[Rasch] Facet parameter separability

Stuart Luppescu slu at ccsr.uchicago.edu
Fri Jun 17 08:15:21 EST 2011

Fellow Raschies, I'm doing an analysis of teacher quality based on
classroom observations using Facets. One of the measurement facets is
the grade level of the classroom. An interesting, unexpected result of
the analysis is that among the grades, 4th grade has the highest
difficulty by far. One of my colleagues has suggested that since 4th
grade if furthest away from one of the main accountability grades (3, 6,
8) the principals are likely to put their worst teachers in that grade.

I thought that because of the separability of the estimation of the
parameters, the distribution of person measures should not effect the
item (or whatever facet) difficulties. And, in fact, the teacher
measures for the 4th grade teachers are much *higher* than for the other

I would appreciate it if someone could explain how this could be.
Stuart Luppescu -=- slu .at. ccsr.uchicago.edu        
University of Chicago -=- CCSR 
才文と智奈美の父 -=-    Kernel 2.6.38-gentoo-r6                
The only people who should use the assign function 
 are those who fully understand why you should
 never use the assign function.    -- Greg Snow    
 R-help (July 2009)  

-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/x-pkcs7-signature
Size: 3573 bytes
Desc: not available
Url : https://mailinglist.acer.edu.au/pipermail/rasch/attachments/20110616/986cf074/attachment.bin 

More information about the Rasch mailing list