Why Comparative Social Policy Research Needs Better Theory and Measurement

ABSTRACT A central question in comparative social policy research is what the effect of a policy (change) is on outcomes such as employment, poverty or inequality. This is the part of the field with the most direct bearing on policy, yet much of its evidence still takes the form of macro–macro associations that lack theoretical clarity. In this article, we argue that better insight into mechanisms requires better theorising, and that better theorising also makes research more valid. Drawing on Coleman's macro–micro–macro model (‘Coleman's boat’), we set out what a complete mechanism‐based account of a policy effect involves: how a policy alters individuals' beliefs, opportunities and constraints, how individuals act in response, and how their actions aggregate into a macro‐level outcome. We argue that this mechanistic thinking matters for three forms of validity. It disciplines the choice of research design and estimand, and provides the explanation that an unbiased estimate alone does not (internal validity); it identifies the conditions under which findings generalise (external validity) and it tells us which policies to measure and how, and at which point in the causal chain (construct validity). Applying the model to Korpi and Palme's ‘Paradox of Redistribution’, we show that comparative research on policy effects has paid sustained theoretical attention to the first step, but rarely sets out the full causal chain and typically leaves aggregation implicit. Theoretically informed, mechanism‐based approaches would enhance the credibility, cumulative progress and policy relevance of comparative social policy research over the next few decades.