Tahir,

just use dummy or contrast-coded variables as representatives for your groups in your regression equation. For two groups, you only need one code variable, for three you need two, and for four groups you need three. Include the code variables in the regression equation together with the products of each code variable with your predictor as well as with the predictor itself.

For three groups it is explained in - this is a step-by-step tutorial:

West, S.G., Aiken, L.S., & Krull, J.L. (1996). Experimental personality designs: Analyzing categorical by continuous variable interactions. Journal of Personality, 64, 1-48.

The generalization for more than three groups is then straightforward.

If you have access to books, these are very informative and my favorites:

Aiken, L.S., & West, S.W. (1991). Testing and interpreting interactions. Newbury Park, CA: Sage.

Cohen, J., Cohen, P., West, S. G., & Aiken, L. S. (2003). Applied multiple regression/correlation analysis for the behavioral sciences, 3rd ed. Hillsdale: Erlbaum.

Regards, Karin

Thanks Karin, but I was trying to do it in AMOS after running SEM. So now what I did was to do it two groups at a time. So if I have 4 groups in education I run analysis of two at a time making it a total of 6 times. Hope Im doing ok?

Tahir, I am not familiar with AMOS, but all programs for the analysis of structural equation models are able to deal with more than two groups. You need to specify the comparison of four groups. There is an article by Barbara Byrne in which three groups are compared using AMOS - four groups should then be straightforward (I hope):

Barbara M. Byrne (2004): Testing for Multigroup Invariance

Using AMOS Graphics: A Road Less Traveled, Structural Equation Modeling: A

Multidisciplinary Journal, 11:2, 272-300.

A Google search will provide a link to the article.

HTH

Karin

Thanks Karin, will check it and let you know if it works :)