The Association Between PTSD Symptoms and IPV Perpetration Across 6 Years

Abstract

A large literature documents that posttraumatic stress disorder (PTSD) symptoms are associated with intimate partner violence (IPV) perpetration among adults. However, research on this relationship among adolescents and young adults has been plagued by methodological flaws (e.g., cross-sectional designs). Thus, the purpose of the present study was to examine the longitudinal and bidirectional associations between PTSD symptoms and psychological and physical IPV perpetration from adolesence to young adulthood. A sample of racially and ethnically diverse high school students (N = 1,042; 56% female) were assessed annually for 6 years (from 2010 to 2015 in Southeastern Texas). At each assessment, participants completed measures of PTSD symptoms and psychological and physical IPV perpetration. The mean age of the sample at the first assessment was 15.09 (SD = .79). Structural equation modeling demonstrated that PTSD symptoms at Years 2, 3, and 4 predicted increases in psychological IPV perpetration in the subsequent year. In turn, psychological IPV perpetration at Years 1 and 4 predicted increases in PTSD symptoms in the subsequent years. In addition, psychological IPV perpetration mediated the association between PTSD symptoms and physical IPV perpetration over time. Results were consistent across gender and race/ethnicity. Findings provide initial evidence that PTSD symptoms are associated with IPV perpetration across time from adolescence to young adulthood. Prevention and intervention programs for adolescent and young adult IPV perpetration may benefit from screening for, and potentially treating, PTSD symptoms.

Keywords

posttraumatic stress disorder intimate partner violence adolescence

Due to the high prevalence of intimate partner violence (IPV) perpetration among adolescents, and the significant implications it has for mental and physical health and long-term adjustment, it is essential that studies focus on IPV prevention and intervention programs. A key step to this endevour is to arrive at a clearer understanding of the longitudinal predictors of IPV perpetration. One potential risk factor for IPV perpetration, which has received minimal attention among adolescents, is posttraumatic stress disorder (PTSD) symptoms. Understanding the contribution of PTSD symptomatology on IPV perpetration can have crucial ramifications for intervention and prevention programs that may need to address this risk factor. Indeed, some researchers have advocated for the treatment of PTSD as one avenue for the treatment of IPV perpetration (Taft, Macdonald, Creech, Monson, & Murphy, 2016). Therefore, in the present study, we examined the longitudinal and bidirectional realtionships between IPV and PTSD symptoms from adolescence to young adulthood.

In the current study, IPV refers to psychological (e.g., threats made toward partner, insulting comments toward partner) and physical (e.g., hitting, slapping, kicking a partner) violence. The annual prevalence rate for psychological IPV perpetration is approximately 80% and the annual prevalence rate for physical IPV perpetration is 20% among adolescents and young adults who are in a dating relationship (Shorey, Cornelius, & Bell, 2008). Moreover, males and females perpetrate IPV at similar rates in adolescence and young adulthood (Shorey et al., 2008; Vagi et al., 2013). Adolescent victims of IPV experience a range of negative mental and physical health consequencs, including depressive symptoms (Exner-Cortens, Eckenrode, & Rothman, 2013), anxiety (Holt & Espelage, 2005), substance use (Temple & Freeman, 2011), suicidal ideation (Exner-Cortens et al., 2013), and physical injuries (Foshee, 1996). Thus, IPV is a prevalent public health problem, and continued research is needed to determine what predicts IPV perpetration across time.

PTSD Symptoms and IPV: Theoretical Considerations

Theoretical models that attempt to explain the relationship between PTSD symptoms and aggression largely draw on information processing theories (e.g., Chemtob, Novaco, Hamada, Gross, & Smith, 1997; Murphy, 2013; Taft, Creech, & Murphy, 2017). The social information processing (SIP) model, first proposed by McFall (1982) and later expanded by Crick and Dodge (1994) and Dodge (1986), proposes that social information processesing occurs through the sequential stages of encoding, interpretation, response generation, response evaluation, and behavioral response. Individuals with PTSD may be at elevated risk of aggression when SIP mechanisms are altered, for example, through the misattributions of one’s partner’s intent (Murphy, 2013) or through hostile attributional biases. This model has been successfully applied to understand IPV perpetration. For example, IPV perpetrators have been found to exhibit hostile cognitive biases (Holtzworth-Munroe, 2000), anticipate experiencing more anger than nonviolent men in response to negative hypothetical partner behaviors (Holtzworth-Munroe & Smutzler, 1996), and to attribute more negative intentions to female partners in hypothetical scenarios (Holtzworth-Munroe & Hutchinson, 1993).

With respect to PTSD and IPV perpetration, researchers have studied the role of anger regulation (e.g., Chemtob et al., 1997) and threats to the self, or shame (e.g., Sippel & Marshall, 2011), in understanding the role of PTSD symptomatology and IPV perpetration within an SIP framework. These theories suggest that individuals with PTSD are likely to experience deficits in anger regulation and are more likely to perpetrate violence due to hypervigilance and the activation of a survival mode, which leads to biased threat perception, such as overgeneralized negative appraisals of social information (Chemtob et al., 1997). Consistent with this theory, several studies have found that hyperarousal symptoms play a unique role in contributing to aggressive behavior (e.g., Taft, Kaloupek, et al., 2007). Shame-related SIP theories propose that elevated levels of shame in indiviudals with PTSD (e.g., Leskela, Dieperink, & Thuras, 2002) may be associated with negative social evaluations, such as anticipated rejection and maladatpive relationship behaviors (Sippel & Marshall, 2011). These explanations also focus on maladaptive coping strategies to address shame-related cognitions, which may lead to externalization of blame, hostility, and aggressive behavior (Sippel & Marshall, 2011).

In addition to SIP theories, stress and strain theories have also been utilized to explain this association in adolescence. Strain theories (e.g., Agnew, 1992) suggest that strain in interpersonal relationships, such as among friends and romantic partners, may increase negative affect, leading to a higher risk of engaging in aggressive behavior. Specifically, Agnew (1992) identified three types of strain: when events (a) prevent or threaten to prevent one from achieving positively valued goals, (b) remove or threaten to remove positively valued stimuli that one possesses, or (c) present or threaten to present one with noxious or negatively valued stimuli (Agnew, 1992). These forms of strain may lead to experiencing negative emotions (e.g., disappointment, depression, fear, and anger), and aggression may be one way for alleviating this strain. According to Agnew (1992), the association between strain and aggression is most likely to hold when there are constraints to other forms of coping or the adolescent has a disposition for delinquent coping. Thus, for individuals with elevated PTSD symptoms, this strain may exacerbate the risk of IPV perpetration in the absence of adaptive coping resources or ineffective emotion regulation strategies, particularly, for negative emotions (Ehring & Quack, 2010; Jakupcak et al., 2007). In conjunction with SIP theory, these findings suggest that adolescence is an important developmental period to understand how PTSD symptomatology may play a unique role in the context of close relationships by presenting as a risk factor for IPV perpetration.

PTSD Symptoms and IPV: Empirical Associations

Following a body of research demonstrating an association between PTSD and general aggression (e.g., Taft, Vogt, Marshall, Panuzio, & Niles, 2007), researchers have become increasingly interested in understanding the association between PTSD and IPV perpetration. A large literature in adults has demonstrated PTSD symptoms to be positively associated with psychological and physical IPV perpetration among men and women (Bell & Orcutt, 2009; Taft, Watkins, Stafford, Street, & Monson, 2011). These studies suggest that individuals with higher levels of PTSD symptomatology also report higher levels of anger, hostility, and aggressiveness (e.g., Jakupcak et al., 2007). The vast majority of studies examining the relationship between PTSD and IPV perpetration have been conducted with veteran samples (e.g., Jakupcak et al., 2007; Taft, Kaloupek, et al., 2007). However, similar associations have also been demonstrated in civilian samples (e.g., Rosenbaum & Leisring, 2003) and individuals arrested for domestic violence (Shorey, Febres, Brasfield, & Stuart, 2012).

Despite the growing interest in the association between PTSD symptomatology and IPV perpetration, only a handful of studies have examined this association among adolescents. In a cross-sectional study of high school and child protective service (CPS)-involved adolescents, Wekerle and colleagues (2001) demonstrated a positive association between trauma symptoms and IPV perpetration for males and females. However, this study did not specify the type of IPV examined. In a similar cross-sectional study with CPS-involved adolescents, Wekerle and colleagues (2009) demonstrated positive associations between trauma symptoms and IPV perpetration for males and females. However, the IPV measure combined different types of aggression into a single variable. In a 1-year longitudinal study of high school students, Wolfe, Wekerle, Scott, Straatman, and Grasley (2004) demonstrated that trauma symptoms predicted psychological IPV perpetation across the year for boys. For girls, anger related to trauma predicted a combined IPV perpetration score (i.e., physical and psychological) over the course of the year.

The small body of research conducted to date on PTSD symptoms and IPV perpetration among adolescents has been limited in several important ways. Notably, studies have failed to examine distinct types of IPV individually. Because prior research suggests that psychological IPV perpetration precedes and increases the risk of physical IPV perpetration (Murphy & O’Leary, 1989), and mediates the association between other risk factors (e.g., alcohol) and physical IPV perpetration (Ortiz, Shorey, & Cornelius, 2015), it is plausible that PTSD symptoms are associated with physical IPV perpetration through psychological IPV perpetration. In addition, it is unclear whether PTSD symptoms are associated with IPV perpetration across periods of time longer than 1 year among adolescents. Because IPV perpetration among adolescents is remarkably stable across time (Capaldi, Shortt, & Crosby, 2003; O’Leary & Smith Slep, 2003), knowledge of predictors of IPV perpetration over time could inform adolescent and young adult IPV prevention and intervention programs on risk factors to target. In addition, although theory and research with adult populations suggest that PTSD symptoms precede and increase the risk of IPV perpetration (Bell & Orcutt, 2009), it is plausible that violence perpetration would increase risk of PTSD symptoms due to the violent and traumatic nature of IPV (e.g., Hecker et al., 2013). Finally, prior research on the association between PTSD symptoms and IPV perpetration among adolescents has failed to consider potential racial/ethnic differences in this relationship. Knowledge on whether this association varies across racial/ethinic groups would provide important information for culturally relevant and sensitive IPV prevention programming.

Current Study

The purpose of the present study was to, therefore, examine the longitudinal, bidirectional associations between PTSD symptoms and psychological and physical IPV perpetration from adolesence to young adulthood. Across 6 years, we hypothesized that (a) PTSD symptoms would be associated with both forms of IPV perpetration over time and (b) psychological IPV would mediate the association between PTSD symptoms and physical IPV perpetration. Finally, we examined whether associations varied across gender and race/ethnicity. However, we did not provide any a priori hypotheses concerning gender or racial/ethnic differences due to limited research on demographic differences in this association among adolescent populations.

Method

Participants

Participants for the current study are from an ongoing longitudinal study of adolescent health (Shorey, Fite, Torres, Stuart, & Temple, 2019; Temple, Shorey, Fite, Stuart, & Le, 2013). A total of 1,042 high school students from multiple public schools in southeast Texas participated. Participants in the current study were followed annually starting in the spring of 2010. At the first assessment, the sample had a mean age of 15.09 years (SD = .79), were 56% were female, and self-identified as Hispanic (31.4%), White (29.4%), African American/Black (27.9%), and other (11.3%). At the first assessment, the sample was primarily in ninth (n = 781) or 10th (n = 250) grade, followed by 11th grade (n = 11). At the first assessment, 17.8% indicate their mother had not graduated from high school, 18.7% indicated their mother finished high school, 24.4% had some college or technical school education, 26.4% finished college, and 12.8% did not know their mother’s education level. For fathers, the rates were 16.4% (did not finish high school), 18.8% (finished high school), 13.7% (some college/technical school), 22.1% (finished college), and 29.0% (did not know). We examined data from Wave 1, Wave 2 (n = 964), Wave 3 (n = 894), Wave 4 (n = 776), Wave 5 (n = 698), and Wave 6 (n = 758).

Procedure

To ensure a representative sample of adolescents, recruitment occurred during school hours in courses with mandated attendance, including English (four schools), world geography (one school), and health (two schools). All students in the selected classes were eligible to participate. Research staff attended each class twice prior to the assessment to describe the purpose of the study, general design, and answer any questions. The same research staff conducted all recruitment sessions. Detailed information about the study along with a parental permission form was sent home with the students for their parents to review, sign, and return (response rate = 62%). Assessments occurred during school hours. Participants were “pulled” from the class from which they were recruited, escorted to a schoolroom reserved for the study, and completed the survey in a group setting with ample space between participants to ensure confidentiality. For participating, students received a US$10 gift card at Waves 1, 2, and 3, a US$20 gift card at Waves 4 and 5, and a US$30 gift card at Wave 6. To increase reliability of self-report, teachers and other school administrators were not allowed to be present during questionnaire administration. After each survey, participants received a referral list for mental health resources in their local area, and also could indicate whether they wanted to speak with a school counselor or the principal investigator of the study, who is a licensed psychologist. When participants graduated high school, administration moved from paper-pencil to web-based surveys. Parental consent and student assent/consent were obtained and all procedures received institutional review board approval.

Measures

IPV perpetration

The perpetration of psychological and physical IPV was assessed each year using the Conflict in Adolescent Dating Relationships Inventory (CADRI; Wolfe et al., 2001). Participants indicated whether each behavior occurred utilizing a yes/no response format. All questions at the first assessment referred to their entire lifetime and questions during Years 2 to 6 referred to the previous year (i.e., since the last survey). Ten items assessed psychological IPV perpetration and four items assessed physical IPV perpetration. A total score for each type of IPV is calculated by summing all items for each subscale, with higher scores corresponding to a greater number of violent acts perpetrated. In prior research, the CADRI has shown good psychometric properties (Wolfe et al., 2001). Internal consistencies in the current sample for psychological IPV perpetration ranged from .80 to .85 and physical IPV perpetration ranged from .76 to .86 across all years.

PTSD symptoms

The primary care PTSD screen (PC-PTSD; Prins, Ouimette, Kimerling, & et al, 2003) was used to measure PTSD symptoms at each assessment. The PC-PTSD includes four items that each assess a core feature of PTSD, including re-experiencing, numbing, hyperarousal, and avoidance. Each item is rated as yes/no and all items are summed to create a total score. The PC-PTSD has demonstrated good diagnostic capabilities for PTSD, with sensitivity ranging from .78 to .91 and specificity ranging from .72 to .87 (Bliese et al., 2008; Prins et al., 2003). Moreover, prior research suggests the PC-PTSD has similar sensitivity/specificity as the PTSD Checklist (Bliese et al., 2008) and has demonstrated good psychometric properties in adolescent samples (M. J. Mason, Mennis, & Schmidt, 2011). In the current sample, internal consistencies ranged from .73 to .82 across all years.

Data Analytic Strategy

Analyses were conducted using structual equation modeling (SEM) in Mplus version 7.0. Full information maximum likelihood estimation (FIMLE) was utilized, as FIMLE does not exclude observations with missing data and thus uses all data to estimate parameters (Kline, 2011). In addition, robust maximum likelihood estimation was used due to study variables being non-normally distributed (i.e., IPV), as this estimate employs maximum likelihood parameter estimates with standard errors and a chi-square statistic that are robust to non-normality (Kline, 2011).

Model fit was evaluated using the chi-square statistic (χ²), the root mean squared error of approximation (RMSEA), and the comparative fit index (CFI). The chi-square fit index is calculated by dividing the chi-square estimate by the degrees of freedom, with values of less than 2.0 indicative of good fit (Hu & Bentler, 1999). The RMSEA is an indicator of model error per degrees of freedom, with values close to, or less than, .06 indicating that the model fit the data well (Hu & Bentler, 1999). The CFI contrasts the estimated model’s fit to that of the null, or “independence” model, with a value of .95 or higher indicative of good fit (Hu & Bentler, 1999).

Cross-lagged associations between PTSD, psychological IPV perpetration, and physical IPV perpetration were examined. This model included stability paths for each construct and cross-lagged associations between all variables across all 6 years. Mediation was examined using the bias-corrected bootstrap method with 500 bootstrap samples and 95% bias-corrected confidence intervals (CIs), which provides a more optimal balance between Type I and II errors relative to other methods for examining mediation (MacKinnon, Lockwood, & Williams, 2004). We examined potential-mediated pathways across time to establish temporal precedence. For example, PTSD symptoms at one assessment (e.g., Year 1) predicted psychological IPV perpetration at the next assessment (e.g., Year 2), which then predicted physical IPV perpetration at the next assessment (e.g., Year 3).

To examine whether results varied across gender or race/ethnicity, the multiple group model (MGM) approach was utilized. The MGM approach is conducted in two parts. In the first step, a model where structural paths are free to vary across levels of the moderator (e.g., gender) is estimated. In the second step, a model that specifies structural paths across levels of the moderator to be constrained to be equal is estimated. A chi-square difference test (Δχ²) between the unconstrained and constrained models is utilized to test whether constraining paths across levels of the moderator results in a significant decrement in the model chi-square (Muthén & Muthén, 2006). If the model chi-square evidences a significant decrement, it can then be assumed that paths varied across levels of the moderator. We conducted the MGM analyses in an iterative process, whereby we examined potential gender differences in stability paths for each construct, one at a time, followed by predictors of each construct, one at a time. This same process was repeated to examine potential racial/ethnic differences (Hispanic, White, African American/Black, and “other”).

Results

Descriptive Statistics

Bivariate correlations, means, and standard deviations among study variables are displayed in Table 1. With a few exceptions, study variables were positively and significantly associated with each other over time. Inspection of the means for PTSD symptoms showed relatively consistent ratings across time, with a slight decrease in symptoms at the final assessment. The prevalence of physical IPV perpetration ranged from 15.8% (Wave 5) to 21.5% (Wave 1) and psychological IPV perpetration ranged from 68.2% (Wave 6) to 81.7% (Wave 1). We also examined whether the data were missing completely at random (MCAR). Little’s MCAR test, for which the null hypothesis is that the data are MCAR, was nonsignificant χ²(2008) = 1909.09, p = .94. Thus, our use of FIMLE estimates in subsequent analyses was supported as data were not missing in any systematic fashion.

Table 1.

Means, Standard Deviations, and Correlations of Study Variables.

	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15	16	17	18
1. PTSD Y1	—	.41*	.33*	.30*	.28*	.25*	.08*	.12*	.12*	.12*	.12*	.03	.23*	.17*	.15*	.19*	.08*	.08*
2. PTSD Y2		—	.45*	.34*	.34*	.36*	.09*	.17*	.12*	.12*	.16*	.10*	.19*	.25*	.20*	.19*	.16*	.16*
3. PTSD Y3			—	.40*	.36*	.33*	.11*	.13*	.20*	.15*	.22*	.15*	.18*	.18*	.25*	.22*	.19*	.17*
4. PTSD Y4				—	.41*	.41*	.07	.05	.08*	.09*	.17*	.12*	.10*	.12*	.13*	.17*	.18*	.15*
5. PTSD Y5					—	.44*	.05	.05	.10*	.12*	.22*	.15*	.11*	.08	.14*	.16*	.23*	.17*
6. PTSD Y6						—	.07	.08*	.14*	.12*	.16*	.21*	.09*	.17*	.15*	.20*	.15*	.26*
7. Physical Y1							—	.46*	.34*	.23*	.16*	.26*	.46*	.33*	.26*	.19*	.14*	.23*
8. Physical Y2								—	.51*	.33*	.14*	.24*	.31*	.47*	.28*	.27*	.14*	.24*
9. Physical Y3									—	.46*	.32*	.28*	.27*	.30*	.47*	.33*	.24*	.32*
10. Physical Y4										—	.37*	.36*	.19*	.19*	.34*	.48*	.24*	.33*
11. Physical Y5											—	.42*	.23*	.19*	.28*	.32*	.47*	.30*
12. Physical Y6												—	.23*	.21*	.20*	.29*	.31*	.52*
13. Psych. Y1													—	.54*	.48*	.43*	.34*	.35*
14. Psych. Y2														—	.55*	.46*	.38*	.36*
15. Psych. Y3															—	.57*	.47*	.44*
16. Psych. Y4																—	.58*	.52*
17. Psych. Y5																	—	.53*
18. Psych. Y6																		—
M	1.20	1.12	1.17	1.02	1.09	0.96	0.41	0.37	0.38	0.30	0.31	0.39	3.45	3.23	3.26	3.06	2.94	2.99
SD	1.35	1.37	1.46	1.38	1.41	1.38	0.93	0.89	0.98	0.81	0.82	0.95	2.70	2.72	2.75	2.86	2.89	2.92

Note. Y = year; PTSD = posttraumatic stress disorder; Physical = physical intimate partner violence perpetration; Psych. = psychological intimate partner violence perpetration.

p < .05.

Cross-Lagged Analyses

The initially estimated cross-lagged path model demonstrated adequate fit to the data, χ²(90) = 445.94, p < .05, RMSEA = .06; CFI = .89. However, modification indicies were consulted to determine which additional parameters, if included, would improve model fit while also being theoretically and conceptually sound. Modification indices indicated the need to include residual covariances between PTSD variables over time (i.e., Year 2 with Years 4, 5, and 6; Year 3 with Year 5; Year 4 with Year 6). This modified model provided a good fit to the data, χ²(85) = 367.32, p < .05, RMSEA = .05; CFI = .91.

As displayed in Figure 1 and Table 2, PTSD symptoms remained stable across the 6 years. As reported on elsewhere, physical IPV perpetration and psychological IPV perpetration remained stable across the 6 years (Shorey et al., 2019). Examination of the associations between constructs over time (i.e., cross-lagged paths) revealed some interesting patterns (see Figure 1 and Table 3). Not surprisingly, psychological IPV perpetration consistently predicted increases in physical IPV perpetration the following year across the duration of the study. However, only Year 1 physical IPV perpetration predicted increases in psychological IPV perpetration at Year 2. Year 4 PTSD symptoms predicted increases in physical IPV perpetration at Year 5.

Figure 1.

Longitudinal associations between PTSD symptoms and IPV perpetration.

Table 2.

Standardized Betas and Standard Errors for Stability Paths.

Predictor →	PTSD Year 2β (SE)	PTSD Year 3β (SE)	PTSD Year 4β (SE)	PTSD Year 5β (SE)	PTSD Year 6β (SE)
Prior year PTSD	.33 (.03)***	.41 (.03)***	.34 (.04)***	.35 (.04)***	.36 (.04)***
Prior year PTSD	Physical IPV Year 2β (SE)	Physical IPV Year 3β (SE)	Physical IPV Year 4β (SE)	Physical IPV Year 5β (SE)	Physical IPV Year 6β (SE)
Prior year physical IPV	.39 (.05)***	.47 (.06)***	.39 (.06)***	.32 (.07)***	.36 (.06)***
Prior year physical IPV	Psychological IPV Year 2β (SE)	Psychological IPV Year 3β (SE)	Psychological IPV Year 4β (SE)	Psychological IPV Year 5β (SE)	Psychological IPV Year 6β (SE)
Prior year psychological IPV	.49 (.03)***	.53 (.03)***	.53 (.04)***	.58 (.04)***	.49 (.04)***

Note. PTSD = posttraumatic stress disorder; IPV = intimate partner violence.

***

p < .001.

Table 3.

Standardized Betas and Standard Errors for Cross-Lagged Paths.

Predictor →	PTSD Year 2β (SE)	PTSD Year 3β (SE)	PTSD Year 4β (SE)	PTSD Year 5β (SE)	PTSD Year 6β (SE)
Prior year psychological IPV	.12 (.03)*	.07 (.04)	.04 (.04)	.08 (.04)*	.03 (.04)
Prior year physical IPV	.00 (.04)	.03 (.04)	.00 (.04)	.07 (.04)	.03 (.05)
Prior year physical IPV	Physical IPV Year 2β (SE)	Physical IPV Year 3β (SE)	Physical IPV Year 4β (SE)	Physical IPV Year 5β (SE)	Physical IPV Year 6β (SE)
Prior year PTSD	.06 (.04)	.03 (.04)	.05 (.04)	.14 (.04)**	.03 (.04)
Prior year psychological IPV	.12 (.04)**	.09 (.04)*	.16 (.04)***	.18 (.04)***	.13 (.04)**
Prior year psychological IPV	Psychological IPV Year 2β (SE)	Psychological IPV Year 3β (SE)	Psychological IPV Year 4β (SE)	Psychological IPV Year 5β (SE)	Psychological IPV Year 6β (SE)
Prior year PTSD	.05 (.03)	.08 (.03)*	.09 (.03)*	.11 (.04)*	.04 (.04)
Prior year physical IPV	.10 (.04)*	.04 (.04)	.07 (.04)	−.02 (.04)	.07 (.05)

Note. PTSD = posttraumatic stress disorder; IPV = intimate partner violence.

p < .05. **p < .01. ***p < .001.

However, physical IPV perpetration did not predict changes in PTSD symptoms at any year. In contrast, PTSD symptoms and psychological IPV perpetration were related across time. Specifically PTSD symptoms at Years 2, 3, and 4 predicted increases in psychological IPV perpetration in the subsequent year. In turn, psychological IPV perpetration at Years 1 and 4 predicted increases in PTSD symptoms in the subsequent year.

Mediation Analyses

Mediation analyses demonstrated no significant mediation effect when examining psychological IPV perpetration at Year 2 as a mediator between PTSD symptoms at Year 1 and physical IPV perpetration at Year 3, B = .005, 95% CI = [–.001, .012]. However, significant effects were found for psychological IPV perpetration at Year 3 mediating the association between PTSD symptoms at Year 2 and physical IPV perpetration at Year 4 (B = .008, 95% CI = [.003, .014]); for psychological IPV perpetration at Year 4 mediating the association between PTSD symptoms at Year 3 and physical IPV perpetration at Year 5 (B = .009, 95% CI = [.003, .019]); and for psychological IPV perpetration at Year 5 mediating the association between PTSD symptoms at Year 4 and physical IPV perpetration at Year 6 (B = .010, 95% CI = [.004, .021]).

MGM Analyses

Results of MGM analyses demonstrated that constraining the stability paths for PTSD symptoms to be equal across gender and race/ethnicity resulted in nonsignficant changes to the model, Δχ²(5) = 6.17, p > .05 and Δχ²(15) = 9.36, p > .05, respectively. Similar results were found for the stability of psychological IPV perpetration, Δχ²(5) = 3.08, p > .05 and Δχ²(15) = 17.45, p > .05, for gender and race/ethnicity, respectively. Stability paths for physical IPV perpetration also did not vary across gender, Δχ²(5) = 10.01, p > .05, or race/ethnicity, Δχ²(15) = 3.02, p > .05.

Results of MGM analyses further demonstrated that constraining the influence of PTSD symptoms on psychological and physical IPV perpetration to be equal across gender and race/ethnicity resulted in nonsignficant changes to the model, Δχ²(10) = 6.38, p > .05 and Δχ²(30) = 33.55, p > .05, respectively. Similarly, constraining the influence of psychological IPV perpetration on physical IPV perpetration and PTSD symptoms across gender and race/ethnicity resulted in nonsignificant changes to the model, Δχ²(10) = 11.09, p > .05 and Δχ²(30) = 40.86, p > .05. Finally, constraining the influence of physical IPV perpetration on psychological IPV perpetration and PTSD symptoms to be equal across gender and race/ethnicity resulted in nonsignficant changes to the model, Δχ²(10) = 16.06, p > .05 and Δχ²(30) = 35.02, p > .05, respectively.

Discussion

Despite a robust literature documenting PTSD symptoms as a risk factor for IPV perpetration in adult samples (Bell & Orcutt, 2009), and speculation that treatment of PTSD symptoms may result in reduced IPV perpetration (Taft et al., 2016), there has been a dearth of research on whether PTSD symptoms longitudinally predicts IPV perpetration among adolescents and young adults. Thus, we examined PTSD symptoms as a predictor of psychological and physical IPV perpetration across 6 years in a racially and ethnically diverse sample of adolescents. We also explored whether PTSD symptoms and IPV perpetration had bidirectional effects.

Findings demonstrated that PTSD symptoms were a fairly consistent predictor of psychological IPV perpetration over time. Specifically, PTSD symptoms at Years 2, 3, and 4 predicted increases in psychological IPV perpetration in the subsequent year. Importantly, this relationship was evident after accounting for the influence of prior psychological and physical IPV perpetration. This is one of the first studies to longitudinally demonstrate an association between PTSD symptoms and adolescent psychological IPV perpetration and is consistent with findings seen in adult and military veteran populations (e.g., Rosenbaum & Leisring, 2003; Taft et al., 2011). From a strain theory perspective (Agnew, 1992), it is likely that conflicts in intimate relationships lead to anger and frustration, increasing risk of psychological IPV perpetration. In addition, PTSD symptoms, which are associated with increased anger and decreased emotion regulation capabilities (Ehring & Quack, 2010; Jakupcak et al., 2007), may lower the thresehold for psychological IPV perpetration. It is also likely, similar to information processing theories (e.g., Chemtob et al., 1997), that elevated PTSD symptoms are associated with a state of hypervigilance and a “survival mode” of functioning, leading to risk of IPV perpetration. Additional research is needed that explores the mechanisms responsible for this association.

In contrast to the findings for psychological IPV perpetration, PTSD symptoms were only directly related to increases in physical IPV perpetration in 1 of the 6 years. This finding is in contrast to research with adult and military veteran populations, which consistently find associations between PTSD symptoms and physical violence perpetration (Bell & Orcutt, 2009). However, our findings showed that PTSD symptoms were indirectly related to physical IPV perpetration through psychological IPV perpetration. That is, PTSD symptoms predicted increases in psychological IPV perpetration which, in turn, predicted increases in physical IPV perpetration. Indeed, prior research suggests that one of the strongest predictors of physical IPV perpetration is psychological IPV perpetration (Baker & Stith, 2008; Murphy & O’Leary, 1989). Thus, PTSD symptoms, though not directly associated with physical IPV perpetration, impact risk of physical IPV perpetration by increasing the risk of psychological IPV perpetration.

Our findings also provided evidence that the association between PTSD symptoms and IPV perpetration is largely not bidirectional. That is, physical IPV perpetration did not predict increases in PTSD symptoms over time, and psychological IPV perpetration only predicted increases in PTSD symptoms in 2 of the 6 years. Still, these preliminary findings suggest that psychological IPV perpetration is a risk factor for PTSD symptoms, and additional research is needed to understand the mechanism for this relationship. It is plausible that this link is mediated by IPV victimization, as perpetrating IPV is associated with increased risk of IPV victimization (Baker & Stith, 2008). Future research is needed to explore this possibility.

In addition, our results demonstrated high levels of stability across 6 years for PTSD symptoms, psychological IPV, and physical IPV. The extant literature suggests that IPV perpetration is remarkably stable across time for adolescents and young adults (Capaldi et al., 2003; O’Leary & Smith Slep, 2003). However, this literature has been restricted to relatively short-term follow-ups for adolescent samples (e.g., 3 months; O’Leary & Smith-Slep) and young adults (e.g., 2.5 years; Capaldi et al., 2003). Thus, these findings add to the literature by documenting moderate levels of stability for PTSD symptoms, physical and psychological IPV perpetration across 6 years from adolescence to young adulthood. This suggests that IPV prevention and intervention programs are needed throughout the course of adolescence, and into young adulthood, and that PTSD symptoms are likely to persist unless treated, which will increase the risk of IPV perpetration.

Diversity Considerations

Our study was unique in its ability to examine the longitudinal relationships between PTSD symptoms and IPV perpetration in a large sample of racially and ethnically diverse adolescents. Moreover, we examined whether our findings varied between males and females. Across all findings, there were no differences based on race/ethnicity or gender, suggesting that PTSD symptoms are a predictor of IPV perpetration irrespective of these characteristics and, therefore, may be a universal risk factor for IPV perpetration. These findings should be considered preliminary until replicated, however, as the majority of prior research on PTSD symptoms and IPV perpetration has been largely restricted to samples comprised primarily of White, non-Hispanic males. Furthermore, as mentioned previously, minimal research has examined PTSD symptoms as a predictor of IPV perpetration in adolescence, particularly among diverse adolescents. Although our sample was highly diverse with respect to race and ethnicity, we were underpowered to examine whether findings varied across certain racial and ethnic groups (e.g., Asian Americans, Alaskan Natives). Additional research with more diverse samples is needed to continue to examine the longitudinal associations between PTSD symptoms and IPV perpetration in adolescence.

Future Directions

Pending replication, these findings suggest that IPV prevention and intervention programs for adolescents and young adults could focus efforts on reducing PTSD symptoms and psychological IPV perpetration, which may have concurrent benefits of reducing physical IPV perpetration. Indeed, researchers have advocated for IPV prevention programming to focus their efforts, in part, on reducing psychological aggression (e.g., Lawrence, Yoon, Langer, & Ro, 2009; Shorey, Febres, Brasfield, & Stuart, 2012) due to its strong association with physical IPV perpetration. Our findings provide additional support to this notion, with psychological IPV perpetration mediating the association between PTSD symptoms and physical IPV perpetration. Adaptive conflict resolution skills, including communication skills training, may be a fruitful area for IPV prevention programs to focus on with adolescents in an effort to reduce psychological IPV perpetration.

Moreover, our findings also suggest that targeting PTSD symptoms among adolescents may also be an important area for IPV prevention programs. Recent adolescent trauma-informed IPV screening initiatives have suggested the need to target modifiable risk factors stemming from trauma as opposed to static, past adversities themselves (Cohen, Shorey, Menon, & Temple, 2018; Thurston & Howell, 2018). Our findings that PTSD confers risk of IPV perpetration suggests that trauma reactions may be useful to include in IPV prevention protocols and to help connect vulnerable youth to necessary services such as existing interventions for adolescent PTSD (e.g., trauma-focused cognitive behavioral therapy). In addition, our findings suggest that IPV prevention and intervention programs are needed throughout the course of adolescence, and into young adulthood, as IPV perpetration remained stable across time.

Limitations

The current study has several limitations. The measure of PTSD symptoms used, while psychometrically sound, is intended to be a screening tool for PTSD and not a comprehensive measure of the diverse symptoms that comprise this disorder. Future research should utilize a comprehensive measure of PTSD symptoms, or structured diagnostic interviews, to examine associations between different symptom clusters of PTSD (e.g., hypervigilance, avoidance) and IPV perpetration. The mechanisms for the association between PTSD symptoms and IPV perpetration (e.g., anger, shame) was not assessed, and future research should examine these potential mechanisms. The sample was obtained from a specific region of the United States, and while the prevalence of measured variables were comparable to national rates, generalizability of findings remains a concern. Nationally representative samples should be utilized in future research. We also have no information on the adolescents who did not participate in this study, and thus are unable to determine if they differed from adolescents who did participate in the study on key variables (e.g., PTSD symptoms). It is possible that the method of survey administration may have affected results (i.e., in-person, online), although prior research suggests measures of IPV administered in these different formats provide equivalent findings (e.g., Brock et al., 2015). Finally, social desirability may have affected study findings (e.g., underreporting of IPV) and future research should assess for this possibility.

Summary

In all, our findings add to the body of literature on PTSD symptoms and IPV perpetration, as this was one of the first studies to examine longitudinal associations between PTSD symptoms and psychological and physical IPV perpetration from adolescence into young adulthood. Findings demonstrated that PTSD symptoms are associated with psychological IPV perpetration over time, and indirectly associated with physical IPV perpetration through psychological IPV perpetration. Importantly, these results were consistent across gender and race/ethnicity, suggesting PTSD symptoms may be a universal risk factor for adolescent IPV perpetration. Given the serious public health consequences of IPV, prevention and intervention programs for adolescent and young adult IPV perpetration may benefit from screening adolescents for PTSD symptoms and providing trauma-informed care for individuals with elevated symptomatology.

Footnotes

Authors’ Note

Ryan C. Shorey is now affiliated with University of Wisconsin-Milwaukee, USA.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The current manuscript was supported, in part, by grants 2016-R2-CX-0035 and 2012-WG-BX-0005 from the National Institute of Justice (NIJ) awarded to Drs. Shorey and Temple, respectively. This work was also supported, in part, by grant K24AA019707 from the National Institute on Alcohol Abuse and Alcoholism awarded to Dr. Stuart. This work was also supported, in part, by grant K23HD059916 from the Eunice Kennedy Shriver National Institute of Child Health & Human Development (NICHD) awarded to Dr. Temple. The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH or NIJ.

Author Biographies

Ryan C. Shorey, PhD, is an assistant professor of Psychology at the University of Wisconsin-Milwaukee. His research focuses on intimate partner violence, substance misuse, and the intersection between these two public health problems.

Paula J. Fite, PhD, is a Dean’s professor at the University of Kansas. Her research focuses on the etiology and developmental progression of problem behavior, including aggression and substance use.

Suvarna V. Menon, MA, is a doctoral candidate at the University of Illinois at Urbana-Champaign and a current predoctoral clinical intern at SUNY Upstate Medical University. Her research focuses on family violence, the mental health consequences of trauma, and community-based intervention and prevention programs for survivors of violence.

Joseph R. Cohen, PhD, is an assistant professor of Psychology at the University of Illinois at Urbana-Champaign. His research focuses on the development of psychological distress and behavioral problems in children and adolescents, and how to translate these findings into improved mental health services for vulnerable youth.

Gregory L. Stuart, PhD, is a professor of Psychology at the University of Tennessee-Knoxville and the Director of Family Violence Research at Butler Hospital. He is also an adjunct Professor in the Department of Psychiatry and Human Behavior at the Warren Alpert Medical School of Brown University. His research focuses primarily on the comorbidity of intimate partner violence and substance misuse. He is particularly interested in interventions that address both substance use and relationship aggression.

Jeff R. Temple, PhD, is a professor and director of Behavioral Health and Research in the Department of Obstetrics and Gynecology at the University of Texas Medical Branch. His research focuses on interpersonal relationships, with a particular emphasis on understanding factors related to the onset, course, consequences, prevention, and intervention of intimate partner and teen dating violence.

References

Agnew

(1992). Foundation for a general strain theory of crime and delinquency. Criminology, 30, 47-88.

Baker

C. R.

Stith

S. M.

(2008). Factors predicting dating violence perpetration among male and female college students. Journal of Aggression, Maltreatment & Trauma, 17, 227-244.

Bell

K. M.

Orcutt

H. K.

(2009). Posttraumatic stress disorder and male-perpetrated intimate partner violence. Journal of the American Medical Association, 302, 562-564.

Bliese

P. D.

Wright

K. M.

Adler

A. B.

Cabrera

Castro

C. A.

Hoge

C. W.

(2008). Validating the primary care posttraumatic stress disorder screen and the posttraumatic stress disorder checklist with soldiers returning from combat. Journal of Consulting and Clinical Psychology, 76, 272-281.

Brock

R. L.

Barry

R. A.

Lawrence

Rolffs

Cerretani

Zarling

(2015). Online administration of questionnaires assessing psychological, physical, and sexual aggression: Establishing psychometric equivalence. Psychology of Violence, 5, 294-304.

Capaldi

D. M.

Shortt

J. W.

Crosby

(2003). Physical and psychological aggression in at-risk young couples: Stability and change in young adulthood. Merrill-Palmer Quarterly, 49, 1-27.

Chemtob

C. M.

Novaco

R. W.

Hamada

R. S.

Gross

D. M.

Smith

(1997). Anger regulation deficits in combat-related posttraumatic stress disorder. Journal of Traumatic Stress, 10, 17-36.

Cohen

J. R.

Shorey

R. C.

Menon

S. V.

Temple

J. R.

(2018). Predicting teen dating violence perpetration. Pediatrics, 141(4), e20172790.

Crick

N. R.

Dodge

K. A.

(1994). A review and reformulation of social information-processing mechanisms in children’s social adjustment. Psychological Bulletin, 115, 74-101.

10.

Dodge

K. A.

(1986). A social information processing model of social competence in children. In Perlmutter

(Ed.), Minnesota symposium on child psychology (Vol. 18, pp. 77-125). Hillsdale, NJ: Lawrence Erlbaum.

11.

Ehring

Quack

(2010). Emotion regulation difficulties in trauma survivors: The role of trauma type and PTSD symptom severity. Behavior Therapy, 41, 587-598.

12.

Exner-Cortens

Eckenrode

Rothman

(2013). Longitudinal associations between teen dating violence victimization and adverse health outcomes. Pediatrics, 131, 71-78.

13.

Foshee

V. A.

(1996). Gender differences in adolescent dating abuse prevalence, types and injuries. Health Education Research, 11, 275-286.

14.

Hecker

Hermenau

Maedl

Hinkel

Schauer

Elbert

(2013). Does perpetrating violence damage mental health? Differences between forcibly recruited and voluntary combatants in DR Congo. Journal of Traumatic Stress, 26, 142-148.

15.

Holt

M. K.

Espelage

D. L.

(2005). Social support as a moderator between dating violence victimization and depression/anxiety among African American and Caucasian adolescents. School Psychology Review, 34, 309-328.

16.

Holtzworth-Munroe

(2000). Social information processing skills deficits in maritally violent men: Summary of a research program. In Vincent

J. P.

Jouriles

E. N.

(Eds.), Domestic violence: Guidelines for research-informed practice (pp. 13-36). London, England: Jessica Kingsley.

17.

Holtzworth-Munroe

Hutchinson

(1993). Attributing negative intent to wife behavior: The attributions of maritally violent versus nonviolent men. Journal of Abnormal Psychology, 102, 206-211.

18.

Holtzworth-Munroe

Smutzler

(1996). Comparing the emotional reactions and behavioral intentions of violent and nonviolent husbands to aggressive, distressed, and other wife behaviors. Violence and Victims, 11, 319-339.

19.

L. T.

Bentler

P. M.

(1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling, 6, 1-55.

20.

Jakupcak

Conybeare

Phelps

Hunt

Holmes

H. A.

. (2007). Anger, hostility, and aggression among Iraq and Afghanistan war veterans reporting PTSD and subthreshold PTSD. Journal of Traumatic Stress, 20, 945-954.

21.

Kline

R. B.

(2011). Principles and practice of structural equation modeling. New York, NY: Guilford Press.

22.

Lawrence

Yoon

Langer

(2009). Is psychological aggression as detrimental as physical aggression? The independent effect of psychological aggression on depression and anxiety symptoms. Violence and Victims, 24, 20-35.

23.

Leskela

Dieperink

Thuras

(2002). Shame and posttraumatic stress disorder. Journal of Traumatic Stress, 15, 223-226.

24.

MacKinnon

D. P.

Lockwood

C. M.

Williams

(2004). Confidence limits for the indirect effect: Distribution of the product and resampling methods. Multivariate Behavioral Research, 39, 99-128.

25.

Mason

M. J.

Mennis

Schmidt

C. D.

(2011). A social operational model of urban adolescents’ tobacco and substance use: A mediational analysis. Journal of Adolescence, 34, 1055-1063.

26.

McFall

R. M.

(1982). A review and reformulation of the concept of social skills. Behavioral Assessment, 4, 1-33.

27.

Murphy

C. M.

(2013). Social information processing and the perpetration of intimate partner violence: It is (and isn’t) what you think. Psychology of Violence, 3, 212-217.

28.

Murphy

C. M.

O’Leary

K. D.

(1989). Psychological aggression predicts physical aggression in early marriage. Journal of Consulting and Clinical Psychology, 57, 579-582.

29.

Muthén

L. K.

Muthén

(2006). The comprehensive modeling program for applied researchers. Los Angeles, CA: Author.

30.

O’Leary

K. D.

Smith Slep

A. M.

(2003). A dyadic longitudinal model of adolescent dating aggression. Journal of Clinical Child and Adolescent Psychology, 32, 314-327.

31.

Ortiz

Shorey

R. C.

Cornelius

T. L.

(2015). An examination of emotion regulation and alcohol use as risk factors for female-perpetrated dating violence. Violence and Victims, 30, 417-431.

32.

Prins

Ouimette

P. C.

Kimerling

Cameron

R. P.

Hugelshofer

D. S.

Shaw-Hegwer

. . . Sheikh

J. I.

(2003). The primary care PTSD screen (PC-PTSD): Development and operating characteristics. Primary Care Psychiatry, 9, 9-14.

33.

Rosenbaum

Leisring

P. A.

(2003). Beyond power and control: Towards an understanding of partner abusive men. Journal of Comparative Family Studies, 34, 7-22.

34.

Shorey

R. C.

Cornelius

T. L.

Bell

K. M.

(2008). A critical review of theoretical frameworks for dating violence: Comparing the dating and marital fields. Aggression and Violent Behavior, 13, 185-194.

35.

Shorey

R. C.

Febres

Brasfield

Stuart

G. L.

(2012). The prevalence of mental health problems in men arrested for domestic violence. Journal of Family Violence, 27, 741-748.

36.

Shorey

R. C.

Fite

P. J.

Torres

E. D.

Stuart

G. L.

Temple

J. R.

(2019). Bidirectional associations between acceptability of violence and intimate partner violence from adolescence to young adulthood. Psychology of Violence, 9(1), 108.

37.

Sippel

L. M.

Marshall

A. D.

(2011). Posttraumatic stress disorder symptoms, intimate partner violence perpetration, and the mediating role of shame processing bias. Journal of Anxiety Disorders, 25, 903-910.

38.

Taft

C. T.

Creech

S. K.

Murphy

C. M.

(2017). Anger and aggression in PTSD. Current Opinion in Psychology, 14, 67-71.

39.

Taft

C. T.

Kaloupek

D. G.

Schumm

J. A.

Marshall

A. D.

Panuzio

King

D. W.

Keane

T. M.

(2007). Posttraumatic stress disorder symptoms, physiological reactivity, alcohol problems, and aggression among military veterans. Journal of Abnormal Psychology, 116, 498-507.

40.

Taft

C. T.

Macdonald

Creech

S. K.

Monson

C. M.

Murphy

C. M.

(2016). A randomized controlled clinical trial of the strength at home men’s program for partner violence in military veterans. Journal of Clinical Psychiatry, 77, 1168-1175.

41.

Taft

C. T.

Vogt

D. S.

Marshall

A. D.

Panuzio

Niles

B. L.

(2007). Aggression among combat veterans: Relationships with combat exposure and symptoms of posttraumatic stress disorder, dysphoria, and anxiety. Journal of Traumatic Stress, 20, 135-145.

42.

Taft

C. T.

Watkins

L. E.

Stafford

Street

A. E.

Monson

C. M.

(2011). Posttraumatic stress disorder and intimate relationship problems: A meta-analysis. Journal of Consulting and Clinical Psychology, 79, 22-33.

43.

Temple

J. R.

Freeman

D. H.

(2011). Dating violence and substance use among ethnically diverse adolescents. Journal of Interpersonal Violence, 26, 701-718.

44.

Temple

J. R.

Shorey

R. C.

Fite

Stuart

G. L.

V. D.

(2013). Substance use as a longitudinal predictor of the perpetration of teen dating violence. Journal of youth and adolescence, 42, 596-606.

45.

Thurston

I. B.

Howell

K. H.

(2018). To screen or not to screen: Overreliance on risk without protective factors in violence research. Pediatrics, 141(4), e20180075.

46.

Vagi

K. J.

Rothman

E. F.

Latzman

N. E.

Tharp

A. T.

Hall

D. M.

Breiding

M. J.

(2013). Beyond correlates: A review of risk and protective factors for adolescent dating violence perpetration. Journal of Youth and Adolescence, 42, 633-649.

47.

Wekerle

Leung

Wall

A. M.

MacMillan

Boyle

Trocme

Waechter

(2009). The contribution of childhood emotional abuse to teen dating violence among child protective services-involved youth. Child Abuse & Neglect, 33, 45-58.

48.

Wekerle

Wolfe

D. A.

Hawkins

D. L.

Pittman

A. L.

Glickman

Lovald

B. E.

(2001). Childhood maltreatment, posttraumatic stress symptomatology, and adolescent dating violence: Considering the value of adolescent perceptions of abuse and a trauma mediational model. Developmental Psychopathology, 13, 847-871.

49.

Wolfe

D. A.

Scott

Reitzel-Jaffe

Wekerle

Grasley

Straatman

A. L.

(2001). Development and validation of the Conflict in Adolescent Dating Relationships Inventory. Psychological Assessment, 13, 277-293.

50.

Wolfe

D. A.

Wekerle

Scott

Straatman

A. L.

Grasley

(2004). Predicting abuse in adolescent dating relationships over 1 year: The role of child maltreatment and trauma. Journal of Abnormal Psychology, 113, 406-415.