Re: Justification for the SAT

[Bob Schaeffer <bobschaeffer@EARTHLINK.NET>]

  Like most pro-testing ideologues, Stephen McGinnis doesn't bother to
get his facts correct.

Stephen McGinnis wrote:
  There is an unquestioned belief that somehow teachers assign
> grades in a perfect manner that never contains bias.

   Where is the evidence for this claim. I know that FairTest has
repeatedly warned that outside checks are needed to ensure that teacher
grading bias doesn't have negative impacts on equity or educational
quality.

  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.

   Another claim without an iota of data. In what ways is it true that
"the teaching corps is very homogeneous" (beyond the fact tht nearly all
have college degrees)? In my school career -- and in my son's -- we
encountered some teachers who were very strict graders and some who were
not, some who relied heavily on multiple-choice tests and some who did
not . .

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.

   More nonsense: in fact, every form of the SAT (and the ACT) is
designed to be closely similar to every other form in content coverage,
format, difficulty level, etc.. Grades can be based on essays, class
participation, projects, performances, portfolios and, yes, even tests.


>
> 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.

  Extra credit here for getting one point right for a change. Yes,
FairTest has campaigned against reliance on the PSAT, just as we have
against the SAT, because of manifest gender bias. I say _manifest_
because the sole scientific claim of these exams is their capacity to
predict first year college grades -- and despite their lower scores on
the SAT (42 points on average in the high school class of 2001), females
earn higher grades in college than their male peers when matched for
identical courses.
  FairTest did not invent the criterion of first year GPA to judge the
fairness and accuracy of the SAT -- that's what the test-makers specify
in their literature. And both the College Board and ACT admit that high
school performance (expressed as either grades or class rank) is a
better predictor of first year college GPA than their tests are.
  The gender bias of the SAT is not just an idle charge: it is something
FairTest and its lawyers have demonstrated in a series of legal cases.
Here's what one federal judge concluded in striking down a New York
State program based solely on SAT scores as unconstitutional, "After a
careful review of the evidence, this Court concludes that SAT scores
capture a student's academic achievement no more than a student's
yearbook photograph captures the full range of her experiences in high
school." The writer was U.S. District Court Judge John H. Walker, a
Ronald Reagan appointee and first cousin of President George Herbert
_Walker_ Bush.

>   (This was of course done by adding a subjective component.)

   This parenthetical phrase reveals the depth of Mr. McGinnis'
ignorance and his willingness to twist the truth to support his
ideology. In fact, the "Writing Skills" section added to the PSAT to
settle FairTest's civil rights complaint (charging the College Board and
Educational Testing Service with illegally assisting gender bias in
manufacturing and administering the exam used as the sole criterion to
select National Merit Scholaship semifinalists) is entirely
_multiple-choice_! If this part of the exam is "subjective" by Mr.
McGinnis definition, so are the Verbal and Math portions of the exam,
which are in precisely the same format (save for a handful of
"open-ended" Math items, which allow only one correct answer).


Bob Schaeffer, Public Education Director
FairTest: National Center for Fair & Open Testing

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