STRUCTURAL EQUATION MODELING IN APPLYING FOR THE CROSS-CULTURAL DATA-ANALYSIS

Practical and theoretical issues in applying structural equation model ing to analysis of multiple-group (e.g., crosscultural) data are discu ssed. Two main problems are considered: a) testing the comparability, or equivalence of psychological constructs and b) detecting possible b etween-group differences on the constructs. The strong factorial invar iance (W. Meredith, 1993) of each variable's loading and intercept (me an-level) parameters implies that constructs are fundamentally the sam e in each sociocultural group, and thus comparable. Under this conditi on, hypotheses about the nature of between-group (e.g., sociocultural) differences and similarities can be confidently and meaningfully test ed among the constructs' moments. The general overview of some of the issues involved in making such tests were provided to explicate them w ithin the framework of multiple-group mean and covariance structures a nalyses (MACS) using a detailed example comparing four sociocultural s ettings.