Hi allstat,
we are clinicians with an interest in research. At the moment we
are looking at the factor structure of the Beck depression inventory
in the brain injured population. After running a factor analysis using
principal components with an oblim & varimax rotation we found
interpretation was impossible. We decided to look at our basic
stats and found that our data was (A) positivly skewed even after
several attempts to fix this with varying types of transformations
such as square root and log!. The tansformations made the second
factor analysis even less interpretable.
(B) linerarity was not present evident in the scatterplots between
two variables randomly selected. In fact our scatterplots were
rather rectangular.
Our sample size = 118
If anyone could help with advice that would be great,
I can be contacted at [log in to unmask]
thanks.
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|