I've used factor analysis on svy data. I did the factor analysis
without weights, then used the factor scores in regression anlysis using
svyreg. My rationale is that the factor analysis allows you to obtain
the most reliable index (factor score) of the items you have (i.e., in
their unweighted form). When the index is used in an analysis it should
be weighted to reflect its distribution in the population and to get
proper standard errors of the coefficients.
On Sun, 2003-05-18 at 08:51, abianchi wrote:
> Hello!
>
> Has anyone attempted to do a principle component factor analysis on a large
> national data set that requires stratified weighting (STRATA) and cluster
> weighting (PSU)? In Stata, when these weights are activated, you have to use
> commands that start with SVY. However, I noticed that there were no SVY
> commands for factor analyses ... any suggestions?
>
> I guess my real question is: can you use these weights and do a factor
> analysis, or do you simply run the factor analysis without the weights, run
> another test (like a reliability test) on the variables, then create a scale
> or index to be used as an independent variable in, say, a regression? Of
> course, the regression will have to use weights, but perhaps you don't need to
> use stratified and cluster weights in a factor analysis.
>
> Please help!
>
> Thanks, Alison
>
> *
> * 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/
--
Gene A. Fisher
Department of Sociology
University of Massachusetts
200 Hicks Way
Amherst, MA 01003
Phone: (413) 545-4056
Fax: (413) 545-3204
email: [email protected]
*
* 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/
