I haven't followed this whole thread, but has anyone mentioned
the possibility that getting a higher score may not necessarily
be a good goal? The purpose of the test is to determine how
much a person knows about a content area. If the person guesses
and gets it correct, their score incorrectly indicates that
they know more than they do. If "guessing success" is normally
distributed (as is beginning to emerge in Marc's small sample),
the "lucky guessers" will end up with higher scores
than the "unlucky guessers" of similar ability. If these scores
are used for admissions decisions, and the tests are valid predictors
of grad school performance, the "lucky guessers" will be in over their
heads and the "unlucky guessers" will not get into better
programs that match their abilities.
Maybe guessing success isn't distributed normally, but the bottom-
line question is: Why do we want to help students get the highest
score instead of the score that best reflects their ability?
TT
================================================
Thomas A. Timmerman, Ph.D.
Assistant Professor
Psychology Department
Austin Peay State University
Clarksville, TN 37044
931-221-1248
[EMAIL PROTECTED]
================================================
