Replication Data for: Navigating Potential Pitfalls in Difference-in-Differences Designs: Reconciling Conflicting Findings on Mass Shootings' Effect on Electoral Outcomes

Recent work on the electoral effects of gun violence in the United States relying on the difference-in-differences design has produced conflicting findings that range from null effects to substantively large effects. At the same time, as the difference-in-difference design on which this research has relied has exploded in popularity, scholars have documented several methodological issues this design faces---including potential violations of parallel-trends and unaccounted for treatment effect heterogeneity. Sadly, these pitfalls (and their solutions) have not been fully explored in political science. In this paper, we apply these advancements to the unresolved debate on the effects of gun violence on electoral outcomes in the United States. We show that studies that find a large positive effect of gun violence on Democratic vote share are a product of a failure to properly specify difference-in-differences models when underlying assumptions are unlikely to hold. Once these biases are corrected, shootings show little evidence of sparking large electoral change. Our work clarifies an important unresolved debate and provides a cautionary guide for the many scholars currently employing difference-in-differences designs.

近期关于美国枪支暴力选举效应的研究依赖双重差分设计（difference-in-differences design），但得出的结果相互矛盾——从无显著效应到具有实质意义的显著效应不等。与此同时，随着此类研究依赖的双重差分设计愈发流行，学者们已记录下该设计面临的若干方法论问题——包括可能违反平行趋势假设以及未考虑处理效应异质性等。遗憾的是，这些缺陷（及其解决方案）在政治学领域尚未得到充分探讨。在本文中，我们将这些方法论进展应用于美国枪支暴力对选举结果影响这一悬而未决的争论。我们发现，那些认为枪支暴力对民主党得票率存在显著正向影响的研究，实则是在基本假设难以成立的情况下，未能恰当设定双重差分模型所致。一旦修正这些偏差，枪击事件几乎不会引发显著选举变化的证据便清晰呈现。我们的研究澄清了这一重要的悬而未决争论，并为当前使用双重差分设计的众多学者提供了一份警示性指南。