|
|  |
[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]
Re: st: Creating Scale with Ordinal Data using IRT
Thank you Ronan,
To respond to the questions of an anonymous statalister
([email protected]) [It always is unpleasant to respond to anonymous
persons !] and several remarks posted today:
- Item Responses Models allows considering binary AND ordinal items.
The Rating Scale Model, The Partial Credit Model and the Graded
Responses Model are the more famous models for ordinal responses
- All the models of the Item Responses Theory (IRT) can be
considered as Generalized Linear Mixed Models, and the parameters can be
estimated with GLLAMM, or for the Rasch model by -xtlogit- or -clogit-
(for the fixed effets Rasch model) [see
http://www.stata.com/support/faqs/stat/rasch.html for details on
technical aspects)
- There is some specific modules under Stata for the Item Response
Theory, for example -raschtest- (ssc describe raschtest) wich allows
estimating the parameters of the Rasch model AND to evaluate the fit of
the model to the data (IRT requires specific fit tests)
- I don't know specific Stata modules to run polytomous IRT models
but the GLLAMM manual explains how the parameters of a Partial Credit
Model can be estimated with GLLAMM and the ologit link
(http://www.gllamm.org)[I admit to using SAS to do that ! Bouuuuh]
- In IRT, there is full and full specific software. For several
years, more and more works are realised on generalist software like
Stata, R and SAS. The module -raschtest- certainly made Stata the more
powerfull software (among these 3 ones) to run and test a Rasch model,
but concerning polytomous models, there is no specific modules under
Stata and SAS (proc NLMIXED) is more easy to use for that !! For a
database of existing Stata modules, SAS macro-programs and R library in
the field of IRT, see the FreeIRT website at http://freeirt.anaqol.org
Best regards,
Jean-Benoit Hardouin
--
Jean-Benoit Hardouin
University of Nantes - Faculty of pharmaceutical sciences
France
[email protected]
http://www.anaqol.org
Ron�n Conroy a �crit :
On 11 Ean 2007, at 15:57, [email protected] wrote:
What commands might I use to create one scale for each person using IRT
based on the responses to the 5 items?
Have a look at Jean-Benoit Hardouin's excellent site, http://anaqol.org/
He's got a treasure trove of Stata resources for IRT.
=========
Ron�n Conroy
Royal College of Surgeons in Ireland
[email protected]
+353 (0) 1 402 2431
+353 (0) 87 799 97 95
http://www.flickr.com/photos/ronanconroy
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
begin:vcard
fn:Jean-Benoit Hardouin
n:Hardouin;Jean-Benoit
org;quoted-printable;quoted-printable:Universit=C3=A9 de Nantes - Facult=C3=A9 des sciences pharmaceutiques;D=C3=A9partement de Biomath=C3=A9matiques et Biostatistique
adr:BP 53508;;1, rue Gaston Veil;Nantes Cedex 1;;44035;France
email;internet:[email protected]
title:ATER - PhD
url:http://anaqol.free.fr
version:2.1
end:vcard
