I just gave my students a hard exam. I have
always believed that students should be indifferent between hard and easy exams.
After all, no matter what a professor says (and "the curve" is the biggest fraud
out there), grades ultimately depend on a student's relative standing in the
class. So what does it matter if the average score is 80% and you're in the 60th
percentile, or the average is 40% and you're in the 60th percentile? One might
even argue that hard exams are better: you leave more questions unanswered and
at least everyone gets to leave without writer's cramp. Turns out I'm an idiot. In a new paper* Luis
Santos-Pinto and Joel Sobel have developed a neat explanation for why rational
students prefer easy tests. Imagine you need two skills to perform well in one
of my exams. For the sake of generality, let us call these skills mathematical
wanking (W) and pointless regurgitation (R). A student can spend time studying
to improve her ability in each skill, but she has limited time. So what should
she concentrate on? The answer is obvious. If she thinks increasing her ability
in W is more effective in raising her grade than effort spent mastering R, she
will concentrate on W.
Not every student makes the same choice. Some
students think W is most important, while others think R is most important. So
some students come to the exam well prepared at W, others well prepared at R.
Students skilled in W will look down on those skilled in R, because they believe
W is the most important skill. Similarly, students skilled in R will look down
on their colleagues skilled in W. So why the preference for easy tests? One way to
capture the idea that a test is easy to imagine that for easy exams a student
can increase her effective ability more with any given effort. So, when
tests are easy, a W student will perceive a bigger gap between her own
near-genius skills in W and the skills of R students, and vice versa. Imagining
they are that much better than the others, every student believes he or she will
do relatively better, and so get a higher grade. Santos-Pinto and Sobel can cite plenty of
evidence for their theory. In experiments, people ranked their ability relative
to others much higher for easy tasks (e.g. manipulating a computer mouse) than
for difficult tasks(e.g. juggling). People are also more likely to choose a
payment based on their relative performance in a test instead of a toss of a
coin when the test is easy. Evidence tells us that students prefer easy
tests, and now theory tells us why. So, should I make my exams easier? Of course
not. Giving hard exams is an important characteristic of above-average
professors. That's why I give them. ____________________________
* Santos-Pinto, Luis, and Joel Sobel (2005): "A
Model of Positive Self-Image in Subjective Assessments." American Economic
Review, 95(5):1386-1402.
8:39:05 AM 
|
|
