[Rasch] Removing misfitting items - Ben W. is right!

Svend Kreiner svkr at sund.ku.dk
Thu Jul 27 05:10:33 AEST 2017


I am sorry, but I have to disagree. Crisan et.al.’s results do not tell us that the Rasch model is robust to item misfit.
 
What they have shown is that the 2PLM model appears to be relatively robust against a “2-dimensional” alternative model.
However, the misfit model is is not a standard two-dimensional model. It is a 2PLM model with the same difficulty but a little stronger item discrimination than the true 2PLM and with a random error added to the person parameter.
 
It goes almost without saying that responses from the misfit model will be similar to responses from the true 2PLM. 
 
For this reason, it is misleading to suggest that their results show that the Rasch model is robust against the kind of misfit (item misfit, DIF and local dependence) that we are concerned about when we test the Rasch model.
 
I agree that Ben Wright deserve a happy afterlife, but it won’t be because of this paper.
 
Cordially
 
Svend

PS 
I tried to include the fomulas showing you the models, bu my email-system at the university would not include them. nInstead, I have sent them from my private e-mail system, but I am not sure that they will turn up at the Rasch-list. if not, I suggest that you either read the paper yourself or send me an email, so that I can return the formulas to you. 



________________________________________
From: Rasch [rasch-bounces at acer.edu.au] on behalf of Gerardo Prieto [gprieto at usal.es]
Sent: Wednesday, July 26, 2017 9:12 AM
To: rasch at acer.edu.au
Subject: Re: [Rasch] Removing misfitting items - Ben W. is right!

Thanks, Mike.
But, removing misfitting anchor items in equating it is convenient?

Regards

Gerardo
> El 26/7/2017, a las 7:52, Mike Linacre <mike at winsteps.com> escribió:
>
> ".   removing the misfitting items only improved the results in the case of severe multidimensionality and large proportion of misfitting items, and deteriorated them otherwise."
> in "Investigating the Practical Consequences of Model Misfit in Unidimensional IRT Models",
> Daniela R. Crisan, Jorge N. Tendeiro, Rob R. Meijer, Applied Psychological Measurement, 41, 6, 439–455, July 2017
>
> Once again, the Rasch model is proved to be robust. Ben Wright is smiling .....
>
> Mike L.
>
> ________________________________________
> Rasch mailing list
> email: Rasch at acer.edu.au
> web: https://mailinglist.acer.edu.au/mailman/options/rasch/gprieto%40usal.es

________________________________________
Rasch mailing list
email: Rasch at acer.edu.au
web: https://mailinglist.acer.edu.au/mailman/options/rasch/svkr%40sund.ku.dk


More information about the Rasch mailing list