PROC GLM for multiple linear regression

[Ruth]

Hi Sir,

I used PROC GENMOD to run a linear regression analysis. The reason that I want to use PROC GENMOD is that it provides the effect coding, while PROC GLM does not have this option.

As the analysis is linear regrssion (DIST=normal), can I get the R-Squared value from PROC GENMOD? It seems that there is no option available for this request.

Many thanks!

[Doc_Duke]

If you are just looking for the R-squared, you can re-run your model in GLM.  The R-squared is invariant to the coding method for categorical variables.