But this morning I wondered, do multiple choice tests accurately reflect student learning?
I give my students access to a test bank as a study tool. Most textbooks now come with electronic files: instructional materials, lecture slides, movies, test generators, animations, etc. I make some of these available to my students, because the main resource they have at their disposal when studying is their textbook. I call the test bank "Chapter Preps" and encourage them to use them to prepare for upcoming lectures. Students who use these score higher on their daily reading quizzes and higher on their exams. It's practice, and practice makes perfect.
This year we changed editions of our biology textbook, and the test bank has changed, but not in the way you would expect. It's still multiple choice, but the number of potential answers has decreased from five (A-E) to four (A-D). This increases the probability of randomly selecting a correct answer by 25% (from 1/5 to 1/4). I've yet to find justification for this by the publisher, and can only assume it is meant to increase student success. Is it appropriate to increase a student's chances of getting it right by decreasing the number of ways they could get it wrong?
When I create a multiple choice test I try to ensure that answers require critical thinking, not just rote memorization. Anybody can memorize a list of names; nucleus, ribosomes, mitochondria - but the most successful students will understand how those organelles function in context. A student may know that the mitochondria is for energy, but a better student will know why the shape of the mitochondria is critical to the production of ATP, a molecule that is essential for proper cell function. Assessing this difference on a multiple choice test is tricky, and I don't always get it right.
When I was an undergraduate I relished the opportunity to show off on an exam. I know, I know, it's obnoxious, but it's true. I hated multiple choice tests. They were boring and too easy and I often pretentiously considered the choices to be substandard to the answer I would have given if I had the freedom to do so. I know, I know, I was an arrogant jerk. If I had to teach College Breana I'd relish the chance to flunk her, mostly because I still am an arrogant jerk.
None of my students have the kind of confidence that I remember from college. I wasn't the only one who looked forward to exams with a nervous excitement - there were others. Many of us have gone on to be professors, and are in no way representative of "normal" college students. We were nerds. NERDS. And proud of it. I don't expect a classroom full of nerds, but I would like it if at least some of my students were interested in more than just the bottom line. They want to know exactly what they have to do to get a passing grade, and they invariably think the standards are set much too high. They like multiple choice, they rely on multiple choice, because they wouldn't be able to craft an answer in their own words.
How do I know? Because I've asked them to, and they can't. Student failure rates increase exponentially when I administer a short answer exam. It's more work for me and worse grades for them. I often look on this as a failure on my part, but I can try teaching the material six ways from Sunday and they still do better on multiple choice exams than on written ones. It's frustrating, and it makes a multiple choice test seem even more inviting. But it is fair? Is it rigorous enough? Is it appropriate for college? I'd love to know how other educators make this decision.