NetNews Usenet Archive 1992 #16/NN_1992_16.iso/spool/misc/educatio/2992

Internet Message Format | 1992-07-22 | 3.5 KB

Path: sparky!uunet!stanford.edu!agate!garnet.berkeley.edu!ewanders
From: [e-mail] ()
Newsgroups: misc.education
Subject: Performance Clumping (was Re: If you have to
Date: 23 Jul 1992 00:47:44 GMT
Organization: University of California, Berkeley
Lines: 54
Message-ID: <[e-mail]>
References: <[e-mail]> <[e-mail]> <[e-mail]>
NNTP-Posting-Host: garnet.berkeley.edu

In article <[e-mail]> [e-mail] (Alan McConnell) writes:
>
>But I wish to raise a question arising from Prof Rubin's post.  Why is it
>that one class in, say, Trigonometry(or American History), will be a "good"
>class, with many A-deserving students, while the next term the class(also
>containing 35 students, say) will be miserable, with many flunking?  Does
>not the Law of Large Numbers mandate pretty much an equal distribution of
>talents in the two classes? (this will, I hope, be a Red Flag to certain
>posters in this topic ;-) ).
>Alan McConnell
>--
>Alan McConnell             "What wondrous life is this I lead!
>Pixel Analysis              Ripe apples drop about my head . ."

I have seen the same trend in several of the classes I have taught.
(all college-level engineering type classes).  While it is reasonable
to assume that students enter with a random distribution of talent and
prior preparation, it is clear that CLASS performance varies by more
that one would expect based on the variation in individual student
performance.  The obvious explanation is that performance of students
within a class is not distributed in a Gaussian (or whatever) fashion
because students are not "independant" variables.

The main factor at work is that students get an idea of what constitutes
acceptable performance by watching each other.  When they come to hand
in an assignment, if one of a group of students has spent 4 hours on
a problem set but finds his buddies spent only 2 hours, he is likely
to decide he is taking it too seriously and slack off.  On the other
hand, if he finds out that they spent 8 hours and wrote a computer
program to solve the last problem he is likely to try harder next week.
Grading on the curve only encourages this sort of thing, because there
is no incentive to try any harder if you are already at the top of
your group.

Another factor in engineering-type classes is that the "good" classes
tend to work cooperatively to a much greater extent than the others.
I'm not talking about dividing up the work to get the most problems
solved with the least effort.  I'm talking about study groups in which
each member of the group makes an extended effort to solve all the
problems alone, but then the group gets together to compare solutions,
study them as alternative problem-solving strategies, and then
brainstorm about the problems none of them could get.  The great groups
then go on to the library, pick the instructor's brains, and actually
go out of their way to learn something while solving the problems.

In weak classes, given a challenging problem set they quickly realize
that no one got problem five, so there is no need to really try.  If
it matters the instructor will go over it in class.  Sigh.

--
Eric W. Anderson      |  [e-mail]  | (510)-486-4986
Chem. Eng. Dept.      |-------------------------------------------------------
Univ. of California   |  I don't speak for the University, the city or the
Berkeley, CA 94720    |  people of Berkeley, and they don't speak for me.