Path: utzoo!utgpu!attcan!uunet!lll-winken!lll-ncis!helios.ee.lbl.gov!pasteur!ucbvax!husc6!cs.utexas.edu!ut-emx!nather
From: nather@ut-emx.UUCP (Ed Nather)
Newsgroups: comp.edu
Subject: Re: Student and Course Integrity
Keywords: losers, responsibility, 90%, crap, 98%
Message-ID: <9286@ut-emx.UUCP>
Date: 7 Jan 89 07:12:33 GMT
References: <4550@homxc.UUCP> <4847@phoenix.Princeton.EDU> <542@mccc.UUCP> <548@mccc.UUCP>
Distribution: na
Organization: The University of Texas at Austin, Austin, Texas
Lines: 62

In article <548@mccc.UUCP>, pjh@mccc.UUCP (Pete Holsberg) writes:
> 
> =I won't get started on multiple-choice testing ...
> 
> It's possible to write good MC tests.  The benefit is that they are easy
> to grade.  But it's probably harder to make up good ones.  When I have a
> good one, I collect it from the students after returning it for
> discussion.  That way, I can use it again in the future.

OK, you've got me started.

In my view, there are NO good multiple-choice tests -- the very format
precludes its use as a sensitive probe of a student's understanding.
When I was in college I found I could usually pass any MC test, even
if I knew nothing of the subject. (We bet bottles of beer in those
days, and I won many more than I lost.)  The procedure is to read
each response and see which is the most likely, semantically, to fit
with the question or statement it is supposed to complete.  In more
than half the cases the semantics alone show the correct -- i.e. the
"original" completion first composed by the test originator -- with
the other possible responses showing their "tacked-on" quality.

But this is beside the point, really.  The ability the phrase an
answer properly is quite different from the ability to recognize one
that is properly phrased (or contains the appropriate "buzz-word.")
Much muddled thinking can be hidden behind a properly-checked box
on a test, which would show up right away if the student had to
choose the words to answer the question.  Eddington said, "If you
can't explain what you're doing in five minutes to a bar-maid, you
don't know what you're doing."  Presumably bar-maids had more time
to listen in Eddington's time than they have now, but the point is
valid, and applies very strongly to communication of understanding.
The appropriate choice of appropriate words mirrors the internal
thought processes -- "understanding" if you will -- information
which it totally lost by putting an "X" in the appropriate box.

Multiple-choice tests can have a harmful effect as well.  On the
first test of a recent semester I graded a test in which the student
had written "shawod" for "shadow."  I asked her if she was familiar
with the term "dyslexia"; she was not.  After some investigation she
was diagnosed as a classical dyslexic by a local physician. She was
a junior in college and, according to her own testimony, had never
before taken an essay exam -- she got all the way through high
school and half of college before she encountered one.  She had only
been tested by multiple choice exams.

She was much too old for the dyslexic training at our university -- 
they accept no one older than 12 on the assumption all children will
have their disability diagnosed before that age -- but managed to 
get help and did much better as a result. I gave her oral exams for 
the rest of my couse, which she passed almost effortlessly.  She was 
very capable compared with most students.  She had to be, I guess,
to learn despite her disability.

On graduation, she stopped by my office to say goodbye, and I asked
her if learning of her disability had discouraged her.  With a huge
grin, she said "Oh, no.  It made me feel ever so much better.  You
see, I thought I was just dumb."

-- 
Ed Nather
Astronomy Dept, U of Texas @ Austin