Abstract
School shootings and other tragedies involving gun-violence have become commonplace in the United States. While much research has been conducted into the direct effects on affected individuals, less is known about community-level effects. We use data on grocery-store sales at the county-level to test the hypothesis that the exposure to a traumatic event induces short-run changes in purchases of alcohol and tobacco, well-known and highly accessible substances often linked to coping with traumatic events. Leveraging a specification curve design, we find that although some specifications yield statistically significant effects, there is no robust support for the main hypothesis. We close with a discussion of potential explanations and draw some methodological conclusions for future research