A Note on Multigroup Comparisons Using SAS PROC CALIS

Abstract

Although SAS PROC CALIS is not designed to perform multigroup comparisons, it is believed that SAS can be “tricked” into doing so for groups of equal size. At present, there are no comprehensive examples of the steps involved in performing a multigroup comparison in SAS. The purpose of this article is to illustrate these steps. We demonstrate procedures using an example to evaluate the measurement invariance of communication satisfaction and organizational justice across 2 groups. Following the approach outlined in Byrne (2004), we conduct the same analysis in AMOS and compare the results to those obtained in SAS. We show that the sample size must be input correctly and the degrees of freedom must be adjusted in order for the standard errors and goodness-of-fit statistics to be correct. In addition, several of the fit indexes must be modified to obtain the correct values.

ACKNOWLEDGMENTS

The authors thank Yali Li for her assistance in developing a portion of the AAS code included in this article. Additionally, we thank the editor, George Marcoulides, and two anonymous reviewers for helpful suggestions that improved the quality of the final version of the article.