We're back from spring break today. On Woodberry's strange trimester schedule, we took a set of exams in the first week of March, and compiled trimester grades right before we left.
In our Physics C - mechanics seminar, I'm responsible for testing, while my colleagues are responsible for daily problems and discussions. I showed everyone the exams, which were quite strong. I gave some suggestions for where to pilot the course in the next few weeks, so that the hard stuff (rotation) would be covered thoroughly, and so the easy stuff (harmonic motion, linear momentum) would be glossed over for the sake of time.
One colleague asked the reasonable question, "Shouldn't we go over the exam?" To which my answer was probably a too-strong "NO NO NO!"
Any major exam comes at the end of a significant, pressure-packed period of preparation, causing a "thank the almighty Bob" catharsis when it's over. Our March exam period is followed by a 2.5 week break. All continuity has been lost -- I doubt that anyone in my honors class could even recall a single question on the exam right now.
I don't generally "go over" even regular-season classroom tests, opting to assign corrections instead. But on the day after a test, I will frequently mention briefly one or two major issues that caused trouble. How do I know whether to mention something? The day after an in-class test, I can feel a buzz through the class when there was an interesting problem. They want to know whether they got it right -- if they got it wrong, they just want to know why their approach was incorrect.* If I hear that buzz, I throw out a point of discussion.
*Note that on the day after, students don't care about the solution to the problem. All they care about is whether they did or could get credit.
What's to gain, then, by "going over" the exam? There's no buzz about any of the long-forgotten questions. No one really remembers how they approached a problem; to discuss the exam effectively, a student would have to re-solve each problem from scratch.
Instead of an in-class discussion, I merely give a test corrections packet. We take two class periods plus a no-other-homework night for students to correct every answer they got wrong initially. The corrections count as the first test grade of this trimester.
By not "going over" the exam, I'm getting two tests for the price of one. They took this exam under test conditions. Now, after so long, the questions have that new-car sheen back; all the students know is which parts were marked wrong. So the corrections become effectively a brand new take-home test.* One way or the other, everyone has to approach these problems again very seriously, gaining the benefit of test taking practice one more time.
* Sure, they're allowed to collaborate -- call it a group test. I've asked enough additional questions about each problem that just copying a friend's answer would do little good.
Over the years I've observed classes in which a teacher discussed a long-ago exam, point by point. Watch such a class some time. Look at the students' body language. See who is truly engaged with what the teacher is saying, and who's staring off into space. When the teacher shows the question about the relationship between weight and mass, are the two students who missed that question (again) paying attention? Or are those two students looking at their laps, while the rest of the class sits bored because that question was so obvious? If you gave the same exam the day after going over it, would the class truly do better?
And most importantly, why rehash an old exam when you could be discussing new physics? Move on, get the class excited about something, get them solving new problems... then make them correct their mistakes on their own time.