Model this

As we have written before, private colleges and universities are by far the biggest offenders on grade inflation, even when you compare private schools to equally selective public schools.

There are charts and further information at the link.  By the way:

…about 43 percent of all letter grades given were A’s, an increase of 28 percentage points since 1960 and 12 percentage points since 1988. The distribution of B’s has stayed relatively constant; the growing share of A’s instead comes at the expense of a shrinking share of C’s, D’s and F’s. In fact, only about 10 percent of grades awarded are D’s and F’s.

It’s worth trying to model that too.


I'm not sure why this is hard to model. Every professor I've ever seen who explicitly announces a grade distribution usually says something like "I give out 40% Bs". As grade inflation occours, professors are probably just targetting a quota of Bs, but chosing not to give out very many Cs and Ds.

How does your model explain why grade inflation is more severe at private schools, even after control for equally selective public ones?

That's not particularly hard to model either. Public schools likely have more oversight over grading policy / less instructor independence (I bet if you control for size and public/private) much of this effect goes away.

I posit that much of grade inflation is the result of instructor independence in grading. Its very hard to be the one instructor to grade more harshly, and there are some returns to grading more leniently then your peers.

Every once in a while, the department will attempt to as a group become stricter about grading, but will push the group to a more lenient standard than they used in the past. Larger schools likely are forced to do this more often/enforce this more strictly.

Lo and behold, grade inflation.

Equally selective schools aren't equal quality nor are they selected equally among accepted freshman. Private schools occupy 1st through 24th place in the US News rankings. UC Berkeley, the top public school, is 25th.

With objective course standards, a good teacher, and very bright students, high grades are not extraordinary. I would expect nearly 100% As in a freshman Calculus class at top schools. College doesn't get tough until Junior year, and Senior year is only as tough as you choose to make it.

That said, there are a great number of extremely poor professors and graduate student teachers. Many of these people camouflage their poor teaching by assuaging student angst with generous curves. Most of these teachers are researchers who consider teaching a distraction from their real job: research, blogging, political activism, and ego boosting.

I would expect nearly 100% As in a freshman Calculus class at top schools


We clearly need to return to return to the gold standard. Hard currency, even for grades.

The essence of grading is separating good, better, and best in a classroom. Anybody who has ever taught knows that grading is a costly activity. Examine the cost function of teachers who grade. When the cost of this 'separating' is low, as in Math classes, you can expect grades to cover a wider ranch from A to F. Costs are low for math teachers because it is easier to handle the separating in a purely objective manner. Where costs are high and grading is highly subjective, teachers are hesitant to separate students into as many buckets, so grades tend to bump up against the top of the spectrum.

What's interesting about this study is that it highlights a key factor that drives the costs of grading, in addition to the nature of the subject matter. Presumably, in private schools, there is a tighter relationship between the teacher and student (smaller class sizes, greater availability, etc). Hence, by building a deeper relationship with their teachers, students effectively raise the cost of bad grades. This has exactly the same effect as moving from a subject matter area of high objectivity to one of high subjectivity. Result: Higher grades!

Or it could be just that the cost of assigning a low grade is higher for more entitled kids.

Potentially true. Or do private schools have different incentives than equally selective public schools?

Well are the private schools taking in better students?

I had a friend who went to Brown University and he said "every one gets A's in Brown". Everyone who went to Brown University with him was really smart. On the other hand when my uncles went to Brown and Yale it was not so hard to get in. I think that my brothers were as smart as my uncles but Brown and Yale which took my uncles would have laughed at my brother applications.

Why are private schools inflating grades more than equally selective public schools?

[I.] Many good grades:
Better grades increase the demand for the course, student satisfaction (up to a certain point) etc. and thus the profits of private schools.
There will be more applications and fewer people who leave the university.
Way to maximize profits: Always increase the proportion of good grades up to the highest acceptable/attainable point.

[II.] Almost no fails:
Failures are necessary since society/students expect them. But failures are expensive from an administrative point of view, as are students who leave the university.
Way to maximize profits: Always reduce the proportion of failures up to the lowest acceptable point.

--> Nash equilibrium: almost no failures, up to the point where the marginal benefit of letting a student fail begins to exceed the marginal cost of letting him fail.

-- S. O. --

Whatever explanation you come up with needs to two address the fact that:

(a) before 1950, private and public schools had similar average GPAs; and
(b) much of the divergence in average GPA between private and public colleges occurred during the 50s and 60s. The difference between the two has not changed much since 1970.

You're largely correct, but, according to the charts, the gap opened in just the 60s, not the 50s; there was no grade inflation in the 50s. The gap between public and private has been narrowing recently.

Private schools are not necessarily for-profit schools. Public and non-profit private schools have essentially the same financial incentives. Public schools may have a few more pressures from state legislatures to accept students or offer certain classes, etc. Private for-profit schools like DeVry and Strayer are a completely different model.

Also I think you're referring to just a regular old equilibrium where marginal benefits and costs are for an individual or firm equal, not a Nash equilibrium where multiple players choose a strategic option given the response of each of the others.

Players will increase grades beyond their competitors' grades, step by step. They can't immediately adjust to the final MB-MC equilibrium, but they gradually move towards that equilibrium via a mechanism of multi-player equilibria.

And no selective school is for-profit.

My first thought was that private universities are running a business, and thus are more likely to "give the customer what they want" (validation). There will thus be a "drift" parameter which is stronger in the private case. I'll admit that I'm not sure what the actual institutional mechanism for this would look like, however. (Perhaps there are some indirect carrots for professors who give their students high marks)

Alternatively, if college is about the inculcation of bourgeois values, perhaps rising tuition implies that private school students are simply becoming more bourgeois and need less reprimanding in the form of bad grades. Public school students are probably also becoming more bourgeois but at a slower rate.

And some other things:
Maybe teaching has gotten better and so the students learn more.
There is evidence people are smarter now due to the Flynn effect.

So people should get better grades.

If schooling is good and economical, due to what is imparted there rather than mostly signalling which I think is your position Tyler, then we should try to make learning at school easier, more fun and cheaper. We should ask, are the students learning what will benefit them in not is it hard enough. Perhaps everyone should get A's. Perhaps subjects should be taught with entertainment quality of the history channel.

Want to bet that B-quality work 60 years ago wasn't better than B-quality work today? Because I'd love to bet the other side.

I tried once to tighten up on the distribution. To warn the students I handed out an approximate grade distribution that had only about 15% As. A student complained to the Director of Undergraduate Studies that this deviated from the distribution that the Department "recommends" and that other professors use. So I was accused of being "unfair." Then, the Director accused me of doing this to reduce the number of students in my class so as to reduce my load. I then found it easier to just go along with the inflation.

I am reminded, however, of a Harvard(?) professor who recently was giving students two grades -- an inflated one that was reported to the registrar and another one that reflected the professor's true evaluation of the student (low). I am tempted to do this. However, I wonder if students will complain even about this.

That was Harvey Mansfield, I believe.

If I had taken a class where the Professor announced that 15% of the class would get an A in a department or even university where 40% of the students get A's, I would be pissed and rightly so. It's blatantly unfair to the students in your class. Especially if you're only trying to tighten the distribution starting that year. People are evaluated on their grades and, "well, that B or C was in a year when that professor was grading really tough", just isn't going to cut it.

If you want to fight grade inflation you have to do it at the department level and you do it by incentivizing staff to conform to a department wide grading policy.


The individual professor can't tighten up until the department does. But the department doesn't want to lose students (and majors) to other departments. So Economics must convince all other departments. (But there will incentive to cheat from this particular cartel.) But now the university doesn't want to lose students to other universities and so forth.

So maybe it should start with the individual professor and students could simply not sign up for his course, reducing his burden. Then other professors seeing this might tighten up too. And so forth. Until, of course, departments lose students or the university loses students. They will then want professors to losen up.

And so it goes.

No, I think it is best for the professor to grade as he wants and let the chips fall where they may.

So you're grade rationing as if high grades are a scarce resource?

If you have certain learning objectives, you teach well, and your students work hard, they should be able to get any grade their performance deserves.

Grades should measure an individual's performance relative to learning objectives, not the performance of peers.

Your grading system is precisely what school's consider "arbitrary and capricious," it's shameful you think being told that it's patently unfair means you have to play some childish psychological game.

In my experience, teachers who rely on a curve are usually the worst teachers hiding their incompetence, laziness, and lack of commitment to student learning.

One thing I am curious about is who is getting the additional As, the former B students, or the former C/ D students?

Common sense suggests that students who used to get As are still getting As, students who used to get Bs are getting As, and students that used to get Cs are now getting Bs, and so on. So you just start treating B as a C, though its harder to sort out who should really be at the top.

But the process of grade inflation might have gone deeper and affected what the professors are looking for when assigning a high grade. Maybe good, unspectacular work is getting Bs as before, and all sorts of crap is now getting As. Is inflation simply shifting the scale or making a complete mess of the process?

To my knowledge, Stuart Rojstaczer has never addressed the clustering effect. In the early days when average grades were lower (pre-1960s) very few schools were national or even regional in any meaningful sense. People frequently selected schools based on geography and there was not nearly as much of a hierarchy of schools based on perceived quality as there is today. What this has meant (and Stuart's data confirm this) is that the range of abilities at schools was quite high, even at the most elite schools. Today, geography is of much lesser importance, especially among elite public and private schools, and there is a much stronger perception that schools greatly differ in quality. As a result, the range of abilities at any given school is much narrower than it used to be. Again, the data confirm this.

I suspect that this, in combination with professors' tendency to grade up when in doubt, has caused some of the grade inflation. Please be aware that I do not think that this is the only cause, but rather a significant cause that has not been greatly explored.

As to why public and private schools had inflated grades at different rates, I must admit that I do not know. It does seem, however, that selective public schools--of which there are not a great many--tend to have constraints on the quality of the student body that they may recruit. One condition common across them all is that they are required or feel responsible for recruiting a large number of in state students. Moreover, these schools attempt to obtain some geographical diversity among their in-state students to satisfy state legislatures.

Another point worth making is that public selective schools in California, Virginia, North Carolina and Illinois have typically grown at a much faster rate than their private counterparts. This, again, is done at least partially to satisfy legislators and their constituents. An ever growing class size makes it harder to maintain the tight clustering of high ability students that their private competitors have managed.

This sounds like a pretty good explanation to me. Personally, I don't see anything wrong with a low percentage of D's and F's. Both are generally considered to represent failing or near enough to failing performance. On the other hand, I would prefer not to see more than 35% of the students getting A's. The distribution should be more or less evenly distributed between A's, B's, and C's in my opinion.

If I am correct about clustering or sorting, then the grading system is largely arbitrary now anyway. In an era when all schools had their share of good, average, and poor students, grades were a useful means of comparing students across different schools. An employer or graduate school could use grades to identify the best students at a given school. The employer or graduate school could also assume that if a person was a good student at school X, then there was a good chance that the person would have been a good student at school Y. Obviously, this was not perfect, but it could still be done.

Now that students have sorted themselves so much by ability, grades cannot be compared across programs within the same school let alone across schools. Handing out more Cs will not fix this problem.

Private universities usually cost a lot more for the students and their parents than equally selective public universities, and perhaps part of what the private universities are selling is an easier grading scale. So what S. O. said.

I suspect the addition of more liberal arts courses is the reason for much of the observed grade inflation. Everyone gets an A for expressing their feelings in those classes. The difference mentioned in the article could be because public universities tend to be diversified while many private universities focus exclusively on liberal arts.

The clustering theory is interesting, but if it is the case then shouldn't there be a bunch of schools with extremely low grade distributions since the smart people all went to better schools?

Isn't this just an example of resource dependency? I would assume that the private institution derives a greater proportion of total revenue from net tuition than the comparable public institution, ceteris paribus. So it would not be reasonable to expect that the private institution would have a greater incentive to grade inflate if this inflation is consistent with student preferences.

I have been teaching college students economics for about 8 years. I have taught large classes and small classes. I have taught at large public institutions, and high quality small liberal arts colleges. Over this entire time I have given very, very few F's. This is not because students aren't failing my courses. The reason I don't give F's is that students that are failing almost always withdraw from the course before I am required to give them a grade. The only students that finish the course are the ones that are going to pass.

This doesn't explain grade inflation in general, but it does explain the lower number of D's and F's. Also I would expect this effect to be more pronounced at smaller private schools since the student population is watched much more closely (At one school I was required to notify the student and the students advisor any time a student was on track to get a C- or below).

You are correct. Most students headed toward a D or F drop.

Some schools allow students to drop much later than others. Some schools permit students to "scrub" their transcript by permitting them to retake courses for a replacement or an average grade. That's a school policy, not a teacher policy.

Correction to previous: it should read "So it would be reasonable to expect" rather than "So it would not be reasonable to expect". Apologies.

As a college student, I'd like to see another time series chart: of the ratio between humanities, social science, and natural science majors.

Has anybody ever studied at grade inflation in departments that have to conform to rigorous accreditation standards? I went to engineering school where there were ample C’s, D’s and F’s. This was especially true in the first two years, which led to an attrition of around 50%. (I think most of the weed-outs wound up studying business and economics.) As I recall, samples of our exams, homework assignments, and grades were retained for the auditors.


I believe that the research shows that science, engineering and the social sciences have experienced the same amount of grade inflation as the humanities. The research also shows that these disciplines have, for as long as the data go back, graded lower than the humanities. So while it is true that these disciplines grade lower, this difference has been steady over time and is not the result of grade inflation.


Often when I read the stats, I see math, science and engineering lumped together with the social sciences. Given the softer-gentler nature of the latter, I would love to see disaggregated data that just looked at the hard fields. But then, maybe the sample size is too small.

I know of this paper:

I guess I should attach the abstract, emphasis mine:

We develop a model of strategic grade determination by universities distinguished by their distributions of student academic abilities. Universities choose grading standards to maximize total wages of graduates. Job placement and wages hinge on a firm's productivity assessment given a student's university, grade and productivity signal. We identify conditions under which better universities set lower grading standards, exploiting the fact that firms cannot distinguish between "good" and "bad" "A"s. In contrast, a social planner sets stricter standards at better universities. We show how increases in skilled jobs drive grade inflation, and determine when grading standards fall faster at better schools.

Large public schools have freshman fail out rates on the order of 25%. The school's attitude tends to be they'd rather fail someone who isn't working to make space for someone else who might.

So, oddly, public schools are much more free market minded in their treatment of students, not because they treat students as customers but because they don't.

Many private schools have programs to help student through their early years. Stanford has their Freshman and Sophomore Program.

Private schools want you to succeed to keep tuition money flowing.

It's also correct that many students drop out or change majors or schools to better suit their abilities. I got two Ds my first semester before dropping out. When I returned to a different college, those credits counted as elective credits but none of the grades were part of my GPA calculation. I got straight As from then on except for one graduate level course I took.

So long as private and public employers, as well as grad schools, prefer to penalize those with low grades from rigorous programs in favor of those with higher grades from soft ones grade inflation will continue. Any program that sets a minimum cutoff of grades (say a B avg in many med schools or job interviews) regardless of faculty or school of origin, there will be grade inflation.

This, despite the fact that many a C student in electrical engineering is uniformly better than A students from a lot of liberal arts majors in weak schools. Yet the latter will actually have a higher chance (though perhaps still not a good one) of getting into many programs.

The problem is cross proposes and cross allegiance. The propose of schools is to educate and grade but testing to grade tends sometimes gets in the way of educating.

The private schools have more of an allegiance to the students, the Government schools more to Government/voters/big business. The private schools should therefore be more prone to grade the students high but also to attempt to better educate them.

I think more separation between education and testing might help everyone.

Why is no one ask are the student learning enough.

Comments for this post are closed