Monday, October 27, 2014

Tensions Running High at UVa.

Multiple-choice is the right choice

Multiple-choice tests are valuable teaching tools for large introductory courses

On October 2nd, my fellow columnist Jared Fogel argued University professors should reduce their use of multiple-choice exams. He criticized the test format for preventing professors from effectively judging students’ understanding of the material, eliminating the possibility of partial credit for students whose work is nearly correct and causing students to feel less obligated to study. While these criticisms are valid for some administrations of multiple-choice tests, I disagree that these are inherent drawbacks of the format that recommend against its use. Multiple-choice exams are a valuable feature of many introductory courses at the University.
While I agree with Fogel on the importance of professors promoting a “deeper understanding of material,” I object to his assumption that multiple-choice tests are never a vehicle to achieve that goal. 


  1. I once heard Jacques Barzun disparage multiple choice. I thought it was easy for him to say, since he never had a class of more than 15 students, all of whom had ACT scores over 30.

    That no partial credit is possible I think is a plus. I hate-hate-HATE grade-grubbing pre-meds, particularly when one grows up to be the physician who removed a cyst from my Mother's nose and left a big hole right on the bridge.

    Where's my STAPLE GUN? (TWITCH! TWITCH!) Does Proffie still have it?

    When I taught at Kennedy Space Center, we were encouraged not to be generous with partial credit. The following happened because a dust cap was mistakenly left on an engine nozzle:

    As always, Wally Schirra's sense of humor came through.

    1. Your comment about pre-med students reminded me of something.

      I've been having some dental work done at the student clinic at my alma mater. I met a number of the students there over the years and they were, nice and pleasant, but, on the whole, not terribly bright. I've often left there asking myself the question: "Just how dumb does one have to be to flunk out of dentistry?"

    2. I left the stapler for you right by the water cooler, Frod.

  2. I figure that the MC format balances (to some extent) the loss of partial credit with undeserved credit for guesses.

    You can use machine grading to get a raw score then go back and quickly review students' written calculations. Give everybody the same amount of partial credit so you avoid getting bogged down in "is this chicken scratch worth 2.4 or 2.7 points?" Also, you can deduct all credit if the students don't show any work (i.e., they copied the MC answer). This all takes extra time but it does help MC be more accurate. Besides, what else does a grad student have to do on the weekend?

  3. One of the problem with multiple-choice exams is that one doesn't see how a student obtains the answer.

    While I was teaching, I not only wanted to see how they got their answers, I needed to. Most of the graduates were going into industry, where they were required and expected to understand what they were doing.

    I wanted to see their method and logic. I looked for what equations they used, what assumptions they made, and whether they were careful in doing their calculations. If, say, they dropped a negative sign, I might not have penalized them severely. If, however, they were completely out in space with their method or assumptions, I was far from generous.

    I once gave a mid-term exam and had a discussion with a student afterwards because of what he wrote for one question. He got the right answer but his thinking was completely woolly. He clearly had no proper understanding of what to do but insisted that he get full credit. I didn't budge and I'm sure he trashed me on the subsequent evaluations.

    A multiple-choice exam wouldn't catch something like that.

    Did that student learn his lesson? I certainly hope so. He came to us after flunking out of university or, as we called it, being sent on a dean's vacation. He returned to finish his degree, went on to get a master's, and is apparently working for a company producing aircraft components. After knowing that, each time I get on board a plane, I have a certain uneasiness--I can't imagine why.....

    1. In theory, not giving him full credit makes sense and you probably did the right thing. However, it may have been more fair to give it to him anyway because when a student comes and explains that s/he actually knew the material but just didn't write it down in the answer, you would probably say "I can't grade what's not there". Or, in this case, the right answer was there.

    2. That's not the point. I didn't care whether he knew, or thought he knew the material, or claimed he knew. He didn't demonstrate to me that he did. Getting the right answer is *not* the objective. Getting the right answer the *right* way is.

      In industry, calculation sheets are often included in project files. Imagine inheriting such documents in which the results may be correct, but the actual calculations were complete rubbish. Information like that is often considered in engineering disciplinary hearings and in any resulting lawsuits.

      Getting the right answer the wrong way is not a sign of competence and, yes, one can be suspended or struck from the professional register because of it.

  4. It keeps amazing me how close my students' MC and "subjective" scores are on any particular test. I take that as a sign of item reliability and a reassurance that I'm grading consistently. With 120 tests to grade at once and no student readers, I'd be lost without Scantron forms. It's enough that I grade about 40% of the test items by hand. The little dears still have to demonstrate certain computational skills and explain things clearly with precise use of vocabulary, and I still get to work in my garden for part of the weekend.

  5. Pardon my jumping into teaching mode, but you can actually ask MCQ all across the Bloom Taxonomy, except for "synthesis" which does require the darlings to actually create something. The link to a wonderful example page from South Africa should be but it's not opening in my browser after correcting a gazillion lab reports. I need another drink....