Planning to start your course with a review lecture? Think again

Planning to start your course with a review lecture? Think again

I was recently talking with a new faculty member who was about to start teaching a first year physics course. He mentioned how the course always started with a couple of lectures reviewing material the students should have seen in high school. 

I did this kind of review myself for many years, but have recently collected data indicating that what I was doing was a mistake.  I subsequently found a much better method.  This topic is particularly relevant at the beginning of the term, and I suspect that many faculty make the same mistake as I.   

I always used to start the term with a lecture or two of review, and then, in most classes during the term, I would spend a few minutes giving students a summary of what was covered in the previous class.  When we collected data on this practice, it showed that this form of review was less than useless.  Rather than helping them improve their memory and understanding of the material, it primarily diverted their attention to thinking about things other than the physics class they were in, and this made it harder to get them reengaged when I started to discuss new material.

It is easy to understand why this method of review fails, and in retrospect, I feel stupid that I did not recognize it sooner.  There is a very well established result from cognitive psychology that familiarity with a topic makes people erroneously believe they understand/have-learned it.  There is another result, which I cannot give you references on, but I am sure is in psychology research literature, which is that when a person is being lectured on something they believe they already know, they will become quickly bored and start thinking about other things (or checking email, etc.). (This result can be confirmed at most faculty meetings.) 

The combination of these two effects means that no student who has previously heard about the topic being reviewed is going to pay attention.  And for those few students who have never heard anything about the topic, such a brief review will be largely useless to them as well. 

The better way I found to do review is to replace ALL review lecturing with problems that the students  solve in class that cover the material I want to review.  This is particularly easy to do if they have clickers.  Doing the problem gets them actively thinking about the relevant material and testing their understanding.  If they get the problem wrong, and often even if they don’t, they are then primed to ask questions and listen to responses and explanations to learn why.  Also, if there are things that everyone in the class already knows, I can see that immediately from their problem solutions, and so I can quickly move on and avoid wasting class time talking about that topic.  That leaves more time to spend on the topics where many struggle with the relevant review problem. 

A final benefit is that, at the end of a review carried out in this way, I have a good idea of what topics individual students, and the class as a whole, have and have not mastered, as I move on to the subsequent material.  I have a vastly better sense of their state of mastery than I previously got from review lectures.

How do you know your review is effective?  What will you try next time?

Carl Wieman is a professor in the Department of Physics and Graduate School of Education.


Carl Wieman added these thoughts in an email to Teaching Commons:

In thinking more about how to optimize the review, I came up with an implementation idea that, although I have not tested it, I am pretty sure would work well. It is a variant on the idea of two-stage exams, which are now being used quite a bit at UBC after I introduced them there. We do have good data on student learning and happiness from them, and a couple of papers published on that.

The idea is to use a two-stage (ungraded) test for the review. Give the students a test in class that has the review problems on it, have them do it individually and turn it in, and then have them do a group test (since first day, just require them to have 3 names on the answer sheet). The resulting discussion will provide nearly all the students with the primed and targeted review that they need.

The instructor will only have to worry about dealing with those students whose individual tests indicated they have seriously deficient backgrounds and those topics that listening in on the group discussions and/or a scan of the group test solutions reveal as widespread deficiencies. During the group test portion, the instructor should listen in on the various group conversations. That is likely to reveal any widespread difficulties, that can then be immediately addressed after the completion of the group test. There would also be a variety of more subtle benefits to this exercise having to do with classroom dynamics and discussion and instructor learning about the class.