Despite the lip service which many sociologists pay to Popper’s hypothetico-deductive model (HDM) of theory testing, few if any major social theories have been definitively falsified. The reason is that sociological explanations do not fit the deductivist model of explanation: They do not contain universal or statistical “covering laws” from which falsifiable hypotheses could be deduced. Sociological explanations are better conceived in realist terms as causal models of the social processes that produce certain outcomes. While few models are completely false, some are nonetheless more empirically adequate than others. This essay argues that (1) the CLM is inadequate to sociology and that (2) attempts to reformulate the HDM are therefore destined to fail. It then outlines (3) a constructive realist model (CRM) of sociological explanation and uses it to develop (4) explanatory realist (ER) criteria for evaluating explanations. Deductivist and realist approaches to methodology are then compared through an examination of Skocpol’s States and Social Revolutions.