A burgeoning empirical literature on the effects of conflict on various economic behavioral parameters exhibits mixed results, with respect to both the magnitude and the direction of the effects. By estimating the distribution of estimated effects of violence on risk preferences, rather than just the average effect, we reconcile the discordant results of the prior literature. The distribution also reveals substantial and previously overlooked variation in the effects of exposure. This raises questions about the widespread use of aggregated measures of exposure to violence in conflict literature. We use panel data from northern Uganda, the latest round collected seven years after the violence ended, to explore the heterogeneous effects of different experiences of violence– personal suffering, perpetration, witness, or indirect experience through family members’suffering – on different measures of ambiguity and risk aversion, correcting for many of the methodological shortcomings of previous studies. We find that violence has an adverse near-term effect on mental health, but with heterogeneous effects on risk aversion depending on the natureof one’s experience of violence. We also find that the risk preference effects persist even after arecovery in mental health.