Re: No Mas

["Sherman Dorn" <sdorn@tempest.coedu.usf.edu>]

Art Burke writes:
 
> I have no idea what challenge you think you are posing or what you 
> are getting at with your distinction between "gradations" and 
> "finite classifications." You talk as if you think there are some 
> kind of Platonic essences out there that tests aren't getting at. 

You have it reversed -- those who claim that there is such a thing as
clearly delineated proficiency have the obligation to demonstrate the
technical qualities and common sense of such cut scores. I don't
particularly care whether someone is making some Platonic claim about an
abstract Proficiency or is just claiming that there is some common-sense
standard we can agree on. Since Glass's 1978 article, no one has come up
with a solid technical counterargument to the point that any proficiency
threshold is arbitrary. 

That doesn't mean that cut-scores can't *ever* be justified -- as I've
written before, there's a good case for emergency-room triage procedures.
But those are life-and-death circumstances, and even there, the reliability
can be awful. In short, the burden's not on me to demonstrate the sense of
using a cut-score method as part of an accountability mechanism, and my
challenge still stands.

Sherman Dorn
University of South Florida
http://www.coedu.usf.edu/~dorn/
http://www.shermandorn.com