Mediation and the Estimation of Indirect Effects in Political Communication Research

Much research in the communication and political science fields—political communication research being no exception—seeks to establish the extent to which certain communication-related variables (e.g., exposure to political debates, the tone of news coverage about politicians, or the diversity in opinions of those with whom one talks about politics) have effects on various political outcomes (e.g., evaluation of candidates running for office, political cynicism, or participation in the political process). To be sure, studies to establish or refute the existence of effects are important. But studies, and the investigators who conduct them, are usually much more impressive and helpful in advancing our understanding when they go further by establishing not only whether or not an effect exists or how large that effect is, but also why that effect exists. What is the process at work that produces an association between political discussion and participation, or between exposure to political debates and political cynicism, or between the tone of news coverage and evaluation of candidates running for public office? For example, is it that frequent political talk exposes one to more diverse viewpoints, which in turn makes one feel confused or less confident in one’s beliefs, which in turn translates into a reduced likelihood of converting those beliefs into political action? Or might frequent talk increase one’s sense that one can make a difference in the outcomes of the political process, spawning greater participation in that process?