The Underprovision of Experiments in Political Science

Abstract

Field experimentation enables researchers to draw unbiased and externally valid causal inferences about social processes. Despite these strengths, field experimentation is seldom used in political science, which relies instead on observational studies and laboratory experiments. This article contends that political scientists underestimate the value of field experimentation and overestimate their ability to draw secure causal inferences from other types of data. After reviewing the history of experimentation in the discipline, the authors discuss the advantages and disadvantages of field experimental methods. They conclude by describing a number of research topics that seem amenable to experimental inquiry.