Unintended Consequences of Transparency

In the Spring of 1998, Cornell University made public the typical grades awarded in courses at the school. So, for example, a student could see that in the Spring of 2007 the median, or middle-most, grade for Introduction to Cryptology was a B.

Question: How would this extra information influence the way you pick classes? Would you register for a course that you're interested in regardless of the fact that it could be difficult, or would you pick courses where the median grade was an A in order to help pad your grade point average?

That's what Talia Bar, Vrinda Kadiyali, and Asaf Zussman of Cornell tried to find out.

They documented that web traffic to the site that held the median grade information peaks during registration periods, suggesting that students were indeed using the information to pick courses.

They also found that the percentage of students enrolled in courses with a median grade of A increased by a whopping 50 percent. And this has had an effect on the typical grade awarded at Cornell: the mean grade has increased from 3.1 in 1990 (on a 4.3 point scale -- that extra 0.3 is because Cornell awards A+'s) to 3.3 in 2004.

The usual grade inflation story is that professors are the ones padding grades, but in this case students are very much helping the process. And the researchers found that 60 percent of the rise in the average grade can be attributed to students picking easier classes.

The irony is that Cornell initially planned to include the median grade information on students' transcripts so that prospective employers, graduate schools, and grant-awarding foundations would have better information about a candidate's grades. Seeing that a student received a B in isolation tells you a different story than seeing that the median grade for the course was a C+.

But for reasons unknown, that part of the plan was not put into place -- until Spring 2008, that is. And some students don't seem to like that idea .