Re: intrarater reliability - anchoring stimuli (Brian Gygi )

Subject: Re: intrarater reliability - anchoring stimuli
From:    Brian Gygi  <bgygi(at)EBIRE.ORG>
Date:    Tue, 1 Apr 2003 16:58:10 -0800

Christel, One way you could do it is gather data on one iteration of all the fragments, and then do an item analysis, which basically predicts performance on the full set from a reduced set. You could eliminate items until you get down to an acceptable test size, with acceptable predictive power for the full set. Most statistical packages (I use Statistica) have an item analysis module. Brian

This message came from the mail archive
maintained by:
DAn Ellis <>
Electrical Engineering Dept., Columbia University