Boudon has distinguished between the primary and secondary effects of socialization as a cause of social disparities in education. His explanation of secondary effects, which rests on an analysis of decision-making within opportunity cost constraints, has attracted support from realist sociologists. The empirical evidence, however, suggests that primary effects, largely the result of cognitive socialization in early childhood, may be a more important source of variance than Boudon recognizes. Some implications of this for a general theory of inequality/difference are examined with reference to the character of social explanation and in the context of the realist discussion on the structure-agency problem.