[Rasch] evidence of validity?

Svend Kreiner svkr at sund.ku.dk
Thu Jan 3 06:12:40 EST 2013


Hi Shirin,

Item fit statistics are useful during analysis of validity, but they have poor power against multidimensional alternatives so you have to do more to convince other researchers that you measure a unidimensional trait. Analysis of local dependence is a natural starting point for that purpose.

Item maps are used for analysis of targeting, but have nothing to do with validity. You may have a perfectly valid measure which is completely useless because it is out of target for your study population. Item maps will tell you if that is the case.

Hope this helps.

Svend
________________________________________
From: rasch-bounces at acer.edu.au [rasch-bounces at acer.edu.au] on behalf of shirin shirazi [shirin.shirazi at gmail.com]
Sent: Wednesday, January 02, 2013 7:53 PM
To: Rasch
Subject: [Rasch] evidence of validity?

Dear List members,

Happy New Year to all wishing you the best.

I have a question and I would appreciate if you could help me.

Does information we get about unidimensionality of our construct by using Winsteps interpreting infit values or item map (especially to support the construct validity of a test or questionnaire), we can also obtain through running Facets and instead interpret the information generated in a table called Category Statistics? Are both needed to substantiate validity? or perhaps both provide the same information we need?

Your responses are highly appreciated

Shirin





More information about the Rasch mailing list