False quantification to the nth degree

mike salovesh (T20MXS1@NIU.BITNET)
Mon, 13 Jun 1994 14:52:00 CDT

Most of the following is part of a personal message I sent to someone
on the list. Now I think about it, it may be of interest to the rest
of you. Please remember that I am a STRONG believer in appropriate
uses of statistics and quantification. What makes me mad is GIGO in
the dumbass belief that if it's expressed in numbers it must mean
something. And here's a horrible example, made even worse because
the guy I was talking to has had a sinister influence in trying to
quantify the personnel judgment game in his department.

A few days ago I talked with a former Dean of Students, a mathema-
tician who now heads our Computer Science Department. I asked him
to review some regulations with me. On a grading scale where an A
counts 4 points, a B counts 3 points, a C counts 2 points, a D
counts 1 point, and an F counts zero, I asked him what is the status
of a student whose academic Grade Point Average (GPA) is 2.000. He
replied that the student would be in good standing. I then asked
about the status of a student with a GPA of 1.999. That student,
he told me, would be put on academic probation, and if his GPA did
not improve in one semester he would be "released" for poor scholar-
ship. Fine, I went on, now what grades am I allowed to give? A or
B or C or D or F, no plusses or minuses, isn't that so? In other
words, 4 or 3 or 2 or 1 or 0, in whole numbers? Yes, he said.

Then I exploded. Look, Larry, I told him, you're a mathematician.
You know damned well that there is no earthly difference between a
GPA of 1.999 and 2.000 in mathematical terms. They go to three
significant figures beyond the evidence. They are absolutely in-
distinguishable. The arithmetic may be impeccable, but the math
stinks. Why don't you tell people that truth?

What truth, he asked. Mathematics has nothing to do with it. Our
rules say that a GPA of 1.999 is not satisfactory; students know the
rules; and if they can't perform according to those very clear rules,
we have to dismiss them. That's all there is to it.

Invincible ignorance. What more can I say?

mike salovesh <t20mxs1@niu.bitnet> OR <t20mxs1@mvs.cso.niu.edu>