[Rasch] sufficient statistic
rmt at rasch.org
Mon Jul 16 09:12:58 EST 2012
We can derive the Rasch model from the axiom that the raw scores are the
sufficient statistics for person ability and item difficulty.
This also demonstrates that most of Classical Test Theory, which acts as
though the raw scores are sufficient statistics for decision-making, is
based on the implicit assumption that the Rasch model holds for their data !!
rmt at rasch.org www.rasch.org/rmt/ Latest RMT 26:1 Summer 2012
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the Rasch