Analogical reasoning and evidence transfer in evidence-based policy

Abstract Evidence-based policy often involves a kind of reasoning by analogy, which starts from the established success of an intervention on a given occasion (= source domain S). Policy makers infer that the same intervention is likely to be successful on another occasion (= target domain T). The practical goal of this paper is to develop, defend and illustrate a model detailing the reasoning steps that policy makers should go through in order to perform this type of analogical reasoning properly. Our model also offers tools for diagnosing flaws in the reasoning process for past cases where this kind of analogical reasoning went wrong. We illustrate the practical utility and diagnostic power of the model by means of two well-known failed evidence transfers. The theoretical goal of this paper is to investigate the ramifications of our model for general theorising on analogical reasoning.