Experimental Estimation of Heterogeneous Treatment Effects Related to Self-Selection

Authors


  • Brian J. Gaines (bjgaines@illinois.edu) is an Associate Professor and James H. Kuklinski (kuklinsk@illinois.edu) is the Matthew T. McClure Professor at the University of Illinois, where they both hold appointments in the Department of Political Science and the Institute of Government and Public Affairs, 1007 W. Nevada St., Urbana, IL 61801.

  • We are very grateful to numerous audiences and individuals for helpful advice, including but not limited to Jake Bowers, Jason Coronel, Tiberiu Dragu, Jamie Druckman, Mark Frederickson, Matthew Hayes, Jude Hays, Rebecca Morton, Tom Rudolph, Jas Sekhon, Lynn Vavreck, the editor, and four anonymous referees. We are especially indebted to Don Green for many helpful suggestions. The data analyzed in Table 2 can be accessed as a Stata file at https://netfiles.uiuc.edu/bjgaines/replication%20data/.

Abstract

Social scientists widely regard the random-assignment experiment as the gold standard for making causal inferences about the world. We argue that it can be improved. For situations in which self-selection and heterogeneity of treatment effects exist, an alternative experimental design that retains random assignment to treatment or control and supplements it with some self-selection of condition offers a clear advantage. It reveals the average treatment effect while also allowing estimation of the distinct effects of the treatment on those apt and inapt to experience the treatment outside the experimental context.

Ancillary