Grade Inflation at My Old Stomping Ground, Part 2

What Works

When I first started working on the problem of grade inflation, I exchanged some emails with Henry Rosovsky, a retired Harvard dean. He agreed that grade inflation was a serious problem, but cautioned that it would be a very difficult one to solve. Larry Summers, past-president at Harvard, agreed with Rosovsky in a two sentence email to me. Both are very, very smart people. At the time, I deferred to their wisdom. But now I disagree.

Peace in the Middle East. Now that's a difficult problem. Nuclear disarmament. Another nearly impossible problem. But grade inflation? Please. There is nothing difficult about solving grade inflation. The principal difficulty with approaching this problem is the perception that it's impossible to solve. Because of that perception, university leaders, who as a type tend to be as craven as they are verbally gifted, don't want to touch the issue.

But perception is not reality. Grade inflation can be solved. It has in fact been solved elsewhere.

First off, I know what doesn't work: soft approaches cannot be expected to be successful. Attached is a chart with data from three schools that have tried to employ soft approaches; Duke grades are given as a reference.

Look at Harvard. Larry Summers jawboned the faculty about grade inflation in the early part of this decade and then was more or less fired (mostly because of other issues). Grades might have gone down slightly as a result of this jawboning, but it's doubtful that the effect would have been long lasting.*

Harvard then replaced jawboning with a town crier approach. Every year there is an announcement by the Dean that grades continue to go up. Apparently, this announcement is supposed to strike fear into every living professor's soul. It doesn't. The town crier approach doesn't work.

Dartmouth and Cornell also have employed soft approaches. Since the mid-1990s both have posted median grades of classes and since that time Dartmouth also has included this information on its student transcripts. Cornell will be putting information about median grades on transcripts beginning this year. This additional data does serve to inform potential employers and graduate schools whether a student is taking classes with strict or easy grading. But it's doubtful that most use this information. Regardless, this approach apparently does nothing to curb grades.

What does work? Real policy to inculcate the concept that the letter "A" truly should be reserved for excellent performance. Three schools have tried it. They've all been successful. Here are the data. I put Wellesley on a separate y axis because its grades were extraordinarily high - above 3.5 - at the turn of the millennium (grade inflation at Wellesley was, unlike the Y2K scare, real). This decade, Wellesley implemented a policy to make B+ the average grade. That policy has worked.

About the same time, Princeton set guidelines to limit A's to 35 percent of a class on average, about the same way their science departments already graded in the early 2000s. That policy has worked as well. Finally, Reed has for 20 years used a "mind the till" approach to grading. It seems to be embedded in their culture now that "A" means excellent.

When I hear people say that the problem of grade inflation is intractable, I think of the three examples above and ask, "Why?" What exactly is so difficult about solving grade inflation at Duke (or anywhere else for that matter)? Grades are too high. You have a culture where in the words of a Duke faculty member, "...for me a decent grade in the class is an A." You have a student body with significant motivation problems that spends a mere 11 hours a week studying. They drink out of boredom. One dies as a result of aspirating vomit. That's not an acceptable intellectual or social culture. Something needs to change.

The recipe for change is readily available. Leadership gets together and works with the faculty to encourage a new culture were "A" doesn't mean "decent" but means something quite a bit more. Grades go down. Hopefully, motivation goes up. Students work harder. Problem solved.

OK, that's way too glib. Yes, there are complexities. But please don't make this problem out to be as difficult as Middle East peace.

It's worth noting that Princeton and Wellesley are not quite there yet with regard to fully establishing a culture that distinguishes between the outstanding and mediocre. Princeton is still working on getting average grades down to the 3.1 to 3.2 range (in the range of Reed's grades). Wellesley may be happy to stick with their 3.3 average.

These schools, like Duke, have many outstanding students. It should be expected that they should receive higher grades on average than those at Podunk-Accepts-Every-Body U. But when over half the grades awarded are A's something is not right.

What is keeping schools like Duke from following the examples set by these three? There is certainly the craven leadership problem (sorry, I couldn't resist bringing that issue up again), but what else? I think that there is another perception that isn't grounded in reality: by lowering grades, somehow students will be significantly disadvantaged with regard to their post-graduate prospects. For example, Duke places about 86% of its students who apply to medical school. Won't Duke pre-meds suffer from a drop in placement rates if grades are lowered?

The answer to that question is that it's likely that the impact of lower grades will at most be small. At Reed College, where grades are as noted above, substantially lower than those at Duke, placement rates for medical school were 74% over the time period 2003-2008. That's a 12% difference in comparison to Duke, but Reed students do not, in general, perform as well on standardized tests as Duke students. My guess is that should Duke lower its average grade down to the 3.2 range, acceptance rates into medical school will be about the same as they are now. In fact at Reed, the acceptance rate for those with 3.1 GPAs or better and MCAT scores of 28 or better is 84%.

The Princeton data on placements reinforces the expectation of at most a small impact. Annual medical school acceptance rates have ranged from 88 to 95 percent since Princeton changed its grading policy.

As for other post-graduate opportunities, Reed ranks well above Duke in the percentage of students who go on to Ph.D. study and their students frequently attend all the top business, law and medical schools. Would they have higher acceptance rates and better quality placements if they had grade inflation? My guess is that they would, but the effect would not be dramatic.

The fact is that schools with great reputations like Princeton, Wellesley and Reed can deflate their grades without significant penalty in the post-graduate placement game. They can be known as the exceptions and if these places effectively communicate that they are exceptions, these schools can have the best of both worlds: rigor and post-graduate success.

Duke has international visibility. Should it decide to deflate its grades, word would quickly get around that at Duke, grade inflation is no longer the norm. The same could be said for all the big name colleges in the universities in the US. The problem of grade inflation is real. The obstacles to solving this problem are far less than imagined.

I've been speculating a lot here about solving the grade inflation problem with an emphasis on Duke. Let's move from speculation to real experience. What would it take? Here's a quote from someone who knows:

"You have to work through your own faculty, and that that is hard work (after all, we worked on this issue for the better part of a decade -- collecting data, analyzing data, transmitting reports to the faculty, inviting ideas from the faculty, engaging in exhortation -- before we proposed and our faculty adopted the new policy). You have to have the running room to take it on (if you have other major priorities in undergraduate education that fully occupy your faculty, for instance -- e.g. new general education requirements or new core curriculum -- you aren't likely to be taking this on at the same time; if a dean is new, that dean probably won't start with this issue; and so on."

The quote above comes from Princeton's Dean Nancy W. Malkiel. She notes that you need seasoned leadership. The dean at Duke, George McLendon, has been at his post for a long time. Having spent many years at Princeton, Dean McLendon likely knows Dean Malkiel reasonably well. I'm sure she'd happily provide advice.

Malkiel also notes that you can't deal with other significant issues concerning undergraduates at the same time. At Duke, I know of no major items being discussed with regard to general education or the core curriculum.

There is nothing stopping Duke from tackling grade inflation. Like McLendon, Duke's provost and president also have been at their jobs for many years. It would seem that the only thing standing in the way of Duke dealing with its grade inflation problem right now is a lack of will on the part of its leadership.

Dean McLendon has said about solving the problem of grade inflation, "I am not a fan of micromanaging grading-grading can only be done well by the professor closest to the class members and their work." This is entirely beside the point. No one is asking him to micro-manage anything. Duke has an obvious problem. McLendon needs to solve it. If he doesn't, he's not doing his job. I could include similar nonsensical quotes on the issue of grade inflation from President Brodhead and Provost Lange, but it would be too depressing to type them and probably even more depressing to read them.

Above all, solving grade inflation requires leadership that is earnest, principled, and hard working. Unfortunately, that's what is missing at Duke right now. It's missing in its dean. It's missing in its provost. It's missing in its president. To solve this problem you need to appeal to people's better side. You can't do that if you don't have the good will and trust of the faculty. Grade inflation is a problem that can't be finessed. It doesn't require savvy. It requires integrity.

Of course, Duke's grade inflation problem isn't unique. Neither is its leadership credibility problem. My hope is that there are a couple dozen quality leaders out there in academe who will eventually take on and solve the problem of grade inflation locally. If that happens, places like Duke will look silly for not having done anything and will belatedly make changes.

I want to finish this post on a happy note. It doesn't have a thing to do with grade inflation. But it does have to do with Duke leadership. In my 15 years at Duke, I did meet one person in a position of leadership who possessed integrity: Kristina Johnson. She will soon play a major role in the Obama administration. Sometimes good things do happen to good people. And that's a wonderful thing.

*Summers has more important things to do now: saving the world from financial collapse. That's another problem more difficult than grade inflation.