[Rasch] item position effect
Stephanou at acer.edu.au
Fri Dec 10 06:29:53 EST 2010
Sent to the list on behalf of Shen, Linjun [lshen at nbome.org]
We did some item position analysis for our Comprehensive Osteopathic Medical Licensing Examinations (COMLEX). COMLEX is computer delivered. It has eight 50-item blocks. Items are predetermined and are the same for each candidate. The order of the blocks is fixed but the order of the items within a block is random. In our study, a random effect Rasch model was used assuming item difficulty was random rather than fixed across different locations within a block. In addition, the same random effect methodology was applied to response-time analysis, in order to provide additional information about whether and how examinees respond to different orders of items differently. The analysis found that the location effects on item difficulty, person ability, and speed estimates were minimal. A paper about this study was presented to the AERA this year. Let me know if you need more information.
Shen, Linjun [lshen at nbome.org]
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the Rasch