If one takes children who at an early age had similar scores on tests of their cognitive ability, one still finds significant differences in their later life educational attainment. How far are these differences related to their social origins? This is the overarching research question that we address – in more detail than in previous work and with reference to the twenty-first as well as to the twentieth century. To do so, we will use data relating to cohorts of children born in 1946, 1958, 1970 and the early 1990s. These data contain detailed information on the results of individuals’ performance on tests of cognitive ability at around age 10, on their parents’ social class, status and education – i.e. on their families’ economic, socio-cultural and specifically educational resources – and also on their own educational careers through into adult life and on the qualifications, academic and vocational, that they have obtained. On this basis, we will be able to give an answer to the question posed above: that is, by comparing, for each cohort, the subsequent educational attainment of children who at around age 10 were of similar ability and by ascertaining how far differences in their attainment are associated with their social origins. This question is of major academic interest in the field of social mobility. But it is also highly relevant to a range of current policy issues relating to the most effective means of offsetting the effects of disadvantaged social origins on individuals’ educational and subsequent socioeconomic attainment.