Sem and between-group differences in latent variable means: gender differences in positive and negative affectivity

Abstract

The purpose of this article is to illustrate how to test between-group differences in latent variable means using the structural equation modeling framework. An illustration is provided outlining the use of Amos Graphics. The illustration explains how to use various steps of invariance testing, ranging from less to more restrictive, until the final test of between-group differences is accomplished. Following the outlined logic of the procedure, an example testing for gender differences in the latent variables of Positive and Negative Affectivity is provided. Controlling for measurement error, Serbian females seem to score lower on Positive Affectivity compared to males. The observed difference belongs to the range of small to medium effect size. Limitations and strengths of the procedure are discussed.