| Abstract
Background: Incarcerated women suffer high rates of trauma exposure, suicidal ideation,
and suicidal behavior. Few studies have addressed the role of interpersonal problems and
social support within the context of suicidality among incarcerated women. The present
study evaluated social support and interpersonal problems as mediators of the association
between trauma exposure and suicide risk. In addition, American Indian and Latina
ethnicities were considered as moderators of the indirect effect of social support upon
suicide risk.
Method: Participants were 224 treatment-seeking incarcerated women, aged 19 to 60
years (M = 34.17, SD =9.35). Women completed self-report measures of trauma history,
interpersonal problems, social support, and suicide risk. The mediating influence of
interpersonal problems and social support was evaluated using a product of coefficients
approach via a series of regressions, including the conditional effects of Latina and
American Indian ethnicity.
Findings: Women reported significant histories of trauma, suicidal behavior, and
interpersonal problems relative to women in the general population. There was a
significant direct effect of total number of traumas on total suicide risk (=.312, sr2=.097,
p<.001). However, neither interpersonal problems nor social support mediated this
association. Similarly, there was no significant moderating effect of Latina or American
Indian ethnicities. Physical and Sexual victimization (=.361, sr2=.097, p<.001) emerged
as a unique predictor of suicide risk relative to other forms of trauma (i.e., General
Disaster, Crime-Related Trauma), which did not significantly predict suicide risk.
Conclusion: These findings replicated previous work concerning the robust association
between trauma exposure and risk of suicide. These findings also raise important issues
for future research in this area, given the non-significant influence of interpersonal
problems and social support within the present model. Future research should better
address the temporal sequence of these predictors via prospective and longitudinal
designs. |