Experimental Triangulation of Coalition Signals: Varying Designs, Converging Results

Abstract

It is probably fair to say that political science has not been a welcoming discipline for experimental research (McDermott, 2002). Our discipline has always expressed skepticism about the usefulness and the prospects of experimental designs to address the key research questions we care about. But the more political scientists have started to think carefully about causal relationships and what is required to test them, the more they came (or should come) to realize that our traditional methodologies and research designs are also not sufficient. The latter have serious limitations as well, and some of these limitations can be addressed by experimental methods. Because experimental designs have unique strengths compared to other research designs, it is not surprising that the use of experiments has evolved and increased over time (Morton and Williams, 2010). Put simply, experiments are flexible tools for theory testing that allow us to establish causality by clearly separating causes and effects.