Re: Exit tests

[Doug Craigen <dcc@ESCAPE.CA>]

Michael Edmiston wrote:
> Arnulfo Castellanos-Moreno says:  "I did not understand. Why do you
> need exit tests?"
> ...
> There is a fairly strong push in the United States to prove to the
> public that the students in colleges and universities get what they
> paid for....
>
> Some people believe one of the best ways to demonstrate this is by the
> use of exit exams.  If there would be a standard exam that all physics
> graduates would take, then if our average physics graduate performs
> near the national average on that exam, we might be able to sleep be
> tter at night knowing that our physics program is okay.

People do some pretty ridiculous things in their attempts to reduce the
quality of education from an institution down to a single number.  Our
annual "Maclean's" survey here in Canada is a good example.  When
Maclean's started I was teaching at Acadia University - generally ranked
in the top 2 or so in its class.  I now teach at the University of
Winnipeg which is in the same class but ranked quite a bit lower.  All I
can say is the rating means squat - they are different Universities with
different mandates, period.

I think that significant individual numbers can be found, but they must
be taken in a broad context of other numbers and there is no single
formula that will reduce them to a universal indicator of quality.  As
Michael points out, the exit exam scores are worthless without knowing
entrance exam scores for the same students.  However, how would anybody
be able to say which is better between a department with entrance/exit
percentiles of 90/95 as compared with another with 50/60 or another with
20/40?  (90/20 or 20/90 are obvious enough, but in most cases we will
probably be looking at comparisons like the first three).  Different
departments would provide fine arguments for why the figure of merit
should be the absolute difference, or the ratio of the diffenence and
the entrance number, or the ratio of the entrance and exit numbers, or
the log of ... plus there would probably be different results depending
on whether the figure of merit was calculated based on average results,
or averaged from individual student's values.

I don't think we should give up - it is very reasonable to ask for the
proof of quality, and it is very self-serving to say that the only thing
that can't be measured is my job performance.  I do think that we need
to keep hammering away at the question "but what do these numbers really
tell us anyways?".

()-()-()-()-()-()-()-()-()-()-()-()-()-()-()-()

Doug Craigen
    http://www.uwinnipeg.ca/~dcraigen