[Rasch] Two Groups, Pre & Post

Nianbo Dong dong.60 at gmail.com
Fri May 5 11:14:36 EST 2006


Hi Greg,

I have a question for you.
To my knowledge, we can use Graded Response Model (GRM) or 
Partial Credit Model (PCM) to calibrate ordinal scaled data.
GRM is an indirect IRT model, which includeds 2-step process.
PCM is generaized from 1PL, i.e. Rasch model (however, PARSCALE uses 2PL).
To decide which model we should use, we need to compare these models
based on -2 log likelihood and item fit statistics to see which model fit
the data better. My question is what the reasons are to "apply Rasch scoring 
for the obvious reasons". I'm sorry that I don't know what "Rasch scoring" 
means here.

Thank you.

Nianbo Dong

First year Ph.D. student
Policy Research, Evaluation, and Measurement (PREM)
Graduate School of Education
University of Pennsylvania
 
  
= = = = = = = = = = = = = = = = = = = = = = = = = =

Original sender's name:    Petroski, Greg 
Original sender's address: PetroskiG at health.missouri.edu

> 
>I will have data from a typical design -- two groups (standard treatment
>and an enhanced treatment) observed at two time points (admission and
>discharge from treatment).  The primary research question concerns group
>differences, pre to post, but within group changes will be of interest
>too.  
>
>The outcome is an ordinal scaled instrument that has been calibrated
>with the Partial Credit in other studies.  I would like to apply Rasch
>scoring for the obvious reasons and use each subjects' Rasch logits in
>place the ordinal scores for subsequent analysis.   
>
>Assume that the two groups do not differ at admission with respect to
>the outcome of interest.  I can see at least two ways to apply Rasch
>scoring and would like the advice of those with more experience with
>Rasch methods.  
>
>1) Pool all the data (both groups, pre and post) and do a single
>calibration.  Each subject is used twice in this scheme (pre and post)
>but as we are only scoring the tool in this step and not testing any
>hypotheses I do not see that the lack of independence is an issue.  
>
>2) Pool both groups at baseline, calibrate.  Also calibrate both groups'
>discharge scores separately and use equating methods to place pre and
>post Rasch scores on a common scale prior to analysis.    It seems like
>the equating step introduces some error that we do not have in #1 above.
>
>Comments on the merits of 1 vs 2, or other approaches will be
>appreciated.  Any citations that use Rasch methods for this type of
>design would also be of interest.
>
>
>Greg Petroski, Ph.D.
>Health Management and Informatics &
>Office of Medical Research/Biostatistics
>137 Hadley Hall (DC 018)
>University of Missouri - Columbia,
>Columbia, Mo.  65212
>ph:  (573)882-3373
>fax: (573)882-7909
>
>   
>_______________________________________________
>Rasch mailing list
>Rasch at acer.edu.au
>http://mailinglist.acer.edu.au/mailman/listinfo/rasch

= = = = = = = = = = = = = = = = = = = = = = = = = =




More information about the Rasch mailing list