Causal Inference Without Control Units

• Activity Number: 691
• Type: Contributed
• Date/Time: Thursday, August 13, 2015 : 10:30 AM to 12:20 PM
• Sponsor: Social Statistics Section
• Author(s): Konstantin Kashin* and Adam Glynn
• Companies: Harvard University and Emory University
• Keywords: causal inference ; front-door ; post-treatment ; research design ; program evaluation ; political science

Abstract

A fundamental tenet of research design is the necessity of control units: units of analysis that have not been assigned the treatment / program of interest. However, withholding treatment from some units can introduce a number of problems: ethical concerns, increased cost, disruptions to business practice, and political fallout. In this paper, we explore the possibility of making causal inferences without control units by using front-door and front-door difference-in-differences estimators. Using data from 31 get-out-the-vote (GOTV) experiments and 30 outcomes from the Oregon Health Insurance Experiment, we demonstrate circumstances where one can reliably bound experimental benchmarks using front-door techniques on only treated units. We further show that we can bracket experimental benchmarks using a combination of front-door and front-door difference-in-differences estimators on the treated units.