Princeton’s report on its decade-long grade-deflation experiment had a lot of fascinating findings in it, only some of which I had room to discuss in my column Tuesday. Here’s a look at what the committee found about grading by department, particularly in STEM (science, technology, engineering and math) fields vs. humanities.

To recap, in 2004, Princeton set a nonbinding target — 35 percent — for the share of grades awarded within each department that should be in the A range (A-, A, A+). Students griped about the stingy allotment of A’s, but the report found that almost no departments actually met the target.


Source: Princeton University. Percentage of A-range grades by department, for all graded work except junior papers
and senior theses, shown as 3-year averages for Academic Years 2011–13, versus Academic
Years 2002–4. The area of each symbol is proportional to the total enrollment in the AY11–13
period, and the departments are color-coded by division.

The chart above shows the percentage of A-range grades awarded by department in the three years immediately preceding the grade deflation efforts, and in the most recent three years; the the 45-degree line cutting diagonally across the chart represents where the dots would fall if there had been zero change in each department’s grades.

A couple of departments (such as Slavic Languages and Literatures, where two-thirds of grades these days are A’s) are above that line, meaning they defiantly decided to dole out more A’s after the university encouraged everyone to tighten the spigot. Most of the dots are below the diagonal line, though, which means that most departments reduced the share of A’s they awarded.

But they didn’t all bring their share of A’s quite  to 35 percent. The horizontal line toward the bottom of the graph shows where dots would land if departments met the targets. Note that only two departments are below the target: economics and physics. A few others, including molecular biology and the department of ecology and evolutionary biology, sit just above the line.

You might also notice that the purplish dots — which represent the sciences — are clustered in the bottom-left quadrant, meaning they were lower-graders both before and after the new policy. The aqua dots, representing the humanities, tend to have much higher grades.

This disparity in grading between STEM fields and the humanities was one of the reasons that the grade targets were first put in place, to develop “consistent standards across departments.” Yes the gap in grading standards remains quite large.

So why do humanities departments still give much higher grades? A cynic might argue that smaller departments that don’t focus on skills in high demand on the job market are trying to lure students in with easy grades. I think that’s part of the story. But I also think it has to do with the nature of the assignments used to determine grades, the typical size of classes, how specialized or advanced the work is and the composition of the students taking the classes.

In big, quantitatively-focused survey classes, where there are discrete right or wrong answers on the exams and problem sets, professors and teaching assistants can look at the distribution of points awarded and draw sharps lines where A’s, B’s, C’s, etc., start and end. With humanities classes, which rely more on essays and papers, it’s likely harder to draw an arbitrary line and say “this paper clearly falls in the B+ section of the distribution.” But it’s not just about quantitative vs. nonquantitative work.

There’s also probably greater variance in achievement in lower-level classes, which tend to be bigger and to draw in students with varying levels of interest and aptitude, especially in certain disciplines, such as economics and math. (A lot of students who end up majoring in classics or philosophy take big, intro-level economics or math courses during freshman year; the reverse probably isn’t as common.) As coursework gets more advanced and specialized, and class sizes become smaller in the process, there’s probably less obvious variation in the quality of students’ work.


And actually if you look at the grading data by lower-level vs. upper-level classes, you’ll see that lower-level classes were much more likely to adhere to the grading targets than were upper-level classes:


Source: Princeton University. Percentage of A-range grades by department,
separated into 100–200 level courses (top plot), and 300–400 level courses, excluding junior
papers and senior theses (bottom plot). The area is again proportional to the enrollment, and
the departments are color-coded by division.

For less advanced classes, a lot of departments average below 35 percent in A-grades, including disciplines in the hard sciences, engineering, social sciences; but for upper-level courses, there are no departments at all that meet the target. Even in quantitative fields, high grades are given out more generously in the advanced courses.

In math, for example,  the average share of A-range grades awarded in lower-level courses in recent years was 32.4 percent. But the share of A-range grades in upper-level math courses was 59.9 percent.

Maybe this is a reflection of the nature of the work at higher levels, or of a narrowing of the range of aptitude of the students who choose to take advanced courses, or of what happens when classes get smaller and students become less anonymous to the professor. One way to interpret these charts, though, is that departments with a lower share of A’s overall may tend to be those that have big survey-level classes taken by students from other departments, students upon whom they can dump a lot of their B’s and C’s.  The university-set grading targets, after all, are for all courses offered by a given department, not for individual classes, so lower grades in gigantic, low-level classes can balance out higher grades in tinier, high-level classes.

Catherine Rampell is an opinion columnist at The Washington Post.