Dear all,
Sorry to bother you again but I'm a bit stuck with this so I'd like to pose
my question on pcf again hoping somebody likes to share her or his
knowledge.
In general my question is on the choice of factor methods. Stata offers four
methods for factor analysis: pcf, pf, ipf, and ml that have different
features. The latter three mostly yield very similar results, so to choose
one out of these seems not to be crucial (I may be wrong on this), though
one would want to avoid arbitrariness (how?).
pcf is different. In any factor analysis I did pcf yields much higher factor
loadings and also often a clearer and more stable factor structure. Just
telling from the results pcf would be the method of choice. But can I use
results obtained by pcf the same way than results obtained by another
method? Or is a factor loading of, say, .5678 in pcf somewhat less than in
ml?
To get an idea of how pcf results are to be interpreted in comparison with
other results one would like to know what makes pcf different. Stata's
reference guide states that the pcf method assumes the communalities to be
1. In my understanding this is a quality of principal components analysis
not of factor analysis. In addition to that pcf normally finds its own
assumption to be wrong, i.e. the items do have some unique variance. Do pcf
results contradict pcf assumptions? What am I missing here?
Any help highly appreciated!
Best, Kai
_____________________________________
Kai M�hleck
International Social Justice Project (ISJP)
Institut f�r Sozialwissenschaften
Humboldt-Universit�t zu Berlin
Unter den Linden 6
10099 Berlin
+49-30-2093-5382
+49-30-2093-4430 (Fax)
[email protected]
www.isjp.de
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
