In sci.stat.edu Johannes Hartig <[EMAIL PROTECTED]> wrote:
> Hi all,
> I have a (hopefully simple but not stupid) software question
> related to factor analysis of dichotomous variables:
> Christofferson (1975) described a GLS estimator for the
> factorization if dichotomous data based on the marginal
> distributions for single items and item pairs. Muthen (1978)
> suggested another method based on the same information.
The NAG library contains a routine for the full ML single factor
solution under the threshold model. Mx (http://griffin.vcu.edu/mx - a
free SEM program) will also do this, but you would have to write a
script (it does have the advantage of dealing with missing (MCAR) data
in its "raw data" approach) and are probably limited to 20-30 variables
tops. LISREL allows you to fit the WLS ("ADF") model to the tetrachoric
correlation matrix where the weight matrix is for all pairs of variables
(for single factor models, this is very close to that from the NAG
routine).
David Duffy.
--
| David Duffy. ,-_|\
| email: [EMAIL PROTECTED] ph: INT+61+7+3362-0217 fax: -0101 / *
| Epidemiology Unit, The Queensland Institute of Medical Research \_,-._/
| 300 Herston Rd, Brisbane, Queensland 4029, Australia v
=================================================================
Instructions for joining and leaving this list and remarks about
the problem of INAPPROPRIATE MESSAGES are available at
http://jse.stat.ncsu.edu/
=================================================================
