[Rasch] Linking to compare
mschulz at pacificmetrics.com
Mon Dec 20 02:58:24 EST 2010
It sounds like your two versions already have all items in common. If all items are in common and you center the scale on the items in both analyses, the item calibrations should (theoretically) be the same. No linking required.
From: rasch-bounces at acer.edu.au [mailto:rasch-bounces at acer.edu.au] On Behalf Of Anthony James
Sent: Sunday, December 19, 2010 3:33 AM
To: Rasch at acer.edu.au
Subject: [Rasch] Linking to compare
I have developed two versions of the same test. The two tests contain the same items and are only different in one minor non-test factor. I want to compare the characteristics of the items under the two conditions.
Do I need some common items and link the two versions so that I can compare the item difficulties under the two conditions? Can I argue that since both versions contain the same items no linking is required?
Theoretically speaking if we give one test to two groups of persons and run two analyses, the items should have identical measures. Therefore, in my case can I say that any variation in item calibrations is attributed to the condition that I have introduced, without any linking?
Rasch mailing list
Rasch at acer.edu.au
More information about the Rasch