[Rasch] splitting items - DIF - more than 2 groups

Randy and Shelley MacIntosh srmac at fastmail.com
Mon Mar 6 07:30:33 AEDT 2017


Rasch list members:
I have found an item with DIF based on 3 educational attainment groups.
The F from the ANOVA of standardized residuals is large and highly 
significant.
However, a post-hoc test shows only 1 of the groups is different from 
the other two.
Is it appropriate to split this item into 3 pseudo-items or just 2 (one 
for the group that differs and one for the 2 groups are are the same)?

More generally, what is the rule to decide how many pseudo-items to 
create when DIF is identified using more than 2 groups?

Thanks,
Randy MacIntosh

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailinglist.acer.edu.au/pipermail/rasch/attachments/20170305/b4259103/attachment.html>


More information about the Rasch mailing list