[Rasch] Estimating Rasch Measures for Extreme Scores
trevor.bond at jcu.edu.au
Sun May 1 13:14:48 EST 2011
Examiners sometimes give conflicting and/or complementary advice.
I would carefully look to see which examiners are saying what and
To look for conflicts complements. The candidate is unlikely to have to
change / defend everything.
Seems that issue 1 neglects two crucial issues:
The size of the SE of person measures in relations to those estimates; they
must be large. How many distinct person strata exist?
And the role of the teachers, nested within schools and the role of schools
Given that the object is some sort of regression, it seems that Facets
software could resolve most of the issues and ConQuest might also resolve
them...and provide the regression as well. N'est-ce pas?
So how does this make operational sense:
These scores need to be appropriately dealt with by Rash (if this is
possible), or they need to be removed from the analysis (with an
assessment made about the impact of the resulting loss of data). ?
Anyhow, while the data collection design might have been improved upon a
priori, it seems that the analyses MUST be improved upon a posteori. Can't
ignore, raters, schools and nesting.
More information about the Rasch