Realism in the Sociology of Education: `explaining' social differences in attainment.

The article examines the relationship between techniques of quantitative analysis and explanatory models of the association between social class and educational success in the context of empirical data from the New Zealand Progress at School project. The data discusses in this paper are from the Progress at School project which followed the educational careers of students who entered 37 New Zealand secondary schools in 1991. The research was designed as a school effects study, but few differences in student attainment that could be attributed to any property of the school attended were observed, and the data examined here with other interests in mind. The Progress at School project was structured by a family resource framework that might best be described as an explanatory sketch, or a set of connected hypotheses, for the purposes of directing the research and interpreting empirical data of both a quantitative and ethnographic kind. The specific techniques of path analysis, and of correlational analysis, more generally, should not be ignored by studies with access to quantitative data on social difference in educational attainment. At the same time, Boudon's argument that the mode of presentation of the data, what he calls the syntax of explanation, has a non-trivial significance in constructing causal accounts of social differences in education should be admitted as on with considerable force. This does not commit realists to the methodological individualism of Boudon's empiricism, or to accounts in terms of a model homo sociologicus acting within deterministic social structures, but neither are realists committed to the habitus-driven agent of an ultimately functionalist, and probably formally determinist, theory of social and educational reproduction.