Background: Recent studies find a correlation between attentional threat avoidance under stress and posttraumatic stress symptoms. In this study, we assessed this association longitudinally in exposed and unexposed individuals. The degree of threat avoidance during exposure was expected to predict levels of posttraumatic stress symptoms 1 year later. Methods: Thirty-two participants were recruited and followed for 12 months, including 18 subjects exposed to rocket attacks and 14 nonexposed subjects. At 1-year follow-up, participants completed self-reports and an attention dot-probe task assessing threat-related bias. Results: State anxiety decreased at follow-up in exposed participants, though posttraumatic stress disorder (PTSD) and depression symptoms remained higher in exposed than in the nonexposed group. Attentional threat avoidance during imminent danger in the exposed group changed to threat attendance a year later, such that both the exposed and the nonexposed group exhibited similar threat bias patterns. As hypothesized, in the exposed group, stronger attentional threat avoidance during stress exposure predicted higher levels of PTSD symptoms 1 year later. Conclusions: Attention bias away from threat during acute stress may relate to risk for PTSD. This suggests that neurocognitive measures may index risk for PTSD. Depression and Anxiety, 2011. © 2011 Wiley-Liss, Inc.