R help - Dec 2004 - Re: Tetrachoric and polychoric correlations, Polycor package

A bit late, but you might like to look at

http://www.qimr.edu.au/davidD/polyr.R

Regarding the original posters queries:

You can analyse polychoric correlations as if they were Pearson
correlations using standard software (eg sem), and this usually doesn't do
too badly, or go to AWLS (Browne) in LISREL etc, or ML analysis
of the full multidimensional contingency table using programs such as Mx,
or as you noted, mvtnorm (Mx uses Alan Genz's algorithms).

You can check model assumptions, and compare the results to those
from similar loglinear models.  For example, for a 3-way table, a single
factor model based on polychoric correlations should fit "perfectly",
if
the no higher order interaction assumption is right,


| David Duffy (MBBS PhD)                                         ,-_|\
| email: davidD at qimr.edu.au  ph: INT+61+7+3362-0217 fax: -0101  /     *
| Epidemiology Unit, Queensland Institute of Medical Research   \_,-._/
| 300 Herston Rd, Brisbane, Queensland 4029, Australia  GPG 4D0B994A v

Dear David,

Your code looks very similar to what's in the polycor package, although the
latter will also calculate quicker estimates with thresholds based on the
marginal distributions of the variables (and also polyserial correlations).
I may add WLS estimates to the sem package -- suggestions would be
appreciated.

Regards,
 John

--------------------------------
John Fox
Department of Sociology
McMaster University
Hamilton, Ontario
Canada L8S 4M4
905-525-9140x23604
http://socserv.mcmaster.ca/jfox 
-------------------------------- 
> -----Original Message-----
> From: r-help-bounces at stat.math.ethz.ch 
> [mailto:r-help-bounces at stat.math.ethz.ch] On Behalf Of David Duffy
> Sent: Thursday, December 09, 2004 1:29 AM
> To: r-help at stat.math.ethz.ch
> Subject: [R] Re: Tetrachoric and polychoric correlations, 
> Polycor package
> 
> A bit late, but you might like to look at
> 
> http://www.qimr.edu.au/davidD/polyr.R
> 
> Regarding the original posters queries:
> 
> You can analyse polychoric correlations as if they were 
> Pearson correlations using standard software (eg sem), and 
> this usually doesn't do too badly, or go to AWLS (Browne) in 
> LISREL etc, or ML analysis of the full multidimensional 
> contingency table using programs such as Mx, or as you noted, 
> mvtnorm (Mx uses Alan Genz's algorithms).
> 
> You can check model assumptions, and compare the results to 
> those from similar loglinear models.  For example, for a 
> 3-way table, a single factor model based on polychoric 
> correlations should fit "perfectly", if the no higher order 
> interaction assumption is right,
> 
> 
> | David Duffy (MBBS PhD)                                         ,-_|\
> | email: davidD at qimr.edu.au  ph: INT+61+7+3362-0217 fax: 
> -0101  /     *
> | Epidemiology Unit, Queensland Institute of Medical Research 
>   \_,-._/
> | 300 Herston Rd, Brisbane, Queensland 4029, Australia  GPG 4D0B994A v
> 
> ______________________________________________
> R-help at stat.math.ethz.ch mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide! 
> http://www.R-project.org/posting-guide.html