Can We Be Casual about Being Causal?

The demonstration of causal relationships among variables has been central to the social sciences for their entire existence, but there has been an upsurge of concern about causal inference over at least the past decade This increased interest in causation has coincided with the increased use of experimental methods, in all the social sciences, and especially in political science and economics. These two trends are wedded because, at least in part, the advocates of experimental methods argue that they are the best, if not the only way, to demonstrate causation. The argument of this paper is that although experimentation is an important weapon in the armamentarium of scholars, it should (like any technique) be considered with some skepticism. Skepticism about any method is always warranted, but this study is particularly concerned with the difficulties in demonstrating causation through experiments – or any other method – in comparative policy analysis. Hence, the title of this paper asks whether we have been too casual in thinking that our research problems are solved by selecting this technique, although the same can be said for advocates of studying causation through regression-based methods. In short, we need to be very cautious, rather than casual, when making claims about causation.