Replication Data for: Do We Really Know the WTO Cures Cancer?

We use a replication experiment of 94 specifications from 16 different studies to show the severity of the problem of selection on unobservables. Using a variety of approaches, we show that membership in the GATT/WTO has a significant effect on a surprisingly high number of dependent variables (34%), variables which have little to no theoretical relationship to the WTO. To make our exercise even more conservative, we demonstrate that membership in a lowimpact environmental treaty, CITES, yields similarly high false positive rates. We advocate for theoretically informed sensitivity analysis, showing how prior theoretical knowledge conditions the crucial choice of covariates for sensitivity tests. While we focus on international institutions, the arguments apply to other subfields and applications.