I am still running stata 8. that was the old way then. I'm outdated!!!
1)What kind of factor analysis are you trying to perform: exploratory or confirmatory?
2)If we are talking about exploratory, you should try to rotate the coefficients. It causes the item loads to be best distributed and eases the interpretation of your 'new' factors. In other words, the structure of the factors become way more clear when the vectors are rotated. If you have stata manuals, check them. The factor analysis section is pretty helpful when you are starting.
Also, you should only score the factors after rotation - or score rotated factors, it appears, in stata 9.
D
-----Original Message-----
From: [email protected]
[mailto:[email protected]]On Behalf Of Paley, Irina
Sent: April 5, 2006 5:34 PM
To: [email protected]
Subject: st: RE: RE: coefficients in factor analysis
Diego,
thanks so much for your reply!
yes, you are right about the number of factors. This was just an example and I mistyped.
Regarding the score command: using stata 9, when I type in: score f1 right after factor, stata tells me: score is not valid after factor.
In the help section, it mentions: predict f1 f2: score first two, rotated factors
1) Am I doing something wrong with the score command?
2) Do you think that the (rotated or unrotated) coefficients that are displayed after factor command are the ones that should be used to actually create the factors? Would you be able to refer me to how these are related to their "regressions scores"? I am sorry I can't seem to find what a regression score is...
Thanks so much for your help!
Irina
-----Original Message-----
From: [email protected]
[mailto:[email protected]]On Behalf Of Diego Bassani
Sent: Wednesday, April 05, 2006 5:03 PM
To: [email protected]
Subject: st: RE: coefficients in factor analysis
I guess you can't 'predict' after factor. you can score. if that is the scenario, what you are seeing are regression scores of the unrotated coefficients you obtained in the first run. Check the eigen values to see how many factors you should retain. also, you mention 'predicting' three factors but your coefficients seem to fit two at max...
maybe I misunderstood something, if I did, give me more details and I'll try to help you.
Diego
Diego G. Bassani, M.Sc., Ph.D.
Centre for Addiction and Mental Health
Health Systems Research & Consulting Unit
33 Russell Street, Tower - T308
Toronto, ON - M5S 2S1
[email protected]
http://www.camh.net/hsrcu
Phone: +01 (416) 535 8501 # 6465
Fax: +01 (416) 979 4703
-----Original Message-----
From: [email protected]
[mailto:[email protected]]On Behalf Of Paley, Irina
Sent: April 5, 2006 4:50 PM
To: [email protected]
Subject: st: coefficients in factor analysis
Dear Statalist,
I apologize in advance for probably a very uneducated question.
I am using factor analysis to create factors.
when I type: factor x1-x3, I get an output which looks like this:
Factor1 Factor2
x1 0.4 0.7
x2 0.9 0.2
x3 0.6 0.5
When I then type: predict f1 f2 f3, I get another output which has the same form but the actual "coefficients" are completely different. For example, they are not 0.4, 0.9 and 0.6 for xs contributing to Factor 1.
Could you explain the correspondence between the contributions of x1, x2 and x3 to, say, Factor 1, reported after "factor" and those reported after "predict"? again, my apologies but I couldn't find it in readily available factor analysis references.
Thank you so much,
Irina
*
* 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/
*
* 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/
*
* 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/
*
* 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/
