[Rasch] Person Misfit & Item Fit
Mike Linacre (RMT)
rmt at rasch.org
Tue Aug 7 08:58:48 EST 2007
Yes, your experience with trimming misfitting persons matches mine.
You are removing the noisiest part of the data, so the remaining data must
have a slightly closer-to-Guttman pattern. Consequently we expect to see a
slightly increased measure dispersion in the estimates from the trimmed data.
However, I have yet to see this change of measure dispersion have any
substantive implications except where person measures are adjacent to
At 8/6/2007, Petroski, Greg wrote:
>I am fitting a PCM and have a fairly large sample (5k). As an experiment
>I ran the calibration with all subjects and then again with the 500 worst
>fitting (OUTFIT range from 2 to 9.9) individuals excluded. There was
>change in parameter estimates and item fit, but not huge, not what I
>expected. This is comforting, but has this been the experience of others
>or is it likely just a quirk of my data and the fact that the sample size
>is large to begin with?
Editor, Rasch Measurement Transactions
rmt at rasch.org www.rasch.org/rmt/ Latest RMT: 20:4 Spring 2007
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the Rasch