 |
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
RE: st: Principal Component Analysis, First Steps
Confirmatory factor analysis is rather different than PCA. CFA isn't in official Stata, but Stas Kolenikov wrote one, -findit cfa-.
I assume you want to create a composite variable of some sort from your proxies? If so, might I suggest looking at -alpha-?
-----Original Message-----
From: "Ulrich Atz" <[email protected]>
To: [email protected]
Sent: 4/19/2009 2:48 PM
Subject: st: Principal Component Analysis, First Steps
Dear StataList!
I am working on a probit model for self-employment. So far the usual
covariates are straightforward, but I'd like to include some more
unusual variables such as e.g. the "social ties" of an individual.
Because of my data set, this is a latent variable, though I have many
potential acceptable proxies. To reduce the structure my idea is to
use a confirmatory factor analysis and boil them all down to one
"index".
I know STATA includes the pca command, but my knowledge is limited,
hence, I wondered if someone can provide me a link for an example of
this kind. The Stata help ones are not particularly enlightening.
Especially the postestimation is not clear, so e.g. how to create a
new variable. I hope I confused not to many things and appreciate your
answer.
Best regards,
Ulrich
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/