How Robust are Subgroup Analyses in Political Science? Insights from Dozens of Replications

• In Event: Experimental Methods

Brief Overview

We reanalyze dozens of experimental subgroup analyses from the top political science journals using a sample-splitting ML method. We find numerous examples of spurious results leading us to recommend the use of ML methods for experimental analysis.

Authors

• Sam Fuller, University of California-Davis

• Jack Rametta, University of California-Davis

• Giulia Venturini, Florida State University