Jas Wer posted on Friday, January 16, 2015 - 6:04 am
I'm applying an EFA to categorical items in a clustered dataset. I noticed that when I use the Type = Complex EFA command, the factor loadings change slightly. I'm unsure what could explain this, since I thought that correcting for clustering would only affect the size of standard errors and therefore p-values. Is there any explanation why the size of factor loadings would change when using complex EFA? Thank you for your help.