I am newbie for R and Stats. I am using EFA in R [psych library] to identify the latent factors underlying my data. I got a reasonable amount of factors matching my purpose.

My question is once I have the factors, is there a way to categorize observations based on the dominant factors, one or more?


1 Answer 1


It might be better to try to classify people before doing the EFA, using e.g. cluster analysis which is designed for that purpose.

However, if, for some reason, you want to classify on the factors, I would first make a bunch of plots: Univariate density plots of each factor, then some bivariate plots, what else I would do depends on the number of factors. It might be obvious where the categorizations are, or it might be clear that there really aren't categories (this often happens; many things are continua). If neither is obvious, then you could try something like kernel densities.

  • $\begingroup$ Hi Peter, Thx for the reply. Actually I did try a k-means clustering beforehand and found out that there are some clusters. But I was looking for a more descriptive definition for cluster centers and thought EFA might help. So I applied EFA for the total data set. Do you mean I should do the EFA for each cluster separately? I am not familiar with the plots you mentioned I will do some reading on that. Thx. $\endgroup$
    – Suranga
    May 1, 2013 at 12:01
  • $\begingroup$ No, I don't think EFA on each cluster would make sense. You might do cluster analysis on the factors.... $\endgroup$
    – Peter Flom
    May 1, 2013 at 12:09

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.