Individual Differences in Structural Equation Model Parameters

Abstract

Individuals may differ in their parameter values. This article discusses a three-step method of studying such differences by calculating and then modeling "individual parameter contributions", making the study of heterogeneity in arbitrary structural equation model parameters as technically challenging as performing linear regression. The proposed approach allows for a contribution to the study of differential error variances in survey methodology that would have been difficult to make otherwise.