Path: utzoo!attcan!uunet!lll-winken!ncis.llnl.gov!helios.ee.lbl.gov!nosc!ucsd!rutgers!att!cuuxb!dlm
From: dlm@cuuxb.ATT.COM (Dennis L. Mumaugh)
Newsgroups: comp.edu
Subject: Multiple Choice Tests (was Student and Course Integrity)
Message-ID: <2394@cuuxb.ATT.COM>
Date: 21 Jan 89 00:15:43 GMT
References: <4550@homxc.UUCP> <4847@phoenix.Princeton.EDU> <542@mccc.UUCP> <548@mccc.UUCP> <9286@ut-emx.UUCP> <407@laic.UUCP>
Reply-To: dlm@cuuxb.UUCP (Dennis L. Mumaugh)
Organization: ATT Data Systems Group, Lisle, Ill.
Lines: 34

In article <407@laic.UUCP> darin@nova.UUCP (Darin Johnson) writes:
>In article <9286@ut-emx.UUCP> nather@ut-emx.UUCP (Ed Nather) writes:
>>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.
>
>Actually, I had a very difficult MC test, that did indeed show student
>understanding.  In order for an answer to count, you had to show your
>work!  Instead of 20 questions filling up a page,  there were maybe 4.
>(I think the reason it was done this way was that students were asking
>for a multiple choice test :-)
>

I had a MC test for Engin 10 (Engineering Measurements) that was
the worst test EVER!

Sample question:  A steel chain calibrated at 70 deg F and 40lb
tension measures 98.010 feet.  The temperature is 83.5 deg F and
the tension is 43.55 lb.

The correct length is:  A. 98.0....

The numbers were all possible answers in which each of the two
correction factors were added and subtracted.  The bear was that
no matter how you worked the problem the correct answer and each
of the possible mistakes were listed as a possible answer.

Other problems had all possible decimal point placements, or
various table interpolation rules.

In essense there was no way to "guess" as any answer you might
provide was there, no way to eliminate wrong answers.
-- 
=Dennis L. Mumaugh
 Lisle, IL       ...!{att,lll-crg}!cuuxb!dlm  OR cuuxb!dlm@arpa.att.com