PCA for categorical data

[mef]

Hi,
In order to avoid multicollinearity in my regression data, I want to do a principal component analysis for purely categorical variables. Does anyone know if there is a standard way to do such an analysis in Eviews?
Thanx!

[trubador]

You can easily apply "Principal Component Analysis" in EViews, but I am not sure if it is the appropriate technique for categorical data. I think you need to try "Multiple Correspondance Analysis", for which EViews does not have a built-in procedure.