Factor & Discriminant Analysis

#1
From a large data set, I've created several groupings of individuals based on factor analysis. In the future, if I want to predict which group an individual fits into based on the same questions, should I use the factor loadings I have, or do a discriminant analysis (or logistic regression - the dependent variable is categorical, the independent variables are dichotomous) on the data I have using the groups I created and use those coefficients? I did a DA on the data I have, and it accurately predicted the group 75% to 90% of the time.

Thanks for any help!
 
Last edited:

CB

Super Moderator
#2
When you say you've created several groupings based on factor analysis, are you talking about groupings of individuals or groupings of variables? (I'd have thought you could only do the latter, but that doesn't seem to be what you have in mind here...?)