The experimental study of politics has grown explosively in the past two decades. Part of that explosion takes the form of a dramatic increase in the number of published articles that use experiments. Perhaps less evident, and arguably more important, experimentalists are exploring topics that would have been unimaginable only a few years ago. Laboratory researchers have studied topics ranging from the effects of media exposure (Iyengar and Kinder 1987) to the conditions under which groups solve collective action problems (Ostrom, Walker, and Gardner 1992), and, at times, have identified empirical anomalies that produced new theoretical insights (McKelvey and Palfrey 1992). Some survey experimenters have developed experimental techniques to measure prejudice (Kuklinski, Cobb, and Gilens 1997) and its effects on support for policies such as welfare or affirmative action (Sniderman and Piazza 1995); others have explored the ways in which framing, information, and decision cues influence voters' policy preferences and support for public officials (Druckman 2004; Tomz 2007). And although the initial wave of field experiments focused on the effects of campaign communications on turnout and voters' preferences (Eldersveld 1956; Gerber and Green 2000; Wantchekon 2003), researchers increasingly use field experiments and natural experiments to study phenomena as varied as election fraud (Hyde 2009), representation (Butler and Nickerson 2009), counterinsurgency (Lyall 2009), and interpersonal communication (Nickerson 2008).

Within the prevailing Fisher-Neyman-Rubin framework of causal inference, causal effects are defined as comparisons of potential outcomes under different treatments. In most contexts, it is impossible or impractical to observe multiple outcomes (realizations of the variable of interest) for any given unit. Given this fundamental problem of causality (Holland 1986), experimentalists approximate the hypothetical treatment effect by comparing averages of groups or, sometimes, averages of differences of matched cases. Hence, they often use (Ȳ|t = 1) − (Ȳ|t = 0) to estimate E[(Yi|t = 1) − (Yi|t = 0)], labeling the former quantity the treatment effect or, more accurately, the average treatment effect.

The rationale for substituting group averages originates in the logic of the random assignment experiment: each unit has different potential outcomes; units are randomly assigned to one treatment or another; and, in expectation, control and treatment groups should be identically distributed. To make causal inferences in this manner requires that one unit's outcomes not be affected by another unit's treatment assignment. This requirement has come to be known as the stable unit treatment value assumption.

Until recently, experimenters have reported average treatment effects as a matter of routine. Unfortunately, this difference of averages often masks as much as it reveals. Most crucially, it ignores heterogeneity in treatment effects, whereby the treatment affects (or would affect if it were actually experienced) some units differently from others.

In his 1909 American Political Science Association presidential address, A. Lawrence Lowell (1910) advised the fledgling discipline against following the model of the natural sciences: “We are limited by the impossibility of experiment. Politics is an observational, not an experimental science…” (7). The lopsided ratio of observational to experimental studies in political science, over the one hundred years since Lowell's statement, arguably affirms his assessment. The next hundred years are likely to be different. The number and influence of experimental studies are growing rapidly as political scientists discover ways of using experimental techniques to illuminate political phenomena.

The growing interest in experimentation reflects the increasing value that the discipline places on causal inference and empirically guided theoretical refinement. Experiments facilitate causal inference through the transparency and content of their procedures, most notably the random assignment of observations (a.k.a. subjects or experimental participants) to treatment and control groups. Experiments also guide theoretical development by providing a means for pinpointing the effects of institutional rules, preference configurations, and other contextual factors that might be difficult to assess using other forms of inference. Most of all, experiments guide theory by providing stubborn facts – that is, reliable information about cause and effect that inspires and constrains theory.