[Rasch] Fit & Dimensionality

Anthony James luckyantonio2003 at yahoo.com
Tue Jul 15 05:51:25 EST 2008


Hi all,
Although Kubinger is an established Rasch researcher in the German world, he appearanatly doesnt see the 2PL and 3PL as violations of  specific objectivity.
Anthony

--- On Mon, 7/14/08, Agustin Tristan <ici_kalt at yahoo.com> wrote:

From: Agustin Tristan <ici_kalt at yahoo.com>
Subject: Re: [Rasch] Fit & Dimensionality
To: "-Rasch" <rasch at acer.edu.au>
Date: Monday, July 14, 2008, 7:52 AM







Hello, concerning the paper mentioned in previous messages:
 
http://www.informaworld.com/smpp/content~content=a785034366~db=all
 
It says that "For data and item pools that fail to fit the Rasch model, various reasons are listed. For instance, the two-parameter logistic or three-parameter logistic models would probably be more suitable."
 
Do we agree with this? If the item pool fails to fit the Rasch model, then the pool is not really suitable to MEASURE something, why the need to have MORE SUITABLE MODELS (2 and 3 PL), more suitable for what?  To FIT (not measure) something?
 
Please give some insight.
 
Regards
Agustin Tristan

FAMILIA DE PROGRAMAS KALT. 
Ave. Cordillera Occidental  No. 635
Colonia  Lomas 4ª San Luis Potosí, San Luis Potosí.
C.P.  78216   MEXICO
 (52) (444) 8 25 50 76
 (52) (444) 8 25 50 77
 (52) (444) 8 25 50 78
web page (in Spanish AND ENGLISH): http://www.ieesa-kalt.com

--- On Mon, 7/14/08, Mike Linacre (RMT) <rmt at rasch.org> wrote:

From: Mike Linacre (RMT) <rmt at rasch.org>
Subject: Re: [Rasch] Fit & Dimensionality
To: rasch at acer.edu.au
Date: Monday, July 14, 2008, 1:03 AM


Anthony quotes:
"Rasch model misfit does not necessarily contradict unidimensionality. The appropriateness of the scoring rule has merely been invalidated in the latter case" (p.381). http://www.informaworld.com/smpp/content~content=a785034366~db=all

Yes, this would happen when an MCQ item has been keyed incorrectly, or a survey item including the word "not" has not been reversed scored. Including different response formats (of different discrimination) in the same instrument can also produce misfit.

And we could imagine other situations in which items operationalizing a unidimensional latent trait would cause the data to misfit the Rasch model. For instance, on an MCQ test, items which cue other items. Or on a survey, items which summarize other items.

When misfit is seriously distorting the measures, the "appropriate scoring rule" could become "Don't score this item!" 

How does this sound?

Mike L. _______________________________________________
Rasch mailing list
Rasch at acer.edu.au
http://mailinglist.acer.edu.au/mailman/listinfo/rasch
_______________________________________________
Rasch mailing list
Rasch at acer.edu.au
http://mailinglist.acer.edu.au/mailman/listinfo/rasch


      
-------------- next part --------------
An HTML attachment was scrubbed...
URL: https://mailinglist.acer.edu.au/pipermail/rasch/attachments/20080714/a71868af/attachment.html 


More information about the Rasch mailing list