how to compare two correlation matrices
Posted on
krzysiek
Joined: 09/05/2013
Forums
Hi,
I'm trying to compare two correlation matrices from a study of two different samples of people from different cultures. These matrices aren't symmetrical because they represent the correlations of two different sets of psychological tests. Is there a way to do this in OpenMx?
Thanks in advance for your help.
Krzysztof
I'm trying to compare two correlation matrices from a study of two different samples of people from different cultures. These matrices aren't symmetrical because they represent the correlations of two different sets of psychological tests. Is there a way to do this in OpenMx?
Thanks in advance for your help.
Krzysztof
Symmetry
If you have two different tests, you should still be able to estimate the correlations within each test, as well as the cross-correlations between tests. True, the last part is not symmetrical, but as I see it the matrices could be set up as
A B'
B C
where B is the across-test block, and A and C are the two within-test blocks. This matrix is symmetrical and could be analyzed in the usual ways. Normally, one would use a covariance matrix (or raw data) instead of a correlation matrix for this analysis.
HTH
Mike
Log in or register to post comments