Abstract
The socio‐economic status of school pupils is best measured by multiple indicators. Some indicators belong to individuals, and some to families; both levels can make an independent contribution, as can separate indicators for mothers and fathers. Multiple indicators allow an efficient and conceptually appropriate treatment of item non‐response. Because this can provide a full description of the SES of one‐parent families, a distinction can be drawn between family structure and family SES. SES is also relational, operating through social groupings such as schools. A statistical analysis of all these propositions is provided using data from Fife, Scotland. The statistical structure of the SES of 16‐year‐old school pupils is investigated, and a statistical model is developed of the relationship between SES and attainment in public examinations. This relationship is estimated both with and without a control for measured ability on entry to secondary school. The modelling shows that several dimensions of SES are needed to model attainment, that SES is an important correlate of attainment over and above ability, and that the SES mix of the school has an effect on attainment over and above individual or family SES. A policy implication is that reliable performance indicators should incorporate measures of SES even if a measure of ability is available.

This publication has 40 references indexed in Scilit: