[Rasch] DK Rasch analysis

Agustin Tristan ici_kalt at yahoo.com
Tue May 18 02:20:05 EST 2010

There are many times that test designers use the dont know option because they think that it is a way to see (or measure? i don't know) if the person is honest in the responses.
It is assumed by those designers that :
a) A correct response implies that the person has the knowledge and is honest
b) A wrong response implies that the person does not have the knowledge and tried to guess and in consequence is not honest
c) A missing response implies that the person does not have the knowledge and is careless, he had no time to answer,  or is probably honest or perhaps was sleeping?
To this regard, it is evident that a correct and a wrong response has nothing to do with honesty, because the designer provides the distractors as plausible competitors to the correct answer, and they may come from a wrong concept, a mistake in a calculation, a bad competency of the person facing the problem. We cannot punish a person for a wrong response assuming that guessing is involved. In fact it is worse to guess an get a correct response because we shall never know if the person tried to guess as we assume that he has the knowledge. The response pattern is useful in this case.
OK, we are interested on honesty! Why don't we produce a test for honesty and measure this trait from low to high honesty, not involving other kind of construct? In fact it is a matter of vailidity of the test: is it designed for knowledge? for honesty? for sleeping capabilities?
Agustin Tristan

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): http://www.ieia.com.mx
web page (in English) http://www.ieesa-kalt.com/English/Frames_sp_pro.html

--- On Mon, 5/17/10, Tom Conner <connert at msu.edu> wrote:

From: Tom Conner <connert at msu.edu>
Subject: Re: [Rasch] DK Rasch analysis
To: "Randy & Shelley MacIntosh" <srmac at bluebottle.com>
Cc: "rasch at acer.edu.au" <rasch at acer.edu.au>
Date: Monday, May 17, 2010, 10:30 AM

The almost unsolvable problem with such an investigation is the interpretation of "don't know" response options.  The issues have been debated multiple times in the literature and I do not see any consensus.  What would the other response options be?  What would the instructions be for filling out the survey?  What would be the placement of "don't know" among the other response options?  How would you distinguish not knowing from "admitting" to not knowing?


Randy & Shelley MacIntosh wrote:
> I am interested in studying the use of  'don't know' responses to survey questions about current events.
> We suspect that women are more willing to admit they don't know something compared to men.
> I was thinking about applying the Rasch model and treating 'don't know' as the 'correct' answer, and then doing a DIF analysis.
> I was wondering if anyone has conducted a similar analysis or can perhaps suggest a published study?
> Thanks,
> Randy MacIntosh
> CSU Sacramento
> _______________________________________________
> Rasch mailing list
> Rasch at acer.edu.au
> https://mailinglist.acer.edu.au/mailman/listinfo/rasch

-- Tom Conner
Professor of Sociology
Michigan State University
office: 517 355-1747
cell: 517 230-0343
"What if there were no hypothetical questions?"

Rasch mailing list
Rasch at acer.edu.au

-------------- next part --------------
An HTML attachment was scrubbed...
URL: https://mailinglist.acer.edu.au/pipermail/rasch/attachments/20100517/7f20f739/attachment.html 

More information about the Rasch mailing list