factor analyses supported one-factor solution
in both the 2003 CA and current study. The item loading the
highest and thus, most exemplar of the scale was “Overall, how
would you rate your health in the past four weeks?” on a 6-point
scale from very poor to excellent.
Five items assessed importance of
spirituality and involvement in and satisfaction with a religious
faith (e.g., “How important is spirituality/faith in your life?”;
“How often do you attend a local church, chapel, synagogue,
mosque, or religious or spiritual community?”). These items were
developed and evaluated in the 2003 CA. Factor analyses from
both the 2003 and 2006 CAs supported a one-factor solution. Items
were rated on a 5-point scale with scores ranging from 1 to 6.
Internal consistency in the current study was ? ? .64.
Five items were drawn from the
Marriage Index (Norton, 1983) were administered. Three items
asked AD members to rate how much they agree with statements
about their relationship on a 6-point scale from 1 to 6. The fourth
item asked AD members to rate their level of happiness in their
relationship, all things considered, on a scale from 1 to 9. Scores
across the 4 items were averaged across items and internal con-
sistency was excellent (? ? .92).
Career support from significant other.
questions asked about partners’ levels of understanding and sup-
port for respondents’ AF jobs and career. Two of the items (“How
does your partner feel about your making a career in the Air
Force?”; “How does your partner feel about your being in the Air
Force now”) were rated on a 6-point scale from extremely unsup-
portive to extremely supportive. The third item, “My partner is
understanding of the demands of my Air Force job” was rated on
a 6-point scale from almost never to most of the time. Scored were
average across items and ranged from 1 to 6 (? ? .77).
Four items from the Quality of
FORAN ET AL.
together as a team, keep positive perspectives during rough times,
and directly confront problems or challenges (e.g., “When my
family has to cooperate to accomplish something, we work to-
gether as a team” rated on a 6-point scale from almost never to
almost always). Scores ranged from 1 to 6 (? ? .89).
Parent-child relationship satisfaction.
fied from the Relationship Satisfaction Scale (Simons, Beaman,
Conger, & Chao, 1993) were used (? ? .68). Items were rated on
a 6-point scale with scores ranging from 1 to 6 (e.g., “All things
considered, how satisfied are you with your relationship with your
child(ren)?” on a scale from very dissatisfied to very satisfied).
Three items assessed family ability to work
Three items modi-
Satisfaction with the AF.
with the AF as a way of life for AD members and their families
(? ? .72). Scores were rated on a 6-point Likert scale (1 to 6) with
higher scores indicating higher levels of satisfaction (e.g., “How
satisfied are you with the Air Force as a way of life?”; “My
husband/wife and I are successful at coping with demands of being
an Air Force family”).
Work group cohesion/preparedness.
adapted from an Army Family Research Program individual read-
iness measure (U.S. Army Community & Support Center, 1989) to
assess cohesion (e.g., working together as a team, having high
morale) on a 6-point scale from strongly disagree to strongly
agree; ? ? .88.
Three items assessed the quality of
relationships with coworkers, supervisors, and supervisees on a
5-point scale from 1 ? extremely poor to 5 ? excellent (? ? .85).
Support from leadership.
Seventeen items assessed the level
of support received from various AF leaders and the preparation
received before and support received after a deployment on a
6-point scale (scores ranging from 1 to 6) (? ? .94).
Five items assessed satisfaction
Six questions were
Soldier and Family Survey (Research Triangle Institute, 1990),
assessed availability of tangible support from varied sources (? ?
.95). For example, AD members rated whether there would be
someone besides their spouse at their current location who would
help with daily chores if they were sick on a scale from 1 ? almost
never to 6 ? almost always.
Six items assessed perceived safety in
their residence, neighborhood, surrounding civilian area, and on
their AF base from very unsafe to very safe from crime and
violence (Princeton Survey Research Associates, 1999). Two ad-
ditional items assessed safety of children. The two latter items
were reverse-scored, so that higher scores indicate more perceived
safety for the measure with scores ranging from 1 to 6 (? ? .77).
Thirteen questions assessed the
availability, quality, and affordability of community resources
(e.g., housing, health care, child care). Three questions assessed
perceptions of the civilian community. Items were rated on a
6-point scale and ranged from 1 to 6 (? ? .89).
Community support for youth.
tunities for youth to use their time well and support for youth by
leadership on a 6-point scale ranging from 1 to 6 (? ? .92).
Five questions, adapted from the 1989 Army
Three items assessed oppor-
senses of shared mission, teamwork, unity, and connectedness in
the community on a 6-point scale ranging from 1 to 6; Bowen,
Mancini, Martin, Ware, & Nelson, 2003; ? ? .95).
Support from neighbors.
Six items assessed support from
people in the neighborhood, as well as one question (adapted from
the Social Capital Benchmark Survey, 2000) asking how often
neighbors get together to fix or improve something. Items were
rated on a 6-point scale and scores ranged from 1 to 6 (? ? .93).
Of the AD members in relationships (n ?
42,744) who participated in the survey, 79.3% completed the
survey (n ? 33,886) and had very little missing data (3.3% item
level data points, on average). The remainder of participants
(20.7%; n ? 8,858) did not complete the survey to the end, but
completed at least the first part of the survey including basic
demographic questions. Multiple imputation was used to address
missing data concerns for participants who did not fully complete
all survey questions, (Allison, 2001). Multiple imputation was
conducted separately by gender using IVEware (Raghunathan,
Solenberger, & Van Hoewyk, 2002). Fifty iterations of multiple
imputation were conducted, with every 10th resulting dataset
saved, and all values that had been imputed for legitimately “not
applicable” data points were then removed from the five resulting
Data was weighted prior to analyses to match the
U.S. military population distributions on demographic character-
istics. Raking, also known as sample balancing (Kalton, 1983),
was used to produce poststratification weights, adjusting for (a)
oversampling at smaller bases, (b) differential nonresponse, and
(c) complex sampling design. Raking is a commonly used tech-
nique in development of survey weights; it uses iterative propor-
tional fitting to match a sample’s marginal distributions to known
population margins. WesVar 4.2 was used to rake on marital
status, base, ethnicity, religion, gender, rank, and job type. Ex-
treme weights (i.e., ? 4) were trimmed (i.e., set to 4) (Potter,
Twenty-one items assessed members’
Data Analytic Strategy
The sample was randomly split into a development sample and
validation sample to permit cross-validation analyses. This re-
sulted in a development sample of 17,247 men and 4,016 women
and a validation sample of 17,466 men and 4,015 women. There
were no statistically significant differences between the develop-
ment and validation samples on any relevant study variables (all
ps ? .01). Further adjustments to the sample were made to max-
imize power for analyses. Although IPV perpetration was mea-
sured on a continuum of severity, a large majority of the sample
did not self-report any acts of IPV perpetration (over 90% of the
sample) and it was necessary to adjust for the skewed distribution
prior to analyses. In fact, the percentages of individuals who did
not report any IPV in the sample were as follows: development
sample ? 94.6% men and 92.0% women, validation sample ?
95.0% men and 91.7% women. Rather than using an excessively
large sample of individuals that did not report any acts of IPV
perpetration, we randomly selected from those that did not report
any acts of IPV perpetration in the past year using a ratio of 1 to
MODERATING RISK FOR INTIMATE PARTNER VIOLENCE
4 IPV perpetrators to those that did report any acts of IPV, as is
typically done in large scale survey studies (Graubard & Korn,
1996). This resulted in an analytical development sample of 4,553
men and 1,442 women that did not report any IPV perpetration.
The validation sample consisted of 4,584 men and 1,437 women
that did not report any IPV perpetration. The randomly selected
individuals that did not endorse any IPV perpetration did not
significantly differ from the rest of the nonaggressive samples on
any of the study variables including hazardous alcohol use (ps ?
.01). This yielded a total sample of 5,397 men and 1,866 women
in the development sample and 5,640 men and 1,902 women in the
validation sample for the analyses with IPV. Demographic infor-
mation is presented in Table 1 for the total sample.
Linear regression was used to test the two-way interactions
between potential moderators of the continuous measure2of haz-
ardous alcohol use in predicting IPV perpetration. All variables
were standardized prior to analyses. Twenty-three separate inter-
actions were tested, of which seven were statistically significant
for men. None of the interactions between hazardous alcohol use
and hypothesized protective factors were statistically significant in
predicting women’s IPV; therefore, all results presented are for
men only (results for women are available on request).
Statistically significant interaction effects (with their concomi-
tant main effects) from the development sample are presented in
Table 2 and the intercorrelations among these predictors is shown
in Table 3. Graphical examination and simple slope analyses
revealed a similar pattern across all interactions. High levels of the
moderator variable reduced the alcohol–aggression association. In
other words, hazardous alcohol use had the weakest association
with IPV at high levels of the protective factor. This pattern of
protection is illustrated graphically, using the interaction between
hazardous alcohol use and parent–child relations (see Figure 1)
and hazardous alcohol use and years in the military as examples
(see Figure 2). Simple slopes at high and low levels of the
protective factors are reported in Table 4. As mentioned above,
variables were standardized prior to analyses, and thus bs reported
in the table can be interpreted as standardized regression coeffi-
cients. Although hazardous alcohol use is still a significant pre-
dictor of IPV when the protective factor is high, the effect size is
reduced. In the case of parent–child relations as a protective factor,
the regression coefficient of hazardous alcohol use predicting IPV
at low levels of parent–child relations is .21, t ? 5.38, p ? .000,
whereas the regression coefficient at high levels of parent–child
relations is only .08, t ? 2.08, p ? .038. Thus, although the impact
of hazardous alcohol use is not completely ameliorated, it is
considerably weakened when parent–child relations is high.
Each of the interaction models presented in Table 2 was tested
in the cross-validation sample using multigroup analysis in Mplus
Version 5.1 (Muthe ´n & Muthe ´n, 2007). The final models were
tested with the regression coefficients and standard errors con-
strained to be equal across the development and validation samples
(total n ? 11,037), then run a second time with these constraints
released. The chi-square difference test of nested models was
used to evaluate differences in model fit across the development
and validation groups. Chi-square difference results were calcu-
lated separately for each of the five imputed datasets and then
combined across datasets using the formula given by Rubin and
Schenker (1991). The cross-validation results for all models are
reported in Table 5. A nonsignificant F indicates that constraining
the parameter estimates to be equal across the development and
validation samples did not significantly worsen model fit. As can
be seen in the table, all the models cross-validated in the hold-out
sample with one exception. The parameter estimates for the inter-
action between spirituality/religiosity and alcohol problems was
significantly different across the development and validation sam-
ples, F ? 4.96, p ? .008. This interaction approached significance
in the validation sample (b ? –0.06, SE ? .03, t ? –1.92, p ?
The models that cross-validated in the validation sample (see
Table 5) were tested to see whether the results generalized across
2Although we elected to use a continuous measure of hazardous drink-
ing to maximize variance and because of concerns about selecting the most
appropriate cutoff given different studies suggesting different criteria (e.g.,
Allen, Litten, Fertig, & Babor, 1997; Berner, Kriston, Bentele, & Harter,
2007; Reinert & Allen, 2007), we did run the analyses with the AUDIT
dichotomized into 1 ? hazardous alcohol use and 0 ? nonhazardous use
based on the WHO recommended cutoff of 8 (Babor et al. 2001) in the
development sample and found similar findings to those presented with a
few minor differences. For men, the interaction between spirituality/
religiousity was not significant (b ? –.14, SE? .08, t ? –1.81, p ? .071)
when using the AUDIT dichotomized measure. This finding is not surpris-
ing, given this interaction did not replicate in the validation sample anal-
yses using the continuous measure. The interaction between community
safety and the AUDIT dichotomized measure also was not significant in
the development sample (b ? –.10, SE ? .07, t ? –1.33, p ? .183). All
other significant interactions with the continuous AUDIT measure were the
same for men (relationship satisfaction, family income, years in the mili-
tary, marital length, and parent-child relations). Results were similar for
women; there were no significant interactions.
Unweighted Participant Characteristics
n (% total)
n (% total)
Junior enlisted (E1 to E4)
Junior NCO’s (E5 to E6)
Senior NCO’s (E7 to E9)
Junior officers (O1 to O3)
Senior officers (? O4)
Single, in a relationship
n ? 11,037 men and 3,768 women. NCO ? Noncommissioned
FORAN ET AL.
geographical regions (Northeastern United States, Midwestern
United States, Southern United States, Western United States,
Asia, and Europe), different city sizes (urban area of 1,000,000 or
more, urban area of 250,000 to 1,000,000, urban area of less than
250,000, rural area of 20,000 or more, rural area of 20,000 or less),
married/unmarried men, and fathers/nonfathers. The development
and validation samples were combined for these analyses, resulting
in 11,037 men for generalizability analyses of region, marital
status, and parental status. The generalizability analyses for city
size combined across development and validation subsamples
were with 9,200 men living in the United States (comparable
information on city size was not available for locations outside the
United States). Mplus Version 5.1 (Muthe ´n & Muthe ´n, 2007) was
again used to conduct multigroup analyses following the same
procedure used for the cross-validation analyses described above.
The unconstrained models were not a significantly better fit than
the constrained models (the chi-square difference test was not
significant, p ? .05) for any of the models tested for region,
marital status, or city size. Hence, this indicates that all the
cross-validated models presented in Tables 4 generalized across
region, marital status, and city size.
In contrast, models were different for fathers and nonfathers for
all significant interactions. Results indicated that risk for IPV was
greater among nonfathers. A significant interaction between haz-
ardous alcohol use and parental status also contributed to the
difference between fathers and nonfathers across the models (B ?
–.165, SE ? .043, t ? –3.78, p ? .001), in that alcohol problems
had stronger associations with IPV among nonfathers than fathers.
A closer examination of this finding indicated that there were
significant three-way interactions between parental status and haz-
Moderators of Men’s Alcohol Problems-IPV Link
Development sample model results
AUDIT ? Relationship satisfaction
AUDIT ? Parent–child relations
Family income/pay grade
AUDIT ? Family income/pay grade
AUDIT ? Marital length
Years in the military
AUDIT ? Years in the military
AUDIT ? Community safety
AUDIT ? Spirituality/religiosity
length: n ? 4,578 men) or parents (parent child relations: n ? 3,353 men).
n ? 5,397 men for all variables except those that were only answerable by married individuals (marital
Intercorrelations Among Significant Interaction Variables For Men
1. Relationship satisfaction
2. Parent–child relations
3. Family income (monthly $)
4. Marital length
5. Years in the military
6. Community safety
8. Hazardous alcohol use
child relations: n ? 3,353 men).
?p ? .001.
n ? 5,397 men for all variables except those that were only answerable by married individuals (marital length: n ? 4,578 men) or parents (parent
MODERATING RISK FOR INTIMATE PARTNER VIOLENCE
ardous alcohol use with relationship satisfaction and marital length
(but not family income, years in the military, or community
safety). As can be seen in Figure 3, a test of the simple slopes
indicated that the association between alcohol and IPV was stron-
gest among men that were not fathers and who had been married
for less time (ps ? .001). The three-way interaction involving
parental status, relationship satisfaction, and hazardous alcohol use
had a somewhat different pattern (see Figure 4). Nonfathers who
reported low relationship satisfaction were at highest risk for IPV;
however, the strength of the alcohol-IPV association in this group
was not significantly different than among men with high relation-
ship satisfaction (whether parents or not; ps for all simple slopes
comparisons ? .05). Among fathers who reported low relationship
satisfaction, on the other hand, hazardous alcohol use did not
significantly increase risk for aggression, and the slope for this
group significantly differed from the other three slopes (ps ?
.001). Thus, being a parent buffered the alcohol-IPV association
among men with low relationship satisfaction.
Alcohol abuse is a well-established risk factor for men’s IPV,
with dozens of studies demonstrating the association (for reviews,
see Foran & O’Leary, 2008; Leonard, 2005). The current study
takes a next step in understanding the alcohol-IPV link by exam-
ining what factors buffer this association in a more systematic and
broader way than has been done in the past. Specifically, past
moderation studies have been limited by (a) examining only a few
moderators at a time, (b) having limited power to detect interac-
tions effects, (c) focusing almost exclusively on men, and (d)
lacking cross-validation of the findings (Whisman & McClelland,
2005). The current study addresses these limitations of the litera-
ture by testing individual, family, workplace, community, and
developmental factors as moderators of the alcohol-IPV link in a
large, representative sample of male and female active duty AF
members. Family, community, and developmental factors were
identified as moderators of men’s alcohol-IPV link, and these
results cross-validated in an independent hold-out sample.
deviation below and above the mean as indicated by “low” and “high,” respectively.
Interaction of hazardous alcohol use and parent-child relations in predicting IPV. Graphed 1 standard
in predicting IPV. Graphed 1 standard deviation below and above the mean
as indicated by “low” and “high,” respectively.
Interaction of hazardous alcohol use and years in the military
Men’s Hazardous Alcohol Use Predicting IPV at High and Low
Levels of Significant Protective Factors
Levels of protective factors
Years in the military
answerable by married individuals (marital length: n ? 4,578 men) or
parents (parent child relations: n ? 3,353 men). For each protective factor,
“low” and “high” indicate 1 standard deviation below the mean and 1
standard deviation above the mean, respectively, The reader is referred to
Table 3 for means and standard deviations of the protective factors.
n ? 5,397 men for all variables except those that were only
FORAN ET AL.
In contrast, there were no significant interactions in predicting
the association between hazardous drinking and women’s IPV
despite the broad range of risk and protective factors assessed and
the large sample size. Other studies have also not found significant
moderator effects of women’s hazardous alcohol use and IPV and
have suggested that it may be due to low power from too few
women drinking heavily in the samples (Schumacher et al., 2008).
However, this explanation cannot adequately explain our results,
because our sample was quite large, with an ample number of
women reporting hazardous drinking in the development sample
(n ? 204 with AUDIT scores ? ? 8, n ? 265 with AUDIT
scores ? ? 7, n ? 371 with AUDIT scores ? ? 6). In addition,
the development sample of aggressive women was sufficiently
large with 323 women endorsing at least one act of IPV perpetra-
tion in the past year. Alternatively, theoretical models of alcohol-
related IPV, having been developed based on men’s samples, may
not generalize to women. In the United States, there is less social
stigma against women’s IPV, in part because it is less likely to
result in serious injury (Cantos, Neidig, & O’Leary, 1994). Hence,
women who engage in IPV may not require the disinhibiting
effects of alcohol to aggress, due to weaker social prohibitions on
female-to-male IPV. Although limited, there is some support that
the association between women’s alcohol use and IPV perpetration
is smaller than for men (Foran & O’Leary, 2008) and is no longer
significant once men’s drinking is taken into account (Leonard &
Senchak, 1996). This finding of a larger effect size for men is also
consistent with experimental studies of alcohol intoxication and
laboratory aggression (Giancola, et al., 2009). More careful atten-
tion to gender differences in models of alcohol-related aggression
is clearly needed.
For men, two family variables (relationship satisfaction, parent–
child satisfaction), one community variable (community safety),
and three developmental variables (years in the military, marital
length, and family income/pay grade) significantly moderated the
alcohol-IPV link. No individual or workplace variables were sig-
nificant moderators. These results replicate past studies in which
high levels of marital satisfaction buffered the alcohol-IPV asso-
ciation (Leonard & Senchak, 1993; Margolin et al., 1998), but
contradict a study by Cunradi (2007) that reported community
Cross-Validation and Generalizability of Interactions of Alcohol and Protective Factors in Predicting Men’s IPV
Interactions of AUDIT with:Cross-validation df ? 3City sizecdf ? 12
Region df ? 15Parental status df ? 3Marital status df ? 3
F ? 1.16
F ? .95
F ? 1.33
F ? 1.06
F ? .33
F ? .99
F ? 4.96??
F ? 1.07
F ? .73
F ? 1.31
F ? 1.37
F ? .96
F ? 1.37
F ? .59
F ? .87
F ? .58
F ? .78
F ? .91
F ? .80
F ? 5.61??
F ? 3.89??
F ? 7.25??
F ? 3.61?
F ? 4.58??
F ? .33
F ? .25
F ? 2.13
F ? .19
F ? .83
Family income/pay grade
an ? 6,951 fathers.
?p ? .05.
F ? Rubin & Schenker’s (1991) F. ?2? mean chi-square difference test results across 5 imputed datasets. n ? 11,037.
bn ? 9366 married men.
??p ? .01.
cn ? 9,200 for all analyses of city size except for fathers (n ? 5,795) and married men (n ? 7,782).
rental status in predicting men’s IPV.
Interaction of hazardous alcohol use, marital length, and pa-
and parental status in predicting men’s IPV.
Interaction of hazardous alcohol use, relationship satisfaction,
MODERATING RISK FOR INTIMATE PARTNER VIOLENCE
safety was not a significant moderator of the men’s alcohol and
IPV association. The Cunradi study (2007) differed from the
present study in terms of the way that alcohol and IPV were
operationalized, which may explain the different findings (Foran &
O’Leary, 2008). Specifically, Cunradi (2007) examined binge
drinking, whereas we used a measure of hazardous drinking. IPV
was measured with one interview question in the Cunradi (2007)
study, whereas the present study utilized an extensive, validated,
anonymous survey questionnaire (Heyman & Slep, 2006).
In terms of the other moderators examined, there is limited
literature to which the current findings can be compared, given that
the majority of variables have not been previously tested as
alcohol-IPV moderators. This was the first study to examine work-
place factors as moderators of the alcohol-IPV link. Results indi-
cate that although workplace factors were bivariately related to
both alcohol and IPV (Foran et al., 2011a; Slep et al., 2010), they
were not significant moderators of risk. We also included several
individual factors theorized to reflect deficits in self-regulation and
coping abilities (e.g., personal coping, perceived financial stress,
depressive symptoms, physical health), based on theoretical mod-
els suggesting that reduced self-regulation may moderate the risk
for alcohol-related aggression (Wills et al., 2002). However, none
were found to significantly moderate the alcohol-IPV link.
Our results provide some support for a relational and develop-
mental perspective to understanding alcohol-related aggression. As
mentioned previously, the finding that high relationship satisfac-
tion is a protective factor adds to a growing body of literature that
has implicated relationship functioning as a moderator of the
alcohol-IPV link (Leonard & Blane, 1992; Leonard & Senchak,
1993; Margolin et al., 1998). This study identified both relation-
ship and parent–child satisfaction as protective factors. Consistent
with the threshold model, both relationship and parental dissatis-
faction may increase the risk for verbal conflict in the home and
place an individual closer to the threshold for IPV, which is
subsequently crossed when combined with hazardous drinking
(Fals-Stewart, Leonard, & Birchler, 2005). Alternatively stated,
high relationship and parenting satisfaction appear to protect
against the effect of alcohol on IPV. Our results imply that if there
is little conflict in the home, a man is unlikely to perpetrate IPV
even if he tends to drink heavily. In terms of treatment implica-
tions, these findings highlight the importance of assessing both
marital and parent–child conflict in treating individuals or couples
at risk for IPV. Interventions should help educate individuals and
couples about situations in which they may be at most risk for
violent behavior and help develop safety plans to prevent these
types of high-risk situations.
However, only modest support for the relational perspective of
alcohol and IPV can be offered by the current results, because
moderation effects were not found for other relationship variables
examined (family coping, career support from significant other,
and spouse deployment support). The two relationship support
factors both referred to work-related support. Findings may have
been different if other types of support (e.g., emotional support,
child-care support) were assessed. Family coping measured the
ability to manage problems as a family; theoretically, one would
expect low family coping to also increase risk for family conflict,
lowering the threshold for IPV as do relationship and parenting
satisfaction. It will be important for future studies to continue to
clarify precisely how relationship functioning impacts alcohol-
Our results highlight the importance of developmental factors in
understanding the alcohol-IPV association (Sher, Grekin, & Wil-
liams, 2005; Zucker et al., 1995; Leonard & Mudar, 2003;
O’Leary, 1999), as all three of the developmental factors studied
cross-validated as significant moderators. Across developmental
factors, the association between alcohol and IPV weakened with
maturation (older years, longer marriage, and greater family in-
come/pay grade). In addition, being a parent also weakened the
alcohol-IPV link. These findings replicate Heyman and col-
leagues’ (1995) longitudinal study of marriage, in which the as-
sociation between alcohol and aggression was found to weaken
over the course of marriage. Hence, addressing hazardous alcohol
use as part of an IPV prevention program may be most effective
among younger men at earlier developmental life stages.
We tested whether results would generalize across marital status
and parental status, as these mark important developmental tran-
sitions that plausibly could be associated with different risk pat-
terns. Results generalized across married and single men but not
across fathers and nonfathers. The transition to marriage has been
theorized to be an important turning point in understanding alcohol
and aggression risk (Leonard & Mudar, 2003). It is possible that to
detect the effects of the transition to marriage, longitudinal data is
needed. However, our finding that parental status did indeed
predict differential risk associations between alcohol and aggres-
sion suggests that the transition to parenthood may be the more
critical transition variable in altering buffering effects. Another
possibility is that marriage may function differently among AD
military personnel compared with civilians. AD personnel tend to
marry earlier than their civilian counterparts, and this may explain
the lack of differences regarding the alcohol-IPV link between
married and single men in the current sample (Hogan & Furst
Finally, there are several limitations of the current study that
deserve mention. The response rate was 45% of those invited to
participant and it is possible that those with IPV and/or hazardous
drinking may have been less likely to participate. Our sample was
cross-sectional and it directionality cannot be inferred from the
current study; it is plausible that hazardous drinking and IPV have
bidirectional relationships, such as IPV, increases risk for subse-
quent hazardous drinking, as well.
Our emphasis was on selecting moderators that were either (a)
potentially modifiable or (b) useful in identifying high-risk groups
that could be targeted in interventions. A range of potential mod-
erators were tested, but not all key risk factors related to IPV were
included, and thus, results may have been more supportive of
individual factors as moderators if factors such as aggressive
personality and anger control problems were included (e.g., Den-
son et al., 2009; Foran & O’Leary, 2007; Giancola, 2002a). Our
measures of work and community functioning were based on
individual perceptions, and results may have differed if group level
responses were assessed instead. In addition, a few of the measures
had low internal consistency and results will need to be replicated
with more extensive measures of each construct. Demographics,
such as education and age, were also not assessed (although
proxies for age and maturation were included, i.e., years in the
military). Deployment-related factors, such as combat exposure,
multiple deployments, and posttraumatic stress disorder (PTSD)
FORAN ET AL.
symptoms, may also impact the hazardous drinking and IPV rela-
tionship. Military personnel with elevated PTSD symptoms report
higher rates of both alcohol abuse (Gewirtz, Polusny, DeGarmo,
Khaylis, & Erbes, 2010) and IPV (Taft, Watkins, Stafford, Street,
& Monson, 2011). It will be important in future studies to replicate
these findings with other types of risk factors of IPV and examine
these relationships longitudinally.
In addition, all of our measures were based on self-report
including the measure of IPV perpetration. It is well documented
that self-report measures of IPV underestimate rates of IPV (Hey-
man & Schlee, 1997) and measures using both members of the
couple’s reports are recommended. Finally, the current findings
highlight types of moderators (developmental and relationship)
that may be relevant to understanding the alcohol-IPV link and
interrelations between these types of moderators in terms of three-
way interactions with parental status. Further work should consider
a more in depth examination of how these types of moderators
combine to amplify or buffer risk for IPV (three- and four-way
Summary and Conclusions
This study adds to a growing body of literature identifying
moderators of alcohol and IPV association. As suggested by the-
oretical models, not all individuals who drink are aggressive, and
different risk and protective factors may increase their threshold
for risk when drinking (e.g., multiple threshold model, Fals-
Stewart et al., 2005). This study identified a set of variables—
specifically, developmental and relationship factors—that may
moderate the alcohol-aggression association for men. Other factors
(e.g., financial stress, depressive symptoms) that have significant
main effect associations with IPV (Slep, Foran, Heyman, & Snarr,
2010) were not moderators of the alcohol-IPV connection in the
present study. Modifiable factors operating at different levels that
buffer the link between alcohol and IPV have tremendous potential
as targets of prevention and intervention. Although causality can-
not be inferred with cross-sectional data, these results suggest that
developmental and relationship factors may be particularly impor-
tant to understanding alcohol-related IPV and may be a fruitful
area for future empirical and theoretical work.
Allen, J. P., Litten, R. Z., Fertig, J. B., & Babor, T. F. (1997). A review of
research on the Alcohol Use Disorders Identification Test (AUDIT).
Alcoholism: Clinical and Experimental Research, 21, 613–619. doi:
Allison, P. D. (2001). Missing data (Sage University Papers Series on
Quantitative Applications in the Social Sciences No. 07–136). Thousand
Oaks, CA: Sage Publications.
Ames, G. M., & Cunradi, C. (2004–2005). Alcohol use and preventing
alcohol-related problems among young adults in the military. Alcohol
Research & Health, 28, 252–257.
Babor, T. F., Biddle-Higgins, J. C., Saunders, J. B., & Monteiro, M. G.
(2001). AUDIT: The Alcohol Use Disorders Identification Test: Guide-
lines for use in primary health care. Geneva, Switzerland: World Health
Berner, M. M., Kriston, L., Bentele, M., & Harter, M. (2007). The alcohol
use disorders identification test for detecting at-risk drinking: A system-
atic review and meta-analysis. Journal of Studies on Alcohol and Drugs,
Bigelow, J. H. (2007). Using linear programming to design samples for a
complex survey. Arlington, VA: RAND Corporation.
Bowen, G. L., Mancini, J. A., Martin, J. A., Ware, W. B., & Nelson, J. P.
(2003). Promoting the adaptation of military families: An empirical test
of a community practice model. Family Relations: An Interdisciplinary
Journal of Applied Family Studies, 52, 33–44. doi:10.1111/j.1741-
Cantos, A. L., Neidig, P. H., & O’Leary, K. D. (1994). Injuries of men and
women in a treatment program for domestic violence. Journal of Family
Violence, 9, 113–124. doi:10.1007/BF01531958
Chassin, L., Pitts, S. C., & Prost, J. (2002). Binge drinking trajectories from
adolescence to emerging adulthood in a high-risk sample: Predictors and
substance abuse outcomes. Journal of Consulting and Clinical Psychol-
ogy, 70, 67–78. doi:10.1037/0022-006X.70.1.67
Conger, R. D., Conger, K. J., Elder, G. H., Lorenz, F. O., Simons, R. L.,
& Whitbeck, L. B. (1993). Family economic stress and adjustment of
early adolescent girls. Developmental Psychology, 29, 206–219. doi:
Cunradi, C. B. (2007). Drinking level, neighborhood social disorder, and
mutual intimate partner violence. Alcoholism: Clinical and Experimen-
tal Research, 31, 1012–1019. doi:10.1111/j.1530-0277.2007.00382.x
Cunradi, C. B., Todd, M., Duke, M., & Ames, G. (2009). Problem drink-
ing, unemployment, and intimate partner violence among a sample of
construction industry workers and their partners. Journal of Family
Violence, 24, 63–74. doi:10.1007/s10896-008-9209-0
Denson, T. F., White, A. J., & Warburton, W. A. (2009). Trait displaced
aggression and psychopathy differentially moderate the effects of acute
alcohol intoxication and rumination on triggered displaced aggression.
Journal of Research in Personality, 43, 673–681. doi:10.1016/j.jrp
Eckhardt, C. I. (2007). Effects of alcohol intoxication on anger experience
and expression among partner assaultive men. Journal of Consulting and
Clinical Psychology, 75, 61–71. doi:10.1037/0022-006X.75.1.61
Eckhardt, C. I. (2008). Effects of alcohol intoxication and aggressivity on
aggressive verbalizations during anger arousal. Aggressive Behavior, 34,
Fals-Stewart, W., Leonard, K. E., & Birchler, G. (2005). The occurrence of
male-to-female intimate partner violence on days of men’s drinking: The
moderating effects of antisocial personality disorder. Journal of Con-
sulting and Clinical Psychology, 73, 239–248. doi:10.1037/0022-006X
Ferrer, V., Bosch, E., Garcia, E., Manassero, M. A., & Gili, M. (2004).
Meta-analytic study of differential characteristics between batterers and
non-batterers: The case of psychopathology and consumption of alcohol
and drugs. Psykhe: Revista de la Escuela de Psicologı ´a, 13, 141–156.
Foran, H. M., & O’Leary, K. D. (2007). Problem drinking, jealousy, and
anger control: Variables predicting aggression against a partner. Journal
of Family Violence, 23, 141–148. doi:10.1007/s10896-007-9136-5
Foran, H. M., & O’Leary, K. D. (2008). Alcohol and intimate partner
violence: A meta-analytic review. Clinical Psychology Review, 28,
Foran, H. M., Slep, A. M. S., Heyman, R. E., & U.S. Air Force Family
Advocacy Program (2011a). Hazardous alcohol use among active duty
Air Force personnel: Identifying unique risk and protective factors.
Psychology of Addictive Behaviors, 25, 28–40. doi:10.1037/a0020748
Foran, H. M., Slep, A. M. S., Heyman, R. E., & U.S. Air Force Family
Advocacy Program (2011b). Prevalences of intimate partner violence in
a representative military sample. Journal of Consulting and Clinical
Psychology, 79, 391–397. doi:10.1037/a0022962
Gewirtz, A. H., Polusny, M. A., DeGarmo, D. S., Khaylis, A., & Erbes,
C. R. (2010). Posttraumatic stress symptoms among National Guard
soldiers deployed to Iraq: Associations with parenting behaviors and
couple adjustment. Journal of Consulting and Clinical Psychology, 78,
MODERATING RISK FOR INTIMATE PARTNER VIOLENCE
Giancola, P. R. (2002a). Alcohol-related aggression in men and women:
The influence of dispositional aggressivity. Journal of Studies on Alco-
hol, 63, 696–708.
Giancola, P. R. (2002b). The influence of trait anger on the alcohol-
aggression relation in men and women. Alcoholism: Clinical and Ex-
perimental Research, 26, 1350–1358. doi:10.1111/j.1530-0277.2002
Giancola, P. R. (2003). The moderating effects of dispositional empathy on
alcohol-related aggression in men and women. Journal of Abnormal
Psychology, 112, 275–281. doi:10.1037/0021-843X.112.2.275
Giancola, P. R., Helton, E. L., Osborne, A. B., Terry, M. K., Fuss, A. M.,
& Westerfield, J. A. (2002). The effects of alcohol and provocation on
aggressive behavior in men and women. Journal of Studies on Alcohol,
Giancola, P. R., Levinson, C. A., Corman, M. D., Godlaski, A. J., Morris,
D. H., Phillips, J. P., & Holt, J. C. D. (2009). Men and women, alcohol
and aggression. Experimental and Clinical Psychopharmacology, 17,
Giancola, P. R., Saucier, D. A., & Gussler-Burkhardt, N. L. (2003). The
effects of affective, behavioral, and cognitive components of trait anger
on the alcohol-aggression relation. Alcoholism: Clinical and Experimen-
tal Research, 27, 1944–1954. doi:10.1097/01.ALC.0000102414
Graubard, B. I., & Korn, E. L. (1996). Survey inference for subpopulations.
American Journal of Epidemiology, 144, 102–106.
Grekin, E. R., Sher, K. J., & Larkins, J. M. (2004). The role of behavioral
undercontrol in the relation between alcohol use and IPV. Journal of
Studies on Alcohol, 65, 658–662.
Heyman, R. E., & Neidig, P. H. (1999). A comparison of spousal aggres-
sion prevalence rates in U.S. Army and civilian representative samples.
Journal of Consulting and Clinical Psychology, 67, 239–242. doi:
Heyman, R. E., O’Leary, K. D., & Jouriles, E. N. (1995). Alcohol and
aggressive personality styles: Potentiators of serious physical aggression
against wives? Journal of Family Psychology, 9, 44–57. doi:10.1037/
Heyman, R. E., & Schlee, K. A. (1997). Toward a better estimate of the
prevalence of partner abuse: Adjusting rates based on the sensitivity of
the Conflict Tactics Scale. Journal of Family Psychology, 11, 332–338.
Heyman, R. E., & Slep, A. M. S. (2003). Active duty member and spouse
feedback on plans for the 2003 AF community assessment and sensitive
behavior supplement. (Report to the U.S. Air Force Family Advocacy
Program). Stony Brook, NY: Stony Brook University.
Heyman, R. E., & Slep, A. M. S. (2006). Creating and field-testing
diagnostic criteria for partner and child maltreatment. Journal of Family
Psychology, 20, 397–408. doi:10.1037/0893-3220.127.116.117
Heyman, R. E., Snarr, J. D., Slep, A. M. S., & U.S. Air Force Family
Advocacy Program. (2011). Validity of a Measure of Clinical Significant
Intimate Partner Violence and Child Abuse. Manuscript submitted for
Hogan, P. F., & Furst Seifert, R. (2010). Marriage and the military:
Evidence that those who serve marry earlier and divorce earlier. Armed
Forces and Society, 36, 420–438. doi:10.1177/0095327X09351228
Kalton, G. (1983). Introduction to survey sampling. Beverly Hills, CA:
Kaufman-Kantor, K. G., & Straus, M. A. (1990). The “Drunken bum”
theory of wife beating. In M. A. Straus & R. J. Gelles (Eds.), Physical
violence in American families: Risk factors and adaptations to violence
in 8,145 families (pp. 203–224). New Brunswick, NJ: Transaction Pub-
Krause, N., & Baker, E. (1992). Financial strain, economic values, and
somatic symptoms in later life. Psychology and Aging, 7, 4–14. doi:
Leonard, K. E. (2005). Alcohol and intimate partner violence: When can
we say that heavy drinking is a contributing cause of violence? Addic-
tion, 100, 422–425. doi:10.1111/j.1360-0443.2005.00994.x
Leonard, K. E., & Blane, H. T. (1992). Alcohol and marital aggression in
a national sample of young men. Journal of Interpersonal Violence, 7,
Leonard, K. E., Bromet, E. J., Parkinson, D. K., Day, N. L., & Ryan, C. M.
(1985). Patterns of alcohol use and physically aggressive behavior in
men. Journal of Studies on Alcohol, 46, 279–282. doi:10.1177/
Leonard, K. E., & Eiden, R. D. (2007). Marital and family processes in the
context of alcohol use and alcohol disorders. Annual Review of Clinical
Psychology, 3, 285–310. doi:10.1146/annurev.clinpsy.3.022806.091424
Leonard, K. E., & Mudar, P. (2003). Peer and partner drinking and the
transition to marriage: A longitudinal examination of selection and
influence processes. Psychology of Addictive Behaviors, 17, 115–125.
Leonard, K. E., & Senchak, M. (1993). Alcohol and premarital aggression
among newlywed couples. Journal of Studies on Alcohol: Supplement,
Leonard, K. E., & Senchak, M. (1996). Prospective prediction of husband
marital aggression within newlywed couples. Journal of Abnormal Psy-
chology, 105, 369–380. doi:10.1037/0021-843X.105.3.369
Littlefield, A. K., Sher, K. J., & Wood, P. K. (2009). Is “maturing out” of
problematic alcohol involvement related to personality change? Journal
of Abnormal Psychology, 118, 360–374. doi:10.1037/a0015125
Margolin, G., John, R. S., & Foo, L. (1998). Interactive and unique risk
factors for husbands’ emotional and physical abuse of their wives.
Journal of Family Violence, 13, 315–344. doi:10.1023/A:
McLeland, K. C., Sutton, G. W., & Schumm, W. R. (2008). Marital
satisfaction before and after deployments associated with the global war
on terror. Psychological Reports, 103, 836–844. doi:10.2466/PR0
Mirowsky, J., & Ross, C. E. (1992). Age and depression. Journal of Health
and Social Behavior, 33, 187–205. doi:10.2307/2137349
Moffitt, T. E., Caspi, A., Harington, H., & Milne, B. J. (2002). Males on
the life-course-persistent and adolescence-limited antisocial pathways:
Follow-up at age 26 years. Development and Psychopathology, 14,
Muthe ´n, L. K., & Muthe ´n, B. O. (2007). Mplus (Version 5.1) [Computer
software]. Los Angeles, CA: Muthe ´n & Muthe ´n.
Norton, R. (1983). Measuring marital quality: A critical look at the depen-
dent variable. Journal of Marriage & the Family, 45, 141–151. doi:
O’Leary, K. D. (1999). Developmental and affective issues in assessing
and treating IPV. Clinical Psychology: Science and Practice, 6, 400–
O’Leary, K. D., & Woodin, E. (2005). IPV and problem drinking across the
lifespan: How much do they decline? Clinical Psychology Review, 25,
Pan, H. S., Neidig, P. H., & O’Leary, K. D. (1994). Predicting mild and
severe husband-to-wife physical aggression. Journal of Consulting and
Clinical Psychology, 62, 975–981. doi:10.1037/0022-006X.62.5.975
Parrott, D. J., & Giancola, P. R. (2004). A further examination of the
relation between trait anger and alcohol-related aggression: The role of
anger control. Alcoholism: Clinical and Experimental Research, 28,
Pernanen, K. (1976). Alcohol in human violence. New York, NY: Guilford
Potter, F. (1988). Survey of procedures to control extreme sampling
weights. Proceedings of the Survey Research Methods Section of the
American Statistical Association, pp. 453–458. Retrieved from
FORAN ET AL.
Princeton Survey Research Associates. (1999). The Community Indicators Download full-text
Survey – National: Topline results. Retrieved from http://www.cpanda
Radloff, L. S. (1977). The CES-D Scale: A self-report depression scale for
research in the general population. Applied Psychological Measurement,
1, 385–401. doi:10.1177/014662167700100306
Raghavan, C., Mennerich, A., Sexton, A., & James, S. (2006). Community
violence and its direct, indirect, and mediating effects on intimate
partner violence. Violence Against Women, 12, 1132–1149. doi:10.1177/
Raghunathan, T. E., Solenberger, P., & Van Hoewyk, J. (2002). IVEware:
Imputation and variance estimation software installation instructions
and user guide. Ann Arbor, MI: The University of Michigan.
Reinert, D. F., & Allen, J. P. (2007). The Alcohol Use Disorders Identi-
fication Test: An update of research findings. Alcoholism: Clinical and
Research Triangle Institute. (1990). AFRP analysis plan (Volume II: Ap-
pendices). Research Triangle Park, NC: Author.
Rubin, D. B., & Schenker, N. (1991). Multiple imputation in health-care
databases: An overview and some applications. Statistics in Medicine,
10, 585–598. doi:10.1002/sim.4780100410
Rumpf, H. J., Hampke, U., Meyer, C., & John, U. (2002). Screening for
alcohol use disorders and at-risk drinking in the general population:
Psychometric performance of three questionnaires. Alcohol and Alco-
holism, 37, 261–268. doi:10.1093/alcalc/37.3.261
Saunders, J. B., Aasland, O. G., Babor, T. F., de la Fuente, J. R., & Grant,
M. (1993). Development of the Alcohol Use Disorders Identification
Test (AUDIT): WHO collaborative project on early detection of persons
with harmful alcohol consumption II. Addiction, 88, 791–804. doi:
Scholz, U., Gutierrez, D. B., Sud, S., & Schwarzer, R. (2002). Is general
self-efficacy a universal construct? European Journal of Psychological
Assessment, 18, 242–251. doi:10.1027//1015-5718.104.22.168
Schumacher, J. A., Homish, G. G., Leonard, K. E., Quigley, B. M., &
Kearns, J. N. (2008). Longitudinal moderators of the relationship be-
tween excessive drinking and intimate partner violence in the early years
of marriage. Journal of Family Psychology, 22, 894–904. doi:10.1037/
Sher, K. J., Grekin, E. R., & Williams, N. A. (2005). The development of
alcohol use disorders. Annual Review of Clinical Psychology, 1, 493–
Simmons, C. A., Lehmann, P., & Cobb, N. (2008). Women arrested for
partner violence and substance use: An exploration of discrepancies in
the literature. Journal of Interpersonal Violence, 23, 707–727. doi:
Simons, R. L., Beaman, J., Conger, R. D., & Chao, W. (1993). Childhood
experience, conceptions of parenting, and attitudes of spouse as deter-
minants of parental behavior. Journal of Marriage & the Family, 55,
Slep, A. M. S., Foran, H., Heyman, R. E., & Snarr, J. D. (2010). Risk and
protective factors for IPV in a large scale survey of the US Air Force.
Journal of Community Health, 35, 375–383. doi:10.1007.s10900-010-
Steele, C. M., & Josephs, R. A. (1990). Alcohol myopia: Its prized and
dangerous effects. American Psychologist, 45, 921–933. doi:10.1037/
Stith, S. M., Green, N. M., Smith, D. B., & Ward, D. B. (2008). Marital
satisfaction and marital discord as risk markers for intimate partner
violence: A meta-analytic review. Journal of Family Violence, 23,
Stith, S. M., Smith, D. B., Penn, C. E., Ward, D. B., & Tritt, D. (2004).
Intimate partner physical abuse perpetration and victimization risk fac-
tors: A meta-analytic review. Aggression and Violent Behavior, 10,
Straus, M. A., Hamby, S. L., Boney-McCoy, S., & Sugarman, D. B. (1996).
The revised Conflict Tactics Scales (CTS2): Development and prelim-
inary psychometric data. Journal of Family Issues, 17, 283–316. doi:
Taft, C. T., Watkins, L. E., Stafford, J., Street, A. E., & Monson, C. M.
(2011). Posttraumatic stress disorder and intimate relationship problems:
A meta-analysis. Journal of Consulting and Clinical Psychology, 79,
The Social Capital Benchmark Survey. (2000). Civil engagement in Amer-
ica. Retrieved from http://www.cfsv.org/communitysurvey/docs/
U.S. Army Community and Support Center. (1989). 1989 army soldier and
family survey: Soldier survey (Report No. SCN: ATNC-AO-89-10A).
Cary, NC: Research Triangle Institute, Caliber Associates, and Human
Resources Research Organization.
U.S. Census Bureau. (2006). Current population survey, annual social and
economic supplement. Retrieved from http://pubdb3.census.gov/macro/
Vickerman, K. A., & Margolin, G. (2008). Trajectories of physical and
emotional marital aggression in midlife couples. Violence and Victims,
23, 18–34. doi:10.1891/0886-6708.23.1.18
Vinokur, A. D., Price, R. H., & Caplan, R. D. (1996). Hard times and
hurtful partners: How financial strain affects depression and relationship
satisfaction of unemployed persons and their spouses. Journal of Per-
sonality and Social Psychology, 71, 166–179. doi:10.1037/0022-3514
Ware, J. E., Jr., Kosinski, M., Dewey, J. E., & Gandek, B. (2001). How to
score and interpret single-item health status measures: A manual for
users of the SF-8TMHealth Survey. Lincoln, RI: Quality Metric Incor-
Whisman, M. A., & McClelland, G. H. (2005). Designing, testing, and
interpreting interactions and moderator effects in family research. Jour-
nal of Family Psychology, 19, 111–120. doi:10.1037/0893-3200.19
Williams, T. S., Connolly, J., Pepler, D., Craig, W., & Laporte, L. (2008).
Risk models of dating aggression across different adolescent relation-
ships: A developmental psychopathology approach. Journal of Consult-
ing and Clinical Psychology, 76, 622–632. doi:10.1037/0022-006X
Wills, T. A., Sandy, J. M., & Yaeger, A. M. (2002). Moderators of the
relation between substance use level and problems: Test of a self-
regulation model in middle adolescence. Journal of Abnormal Psychol-
ogy, 111, 3–21. doi:10.1037/0021-843X.111.1.3
Zucker, R. A., Fitzgerald, H. E., & Moses, H. D. (1995). Emergence of
alcohol problems and the several alcoholisms: A developmental per-
spective on etiologic theory and life course trajectory. In D. Cicchetti &
D. J. Cohen (Eds.), Developmental psychopathology: Vol. 2, risk, dis-
order and adaptation (pp. 677–711). New York, NY: Wiley.
Received May 10, 2011
Revision received December 30, 2011
Accepted December 30, 2011 ?
MODERATING RISK FOR INTIMATE PARTNER VIOLENCE