Re: LCA: model failed to determine a valid solution...why?

Dr. Uebersax,
I conducted a principal component analysis, followed by varimax
rotation with Kaiser normalization, for the 17-item-HRSD and then the
25-item-HRSD. I entered HRSD items as ordinal and dichotomized
variables. I extracted all factors with an eigenvalue greater than
one, and then rotated. The sample size for both the unipolar and
bipolar datasets seemed adequate since communalities after extraction
were above 0.5. The unipolar PCA and bipolar PCA produced different
results, suggesting that the structure of the instrument is different
between the two samples. Could this be the problem with my LCA?

.

Loading