by Joe Miller
Grading was my least favorite part of being a college professor.
I jokingly told my students that I threw their papers down a staircase, giving the highest grades to those that flew the farthest (since obviously those papers were the densest).
I stopped telling that joke after a student turned in a paper with pennies taped to the last page.
My actual grading process was laborious and time-consuming. I learned my process as a first-semester teaching assistant for the late Patrick Croskery, who was then an assistant professor at Virginia Tech. Pat was one of the smartest philosophers I’ve had the pleasure of knowing – and one of the few professors at a research university to put equal effort into teaching and scholarship.
Pat required his teaching assistants to read each essay three times.
The first pass was for identifying patterns, finding the things that most students got right and the things that most students got wrong. Pat’s view was that if there was something that nearly everyone got wrong, that probably indicated a problem with our teaching, not with their learning.
Our second reading was for adding comments to the margins. The bulk of those comments addressed the themes we identified in the first reading. It was also the reading where we divided papers into three stacks.
The largest stack was for average papers – the ones that got the things right that most everyone got right and missed the things that were most commonly missed.
The other two stacks consisted of (1) essays that made additional errors and/or didn’t get common things right and (2) essays that didn’t make the most common sorts of mistakes.
Grades in the middle stack generally ranged from C to B-. The top and bottom stacks held papers with higher and lower grades, respectively.
Grading – especially in the humanities – is an inherently subjective and relative undertaking. There is no such thing as the Ideal Freshman Essay. Grades are really only meaningful by comparison. Pat’s system was designed to measure whether a student performed better or worse than the average Virginia Tech student.
That meant the average paper in our average stack hovered between C+ and B-, as that was the median in an introductory humanities course at Virginia Tech at the time.
At the University of Virginia, that same system put the most average paper at a B+. At West Point, the average dropped back to B-. At Hampden-Sydney, it was a C+ and UNC–Pembroke, a C.
The letters didn’t reflect major differences in the work. The median Hampden-Sydney essay had about the same quality of arguments (and slightly better grammar) than the median UVA essay. The grades simply reflect grading standards relative to the college the student was attending.
Those average grades have gotten much higher over time. (Clearly not all professors use the Pat Croskery method.)
At UVA, the average graduating GPA increased from 3.1 in 1992 to 3.6 in 2021. There is zero evidence that students in 2021 were 16% smarter than their 1990s-era counterparts. Professors are simply assigning higher grades for the same quality of work.
Grade inflation, even at its most benign, makes everyone’s job a little harder. Above average students are compressed into just 0.4 points instead of 0.9. And it means a UVA graduate with a 3.0 GPA performed well below average.
In its more pernicious forms, grade inflation can mask serious deficiencies.
The University of California at San Diego has discovered this the hard way. The number of students testing into remedial mathematics has jumped from 32 in 2020 to over 1,000 in 2025.
In fact, the university has had to introduce a new layer of remedial mathematics when tests showed that 665 students needed to start with elementary school math.
More infuriating, 42% of the students who required remedial elementary school math received good grades in high school precalculus or calculus courses. Indeed, a quarter of them had 4.0 GPAs in math. The average high school math GPA of UCSD students in remedial math was 3.7, a solid A-.
I grumbled a lot about being Pat’s teaching assistant. Grad students who worked with other professors spent way less time grading.
By the time I started teaching my own classes, I was glad that I’d internalized Pat’s rigorous system. It still took a lot of time, but I was also confident that the grades I handed back communicated something meaningful about each student’s performance.
Students would be a lot better off if more teachers graded that same way.
joe.miller@fountaindigitalconsulting.com
