[Rasch] Fan: raw scores and Rasch
Mike Linacre (RMT)
rmt at rasch.org
Tue Oct 30 10:53:46 EST 2007
Quote: "There simply is little appreciable difference between raw score
sums and the Rasch model,"
Except when there is an appreciable difference!
For instance, when there is missing data ...
Or when the floor and ceiling effects of raw score sums are producing
misleading inferences. In one study, a raw-score "ceiling effect" was being
misunderstood as "the therapy has stopped being cost-effective". In fact,
the therapy was increasingly cost-effective.
Or when instruments must be equated. Even the great statistician, John
Tukey, misequated samples based on raw-score-percentages, producing
nonsense results: www.rasch.org/rmt/rmt64m.htm
Editor, Rasch Measurement Transactions
rmt at rasch.org www.rasch.org/rmt/ Latest RMT: 21:1 Summer 2007
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the Rasch