In Salmon's discussion of his account of statistical relevance and statistical explanation there is a peculiarity in the selection of examples. Where he wishes to show that statistical accounts are reasonably treated as explanatory, he draws examples from the social sciences, such as juvenile delinquency. But when he explains the concept of ‘causal’ relevance, the examples are selected from the natural sciences. This conceals difficulties with Salmon's account of causality in the face of multiple causes such as are characteristic of the social sciences. Salmon's account is shown not to escape difficulties associated with Simon's earlier approach.