st: RE: Factor analysis with binary data

["David Radwin" <dradwin@mprinc.com>]

Walt,

The usual advice on this list is to not dichotomize or categorize
continuous variables at all, much less do so based on an "arbitrary
guess."

See, for example:

http://www.stata.com/statalist/archive/2010-02/msg00871.html

http://www.stata.com/statalist/archive/2010-11/msg00443.html

Is there some way to use the variables as continuous?

David
--
David Radwin
Research Associate
MPR Associates, Inc.
2150 Shattuck Ave., Suite 800
Berkeley, CA 94704
Phone: 510-849-4942
Fax: 510-849-0794

www.mprinc.com


> -----Original Message-----
> From: owner-statalist@hsphsun2.harvard.edu [mailto:owner-
> statalist@hsphsun2.harvard.edu] On Behalf Of Data Analytics Corp.
> Sent: Wednesday, March 02, 2011 9:10 AM
> To: statalist@hsphsun2.harvard.edu
> Subject: st: Factor analysis with binary data
> 
> Hi,
> 
> I have to do a factor analysis with binary survey data.  I have no
> problem doing the factor analysis per se (I'll develop a correlation
> matrix using tetrachoric correlations), but I do have a question about
> the predicted scores.  They will be continuous, but I need them for
> other analysis to be binary.  Any suggestions for how I can take the
> scores for a factor and recode them into 0/1 values.  I thought of
> looking at the distributions and making an arbitrary guess for a
> cut-off: anything above is 1; below is 0. A first guess for a cut-off
> would be 0: anything positive is 1; negative is 0.  Does anyone have a
> better suggestion?
> 
> Thanks,
> 
> Walt

*
*   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/