Statalist


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

st: RE: Tetrachoric Correlation for binary and continuous variables


From   "Verkuilen, Jay" <[email protected]>
To   <[email protected]>
Subject   st: RE: Tetrachoric Correlation for binary and continuous variables
Date   Tue, 18 Nov 2008 11:40:41 -0500

You could use -polychoric- as Stas mentions. WARNING: Such matrices are
sometimes not positive definite. 

If you need to do factor analysis with variables of mixed type, you may
want to set things up as a confirmatory analysis and use -gllamm-
instead, or consider using software explicitly set up for analysis of
mixed variables such as Mike Neale's MX (free!) or Mplus. 


-----Original Message-----
From: [email protected]
[mailto:[email protected]] On Behalf Of Juan Julio
Gutierrez
Sent: Monday, November 17, 2008 10:05 PM
To: [email protected]
Subject: st: Tetrachoric Correlation for binary and continuous variables


Dear all, 

I want to perform a factor analysis. My set of variables combine binary
and continous variables. 

I understand that the Tetrachoric Correlation is used to identify
reasonable covariance matrices for binary data, before performing factor
analysis. 

When reading the Stata online help, it seems that the comand 

tetrachoric var1 var2 var3, 

will treat all as binary variables, when for example var3 is actually
countinuous. 

the problem is how to perform tetrachoric only in the subset of binary
vars and then perform factor in both binary and continuous vars?

Thanks a lot

Juan Julio


      
*
*   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/



© Copyright 1996–2024 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index