The average GPA at UNC has increased from 2.99 to 3.21 over the 13 years between 1995 and 2008 (that’s on a 4.0 scale: UNC does not offer the grade of A+). Varieties of “A” are now the most common grade. As and Bs together made up 82 percent of grades in 2008. This last fact tells you a lot about grading as well: everyone gets As and Bs. My bet is that the distribution of GPAs at UNC is skewed left: a few students are responsible for holding down the average GPA by accumulating the lion’s share of Cs and Ds. Anyone who bothers to do the assignments at all can pretty much expect a B- for their efforts – at the very least.
What is UNC doing about this?
It has basically given in to some of the issues identified in the comments section in my earlier post: grade deflation might not be recognized by employers or graduate schools, leaving deserving students out in the cold, or gifted students might prefer to attend an easier university than Stanford in order to achieve inflated grades. That means that targeting a lower GPA is out: UNC faculty rejected a suggestion to aim for an average UNC GPA of 2.6-2.7 (a B-). Instead, the plan, as it is currently envisioned, is the same as the one that I suggested in my original post (and which I’m sure has been suggested many times before that – as evidenced by the fact that Cornell has had this policy in one form or another since 1996): to add the median grade of each class to the transcript, along with an explanatory note about the system.
This system would have several important characteristics. First, using the median would provide a more honest accounting of grading practices. As I mentioned, I think that the current grade distribution is very left skewed. Using the median would mean that classes with 6 As, 6 Bs, 1 C, and 2Ds couldn’t claim a B- average: it would have to be honest that the median student has at least a B (basically, any student who did the work according to directions got a B). Second, this system would enable professors to more effectively use the full range of grades, since students could achieve relative to their classmates rather than relative to the idea that they “deserve” a B+. The other effect is dangerous: if students feel that they need high grades, either for self-esteem or because they don’t believe that outside employers will understand “relative” GPAs, then they might use this reporting of median grades to avoid “hard” professors or classes. I’m not as concerned about this phenomenon for two reasons. First, it’s already happening. If you think that Courserank grade reporting doesn’t discourage the marginal student from taking a class with a lower average, then you should probably think again. Second, if the class medians were just contained on a transcript and not published online, then the information might be slow to disseminate, reducing the impact of this issue (of course, it’s only a matter of time until students have posted all of this information online, so the university might want to preempt this development).
Stanford should follow this lead and also adopt “truth in grading.” The expanded transcript is a low-cost place for this movement to start.