The upcoming October issue of The Teaching Professor contains a thoughtful essay on the value of “modeling stupidity.” I know, it doesn’t sound like something any self-respecting professor should do, certainly not regularly. What the author advocates is being able to honestly admit to students when you don’t know something and regularly acknowledging how much there is to know and how little of it even highly educated professors know.
It takes some courage to admit when you don’t know or haven’t thought of something you should have considered, as I’m finding right now. This weekend I read an article pragmatically titled, “How Many Tests Do We Really Need?” Like most faculty, I opted for more, rather than less. However, as I read the article I realized I’d never really thought about test frequency, never even considered that the question might well have been addressed empirically.
A 1991 meta-analysis discussed in the article compared the results of 40 studies on test frequency published between 1929 and 1989. I’m going to be writing more about the findings in an upcoming issue, but in a nutshell those studies that document a “moderate benefit” on student learning (as measured by the final) favor more frequent than less frequent testing. However, the majority of studies found that test frequency had no effect or a nonsignificant effect on student learning (as measured by the final). I had no idea.
And the research has looked at a whole range of questions beyond the frequency question. So, if two tests improve performance on the final, do four tests improve it even more? How does performance on the midcourse exams and quizzes correlate with performance on the final? What happens when the midcourse exams and the final use different question formats? Do other kinds of feedback (in addition to the score) and debrief activities improve subsequent exam scores?
I am regularly surprised by how many aspects of our instructional practice are based on tradition and what everybody else is doing. I am regularly surprised at how many aspects of instructional practice have been studied and how often we have no idea that our practice could be evidence based. And I am regularly surprised by how much there is about teaching and learning I still don’t know and haven’t even considered.
If you can’t wait for a summary of what this article contains, here’s the reference: Kuo, T., and Simon, A. (2009). How many tests do we really need? College Teaching, 57 (3), 156-160.