"Is it Examinable?"

Today I have spent most of my time writing up the solution set to one homework problem and writing up the problem sheets for the final three problem sheets in the brand new statistical mechanics course that I am teaching at Oxford this term.

I am not yet done writing up these problem sheets, so I really ought to continue work on that, but there is a major issue that arises both in lecture material (the actual lectures as well as the notes that go with them) and in coming up with problem sheets. These can be summarized neatly by the question in the title: Is it examinable?

First, let me say that I hate this question. I completely understand why Oxford students---including the best ones---ask this, because a module's end-of-year exam counts for everything in providing official recognition of what a student did or did not learn in a course. This sets up a whole swath of entirely understandable but rather artificial motivations for our students. Simply, they are in many instances judged almost exclusively on their exam results, so students will concentrate on working in order to maximize those results instead of working to maximize their learning of the material. These two things can coincide (and might even coincide a lot in some cases), but there are crucial situations in which they clash against each other---and this is especially true in advanced classes:

(1) Exam problems are supposed to take 45 minutes and are split into parts that are meant to help guide the student. Real problems don't work that way. The guidance can often be helpful for the learning, so that doesn't particularly bother me, and indeed I think it is desirable to add such things to homework problems as well. The 45-minute thing is a huge problem, though, because to learn most material in any reasonable fashion (and especially once one gets to the third and fourth years of the course), one needs to think about much more difficult problems. And then students understandably start asking about what can be tested and what can't be.

(2) As soon as one indicates that something is "not examinable", most students here (again, even the best ones) will typically not spend any more time trying to learn it, because the incentives they have been given by this system tell them that that is the optimal thing for them to do. So I can't blame them.

(3) Points (1) and (2) come into conflict a lot, as one wants to be reasonable (and, indeed, one has to be reasonable) regarding what can be asked on an exam, but there are things that one should learn from an advanced course in a certain subject---and those things are important for research, general education about problem-solving, and possibly even later classes.

(4) In my third-year and fourth-year courses (and beyond), I like to bring in some ideas from current research and to prepare my students for thinking about problems that have not been completely solved. In fact, at this stage of my students' education, I think that it is not just important but genuinely crucial to do this. One way to help do such things is to have some open-ended issues on some homework problems. This will help them not just for the material related to the course, but for other things, such as new applications of such material, modelling issues, and other really important things that many (and perhaps most) of these students need to learn for research, their future jobs, etc. Of course, if a student knows something is not examinable, the incentives at Oxford make it very difficult for such endeavors to work as intended. Sigh...

(5) For advanced courses, I would rather have examination via some kind of project or an open-ended exam (or both) and that would help get around these issues, but it's very difficult to set up such things here.

(6) I'm sure there are other things, and of course many of the things that I have mentioned above are deeply correlated with each other, but a fundamental observation remains: There are so many ways that examination can just get in the way of education---and the latter is obviously my goal when teaching a course!---and the inevitable "Is it examinable?" question that pops up so frequently around Oxford really underscores these conflicts.


OK, now I am going to stop ranting (at least temporarily) and get back to work.