As we head into a new academic term, I thought I’d do a little reflecting on some reading I’ve been doing lately (nothing like the end of term office clean-up to uncover a few gems saved up over the term!). One article review that struck me from The Teaching Professor was about how many tests are enough to help maximize student learning. When I consult on course designs, this question often arises. So what does the literature have to say?
It appears that more is not necessarily better. Based primarily on an extensive meta-analysis, Kuo and Simon (2009) reported that while students may tend to evaluate us more positively when we give more tests, their actual learning outcomes (as assessed by a final exam) are typically not any better when they are given multiple tests or quizzes versus being tested only 2 or 3 times in a term before the final. The one exception to this appeared to be when there is item overlap (re-testing the same concepts) between tests. This finding makes sense in the case where students acknowledge they have misunderstood or failed to grasp a concept and then attempt to learn or re-learn it in time for the next test or the final exam.
This situation of having students return to a concept reinforces another point made in this article about the value of giving some type of feedback after a test. The authors contend that “testing can serve an instructional function only if students successfully achieve the proficiency level for each section of material through posttest feedback and additional instruction as needed” (p.159). While time pressures may result in us leaving a test largely unexamined once it is marked, this approach only serves an assessment function, not an instructional one. This article gives some interesting food for thought when considering issues such as your philosophy of testing or whether to strive for depth versus breadth in a course design.
If you want to read more about test frequency, you can find Kuo and Simon’s article in College Teaching, 57 (3), 156-160, in CTE’s library in MC.
______________________________________
The Centre for Teaching Excellence welcomes contributions to its blog. If you are a faculty member, staff member, or student at the University of Waterloo (or beyond!) and would like contribute a posting about some aspect of teaching or learning, please contact Mark Morton or Trevor Holmes.
Thanks for the response, Jeffrey. Your idea about seeing the impact of peer evaluated tests (or in-class clicker questions) is an interesting one. That would be a definite time-saver for marking but would still take time in class that would, in my opinion, be very worthwhile to foster students’ learning.
The assignment question is a good one too, although the results may come out to be quite similar to the test frequency article. One result that I didn’t include was that while students often got higher grades on each unit test when there were more frequent tests (the assumption was there was simply less to study), they did not do noticeably better on the final exam.
The bottom line is that this issue relates very well to one of Chickering & Gamson’s 7 Principles of Effective Undergraduate Education: time on task is absolutely critical. The more you work with a concept or idea, the better you can come to understand it.
An intriguing result that I think most professors already do or would be willing to accept, especially given their increasing work load.
It would be nice to see the results of a similar study regarding peer evaluated tests (which would decidedly not count toward the final course grade). I expect that such exercises would foster discussion and (hopefully) constructive feedback.
I would also be very interested in learning what the literature indicates about the most effective number of out-of-class evaluations, such as assignments. I would assume that results of such a study would vary more widely by faculty, or even by program of study.