correlation of factors after rotation (using factormat with polychoric correlation matrix)

[Dharmi Kapadia]

Hi,

I have used exploratory factor analysis using factormat for a polychoric correlation matrix. I extracted two factors and I want to know the correlation between them. I tried

estat common

but this returned that the correlation between the two factors was 0.


I then tried to use this code which I found on Stata' FAQ.

mat L = get(r_L)
mat T = syminv(L'*L)*L'*get(Ld)
mat Phi = T*T'
mat list Phi

However this did not work either. I got the error: "matrix L not found". But when I did

ereturn list

it seemed that this matrix was there. These are the matrices that were available:
e(r_Ev) : 1 x 2
e(r_Phi) : 2 x 2
e(r_T) : 2 x 2
e(r_L) : 12 x 2
e(C) : 12 x 12
e(Phi) : 2 x 2
e(L) : 12 x 2
e(Psi) : 1 x 12
e(Ev) : 1 x 12

Can anyone help with how to get the correlation matrix between the two extracted factors?

Many thanks,
Dharmi

[wbuchanan]

Need to see the commands you used. Did you use any options/commands to get obliquely rotated solutions? If the command you used an orthogonal rotation would you expect there to be a large enough non-zero correlation to be detected? (And yes, I'm aware that orthogonal components should have a zero correlation, but just wanted to include a sanity check).

[Dharmi Kapadia]

Hello,

Thank you for the response.
Here is the full list of commands that I used. I used oblique (oblimax) rotation, therefore the command, estat common, should have returned a non-zero correlation.

polychoric clpersa clpersb clpersc clpersd clperse clpersg clpersh clpersi/*
*/ clpersj clpersm clpersn clperso if respsex==2

display r(sum_w)
global N = r(sum_w)
matrix r = r(R)
factormat r, n($N) factors(2)
rotate, oblimax

mat L = get(r_L)
mat T = syminv(L'*L)*L'*get(Ld)
mat Phi = T*T'
mat list Phi

estat common

I would appreciate if anyone can see where I have made an error?

Many thanks,
Dharmi