Background and objective: This study tested three alternative explanations for research indicating a positive, but heterogeneous relationship between self-reported posttraumatic growth (PTG) and posttraumatic stress symptoms (PSS): (a) the third-variable hypothesis that the relationship between PTG and PSS is a spurious one driven by positive relationships with resource loss, (b) the growth over time hypothesis that the relationship between PTG and PSS is initially a positive one, but becomes negative over time, and (c) the moderator hypothesis that resource loss moderates the relationship between PTG and PSS such that PTG is associated with lower levels of PSS as loss increases. Design and method: A nationally representative sample (N=1622) of Israelis was assessed at three time points during a period of ongoing violence. PTG, resource loss, and the interaction between PTG and loss were examined as lagged predictors of PSS to test the proposed hypotheses. Results: Results were inconsistent with all three hypotheses, showing that PTG positively predicted subsequent PSS when accounting for main and interactive effects of loss. Conclusions: Our results suggest that self-reported PTG is a meaningful but counterintuitive predictor of poorer mental health following trauma.