Re: Justification for the SAT

[Stephen McGinnis <Nanodev@AOL.COM>]

The fundamental problem is the inherent bias in the assessment reform
community.  There is an unquestioned belief that somehow teachers assign
grades in a perfect manner that never contains bias.  The fact that there are
multiple assessments by many individuals is irrelevant.  Since the teaching
corps is very homogeneous and grades are based on almost identical subjective
factors (homework, neatness, behavior,etc.) the inclusion of multiple
evaluations does not improve accuracy.  This argument is in fact identical to
saying that if we gave the SAT and the ACT to every student, every month, it
would somehow become an accurate assessment.

I find it interesting that the PSAT was mentioned.  There has been a
substantial effort led by FairTest to eliminate apparent bias in this
objective test.  (This was of course done by adding a subjective component.)
However, most scholarships and financial aid are awarded based on teacher
assigned grades. This bias has never been challenged.

The latest Department of Education data shows that the college attendance
gender gap is 56%-44% and widening.   As a result, it is particularly
appropriate to begin to examine the gender diversity of the student body (as
many institutions are currently doing) and to examine the use of biased
assessments in the admission process.  (In fact there have been several
reports of institutions adding to male applicants GPA and/or subtracting from
female students GPA to counter the effects of grade bias.)  However, the
optimum solution would be to address the gender bias inherent in the grading
process.

Of course such an analysis will not happen.  The belief in the absolute
accuracy of classroom grades reminds me of something from another era.  Are
we back to the dogma that the earth is flat?  There seems to be an equal
amount of quantitative evidence to support both theories.

Sincerely,

Steve McGinnis

--------------------------------------------------------------------------
To unsubscribe from the ARN-L list, send command SIGNOFF ARN-L
to LISTSERV@LISTS.CUA.EDU.