Factor rotation performed on PCA output

[EViews Glenn]

Your loadings don't appear to be the same (FEVEC). What's the version/date on your copy of EViews?

[bernerth]

Just realized there is a page two in this chat chain.
My Eviews version is 7.1, Standard Edition - Sep 10 2010 build

[EViews Gareth]

Way out of date. You need to update to the latest version by installing the Upgrade Installer:
http://eviews.com/download/download.html

[bernerth]

Thx Glenn, now it worked. Is there a way to perform the varimax rotation on the adjusted factor loadings? By adjusted I mean divide the factor loadings of the principal factors (as derived by the EFA) by the square root of the corresponding eigenvalue? That way I can perform the rotation on the loadings as if they were the loadings of the principal components (as derived by the EFA). I understand that ordinally it won't make a difference, but the actual rotated loadings will be different if I simply perform the rotation on the unadjusted factor loadings.
Thx
T

[EViews Glenn]

I can't think of an easy way. I think that you *might* be able to compute the covariance matrix and save it using the covariance tools. Then scale it by the eigenvalues in such a way that the factor analysis gives you scaled loadings that match the principal components unscaled loadings. Feed that into the factor analysis routines. But that's just a hunch.

[bernerth]

Ok thx Glenn. For interpretation purposes the ordinal ordering works for me.
Thx again for all your help.
T