[Rasch] Standardizing measures and standard errors

Mike Linacre mike at winsteps.com
Fri Mar 11 10:13:09 AEDT 2016


Yes, Stuart.

You would divide the person and item measures and also their standard 
errors (model or misfit-inflated) all by the same number.

Mike L.

On 3/11/2016 2:43 AM, Stuart Luppescu wrote:
> Fellow Raschies, I want to standardize Rasch measures and their
> standard errors. I'm standardizing the measures in the conventional
> way, and putting the standard errors on the standardize scale by
> dividing them by the standard deviation of the measures. The question
> is, is this method valid regardless of whether I use the model standard
> errors or the fit-inflated (real) standard errors, or do I need to
> revise the scaling factor to account for the changes in the standard
> errors due to the fit inflation? My supposition is that I use the same
> scaling factor regardless of the type of standard error. Amiright?



More information about the Rasch mailing list