[Rasch] Concurrent calibration in Rasch and person parameter estimation

Mike Linacre mike at winsteps.com
Tue Dec 31 10:50:51 AEDT 2019


Hi Grace,

thank you for your questions.

You asked:
 >     I have two datasets on the same group. The sample size is 79, the 
number of items was 40 and 55 respectively.  I have 3 questions.
 >     Q1: how do we do concurrent calibration in Rasch?

Reply: This depends on what you want to discover. Simplest is to "rack" 
the two datasets, so that there are 79 persons and 95 items. Put a code 
in each item label so that you can do a Differential Person Function 
(DPF) analysis of person-by-dataset. (or transpose the data and do a DIF 
analysis).

 >     Q2: what are some other ways to estimate person parameters apart 
from the use of item responses?

Reply: Do you have other information about the persons, such as School 
Grade, Age, Health, Occupation, Income? These may give a rough estimate 
of person ability.

 >     Q3: Can TAM (the R package) do Q1 and Q2? If not, what programs 
can do?

Reply: TAM can do Q1, not Q2.

OK, Grace?

Mike L.


More information about the Rasch mailing list