Social support is a robust correlate of posttraumatic stress disorder (PTSD) symptoms and of general psychological distress (Ozer, Best, Lipsey, & Weiss, 2003). The nature of the causal relationship between support and PTSD remains the subject of debate, with 2 models, social erosion and social causation, often used to explain findings. Despite extensive research using these models, no studies of which we are aware have included tests of both models within the same series of analyses, across more than 2 time points, in veterans. These competing models were tested in a sample of National Guard soldiers (N = 521) who completed measures of perceived social support and the PTSD Checklist-Military version (Weathers, Litz, Herman, Huska, & Keane, 1993) at 3 months, 15 months, and 27 months following a combat deployment to Iraq. Analyses were run separately for overall PTSD symptoms and the PTSD components of intrusion, trauma-avoidance, dysphoria, and hyperarousal. Both the social erosion (βs ranging from -.10 to -.19) and social causation (βs ranging from -.08 to -.13) hypotheses were supported. Results suggested PTSD-specific symptom dimensions may both erode and be influenced by social support, whereas general psychological distress erodes social support. Implications for clinical intervention and research are discussed.