Conferences

Multiple indicator models have long been best practice in the measurement of social attitudes and other subjective phenomena. If analysed by models with latent variables and complex error structures, multiple indicators make it possible to identify the sources and size of invalid and unreliable measurement and apply appropriate corrections. This approach is rarely applied to the analysis of social-demographic variables, such as education or occupation, apparently on the mistaken assumption that these variables are measured without much random of systematic error. In this contribution I propose questionnaire formats to raise multiple indicator information for both education and occupation, and show how including this information leads to substantively different conclusions about the comparative structure of status attainment patterns. The appropriate data have been collected in the European Social Survey in the Netherlands and elsewhere.