Clinical Psychology Review 46 (2016) 25–33
Contents lists available at ScienceDirect
Clinical Psychology Review
journal homepage: www.elsevier.com/locate/clinpsychrev
Children's exposure to intimate partner violence: A meta-analysis of
longitudinal associations with child adjustment problems
Nicole L. Vu, Ernest N. Jouriles ⁎, Renee McDonald, David Rosenfield
Department of Psychology, Southern Methodist University, P.O. Box 750442, Dallas, TX 75275-0442, USA
H I G H L I G H T S
• Children's exposure to IPV is linked prospectively to future adjustment problems.
• The link between IPV and children's adjustment problems becomes stronger over time.
• Associations are stronger when IPV is conceptualized broadly, rather than narrowly.
a r t i c l e
i n f o
Article history:
Received 1 September 2015
Received in revised form 22 March 2016
Accepted 11 April 2016
Available online 13 April 2016
Keywords:
Intimate partner violence
Child adjustment
Meta-analysis
Longitudinal
a b s t r a c t
This meta-analysis reviewed 74 studies that examined longitudinal associations between children's exposure to
intimate partner violence (IPV) and their adjustment problems. Results indicated that children's exposure to IPV
is linked prospectively with child externalizing, internalizing, and total adjustment problems. Moreover, the
magnitude of the association between IPV exposure and child externalizing and internalizing problems
strengthens over time. In addition, associations are stronger between IPV exposure and child externalizing and
internalizing problems when IPV is conceptualized broadly rather than narrowly (physical IPV + psychological
and/or sexual IPV versus physical IPV only), and when information on IPV and child adjustment problems is obtained from the same source, rather than independent sources. When IPV exposure is measured at younger ages,
compared to older ages, the association between IPV and child externalizing problems is greater. However, when
child adjustment problems are measured at older ages, compared to younger ages, the association between IPV
and child internalizing problems is greater. Child sex, sample type, and whether only the male partner's violence
or both partners' violence was measured did not predict the association between children's exposure to IPV and
later adjustment problems. The findings have both research and clinical implications regarding the long-term adjustment of children exposed to IPV and the conceptualization and measurement of resilience subsequent to IPV.
© 2016 Elsevier Ltd. All rights reserved.
Contents
1.
2.
Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.1.
Literature search . . . . . . . . . . . . . . . . . . . . .
1.2.
Coding procedures . . . . . . . . . . . . . . . . . . . . .
1.3.
Data analyses . . . . . . . . . . . . . . . . . . . . . . .
1.3.1.
Dependence of data . . . . . . . . . . . . . . . .
Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.1.
Overall average study-level effect size . . . . . . . . . . . .
2.1.1.
Publication bias . . . . . . . . . . . . . . . . . .
2.2.
Predictor analyses . . . . . . . . . . . . . . . . . . . . .
2.2.1.
Lag . . . . . . . . . . . . . . . . . . . . . . .
2.2.2.
Child age when IPV is assessed . . . . . . . . . . .
2.2.3.
Child age at the time of child adjustment assessment .
2.2.4.
Child sex . . . . . . . . . . . . . . . . . . . . .
⁎ Corresponding author.
E-mail address: ejourile@smu.edu (E.N. Jouriles).
http://dx.doi.org/10.1016/j.cpr.2016.04.003
0272-7358/© 2016 Elsevier Ltd. All rights reserved.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
27
27
27
28
28
28
29
29
29
29
30
30
30
26
N.L. Vu et al. / Clinical Psychology Review 46 (2016) 25–33
2.2.5.
Conceptualization of IPV . . . . . . . . . . .
2.2.6.
Whose IPV . . . . . . . . . . . . . . . . .
2.2.7.
IPV measure . . . . . . . . . . . . . . . . .
2.2.8.
Sample type . . . . . . . . . . . . . . . . .
2.2.9.
Common raters . . . . . . . . . . . . . . .
2.2.10.
Interactions among lag, child age, and child sex
3.
Discussion . . . . . . . . . . . . . . . . . . . . . . . . .
Appendix A.
Supplementary data . . . . . . . . . . . . . . . .
References . . . . . . . . . . . . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
Approximately 15.5 million children in the US are estimated to live
in two-parent households in which intimate partner violence (IPV)
has occurred within the previous year (McDonald, Jouriles, RamisettyMikler, Caetano, & Green, 2006), and about 16% have witnessed IPV at
least once during their lives (Finkelhor, Turner, Shattuck, & Hamby,
2015). Meta-analyses suggest that children's exposure to IPV, defined
as children living with parents who report occurrences of IPV, correlates
with a variety of adjustment problems, with an effect size in the small to
medium range (Evans, Davies, & DiLillo, 2008; Kitzmann, Gaylord, Holt,
& Kenny, 2003; Wolfe, Crooks, Lee, McIntyre-Smith, & Jaffe, 2003).
These reviews, however, have focused primarily on cross-sectional associations. Important questions about links between children's exposure to IPV and their adjustment problems, which require longitudinal
data, remain unanswered. For example, it is not clear whether children's
exposure to IPV predicts later adjustment problems, or whether the
strength of the relation fades over time. Since the publication of these
prior meta-analytic reviews, longitudinal research has accumulated
that permits an analysis of prospective associations between exposure
to IPV and child adjustment problems, hopefully providing a clearer
and more conclusive understanding of how children's adjustment may
be related to exposure to IPV. The primary purpose of this study is to
conduct a meta-analysis examining the question of whether children's
exposure to IPV predicts later adjustment problems, and whether the
magnitude of the association changes over time.
On the basis of theory, it could be reasoned that children's exposure
to IPV can set in motion a chain of events that have long-lasting effects
on child development and, in particular, the emergence of adjustment
problems. These events might include processes internal to the child,
such as the development of cognitions about interpersonal relationships
(e.g., enduring negative beliefs about how individuals act toward one
another), or biological changes that influence the child's ability to regulate emotions. They might also include events external to the child, such
as the mother experiencing symptoms of trauma and depression, or the
mother seeking refuge with her children at a shelter because of the IPV.
Many investigators have documented prospective associations between
children's exposure to IPV and later adjustment problems. Indeed, several studies indicate that the relation between exposure to IPV and
child adjustment problems is evident for periods of 10 years or more
(e.g., Narayan, Englund, Carlson, & Egeland, 2013; Yates, Dodds,
Sroufe, & Egeland, 2003).
One form of long-term influence has been denoted a “sleeper effect”:
an effect that is weak or undetectable early on, but which strengthens
and becomes clearly evident at a later point in time. For example, a
child might initially experience some adjustment difficulties (e.g., an
“adjustment reaction”) after exposure to IPV, but rather than gradually
dissipating, the difficulties later worsen. Similarly, children who appear
to adapt well, despite their exposure to IPV, may later develop problems
related to the IPV. Sleeper effects might emerge for a variety of reasons,
including the obvious one that some child problems simply take time to
unfold and crystallize following exposure to IPV. The implications of
sleeper effects can be profound, however, particularly when attempting
to understand resilience among children exposed to IPV (Howell, 2011).
In contrast to sleeper effects, some have suggested that the association between children's exposure to IPV and certain adjustment problems is likely to weaken or fade over time (e.g., Ware et al., 2001). In
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
30
30
31
31
31
31
31
32
32
much of the early research on children's exposure to IPV, the “violent”
or “exposed” samples were often help-seeking samples, many recruited
from domestic violence shelters. Living at a shelter can be highly stressful for both mothers and children. This is due not only to the timing of
many shelter stays—immediately after a recent experience of IPV—but
also due to having to physically move away from home, the general uncertainty about the family's future, and, often for the children, changing
schools and peer groups. Given this convergence of highly stressful
events and circumstances, child adjustment problems might generally
be expected. However, with some distance from the events and circumstances, child behavioral and emotional functioning might be expected
to return to its previous level. Results from several prospective studies
suggest this indeed might be the case (e.g., Lohman, Neppl, Senia, &
Schofield, 2013; Ware et al., 2001).
There are methodological advantages to focusing on longitudinal
studies for making conclusions about the relation between children's
exposure to IPV and their adjustment problems. The temporal ordering
of events is less ambiguous in longitudinal research. Although IPV is
often assumed to have a unidirectional effect on children's mental
health problems, for decades researchers who study family processes
have also suggested that child adjustment problems increase tension
between parents, heightening the risk for interparental conflict and
IPV (e.g., Neiderhiser, Marceau, & Reiss, 2013; Patterson, 1982; Rutter,
1994). Consistent with this are findings that children's behavioral dysregulation (e.g., aggression, misbehavior, or self-harm) is associated
with marital discord one year later (Schermerhorn, Cummings,
DeCarlo, & Davies, 2007).
Longitudinal studies can also mitigate the inflationary effects of
common method variance (CMV)—systematic variance shared among
variables due to the methods used to measure them—on estimates of
the strength of the relation between IPV and child adjustment problems
(Podsakoff, MacKenzie, & Podsakoff, 2012). In research on IPV and child
adjustment, CMV can be especially problematic when information
about both variables is obtained from the same source (e.g., the child's
mother), which is common in much of the research on this topic. Although CMV can be a problem in longitudinal research as well, there
are more potential factors contributing to CMV in cross-sectional research. To offer a concrete example, if data on children's exposure to
IPV and their adjustment problems are both obtained from the child's
mother at one time point, it is plausible that reports on both measures
may be biased by certain transitory variables, such as the mother's
mood at the time of the assessment, her level of fatigue or alertness,
or events that happened the day before. In addition, the mother's responses to an IPV measure can bias her reports of child problems, or
vice versa.
In this review we also examine how conceptualizing and operationalizing IPV in certain ways can influence its association with child
adjustment problems. Specifically, most studies on children's exposure
to IPV conceptualize IPV as acts of physical aggression (Evans et al.,
2008; Kitzmann et al., 2003). However, some researchers have begun
to broaden the conceptualization of IPV to also include acts of psychological and/or sexual IPV (e.g., Huang, Wang, & Warrener, 2010;
Jouriles, McDonald, Vu, & Sargent, 2015; Jouriles, Norwood, McDonald,
Vincent, & Mahoney, 1996; Schnurr & Lohman, 2013; Zarling et al.,
2013). Theoretically, a broader conceptualization of IPV may allow for
N.L. Vu et al. / Clinical Psychology Review 46 (2016) 25–33
a more comprehensive understanding of the negative impact of exposure to IPV on children. If children are at increased risk for adjustment
problems when using a broader conceptualization of IPV, limiting its
measurement to include only physical aggression may result in underestimates of the true impact of IPV exposure on children. Crosssectional research indicates that considering both physical and psychological aggression in the measurement of IPV strengthens the prediction
of child adjustment problems, as compared to only measuring physical
IPV (Jouriles et al., 1996). This review evaluates whether this is also true
for longitudinal data.
A second characteristic of IPV that we examine in this review is
whose violence is assessed: only that of the mother's partner, or both
the mother's and the partner's violence. Many studies in this literature
focus only on the violence directed toward the child's mother by her
male intimate partner. However, cross-sectional studies suggest that
assessing his as well as her IPV improves the prediction of child adjustment problems, as compared to the assessment of only his IPV
(McDonald, Jouriles, Tart, & Minze, 2009). Focusing only on violence
against the mother may underestimate the level of IPV exposure that
children experience, and may limit our understanding of its true impact.
On the other hand, focusing only on violence toward the mother may
suffice if that violence is responsible for the long-term negative impact
on child adjustment.
Related to these measurement considerations, children's exposure to
IPV has been assessed with many different measures. Perhaps the most
commonly used measure of IPV is the Conflict Tactics Scales (CTS) and
versions thereof (Straus, 1979; Straus, Hamby, Boney-McCoy, &
Sugarman, 1996). In a prior meta-analysis, Kitzmann et al. (2003)
found that studies of IPV using the CTS yield larger effect sizes than studies using other measures. We examine whether this finding is replicated
in our review of longitudinal studies.
In addition, in this review we consider the nature of the sample
studied—whether it was a help-seeking sample (e.g., recruited from a
domestic violence shelter) as compared to a sample drawn from the
general community. It is well documented that IPV is more frequent
and severe in help-seeking samples than in general community samples
(Johnson, 1995). In addition, families characterized by frequent and severe IPV often have other co-occurring problems that place children at
risk for adjustment problems, such as child maltreatment and unstable
residences (Jouriles, McDonald, Slep, Heyman, & Garrido, 2008; Turner
et al., 2012). It seems plausible that children in families seeking help,
compared to those in families recruited from the general community,
fare worse because the IPV occurs in the context of multiple risk factors
for child problems, which interact with each other to potentiate adverse
outcomes (Evans, Li, & Whipple, 2013).
In sum, this review examines longitudinal associations between
children's exposure to IPV and their adjustment problems. We examine
whether this association grows stronger or weaker over time. We also
examine whether the magnitude of the association differs depending
on: how broadly IPV is conceptualized (whether psychological or sexual
IPV are considered along with the physical IPV), whether IPV is defined
as men's violence only or men's as well as women's, whether the CTS
(and versions thereof) is used to assess IPV or some other measure is
used, and whether families are from help-seeking or general community samples. Although we do not have specific hypotheses about whether
the association weakens, strengthens, or remains stable over time, we
expect a stronger association when a broader conceptualization of IPV
is used, when both partners' IPV is considered, and when the sample
is help-seeking. We also explore effects of one type of common method
variance (i.e., whether information on IPV and child problems is obtained from the same source) on the association.
Finally, we examine whether findings differ as a function of child age
and sex. Although no evidence for the impact of child age on the association between exposure to IPV and child adjustment has emerged from
previous meta-analyses of primarily cross-sectional studies, there have
been some contradictory findings pertaining to child sex (Evans et al.,
27
2008; Kitzmann et al., 2003; Wolfe et al., 2003). There is also reason
to believe that children's exposure to IPV might interact with child socialization practices, so that relations with adjustment problems become more pronounced with girls, as compared to boys, over time
(Davies & Lindsay, 2004). Specifically the gender intensification hypothesis suggests that boys and girls are increasingly pressured to conform to conventional gender roles with age, with boys becoming more
independent and girls becoming more concerned about interpersonal
connectedness and maintaining social relationships. As a result, girls
may be more sensitive to and adversely affected by parental IPV over
time.
1. Method
1.1. Literature search
We conducted searches of PsycINFO, PsycARTICLES, and MEDLINE
databases through April 2015 to identify studies to include in the
meta-analysis. We used a variety of combinations of the following
search terms: domestic violence, intimate partner violence, family violence, marital violence, marital aggression, interparental conflict,
interparental aggression, partner abuse, wife abuse, battering; child,
youth, adolescent, teen; and longitudinal, prospective, repeated, and
over time. In addition, prominent researchers in the field were identified, and all their publications were evaluated for inclusion. This search
yielded 2017 publications and dissertations.
To be included in the meta-analysis, a study had to meet the following criteria: 1) It reported the results of a quantitative empirical study.
Review papers, qualitative studies, and case studies were excluded.
2) The study included a measurement of parental physical IPV. Thus,
our definition of children's exposure to IPV was children living in families in which occurrences of physical IPV were reported. This definition
is consistent with the definitions used in other reviews (Evans et al.,
2008; Kitzmann et al., 2003; Wolfe et al., 2003). Studies that examined
interparental conflict, but not parental physical IPV, were excluded.
3) The study examined child adjustment problems within the domains
of externalizing and/or internalizing problems. Studies that only included other child adjustment variables (e.g., those related to academic
functioning and physical health outcomes) were excluded. 4) IPV was
assessed at an earlier time point than child adjustment problems.
When IPV and child adjustment problems were measured at the same
time point, the study was excluded. 5) The article included a prospective, zero-order correlation between IPV and later child adjustment
problems. 6) IPV was assessed when youth were 18 years old or younger. Authors of dissertations and publications that met all inclusion
criteria except for the reporting of a zero-order correlation (criterion
5) were contacted to request the correlations; multiple attempts were
made to contact non-responsive authors in efforts to obtain this information. Of the 2017 publications and dissertations identified in the initial search, 1943 were excluded because the studies: did not meet
inclusion criteria, included a correlation that had already been included
(e.g., when studies were conducted with the same sample of participants), or authors did not provide correlations when contacted. A total
of 74 publications and dissertations met the inclusion criteria and
were thus included in the meta-analysis. The first author and a research
assistant reviewed the 2017 studies, overlapping in their review of 202
studies (10%). There was 99% agreement on whether a study should be
included or not. The two disagreements were handled by having a third
person review the two studies to determine whether they met inclusion
criteria.
1.2. Coding procedures
The first author and a research assistant coded the following information from each study: time elapsed between assessment of IPV and
assessment of child adjustment problems, average age of the children
28
N.L. Vu et al. / Clinical Psychology Review 46 (2016) 25–33
in the study when IPV was assessed, average age of the children in the
study when child adjustment was assessed, proportion of male and female children in the sample, IPV conceptualization (i.e., narrow = physical IPV only; broader = physical plus psychological IPV and/or sexual
IPV), whose IPV was reported (partner only versus both mother and
partner), whether IPV was assessed using the CTS or some version
thereof versus IPV that was assessed with a different measure of violence, whether children came from a general community or helpseeking sample, whether data on IPV and child adjustment problems
came from the same source (i.e., reported by the same individual),
whether the publication came from the same sample as another one
of the 74 studies, sample size, type of child adjustment problem (externalizing, internalizing, or total adjustment problems—measures that included both externalizing and internalizing problems), and the zeroorder correlation between IPV and later child adjustment problems. A
total of 376 IPV and child adjustment problem correlations were extracted (201 externalizing problems, 158 internalizing problems, and
17 total adjustment problems) from the 74 studies. Many studies provided multiple correlations that could be included, resulting in more
correlations than studies. Percent agreement on each of the coded variables ranged from 80% to 100%, and disagreements were resolved via
discussion between the two coders.
1.3. Data analyses
Correlation coefficients (Pearson's r) were used as effect sizes in all
analyses. All analyses were conducted in R using the metafor package
(Viechtbauer, 2010).
1.3.1. Dependence of data
The 74 studies in the meta-analysis include 41 distinct samples;
thus, there is a substantial degree of clustering within sample. That is,
many different correlations could come from one sample because
many studies used multiple measures to examine the same construct
(e.g., CBCL Internalizing Scale and Child Depression Inventory for internalizing problems) and/or examined correlations at multiple time
points (e.g., baseline IPV with internalizing problems 6 months and
12 months later). Since different correlations from the same sample
are likely to be related, the assumption of independence of correlations
is violated. Dependence among effect sizes has commonly been addressed in a number of ways, such as by treating the effect sizes as independent (i.e., ignoring the dependence), averaging across multiple
effect sizes, or selecting a single effect size per sample. However, because there are problems with these approaches (see Becker, 2000),
we used multilevel modeling (MLM) to account for the dependence of
effect sizes.
MLM is useful when the assumption of independence of scores is violated, and it is particularly useful when multiple data points within a
meta-analysis come from the same study (Hox, 2010). For our MLM
analyses, the effect sizes (i.e., the correlations between IPV and child adjustment problems) were nested within study sample. In all analyses,
the correlation between exposure to IPV and child adjustment problems
was entered as the dependent variable, and study sample was used as
the cluster. Separate models were then computed for externalizing
problems, internalizing problems, and total adjustment problems.
Separate analyses examined the effect of each of the following predictors on the association of IPV with child adjustment problems
(i.e., effect size): lag, the average age of the children in the study
when IPV was assessed, the average age of the children when child adjustment problems were assessed, the proportion of males in the sample, IPV conceptualization as broad versus narrow, whose IPV was
reported, the IPV measure used (CTS or some version thereof versus a
different measure), the sample type, and common reporters of IPV
and child adjustment (see Table 1 for information on how variables
were coded). In addition, we examined whether lag (the time between
the assessment of IPV and the assessment of child problems) influenced
Table 1
Coding of variables used in meta-analysis.
Variable
Code or unit type
Sample ID
ID based on whether publication came from the same
sample as another publication
Adjustment problem type
Externalizing
Internalizing
Total
Lag
Years elapsed between assessment of IPV and later
child adjustment problems
Child age when IPV
assessed
Years
Child age when child
adjustment assessed
Years
Child sex
Proportion of males in sample
IPV conceptualization
0=narrowa
1=broadb
Whose IPV
0=mother and partner
1=partner only
IPV measure
0=CTS
1=non-CTS
Sample type
0=help-seeking
1=community
Common ratersc
0=none
1=somed
2=absolute
Zero-order correlation
r
Sample size
n
a
Note. Narrow IPV conceptualization was characterized by physical IPV only. bBroader IPV
conceptualization was characterized by physical IPV in combination with at least one
other type of IPV (e.g., psychological IPV, sexual IPV). c“Common raters” indicates the extent that IPV and child adjustment were reported by the same individual. dThe extent of
common raters was coded as “some” when IPV and/or child adjustment were determined
using reports from multiple sources on at least one, but not all, of the measures.
the correlation between exposure to IPV and child adjustment differentially as a function of child age and sex; that is, whether age and sex
moderated the effect of lag on effect size. The final level-1 MLM model
for the analysis for each predictor was:
rkj ¼ b0k þ b1k Predictorkj þ εkj
where rkj is the ith correlation for sample k, and “Predictorkj” is the value
of the specific predictor (i.e., lag, T1 child age, T2 child age, proportion of
male children, IPV conceptualization, whose IPV was reported, IPV measure, sample type, and CMV due to common reporters of IPV and child
adjustment).
The level-2 models for most analyses were:
b0k ¼ γ00 þ μ0k
b1k ¼ γ10 þ μ1k
For the examination of child age or sex as moderators, the main effect of sex or age and the interaction between sex or age and lag were
added to the above models.
2. Results
Table 2 lists the number of correlations in each of the studies included in the meta-analysis. Means, standard deviations, and ranges of
N.L. Vu et al. / Clinical Psychology Review 46 (2016) 25–33
externalizing problems, r = .08 for internalizing problems, r = .16 for
total adjustment problems.
Table 2
Number of correlations by type of child adjustment problems.
Variable
Correlation
p b .05
p N .05
Child sex
Male sample
Female sample
Mixed-sex sample
IPV conceptualization
Narrowa
Broaderb
Whose IPV was measured
Partner-only
Partner and mother
IPV measure
CTS
Non-CTS
Common ratersc
None
Somed
Absolute
Sample
Community
Help-seeking
Community + help-seeking
29
Externalizing
problems
(n = 201)
Internalizing
problems
(n = 158)
Total
problems
(n = 17)
98
103
51
107
13
4
13
13
175
7
5
146
0
0
17
170
31
98
60
6
11
81
107
65
89
6
11
111
90
46
112
9
8
94
19
88
75
15
68
8
0
9
135
51
15
88
54
16
11
3
3
Note. aNarrow IPV conceptualization was characterized by physical IPV only. bBroader IPV
conceptualization was characterized by physical IPV in combination with at least one
other type of IPV (e.g., psychological IPV, sexual IPV). c“Common raters” indicates the extent that IPV and child adjustment were reported by the same individual. dThe extent of
common raters was coded as “some” when IPV and/or child adjustment were determined
using reports from multiple sources on at least one, but not all, of the measures.
scores among the study variables are listed in Table 3. There were more
correlations for externalizing problems (n = 201) and internalizing
problems (n = 158) than for total adjustment problems (n = 17).
Among the bivariate correlations between IPV and child adjustment,
49% of those for externalizing problems (98/201), 32% of those for internalizing problems (51/158), and 76% of those for total adjustment problems (13/17) were statistically different from zero (i.e., p b .05). On
average, the correlations were small in magnitude: r = .13 for
Table 3
Means, standard deviations, and ranges of continuous variables across types of child adjustment problems.
2.1. Overall average study-level effect size
The homogeneity test for the effect sizes indicated between-study
heterogeneity for externalizing problems, Q T(200) = 1171.97,
p b .001; internalizing problems, QT(157) = 314.84, p b .001; and total
adjustment problems, Q T(15) = 33.31, p = .004. Consequently, we
employed a random-effects model to examine overall average effect
sizes. The weighted average correlation differed from 0 for externalizing
problems, r = .15 (SE = .02), z = 8.70, p b .001; internalizing problems,
r = .10 (SE = .01), z = 7.57, p b .001; and total adjustment problems,
r = .15 (SE = 0.03), z = 4.98, p b .001.
2.1.1. Publication bias
To evaluate the extent to which null results from unpublished studies might influence findings, we assessed funnel plot asymmetry for our
effect size using Egger's regression test (Egger, Smith, Schneider, &
Minder, 1998). In a funnel plot, effect sizes are plotted against a measure
of precision (e.g., standard error). The presence of funnel plot asymmetry suggests that the data are biased and do not arise from a single population effect. The funnel plot for our data was not asymmetric for
externalizing problems, t(199) = 0.86, p = .390, internalizing problems, t(156) = − 1.46, p = .147, or total adjustment problems,
t(15) = 0.53, p = .607, which suggests that null results from unpublished studies did not significantly affect our results. We also calculated
Rosenthal's fail-safe N, which provides an estimate of the number of unpublished studies with non-significant results necessary to make an average effect size non-significant (Rosenthal, 1979). In the present metaanalysis, a fail-safe N of 75,022 studies was obtained for externalizing
problems, 14,257 studies for internalizing problems, and 1458 studies
for total behavior problems. It should be noted, however, that when
nonindependent data are used, the fail-safe N can result in unusually
large numbers (Rosenthal, 1979). In short, these analyses converge to
suggest that the present meta-analytic results are robust, and not likely
to be changed significantly by unpublished null results.
2.2. Predictor analyses
To attempt to explain the heterogeneity in correlations, we used a
mixed-effects model to identify potential predictors of the correlation
of IPV and child problems. In a mixed-effects model, the effect sizes
(correlations) are estimated in a random-effects model, and predictors
are tested in a fixed-effects model. Results from these analyses are summarized in Table 4.
2.2.1. Lag
The average length of time between the assessment of IPV and assessment of child adjustment (lag) was 3.99 (SD = 4.10) years, and
ranged from 3 months to 23.5 years. Lag was positively associated
with the correlation of IPV exposure and externalizing problems,
Q B(1) = 8.40, p = .004, and internalizing problems, Q B (1) =
11.94, p b .001. That is, the greater the time between assessment of
IPV and assessment of adjustment problems, the stronger the correlation between IPV and adjustment problems. Lag was not associated
with the correlation between IPV and total adjustment problems,
Q B(1) = 0.41, p = .523.
For visual simplification, and to aid and in interpreting the nature of
the lag effects, we graphed the model-estimated association between
lag and effect size for the three types of adjustment problems (see
Fig. 1). When little-to-no time has passed between IPV and child adjustment problems (lag = 0 years), the correlation between IPV and child
adjustment problems is estimated to be .13 for externalizing problems,
.06 for internalizing problems, and .13 for total adjustment problems
(although our meta-analysis includes only longitudinal studies,
30
N.L. Vu et al. / Clinical Psychology Review 46 (2016) 25–33
Table 4
Results of predictors of IPV exposure and child adjustment association.
Predictor
Lag
Externalizing
Internalizing
Total
Child age when IPV assessed
Externalizing
Internalizing
Total
Child age when child adjustment
assessed
Externalizing
Internalizing
Total
Proportion of males
Externalizing
Internalizing
Total
IPV conceptualization
Externalizing
Internalizing
Total
Whose IPV
Externalizing
Internalizing
Total
IPV measure
Externalizing
Internalizing
Total
Common raters
Externalizing
Internalizing
Total
Sample type
Externalizing
Internalizing
Total
Lag × Age
Externalizing
Internalizing
Total
Lag × Proportion of males
Externalizing
Internalizing
Total
K
201
158
17
b
SE
.004
b.001
.523
201
−.015
158 b−.001
17
.004
.003 −5.427
.003 −0.13
.013
0.31
b.001
.895
.755
201
158
17
.001
.005
.006
.002
.002
.007
0.48
2.94
0.90
.629
.003
.367
188
138
14
−.028
−.033
.340
.023 −1.22
.037 −0.89
.387
0.87
.224
.373
.382
201
158
17
−.083
−.083
.032
.018 −4.53
.019 −4.47
.062
0.52
b.001
b.001
.605
188
154
17
−.002
−.034
.015
.017 −0.12
.026 −1.30
.059
0.25
.908
.193
.800
201
−.023
158 b−.001
17 −.091
.016 −1.40
.022 −0.01
.045 −2.04
.162
.993
.042
201
158
17
.031
.028
−.004
.009
3.59
.007
3.74
.020 −0.17
b .001
b .001
.865
186
142
14
.008
−.046
.027
.042
0.20
.028 −1.64
.027
4.77
.844
.100
b .001
201 b−.001 b .001 −1.37
158 b−.001 b .001 −1.46
17 −.009
.004 −2.16
.171
.144
.031a
−.003
−.039
.103
.002
.002
.007
p
2.90
3.45
0.64
188
138
14
.006
.008
.004
Z
.010 −0.33
.011 −0.37
.096
1.08
.744
.714
.282
Note. aAlthough the slope for Lag × Age and total adjustment problems was statistically significant, the QB test for moderators was not. IPV conceptualization: 0 = broader IPV (physical IPV + psychological and/or sexual IPV), 1 = narrower IPV (physical IPV only). Whose
IPV: 0 = partner and mother, 1 = partner only. Common raters was coded as 0 = none,
1 = some, 2 = absolute. Sample type was coded as 0 = help-seeking, 1 = community.
estimates reported for lag = 0 years are derived from the model). These
estimates are consistent with the small effect sizes derived from at least
one other meta-analysis of primarily cross-sectional data on IPV and
child adjustment problems (Kitzmann et al., 2003). For a 10-year lag,
however, the estimated correlation is .18 for externalizing problems,
.13 for internalizing problems, and .17 for total adjustment problems.
2.2.2. Child age when IPV is assessed
The average age of the children at the time of the IPV assessment in
each study ranged from pre-birth (during mothers' third trimester of
pregnancy) to 18 years. Because many studies recruited children of
varying ages (e.g., 6–10 years old), the average child age at the time of
the assessment of IPV was used in our analyses. Older child age at the
time of the assessment of IPV was associated with lower correlations between IPV and externalizing problems, Q B(1) = 27.73, p b .001. The estimated correlation between IPV and externalizing problems is .21
when IPV was assessed at an average child age of 5 years, .13 at age
10 years, and .06 at age 15 years. Child age at the time of the assessment
of IPV was not associated with the correlation between IPV and
Fig. 1. Lag as a predictor of the association between IPV exposure and child adjustment
problems. The association between lag (i.e., years lapsed between assessment of IPV exposure and child adjustment problems) and the correlation between IPV exposure and 1) externalizing problems, b = .006, SE = .002, p = .004; 2) internalizing problems, b = .008,
SE = .002, p b .001; and 3) total behavior problems, b = .004, SE = .007, p = .523. All studies included in the meta-analysis are longitudinal: lag of 0 years is a model-derived
estimate.
internalizing problems, Q B(1) = 0.02, p = .895, or total adjustment
problems, Q B(1) = 0.10, p = .755.
2.2.3. Child age at the time of child adjustment assessment
The average age of the children at the time of the child adjustment
assessment ranged from 0.75 to 27 years. The age at the time at which
child adjustment was assessed related positively with correlations between IPV and internalizing problems, Q B(1) = 8.65, p = .003. The estimated correlation between IPV and internalizing problems is .06 when
child adjustment was assessed at an average child age of 5 years, .08 at
age 10 years, and .11 at age 15 years. The effect of child age at the time of
the child adjustment assessment was not associated with the correlation between IPV and child externalizing problems, Q B(1) = 0.23,
p = .629, or total adjustment problems, Q B(1) = 0.81, p = .367.
2.2.4. Child sex
The effect of child sex on the association between IPV and child adjustment was examined by using the proportion of males in the study
as a predictor of the association. The proportion of males was not associated with the correlation between IPV and externalizing problems,
Q B(1) = 1.48, p = .224, internalizing problems, Q B(1) = 10.79, p =
.373, or total adjustment problems, Q B(1) = 0.77, p = .382.
2.2.5. Conceptualization of IPV
The correlation between IPV and child problems differed across conceptualizations of IPV (physical plus psychological and/or sexual IPV vs.
physical IPV only) for externalizing problems, Q B(1) = 20.54, p b .001,
and internalizing problems, Q B(1) = 19.98, p b .001. Correlations
were higher when IPV was conceptualized more broadly. The estimated
correlation between IPV and externalizing problems is .21 when IPV is
conceptualized more broadly and .13 when IPV is conceptualized
more narrowly. The estimated correlation between IPV and internalizing problems is .15 when IPV is assessed more broadly and .07 when
IPV is assessed more narrowly. The correlation between IPV and total
adjustment problems did not differ across the two conceptualizations
of IPV, Q B(1) = 0.27, p = .605.
2.2.6. Whose IPV
The correlation between IPV and child adjustment problems did not
differ according to whether only the partner's IPV or both the mother's
and the partner's IPV was assessed for externalizing problems, Q B(1) =
0.01, p = .908, internalizing problems, Q B(1) = 1.70, p = .193, or total
adjustment problems, Q B(1) = 0.06, p = .800.
N.L. Vu et al. / Clinical Psychology Review 46 (2016) 25–33
2.2.7. IPV measure
The correlation between IPV and child total adjustment problems
was greater when IPV was assessed using the CTS or some version
thereof as compared to other measures, Q B(1) = 3.87, p = .049. The estimated correlation between IPV and total adjustment problems is .18
when IPV was assessed using the CTS or some version thereof and .09
when IPV was assessed using other measures. The IPV measure did
not affect the correlation of IPV with externalizing problems, Q B(1) =
1.96, p = .162, or internalizing problems, Q B(1) b 0.01, p = .993.
2.2.8. Sample type
The correlation between IPV and child adjustment problems did not
differ across community samples and help-seeking samples for externalizing problems, Q B(1) = 0.04, p = .844, or internalizing problems,
Q B(1) = 2.69, p = .101. Sample type was not examined as a predictor
of the correlation between IPV and total adjustment problems due to a
limited number of correlations from help-seeking samples (n = 3).
2.2.9. Common raters
The extent that information on both IPV and child adjustment was
obtained from the same reporter was positively associated with correlations for externalizing problems, Q B(1) = 12.92, p b .001, and internalizing problems, Q B(1) = 13.98, p b .001, but not total adjustment
problems, Q B(1) = 0.03, p = .865. Analyses examining whether IPV
was correlated with child externalizing and internalizing problems
across studies using the same or different reporters for the two variables
were conducted using two random-effects models: 1) one that included
studies that used different reporters for the two variables, and 2) one
that included studies that used the same reporter for the two variables.
Results indicated that there was still a statistically significant association
for externalizing problems, r = .12, p b .001, and internalizing problems,
r = .07, p b .001. Not surprisingly, these correlations were stronger
when the reporters of IPV and child adjustment were the same: externalizing problems, r = .18, p b .001; internalizing problems, r = .12,
p b .001.
2.2.10. Interactions among lag, child age, and child sex
Average age of children in the sample did not moderate the association between lag and the correlation between IPV and child externalizing problems, b b −.01, SE b .01, p = .171, internalizing problems,
b b .01, SE b .01, p = .144, or total adjustment problems, b = −.01,
SE b .01, p = .031 (although the slope for Lag × Age and total adjustment
problems was statistically significant, the Q B test for moderators was
not, Q B = 6.07, p = .108). The proportion of males in the sample also
did not moderate the association between lag and the correlation of
IPV with child externalizing problems, b b −.01, SE = 0.01, p = .744, internalizing problems, b = −.04, SE = 0.01, p = .714, or total adjustment
problems, b = .10, SE = 0.10, p = .282.
3. Discussion
A primary goal of this meta-analysis was to examine longitudinal associations between children's exposure to IPV and their adjustment
problems, and to assess whether the strength of the association changes
over time. In general, our findings suggest that there is indeed a prospective association between children's exposure to IPV and child externalizing and internalizing problems. In addition, contrary to what might
normally be expected, the greater the time lag between the measurements of IPV and child adjustment problems, the stronger the association. Although our findings point to a relatively weak association
between IPV exposure and child adjustment problems over time, it
bears noting that the association increases by 40% for externalizing
problems (.13 to .18) and 117% for internalizing problems (.06 to .13)
ten years after the assessment of IPV exposure. This holds potentially
important implications for our understanding of resilience among children exposed to IPV. Specifically, scientists and clinicians who do not
31
observe adjustment problems among children newly exposed to IPV
may find it easy to label them as unaffected or within normal limits of
adjustment. In fact, such children are often referred to as resilient
(Howell, 2011). However, the present findings point to the need for a
more long-term view of resilience. Indeed, from a clinical perspective
it may be important to conduct periodic assessments of children exposed to IPV—to “check in”—in order to detect the emergence and development of problems, and to alert parents about the possibility of
delayed effects. Alternatively stated, caution should be exercised in assuming that a child recently exposed to IPV will not have adverse reactions to it, regardless of how well that child appears to be functioning in
the short term.
Finding that the link between IPV exposure and child externalizing
and internalizing problems strengthens over time is consistent with
the idea of a sleeper effect. This might occur because many child adjustment problems likely do not emerge, nor do they fully develop, immediately following exposure to IPV (or exposure to any stressful life
circumstance). Rather, their emergence is a process: They unfold and
may crystallize over time in the context of events and experiences
that often occur after the IPV. This might include the development of
harmful cognitions, such as insecurity or doubts about the stability of
the family system or the safety of one or both parents (Bergman,
Cummings, & Davies, 2014; Fosco, DeBoard, & Grych, 2007). It might
also include the development of harmful reactions to IPV, such as attempts to intervene and stop their parents' violence (Jouriles,
Rosenfield, McDonald, & Mueller, 2014). Another hypothesis is that it
is not a sleeper effect at all. Rather, the stronger associations are a result
of cumulative effects of IPV over time. Specifically, children's exposure
to IPV rarely happens as a single occurrence; it is often a repeated occurrence and in some families it is chronic (Margolin et al., 2009). It is possible that the longer lag allows for the effects of greater cumulative
exposure to IPV to be captured, and the stronger associations simply reflect this. Unfortunately, this alternative hypothesis cannot be ruled out
from existing longitudinal research on this phenomenon.
Our results also indicate that child age when IPV exposure is
assessed moderates the relation between IPV exposure and child externalizing problems. Specifically, the relation is stronger when IPV exposure is assessed when children are younger. Child age did not moderate
the relation for internalizing problems or total adjustment problems.
This finding is noteworthy because some research suggests that younger, as compared to older, children are at particularly high risk for IPV
exposure (e.g., Fantuzzo, Boruch, Beriama, Atkins, & Marcus, 1997). It
might be argued that because younger children have limited social influences outside the family (e.g., classmates, friends), have less developed problem-solving and coping skills, and are more likely to be
home when the violence occurs, they may be especially vulnerable to
IPV. However, this finding should be interpreted with caution, as the
studies in this meta-analysis did not typically provide information to
examine the age at which children were first exposed to IPV, or the
course or nature of the IPV exposure over time. All that was available
to analyze, in most cases, was whether or not the child had been exposed to IPV during a designated time period (e.g., 12 months) prior
to the onset of the study.
The age at which child problems were assessed moderated the association between IPV exposure and child internalizing problems. Specifically, the association is stronger when children are older at the time
that internalizing problems are assessed. This moderator effect did not
emerge for externalizing or total adjustment problems. Some research
suggests that for more than one third of children, internalizing problems
increase with age (Nantel-Vivier, Pihl, Cote, & Tremblay, 2014). It is possible that the sleeper effect observed in the present meta-analysis can
be at least partially accounted for by this developmental trajectory of internalizing problems over time.
Our results also suggest that conceptualizing IPV more broadly
(i.e., accounting for physical IPV and other forms of IPV) yields a stronger
association between IPV exposure and child adjustment, compared to
32
N.L. Vu et al. / Clinical Psychology Review 46 (2016) 25–33
conceptualizing IPV to only include physical acts of violence. This suggests
that assessing psychological and/or sexual IPV in addition to physical IPV
may prove useful in gaining a more comprehensive understanding of the
impact of IPV on child adjustment. Indeed, many have argued that omitting psychological and/or sexual violence from efforts to estimate IPV
prevalence has resulted in incorrect conclusions about IPV (e.g., Hamby,
2014). Our findings are consistent with the results from a handful of
cross-sectional studies that have demonstrated benefits of measuring
psychological or sexual IPV when examining relations with children's adjustment problems (e.g. Jouriles et al., 1996, 2015). An increased risk for
parental psychological distress has been proposed as one possible mechanism for explaining the greater harm to children that broader conceptualizations of IPV appear to capture (Jouriles et al., 2015). The present
findings provide additional evidence of the possible advantages of
adopting a broader conceptualization of IPV.
Our results also indicate that using the CTS or some version thereof to
measure IPV yields stronger associations between IPV exposure and
child total adjustment problems, but not externalizing or internalizing
problems. This former finding is consistent with findings from a previous
meta-analysis (Kitzmann et al., 2003). The CTS is widely used and wellvalidated, whereas many other measures of IPV (e.g., unstructured interviews, one or two questions assessing violence) have not been wellvalidated and thus may yield attenuated associations between IPV
exposure and child adjustment. However, our findings should be
interpreted with caution, because this finding emerged only with total
adjustment problems, and not externalizing nor internalizing problems,
and there were many more studies specifically examining externalizing
or internalizing problems, compared to total adjustment problems.
In addition, and not surprisingly, our results indicate that obtaining
information on IPV exposure and child adjustment problems from the
same reporter yields a stronger association. This suggests that collecting
information on both variables from the same source may result in an inflated estimate of the association between IPV exposure and child adjustment. However, when data were collected from different reporters, these
two types of problems were still related to one another, which suggests
that the longitudinal association is not solely due to common method
variance. In short, the longitudinal association between children's exposure to IPV and child adjustment problems still emerges after accounting
for method variance due to a single reporter; yet, it still may behoove researchers to obtain reports of IPV exposure and child adjustment problems from different individuals in order to rule out common method
variance effects.
Some of our hypotheses were not supported by the data, and some
of the exploratory analyses yielded null results. The association between
IPV exposure and child problems, for example, did not vary as a function
of whose IPV was measured (male partner only versus both partners).
One interpretation of this result is that mothers' IPV may not be important to the development of child adjustment problems over and above
the partner's IPV. However, this result should be interpreted cautiously,
and additional research may be needed to explore the conditions under
which it holds. For example, there may be some outcomes (e.g., other
types or measures of child adjustment problems), or some particular
measures of IPV, for which this result does not hold. Another null result
pertained to the influence of child sex, which did not affect the association between IPV and child adjustment problems. This particular null result is consistent with the conclusions from previous meta-analyses
(Evans et al., 2008; Kitzmann et al., 2003; Wolfe et al., 2003) of primarily
cross-sectional studies, and it suggests that boys and girls are at similar
risk for developing adjustment problems following IPV exposure.
Several limitations should be considered when interpreting the
present findings. First, even though we focused on studies that evaluated prospective associations between children's exposure to IPV and adjustment problems, we cannot infer causation. Although longitudinal
research circumvents some of the problems of cross-sectional research
for making causal inferences (e.g., temporal sequencing of events), prospective studies are still vulnerable to unmeasured third-variable
explanations for causal processes. For example, IPV often co-occurs
with other risk factors for child adjustment problems, such as child maltreatment (Jouriles et al., 2008); thus, causal inferences should not be
made from these data. Second, we only included studies in which correlations between children's exposure to IPV and adjustment problems
were available (either in the publication or dissertation or in response
to email requests). We believe we conducted a comprehensive review,
but it is not clear how studies that were not included in our analysis
may have influenced our findings. Third, although we focused on externalizing and internalizing problems, there is a growing literature on
children's IPV exposure and its link to other problem areas. Future reviews should consider these associations, including those related to academic functioning and physical health. Fourth, it is likely that children
who have been exposed to IPV at one point in time are exposed to IPV at
other time points as well (Margolin et al., 2009). The ongoing nature of
IPV exposure over time needs to be considered in order to make precise
statements about child's age at the time of IPV exposure, IPV duration
and chronicity, and later adjustment problems. Fifth, our definition of
children's exposure to IPV included simply living in a household
where IPV was reported to have occurred. Different definitions of
children's exposure to IPV may lead to different results.
In summary, our findings indicate that the link between children's
exposure to IPV and children's adjustment problems strengthens over
time. This is the first review of research in this area to focus exclusively
on longitudinal relations and to examine the strength of the association
over time. The findings suggest that it can take months to years for problems to emerge as a result of IPV exposure. As a result, clinicians who
work with children and adolescents may benefit from assessing history
of exposure to IPV, and when possible, assess for further IPV exposure
and for adjustment problems at subsequent points in time. In addition,
data from this review indicate that there is benefit to be gained from enhancing our measurement strategies in other ways. For example, over
half of the correlations used in this review (53%) had at least some overlap in source of data for IPV exposure and child adjustment problems.
Our results suggest that such overlap may lead to an inflation of the association between IPV exposure and child adjustment problems. In addition, most of the correlations on which this review is based
(i.e., 73%) were based on a narrow conceptualization of IPV; however,
our results suggest that broadening the conceptualization of IPV may
add to our knowledge of IPV exposure and its association with child adjustment problems. Future research should examine the utility of
broader conceptualizations as well as the optimal operationalization of
a broader conceptualization (i.e., the relative contributions of physical,
psychological, and sexual IPV, and how best to operationalize them).
In sum, this review provides clear indication that the association of
IPV with child internalizing and externalizing problems is robust and
long-lived, and it points to areas where our knowledge about the association should be strengthened and refined.
Appendix A. Supplementary data
Supplementary data to this article can be found online at http://dx.
doi.org/10.1016/j.cpr.2016.04.004.
References
Becker, B. J. (2000). Multivariate meta-analysis. In H. E. A. Tinsley, & S. D. Brown (Eds.),
Handbook of applied multivariate statistics and mathematical modeling (pp. 499–525).
San Diego: Academic Press.
Bergman, K. N., Cummings, E. M., & Davies, P. T. (2014). Interparental aggression and adolescent adjustment: The role of emotional insecurity and adrenocortical activity.
Journal of Family Violence, 29, 763–771. http://dx.doi.org/10.1007/s10896-014-9632-3.
Davies, P. T., & Lindsay, L. (2004). Interparental conflict and adolescent adjustment: Why
does gender moderate early adolescent vulnerability? Journal of Family Psychology,
18, 170–180. http://dx.doi.org/10.1037/0893-3200.18.1.160.
Egger, M., Smith, G. D., Schneider, M., & Minder, C. (1998). Bias in meta-analysis detected
by a simple, graphical test. BMJ, 315, 629–634. http://dx.doi.org/10.1136/bmj.316.
7129.469.
N.L. Vu et al. / Clinical Psychology Review 46 (2016) 25–33
Evans, S. E., Davies, C., & DiLillo, D. (2008). Exposure to domestic violence: A metaanalysis of child and adolescent outcomes. Aggression and Violent Behavior, 13,
131–140. http://dx.doi.org/10.1016/j.avb.2008.02.005.
Evans, G. W., Li, D., & Whipple, S. S. (2013). Cumulative risk and child development.
Psychological Bulletin, 139, 1342–1396. http://dx.doi.org/10.1037/a0031808.
Fantuzzo, J., Boruch, R., Beriama, A., Atkins, M., & Marcus, S. (1997). Domestic violence and
children: Prevalence and risk in five major U.S. cities. Journal of the American Academy
of Child and Adolescent Psychiatry, 36, 116–122. http://dx.doi.org/10.1097/00004583199701000-00025.
Finkelhor, D., Turner, H., Shattuck, A., & Hamby, S. L. (2015). Prevalence of childhood exposure to violence, crime, and abuse: Results from the National Survey of Children's
Exposure to Violence. JAMA Pediatrics, 169, 746–754. http://dx.doi.org/10.1001/
jamapediatrics.2015.0676.
Fosco, G. M., DeBoard, R. L., & Grych, J. H. (2007). Making sense of family violence: Implications of children's appraisals of interparental aggression for their short- and longterm functioning. European Psychologist, 12, 6–16. http://dx.doi.org/10.1027/10169040.12.1.6.
Hamby, S. (2014). Intimate partner and sexual violence research: Scientific progress, scientific challenges, and gender. Trauma, Violence & Abuse, 15, 149–158. http://dx.doi.
org/10.1177/1524838014520723.
Howell, K. H. (2011). Resilience and psychopathology in children exposed to family violence. Aggression and Violent Behavior, 16, 562–569. http://dx.doi.org/10.1016/j.avb.
2011.09.001.
Hox, J. (2010). The multilevel approach to meta-analysis. In Hox (Ed.), The multilevel approach to meta-analysis: Techniques and applications. New York, NY: Psychology Press.
Huang, C., Wang, L., & Warrener, C. (2010). Effects of domestic violence on behavior problems of preschool-aged children: Do maternal mental health and parenting mediate
the effects? Children and Youth Services Review, 32, 1317–1323. http://dx.doi.org/10.
1016/j.childyouth.2010.04.024.
Johnson, M. P. (1995). Patriarchal terrorism and common couple violence: Two forms of
violence against women. Journal of Marriage and the Family, 57, 283–294. http://dx.
doi.org/10.2307/353683.
Jouriles, E. N., McDonald, R., Slep, A. M. S., Heyman, R. E., & Garrido, E. (2008). Child abuse
in the context of domestic violence: Prevalence, explanations and practice implications. Violence and Victims, 23, 221–235. http://dx.doi.org/10.1891/0886-6708.23.2.
221.
Jouriles, E. N., McDonald, R., Vu, N. L., & Sargent, K. S. (2015). Children's exposure to intimate partner violence: Should sexual coercion be considered? Journal of Family
Psychology.
Jouriles, E. N., Norwood, W. D., McDonald, R., Vincent, J. P., & Mahoney, A. (1996). Physical
violence and other forms of marital aggression: Links with children's behavior problems. Journal of Family Psychology, 10, 223–234. http://dx.doi.org/10.1037/0893-3200.
10.2.223.
Jouriles, E. N., Rosenfield, D., McDonald, R., & Mueller, V. (2014). Child involvement in
interparental conflict and child adjustment problems: A longitudinal study of violent
families. Journal of Abnormal Child Psychology, 42, 693–704. http://dx.doi.org/10.
1007/s10802-013-9821-1.
Kitzmann, K. M., Gaylord, N. K., Holt, A. R., & Kenny, E. D. (2003). Child witnesses to domestic violence: A meta-analytic review. Journal of Consulting and Clinical
Psychology, 2, 339–352. http://dx.doi.org/10.1037/0022-006X.71.2.339.
Lohman, B. J., Neppl, T. K., Senia, J. M., & Schofield, T. J. (2013). Understanding adolescent
and family influences on intimate partner psychological violence during emerging
adulthood and adulthood. Journal of Youth and Adolescence, 42, 500–517. http://dx.
doi.org/10.1007/s10964-013-9923-7.
Margolin, G., Vickerman, K. A., Ramos, M. C., Serrano, S. D., Gordis, E. B., Iturralde, E., ...
Spies, L. A. (2009). Youth exposed to violence: Stability, co-occurrence, and context.
Clinical Child and Family Psychology Review, 12, 39–54. http://dx.doi.org/10.1007/
s10567-009-0040-9.
McDonald, R., Jouriles, E. N., Ramisetty-Mikler, S., Caetano, R., & Green, C. E. (2006). Estimating the number of American children living in partner-violent families. Journal of
Family Psychology, 20, 137–142. http://dx.doi.org/10.1037/0893-3200.20.1.137.
McDonald, R., Jouriles, E. N., Tart, C. D., & Minze, L. C. (2009). Children's adjustment problems in families characterized by men's severe violence toward women: Does other
33
family violence matter? Child Abuse & Neglect, 33, 94–101. http://dx.doi.org/10.
1016/j.chiabu.2008.03.005.
Nantel-Vivier, A., Pihl, R. O., Cote, S., & Tremblay, R. E. (2014). Developmental association
of prosocial behavior with aggression, anxiety and depression from infancy to preadolescence. Journal of Child Psychology and Psychiatry, 55, 1135–1144. http://dx.doi.
org/10.1111/jcpp.12235.
Narayan, A. J., Englund, M. M., Carlson, E. A., & Egeland, B. (2013). Adolescent conflict as a
developmental process in the prospective pathway from exposure to interparental
violence to dating violence. Journal of Abnormal Child Psychology, 42, 239–250.
http://dx.doi.org/10.1007/s10802-013-9782-4.
Neiderhiser, J. M., Marceau, K., & Reiss, D. (2013). Four factors for the initiation of substance use by young adulthood: A 10-year follow-up twin and sibling study of marital
conflict, monitoring, siblings, and peers. Development and Psychopathology, 25,
133–149. http://dx.doi.org/10.1017/S0954579412000958.
Patterson, G. R. (1982). Coercive family process. Eugene, OR: Castalia.
Podsakoff, P. M., MacKenzie, S. B., & Podsakoff, N. P. (2012). Sources of method bias in social science research and recommendations on how to control it. Annual Review of
Psychology, 63, 539–569. http://dx.doi.org/10.1146/annurev-psych-120710-100452.
Rosenthal, R. (1979). The “file drawer problem” and tolerance for null results.
Psychological Bulletin, 86, 638–641. http://dx.doi.org/10.1037/0033-2909.86.3.638.
Rutter, M. (1994). Family discord and conduct disorder: Cause, consequence, or correlate?
Journal of Family Psychology, 8, 170–186. http://dx.doi.org/10.1037/0893-3200.8.2.
170.
Schermerhorn, A. C., Cummings, E. M., DeCarlo, C. A., & Davies, P. T. (2007). Children's influence in the marital relationship. Journal of Family Psychology, 21, 259–269. http://
dx.doi.org/10.1037/0893-3200.21.2.259.
Schnurr, M. P., & Lohman, B. J. (2013). Long term effects of exposure to abuse during
childhood: Longitudinal impact of toddler's exposure to domestic violence. Journal
of Aggression, Maltreatment & Trauma, 22, 1015–1031. http://dx.doi.org/10.1080/
10926771.2013.834019.
Straus, M. A. (1979). Measuring intrafamily conflict and violence: The Conflict Tactics
(CT) Scales. Journal of Marriage and the Family, 41, 75–88. http://dx.doi.org/10.2307/
351733.
Straus, M. A., Hamby, S. L., Boney-McCoy, S., & Sugarman, D. B. (1996). The revised Conflict Tactics Scales (CTS2): Development and preliminary psychometric data. Journal
of Family Issues, 17, 283–316. http://dx.doi.org/10.1177/019251396017003001.
Turner, H. A., Finkelhor, D., Ormrod, R., Hamby, S., Leeb, R. T., Mercy, J. A., & Holt, M.
(2012). Family context, victimization, and child trauma symptoms: Variations in
safe, stable, and nurturing relationships during early and middle childhood.
American Journal of Orthopsychiatry, 82, 209–219. http://dx.doi.org/10.1111/j.19390025.2012.01147.x.
Viechtbauer, W. (2010). Conducting meta-analyses in R with the metafor package. Journal
of Statistical Software, 36, 1–48.
Ware, H. S., Jouriles, E. N., Spiller, L. C., McDonald, R., Swank, P. R., & Norwood, W. D.
(2001). Conduct problems among children at battered women's shelters: Prevalence
and stability of maternal reports. Journal of Family Violence, 16, 291–307. http://dx.
doi.org/10.1023/A:1011190316783.
Wolfe, D. A., Crooks, C. V., Lee, V., McIntyre-Smith, A., & Jaffe, P. G. (2003). The effects of
children's exposure to domestic violence: A meta-analysis and critique. Clinical
Child and Family Psychology Review, 6, 171–187. http://dx.doi.org/10.1023/A:
1024910416164.
Yates, T. M., Dodds, M. F., Sroufe, A., & Egeland, B. (2003). Exposure to partner violence and
child behavior problems: A prospective study controlling for child physical abuse and
neglect, child cognitive ability, socioeconomic status, and life stress. Development and
Psychopathology, 15, 199–218. http://dx.doi.org/10.1017/S0954579403000117.
Zarling, A. L., Taber-Thomas, S., Murray, A., Knuston, J. F., Lawrence, E., Valles, N., ... Bank, L.
(2013). Internalizing and externalizing symptoms in young children exposed to intimate partner violence: Examining intervening processes. Journal of Family Psychology,
27, 945–955. http://dx.doi.org/10.1037/a0034804.