While some studies of conflicts and disasters find these negative collective shocks make people more prosocial, others find they reduce cooperation. These conflicting findings may be a consequence of focusing on a single shock type, a single preference measure, or a single regional or temporal context experiencing shocks. We address these limitations by creating and analyzing a new global dataset of collective shocks and social preferences at the subnational level. We then explore the potential differences in how various shocks (armed conflicts, natural disasters) relate to various social preferences (altruism, reciprocity, trust). Our preliminary analysis shows that, while exposure to wartime violence does not systematically alter prosociality, exposure to natural disasters reduces some social preferences in the short term, reverting to baseline levels in the long term. By comparing local experiences globally, our project helps develop a nuanced view of how shocks influence preferences, with implications for cooperation and governance.