(PS4-B44) Longitudinal Predictors of Intimate Partner Violence Following Deployment: A Comparison Between Soldier-only versus Bidirectional Use of Violence in Romantic Dyads

Intimate partner violence (IPV) includes the use of physical, sexual, or psychological violence in the context of a current or former romantic relationship. Relative to civilians, Veterans, and particularly those with PTSD, are more likely to both experience and use IPV. Although IPV is typically characterized as being enacted by one partner in a romantic relationship (i.e., unidirectional IPV), a growing body of literature demonstrates that bidirectional IPV (i.e., both partners’ use of violence) is the most common pattern of violence within romantic relationships. A dearth of research has examined factors that may differentiate uni- versus bidirectional IPV following a stressor. The present study sought to longitudinally explore soldiers’ self-reported mental health variables and each partner’s prior use of IPV as predictors of soldier-only and bidirectional IPV following deployment (N = 490 dyads). Data from the Readiness and Resilience in National Guard Soldiers (RINGS-2) project were examined for each partner collected 2-5 months prior to deployment and within one year following soldiers’ return.

Hierarchical logistic regression models were used to evaluate predictors of physical, sexual, and psychological IPV. The first set of regression models predicted the presence of soldier-only IPV and the second set of regression models predicted the presence of bidirectional IPV. Across regression models, the same predisposing variables were examined in the first block of the equation (soldier age and prior OEF/OIF deployment status). The second block included pre-deployment use of violence (i.e., both soldier and partner), and soldier self-reported symptoms of PTSD (PTSS). The third block included indicators of soldier post-deployment mental health variables (i.e., PTSS and alcohol use).

After adjusting for multiple comparisons, the full multivariate models predicting post-deployment soldier-only IPV were statistically significant for physical IPV only (p < .001). Partner use of pre-deployment physical IPV was associated with soldier-only physical IPV (p < .001). The full multivariate models predicting post-deployment bidirectional IPV were significant for physical, sexual and psychological IPV (all p’s < .001). Soldier post-deployment PTSS positively predicted bidirectional physical IPV (p < .001). Previous OEF/OIF deployment (p = .003), partner use of pre-deployment psychological IPV (p < .001), and soldier post-deployment PTSS (p = .006) predicted bidirectional psychological IPV.  No significant predictors emerged for sexual IPV.

A number of pre-deployment and post-deployment predictors differed between form of IPV used and whether the violence was uni- or bidirectional. These nuanced findings underscore the importance of longitudinal examination of variables predicting uni- versus bidirectional forms of IPV in the aftermath of a stressor. Full results and implications will be discussed, including limitations in the assessment of bidirectional IPV using available self-report measures.