Research emphasising the importance of parenting behaviours and aspirations for child outcomes has been seized on by policymakers to suggest the responsibility of the worst off themselves for low levels of social mobility. This article provides a critique of the way in which research evidence has been used to support the dominant policy discourse in this area, as well as an empirical analysis. We use the Millennium Cohort Study to interrogate the relationship between social class and attainment in the early years of schooling. We investigate the extent to which social class inequalities in early cognitive scores can be accounted for by parental education, income, family social resources and parental behaviours. We conclude that social class remains an important concept for both researchers and policymakers, and that the link between structural inequalities and inequalities in children’s cognitive scores cannot be readily accounted for in terms of individual parenting behaviours.