Protective Buffering by Service Members and Romantic Partners During Military Deployments: Associations with Mental and Relationship Health

Sarah Carter

Advisor: Keith Renshaw

Committee Members: Todd Kashdan, Tim Curby

David J. King Hall, #2013
June 12, 2017, 10:00 AM to 07:00 AM


Since the beginning of military operations in Iraq and Afghanistan, well over 2 million military service members (SMs) have deployed to combat zones. For SMs and romantic partners, deployments represent one of the most stressful situations in their lives, with one of the largest concerns being the wellbeing of the other partner. In order to shield each other from stressors occurring at home or in combat, both deployed SMs and at-home partners report engaging in protective buffering, or intentionally withholding information or concerns in an attempt to protect the other partner. Despite retrospective, qualitative studies that document the presence of protective buffering, no quantitative research has yet assessed whether protective buffering is effective or what other effects it may have on the couple.

            This dissertation addresses this gap in two separate but related manuscripts. Both manuscripts utilize data from 54 military couples drawn from a larger study of Army couples. These couples were married and experiencing a deployment of the male soldier during either the 5- or 6-year follow-up assessment within the larger study. For these couples, pre- and post-deployment data were derived from the time point that immediately preceded and followed the soldier’s deployment time point.

The first manuscript explores the frequency of protective buffering by SMs during deployment, as well as pre-existing and logistical deployment factors that may be related to levels of protective buffering. More than half (56%) of SMs indicated that they engaged in protective buffering at least some of the time, with an average item response of 4.51 (SD = 1.29) on a 1-7 scale. Of over 10 pre-deployment and deployment variables that were explored, only partners’ psychological distress prior to deployment was significantly associated (positively) with SMs’ protective buffering. I subsequently sought to assess the relationship of SM protective buffering with distress and marital satisfaction in SMs and partners across the deployment cycle. First, I used repeated measures ANOVA to assess levels of marital satisfaction and distress across all three time points. Surprisingly, SMs reported significantly higher marital satisfaction during deployment, relative to both pre- and post-deployment. They also had significantly lower distress post-deployment, as compared to pre-deployment and during deployment. On the other hand, there were no significant differences in partners’ distress or marital satisfaction across the deployment cycle. To assess how levels and trajectories of distress and marital satisfaction varied based on SM protective buffering, a multigroup, cross-lagged, autoregressive model with individual distress and marital satisfaction for partners was modeled across time. A dichotomous high or low protective buffering variable was created using a median split. Each intercept and slope was then systematically constrained across the high and low protective buffering groups to identify the intercepts and slopes that contributed to significantly worse fit when constrained. Factors that significantly worsened the model fit when constrained were allowed to vary in the final model. Relative to those with higher protective buffering, partners of SMs with lower buffering had lower distress during deployment and a more stable slope between deployment and post-deployment distress, indicating that the lower distress at deployment was also likely continued into post-deployment. Contrary to hypotheses, there were no significant differences in levels or slopes of SM distress or marital satisfaction when comparing those with higher and lower protective buffering. Overall, the results indicate that SMs’ protective buffering was not associated with any negative consequences for themselves, but non-deployed partners reported higher levels of distress during deployment when SMs engaged in more buffering. Due to the cross-sectional nature of the deployment factors, it is unknown if more protective buffering might be driving higher partner distress, or if SMs were reacting to higher distress in their partners with more buffering.

            The second manuscript investigates protective buffering by non-deployed partners. Partners have consistently described feeling a responsibility not to burden or distract deployed SMs, fearing that family-related stress or conflict might result in a deployed SM being distracted during dangerous situations. Thus, I first assessed the relationship between level of protective buffering reported by the partner and the degree to which SMs reported that family stresses interfered with work functioning (i.e., family-to-work spillover) during deployment. Contrary to hypothesis, the correlation between partner protective buffering and SMs' report of spillover was small and not significant. Subsequently, I explored the association of partners’ protective buffering with both partners’ and SMs’ psychological distress and marital satisfaction across all three time points, via both hierarchical regressions (for the deployment time period only) and multigroup, cross-lagged autoregressive models. In both types of analyses, higher protective buffering by partners was related to higher partner distress and lower SM marital satisfaction during deployment. Contrary to hypotheses, these effects did not continue post-deployment. In sum, protective buffering was not associated with lower spillover in SMs, but was associated with higher distress in partners. Notably, the significant relationships of protective buffering with partner distress and SM marital satisfaction were cross-sectional. Thus, protective buffering may be causing the negative outcomes, or partners who are distressed or sense lower satisfaction in their deployed partner may be more likely to buffer.