Abstract: This paper reports on the form in which Oxford Finals questions are posed. It criticizes the apparent lack of originality and provides both theoretical and empirical estimates to support this finding. We propose a number of solutions which, if adopted, will substantially improve examination performance and unambiguously raise welfare.
Regardless, that does not stop me from plowing on.
Commentary: I am not as pessimistic as the authors of this tract. E.g. whereas they state "as the level of understanding (sic) in paper i for student j is not observable to the examiner", I would contend that it largely is.
I also demand a more realistic modeling of costs. I would argue on the supply side that creative questions demand more time in creation, but also in correction. Further, students are likely to respond positively to a certain degree of creativity in questions, but at some unknown point student performance is negatively correlated with increasingly creative questions.
We will abstract from potential spillover effects, in that a student exposed to creative questions in principles courses may well be enchanted by subject and continue further studies in economics. Whether this is a positive or negative spillover is unknown, but the presence of such an externality would have obvious effects on optimal behaviour.
This is not to take issue with the author's thrust, as I fully support the call for more creative examinations, but that a more careful modeling of the situation would resolve Dilemma 1 by permitting proof of Lemma 1.
P.S. The best exam question I ever received was in monetary theory, when asked to "describe the effects of Canada-US monetary union", which had never been explicitly addressed during the course. I suspect I spent a highly disproportionate amount of time in answer relative to the weight of the query (12%).