Research That Matters (January 17 - 20, 2008)

Background and Purpose: Considerable research, conducted over the past 20 years, has linked exposure to violence to Post-traumatic Stress Disorder (PTSD) in children, but almost no research has investigated PTSD as a mediating variable between exposure to violence and other problems. Some research has suggested PTSD may result in other psychological problems, such as depression (Kliewer, et al., 1998; Mazza and Reynolds, 1999), and behavioral and academic problems at school (Overstreet, 2000). This research was conducted to determine if PTSD acts as a mediating variable between exposure to violence and depression, and behavioral problems at school. Methods: Two hundred forty-seven children ages eight to thirteen from the surrounding townships of Cape Town, South Africa, participated in the project. Three of the townships are predominantly Coloured and Afrikaans and English speaking, while two are almost totally African Black and Xhosa speaking. Interviews with the children were conducted with bilingual graduate students, although the instruments were translated into both Afrikaans and Xhosa. Exposure to community violence was measured with a revised version of the Nadel, Spellman, and Alvarez-Cannio Victimization Scale (1996), expanded to included police, gang, and individual violence. Depression was measured with ten items from the Richters Martinez, and Valla (1990) instrument, “Levonn: A Cartoon-based Interview for Assessing Children's Distress Symptoms”; PTSD was measured with 13 items from the same scale. Behavioral problems at school were measured with one item, which asked how often the child had been in trouble at school. Correlation and regression analysis (described in Holmbeck, 1997) were conducted to assess PTSD as a mediating variable for the relationship between exposure to violence and depression, and behavioral problems at school. Results: Analysis of both depression and behavioral problems at school met the first three criteria for PTSD as a mediating variable for exposure to violence and the dependent variables. Exposure to violence and PTSD, exposure to violence and depression, exposure to violence and behavioral problems at school, PTSD and depression, and PTSD and behavioral problems at school were all moderately correlated and statistically significant (p < .01). The final test for PTSD as a mediating variable was conducted for both depression and school behavioral problems, controlling for age, race, and gender. As expected, when entered into the regression equation, each form of exposure to violence had a significant effect on depression. However, when PTSD was entered, all standardized Beta coefficients for exposure to violence dropped considerably, but remained statistically significant, indicating both direct and indirect effects of exposure to violence. However, when PTSD was entered into the equation for school behavioral problems, the Beta coefficients for exposure to violence remained approximately the same. Conclusions and Implications: Findings indicated that PTSD partially mediated the effects of exposure to violence on depression, but not behavioral problems at school. This has important implications for interventions: Early intervention in the treatment of PTSD may prevent later development of depression. The findings also suggest exposure to violence may lead directly to externalizing problems, such as behavioral problems at school. Additional implications are discussed.