[Rasch] Test Equating

Rama Guinda guinda36 at hotmail.com
Wed Dec 14 21:36:06 EST 2005


Dear list members,
I have a question regarding test equating. I’d be appreciated if you 
replied.
I have two versions of a test which are supposed to measure one construct. 
One version is more difficult than the other one. Each test has 20 items 
with 5 items common to both versions.
In order to be able to compare people who take different versions of the 
test,
Is it enough to compute the logit scores for each person and directly 
compare them?
I mean do the common items take care of everything? If not, then what other 
steps should be taken in advance to be able to compare the ability of people 
who take different versions of the test?
Cheers
Rama

_________________________________________________________________
A Camera Sells every 2 Minutes on eBay. Get Your's at a Great Price Now! 
http://adfarm.mediaplex.com/ad/ck/4686-26272-10936-699?ck=Cameras




More information about the Rasch mailing list