Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!csd4.milw.wisc.edu!bionet!agate!bizet.Berkeley.EDU!matloff
From: matloff@bizet.Berkeley.EDU (Norman Matloff)
Newsgroups: comp.edu
Subject: Re: Exam Files
Message-ID: <19873@agate.BERKELEY.EDU>
Date: 3 Feb 89 04:14:09 GMT
References: <1461@trantor.harris-atd.com> <19554@agate.BERKELEY.EDU> <27541@bu-cs.BU.EDU> <15993@joyce.istc.sri.com> <429@laic.UUCP> <19810@agate.BERKELEY.EDU> <9388@pasteur.Berkeley.EDU> <19863@agate.BERKELEY.EDU> <9427@pasteur.Berkeley.EDU>
Sender: usenet@agate.BERKELEY.EDU
Reply-To: matloff@iris.ucdavis.edu (Norm Matloff)
Distribution: usa
Organization: EECS, UC Davis
Lines: 54

In article <9427@pasteur.Berkeley.EDU> timlee@ernie.Berkeley.EDU.UUCP (Timothy J. Lee) writes:
>In article <19863@agate.BERKELEY.EDU> matloff@iris.ucdavis.edu (Norm Matloff) writes:
>|*>on if one is willing to put in enough time.  For exams, there is what the
>|*>Taiwan students call the "archaeology method," which is basically the one
>|*>used by American fraternity houses  --  extensive archives of past exams
>|*>(it's really quite amazing how many professors will give essentially the
>|*>same exam year after year, at least in terms of general content).  For

>|I seem to recall seeing exam files in the UCB library.  Was that motivated
>|by a desire to equalize opportunity?

>I don't know about the Moffitt Undergraduate Library file's reasons for
>existing.

This would appear to be the reason; interestingly, this is one of the
reasons cited by Robert Drabek for making his exams available at Arizona.

>|>Perhaps the visibility of some such files ensures that instructors
>|>don't give the same question (or variation with different numbers) twice.
 
>|In my observation, most professors who give similar exams year after year
>|don't even realize that exam files exist and are used.

>Eta Kappa Nu, the Electrical Engineering honor society, maintains an
>exam file of EE and CS exams.  This particular file is highly visible:

Right.  I meant the files in fraternities and "archaeology societies."
In the case of a highly visible file, that would be different, though
I'll bet that there are still some professors who give fairly similar
exams each year, i.e. you can count of their exams having problems
of certain types.

>in it, and one instructor who _did_ recycle problems came to have his
>course's exams removed from the file (his request was honored).

Of course his request has to be honored, but it is highly inequitable,
in my opinion, because it means that some students will get good grades
based on the quality of their "network" rather than based on their insight
into the subject matter.

>It may be of interest to know that the Computer Science division
>encourages students to purchase collections of old preliminary
>exams when they are studying for the preliminary exams.  Certainly
>those who write these exams know better than to clone problems.

This is entirely different.  As far as I know, most graduate programs
in most fields encourage this, so of course the faculty are highly
aware of it, as opposed to the situation I was described.  For the
same reason, the grad students don't EXPECT to see the old problems
resurface (although once in a while they might); they are not using
their access to the old exams as a substitute for insight into the
subject matter.

   Norm