[Rasch] Incredible comments on Rasch

Jakob Bjorner jbjorner at qualitymetric.com
Thu Oct 5 22:44:11 EST 2006


Dear Eric,

In my opinion, the two sentences:

'The proposed validation analysis (using Rasch modelling) is not
appropriate for a questionnaire focusing on attitude and behaviour and
should be revised.'

And 

'... Rasch analysis for the valdiation/reduction of items is not
appropriate because the items on attitude or usage do not have a linear
order, unlike educational assessment ...'

represent a misunderstanding of the Rasch model. Certainly, Rasch models
can be used to analyze questionnaires on attitude and behavior. However,
Rasch analysis can be applied in many different ways, some of which may
be more or less helpful depending on the situation. E.g. some
researchers would estimate the location parameter (mean of the item
thresholds) for each item and exclude items that had location parameters
too similar to other items. That may be a good strategy if you want a
test with a clear hierarchical order of items, but may not be a good
idea in the present situation. It seems to me that the reviewers are
taking issue with such an approach. However, that is only one way to use
the Rasch model. Mathematically, the model can still be valid, even if
all items have the same location (this is the sort of test you would
want in a fail-pass situation).

I agree with Bryce and the reviewers that know group comparison,
concurrent validity, and construct/criterion validity can be valuable
additions to a Rasch analysis. These approaches are not mutually
exclusive.

Hope this helps.

Jakob

 

Jakob Bue Bjorner, MD, PhD
Chief Science Officer
QualityMetric Incorporated
640 George Washington Highway, Suite 201 
Lincoln, RI  02865
Phone: (401) 334-8800, ext. 271
Fax: (401) 334-8801
jbjorner at qualitymetric.com
<blocked::BLOCKED::mailto:jbjorner at qualitymetric.com>  
www.qualitymetric.com <blocked::BLOCKED::http://www.qualitymetric.com/> 
www.sf-36.org

________________________________

From: rasch-bounces at acer.edu.au [mailto:rasch-bounces at acer.edu.au] On
Behalf Of Reeve, Bryce (NIH/NCI) [E]
Sent: Friday, September 29, 2006 1:21 PM
To: Eric Wong; rasch at acer.edu.au
Subject: RE: [Rasch] Incredible comments on Rasch

 

Dear Eric and colleagues,

 

I do not know what you proposed, but from the comments below it seems
like they are saying that a full validation of the IMAQ questionnaire
should include a number of different validation
methodologies/strategies, not just Rasch analysis.  I agree, Rasch
analysis gives you some information but not the full picture.  In
validating the IMAQ, you should consider the questionnaire's construct
validity (e.g., convergent and divergent validity), content validity,
and criterion validity.  These different types of validation require
different methods.  By itself, one method does not address all aspects
of questionnaire validation.

 

Sincerely,

-Bryce

 

________________________________

From: Eric Wong [mailto:mcwong at cuhk.edu.hk] 
Sent: Friday, September 29, 2006 5:16 AM
To: rasch at acer.edu.au
Subject: [Rasch] Incredible comments on Rasch
Importance: High

 

	Dear all,
	
	Unbelievable comments from reviewers of a research grant review
board:
	
	"The questionnaire should be shortened. The proposed validation
analysis (using Rasch modelling) is not appropriate for a questionnaire
focusing on attitude and behaviour and should be revised."
	
	"The methods proposed for the adpatation of the content of the
IMAQ (the questionnaire) are well written, but Rasch analysis for the
valdiation/reduction of items is not appropriate because the items on
attitude or usage do not have a linear order, unlike educational
assessment. More standard methods, such as know group comparison or
concurrent validity, for validating the construct/criterion validity of
the IMAQ should be used."
	
	Any comments?

	 

	Eric 

 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: https://mailinglist.acer.edu.au/pipermail/rasch/attachments/20061005/dc5749ec/attachment.html 


More information about the Rasch mailing list