Causal Mechanisms: Yes, But…

Abstract

In recent years, the importance of mechanism-centered explanation has become an article of faith within the social sciences, uniting researchers from a wide variety of methodological traditions—quantitative and qualitative, experimental and nonexperimental, nomothetic and idiographic, formal models and narrative prose. Despite its many virtues, there are reasons to be skeptical of social science’s newfound infatuation with causal mechanisms. First, the concept of a mechanism-centered (“mechanismic”) explanation is fundamentally ambiguous, meaning different things to different people. Second, the minimal objectives associated with the turn to mechanisms—to specify causal mechanisms and engage in detailed causal reasoning—are not at variance with traditional practices in the social sciences and thus hardly qualify as a distinct approach to causal assessment. Finally, the more demanding goal of rigorously testing causal mechanisms in causal arguments is admirable but often unrealistic. To clarify, this is not a polemic against mechanisms. It is a polemic against a dogmatic interpretation of the mechanismic mission. Causal mechanisms are rightly regarded as an important, but secondary, element of causal assessment—by no means a necessary condition.