Response: Statistical Models of Educational Stratification—Hauser and Andrew's Models for School Transitions

Abstract

Hauser and Andrew marry two good ideas that emerged within quanti tative social science during the 1970s-namely, the multiple-indicator multiple-cause (MIMIC) model for multivariate responses and the logistic response model for transitions between stages of educational attainment. The MIMIC model, now part of standard textbooks for structural equation models with latent variables (e. g., Bollen 1989), is a natural, rigorous, and efficient way to represent the effects of exogenous variables on multiple endogenous variables (Hauser and Goldberger 1971). It is also a tool for testing whether a set of variables behave as a scale in their relationship to other variables in a multivariate model and thus for avoiding ad hoc, unsubstantiated index construction (e. g., Hauser 1973). As Hauser and Andrew note, the logistic response model for school transitions is a widely used tool for the analysis of educational stratification, allowing investigators of family background effects on ed ucational attainment to recognize that schooling is a sequence of events in time rather than a single status, and that the sources of inequality of educational opportunity and outcome may be different at different stages of schooling. It is gratifying to see that these ideas have contin ued to inspire application, discussion, and improvement approximately 30 years after they were introduced. In this comment, I place my