Re: Principal Components Analysis on factor data.

[Dr Murray Finkelstein <finkelm@gov.on.ca>]

THe method of correspondence analysis is designed to handle categorical 
data.  See for example the discussion in Ripley's Complements. 
http://www.stats.ox.ac.uk/pub/MASS3/Compl.shtml
Lucke, Joseph F wrote:

> The factor analysis of dichotomous/polychotomous/categorical data is a major
> research area. Simple principal components analyses (PCA) of binary data can
> generate to artifactual factors. PCA of polychotomous data is simply
> incorrect as the categories are converted into a quantitative scale.
>
> I don't think S-plus or R offer a categorical factor analysis function.
> Check out
> http://ourworld.compuserve.com/homepages/jsuebersax/binary.htm for an
> overview and the Mplus web site http://www.statmodel.com/index2.html for a
> specific software implementation. 
>
> Joe
> Joseph F. Lucke, PhD
> Statistical Scientist
> Family Nursing Care (MC 7951)
> School of Nursing
> University of Texas Health Science Center at San Antonio
> 7703 Floyd Curl Drive
> San Antonio, TX 78229-3900
> Voice: 210 567 0474
> Fax:    210 567-5822 
> www.uthscsa.edu
>
>
>
> -----Original Message-----
> From: Eric Zivot [mailto:ezivot@u.washington.edu] 
> Sent: Wednesday, August 27, 2003 3:54 PM
> To: Augusto Cesar H. Dantas; Splus Discussion List
> Subject: Re: [S] Principal Components Analysis on factor data.
>
>
> Sounds like you need to use a probabilistic discrete choice model (e.g.
> multinomial logit/probit or ordered logit/probit model). Insightful is
> developing a module for such analysis. Contact me if you are interested. ez
>
> -----Original Message-----
> From: s-news-owner@lists.biostat.wustl.edu
> [mailto:s-news-owner@lists.biostat.wustl.edu]On Behalf Of Augusto Cesar H.
> Dantas
> Sent: Wednesday, August 27, 2003 9:50 AM
> To: Splus Discussion List
> Subject: [S] Principal Components Analysis on factor data.
>
>
>
>         Dear all,
>
>         does anybody know a way to extract components in a 'YES/NO/DON'T
> KNOW' data? The data I have consists of answers to a survey, and I would
> like to analyse the relevance of the different questions. (note that a
> simple transformation to 1/0/-1 data would result in real number components;
> I would like a kind of ''binary PCA'')
>
>         Thanks in advance,
>         Augusto.
>
> --
> Augusto Cesar Heluy Dantas
> D.Sc. Student at COPPE/UFRJ
> Rio de Janeiro, Brazil
>
> --------------------------------------------------------------------
> This message was distributed by s-news@lists.biostat.wustl.edu.  To
> unsubscribe send e-mail to s-news-request@lists.biostat.wustl.edu with the
> BODY of the message:  unsubscribe s-news
>
> --------------------------------------------------------------------
> This message was distributed by s-news@lists.biostat.wustl.edu.  To
> unsubscribe send e-mail to s-news-request@lists.biostat.wustl.edu with the
> BODY of the message:  unsubscribe s-news
> --------------------------------------------------------------------
> This message was distributed by s-news@lists.biostat.wustl.edu.  To
> unsubscribe send e-mail to s-news-request@lists.biostat.wustl.edu with
> the BODY of the message:  unsubscribe s-news
>
>
>  
--

|~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
| Murray M Finkelstein PhD MD             Phone: 416-326-7879          
| Medical Consultant,                             FAX: 416-326-7761            
| Ontario Ministry of Labour                   E-mail: finkelm@gov.on.ca    
| 655 Bay Street, 14th Floor                                                                            
| Toronto, Ontario                                                                                            
| Canada M7A 1T7