Beyond exam basics
Posted by Melissa on December 20, 2010
It’s that time of year when most academics on the semester system have just finished giving final exams. What I didn’t realize until I became a faculty member was how challenging writing (good) exams could be, as Doug Natelson noted on his blog a couple of weeks ago. When I first started at Carleton, I think I was much more nervous than my students on exam day… was the exam too hard, too easy, too long, too trivial? Did my exam effectively assess student learning? With experience, I’ve gotten better at gauging the appropriateness of an exam, although I still occasionally miss the mark.
Now that I’ve gotten past the worst of the exam-writing butterflies, I’ve been considering aspects of exams that go beyond the nuts and bolts. In particular, a PHYSLRNR list-serve discussion in October in response to this Boston Globe article coincided with the arrival of a Tomorrow’s Professor post on creating exams that aim to improve metacognitive skills. (Stephanie Chasteen had a timely blog post in response to the PHYSLRNR discussion.) Together the list-serve discussion and the TP’s post provided some interesting food for thought on the topic of exams, serving as a reminder that exams aren’t just summative assessment tools but an integral part of the learning process.
All this brought to mind research reported in the July 2010 issue of the American Journal of Physics, “Do advanced physics students learn from their mistakes without explicit intervention?” by Andrew Mason and Chandralekha Singh, showing that students failed to use exams as learning experiences. Mason and Singh studied an upper-level honors quantum mechanics class where identical exam questions appeared on the midterm and final exams, and students were given the midterm exam solutions. While some students’ performances on the exam questions the second time around were as good or better than they had been the first time, indicating the concepts had been learned and retained, nearly as many students performed worse on questions that were repeats from the midterm. In introductory classes, I occasionally ask students to review their exam solutions, compare them to my exam solutions, and turn in a written evaluation of where they went wrong in their approach, but I generally assume that such explicit interventions are not needed for advanced students. However, Mason and Singh’s work made me question that assumption.
Thinking about all this — the testing effect and retention, testing approaches that enhance metacognition, summative versus formative assessment — is enough to cause me to redevelop exam-writing anxiety.