[Rasch] about invariance
ici_kalt at yahoo.com
Mon Nov 5 04:00:52 EST 2007
Following the standard procedure to get the scale of a test (measures in logits vs number of correct answers), we know that two persons answering correctly M items of a single test with N items, will have the same measure B in logits (but different fit), independently of which items were answered by each person.
I've shown this to my colleagues, but one indicates he gets different results for two persons with 4 correct answers in a 10 items test, if person (a) answers only the 4 hard items and person (b) answers correctly only the 4 easy items. He promised me to send an explanation.
Has somebody found this result that does not correspond to what the Rasch model (or IRT) may explain? How can it be done if possible? This does not correspond to invariance!
I've seen RMT, winsteps manual, the book by Hambleton and the notes from my course with David Andrich, but I don't have a solution.
FAMILIA DE PROGRAMAS KALT.
Mariano Jiménez 1830 A
Col. Balcones del Valle
78280, San Luis Potosí, S.L.P. México
TEL (52) 44-4820 37 88, 44-4820 04 31
FAX (52) 44-4815 48 48
web page (in Spanish AND ENGLISH): http://www.ieesa-kalt.com
Do You Yahoo!?
Tired of spam? Yahoo! Mail has the best spam protection around
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the Rasch