Any variable set that produces valid correlations can be factor
analyzed. FA is just the reduction of a correlation matrix. I even
asked clinicians to generate a correlation matrix for the WAIS subtests
and then factor analyzed the matrices. Everyone was remarkably
consistent in rarifying the factors. They all believed that the
subtests were much more intercorrelated than they are and they were all
apparently strong believers in Spearman's G. IQ reigns! It was very
difficult for people to incorporate noise into their judgements.
Since you have an hypothesis about the item clusters, you should look
into confirmatory factor analyses. You basically give the procedure the
model you have and the analysis tests the hypothetical factor matrix
against the pattern in the data.
Mike Williams
Drexel University
---
You are currently subscribed to tips as: [email protected].
To unsubscribe click here:
http://fsulist.frostburg.edu/u?id=13090.68da6e6e5325aa33287ff385b70df5d5&n=T&l=tips&o=11716
or send a blank email to
leave-11716-13090.68da6e6e5325aa33287ff385b70df...@fsulist.frostburg.edu
