grade inflation is bad, but what’s the option?
Andrew Perrin is right on target in criticizing grade inflation. High grades should be awarded for clearly good work. But here’s the problem with grade inflation criticism. There is legitimate disagreement on the alternative. Consider the following ways to distribute grades:
- A curve. The top 10% get As, the next 20% get Bs, etc.
- A fixed bar. There is a pre-determined level of skill that get’s you the high grade.
- Pass/no pass. We abolish the idea of grades and just work with a “you get it or you don’t” system.
- Contingent. The instructor decides the grades based on the merits and skills of each batch of students.
- A budget. Instructors don’t have to give A’s, but all A’s are capped at 10% of the course.
Each of these grading schemes is justifiable and each is a rational response to grade inflation. The curve “solves” grad inflation by turning grades into a ranking process. The fixed bar accepts grade inflation (all A’s) because there is an independent bar (which itself might be subject to inflation). Pass/no pass abolishes high marks, you only worry about failures.. Contingent embraces grade inflation – if the instructor thinks everyone is doing well (e.g., a basic programming class at MIT) then why not give A’s? The budget allows a little inflation, but makes the instructor choose.
Overall, higher education uses #4 for most undergraduate courses (contingent). Graduate course work tends to be pass/no pass, with A being the de facto “pass.” Graduate exams are almost always just pass, with a few high passes. Many science departments stick with curves and grades are almost always ranks. You see fixed bar systems in some educational settings. Military training is one case. For example, the elite SEAL program doesn’t curve. There’s a bar and if everyone fails, so be it. To my knowledge, very few educational institutions use #5, a fixed budget of A’s to be given at the discretion of the teacher.
The basic point is that we can complain about grade inflation all we want, but nothing will change until there’s a serious discussion of an alternative to the contingent/discretionary system we now use.
orgtheory.net 
If grade inflation is simply a function that raises the grades of all but leaves the distribution intact, no information is lost. One solution would just be to report school rank instead of GPA.
LikeLike
Thorfinn
January 2, 2011 at 1:20 am
FWIW I agree with Perrin. You should get the grade you deserve – full stop.
LikeLike
Michael G.
January 2, 2011 at 4:09 am
Sure, you should get the grade you “deserve”. How should we determine what you deserve?
I think you are suggesting a fixed bar (set of bars, one for each grade). It is certainly a possibility, but can have problems as different people set different bars.
LikeLike
DBonar
January 2, 2011 at 2:00 pm
[…] Rojas over at orgtheory.net has a post about grade inflation in which he lists five possible […]
LikeLike
Grade Inflation « Marc F. Bellemare
January 2, 2011 at 2:02 pm
I did my first undergraduate degree at Hampshire College, where there are no grades: you either pass and get a detailed written evaluation of your work, or you fail. I found this system much more useful to me as a student than any of the grading systems I’ve experienced since then.
However, the more subjective grades are, the more room there is for class bias. Working-class students are better able to compete with their more privileged peers when evaluation is based on objectively measurable standards. Objective tends to mean quantifiable, which means grades.
If we must have grades, why not use system #2, the fixed bar? If nobody gets an A, so be it. This seems to be the system used in France.
LikeLike
Benjamin Geer
January 2, 2011 at 2:45 pm
Your suggestions miss the most practical and obvious: let instructors give whatever grades they want, but change the calculation of GPAs or reporting of grades to account for variations in grading scales by school, department or class. That’s where this is probably going, I think.
LikeLike
Philip Cohen
January 2, 2011 at 4:45 pm
I think it’s difficult to talk about grade inflation without also talking about its endogenous relationship with grade grubbing. Grubbing leads to inflation both directly because there’s a rachet effect where harsh mistakes get corrected but generous mistakes remain and a probably more important second order effect that faculty end up grading defensively, thinking about whether they can justify a relatively low grade to a petitioning student. Conversely, inflation leads to grubbing because students come to expect high grades. I’m thinking that some of Fabio’s grading models could help with this but some of them are going to be inherently unstable unless grubbing is abolished with a firm stare decisis policy.
LikeLike
gabrielrossman
January 2, 2011 at 7:28 pm
Good points Gabriel. There’s clearly an endogenous relationship between grade inflation and grade grubbing. If inflation leads to grade grubbing, could the opposite be true as well, where setting higher standards from day 1 reduces grade grubbing?
The more I talk with grade grubbers, the more I find that they are expecting a good grade because no one explained the difference between an average paper and an excellent one before the paper was due. When the grading standards aren’t explained clearly, students use their “effort” as a proxy for their expected grade. As we might expect, this is a highly problematic proxy that sends a lot of students to our offices.
There are a few times when students complain about their grades due to legitimate mistakes, and I think we need to allow students to petition for a re-grade in the case of real mistakes. But most complaints come from a mismatch between student expectations and student results. I think we can actually do a lot to adjust students expectations and prevent grade grubbing if we let them know exactly what it takes to get an A and what more they need to get an A. The more legitimate our grading schemes appear before students get their grade back, they less they complain, at least in my experience.
LikeLike
Noah
January 2, 2011 at 8:23 pm
I have heard of systems in which instead of a strict curve, courses (at least those with more than X number of students) are required to have a fixed average grade–say an average of B-. Instructors would then be free to shape distributions around this average. While I still don’t love fixing the average, as it does mean that students end up with a grade truly reflecting their accomplishments, it does at least allow a distribution reflecting reality rather than the bell curve. Of course, structuring it so the average grade could be B- *or lower* would also be helpful if the goal is combating grade inflation.
In terms of grade grubbing, I think one of the major current issues is that many students graduate from high school with calculated GPAs well over 4.0 and don’t adjust their expectations downward to reflect the fact that most colleges cap grades at 4.0. Incidentally, many law schools allow grades over 4.0 despite strictly curving grades, so curving alone does not necessarily combat grade inflation.
LikeLike
Mikaila
January 2, 2011 at 8:30 pm
Noah,
>between an average paper and an excellent one before the
>paper was due
yes, and the tricky part about this is that the difference is usually hard to articulate. nonetheless it does seem to help to articulate a standard that merely “hitting all the points” is only a “B” and an “A” requires that as well as showing elegance and insight.
> But most complaints come from a mismatch between student
>expectations and student results.
This is demonstrably true from the fact that in classes with electronic grade posting, grubbing is triggered by the posting itself rather than the return of the work and the attached comments.
LikeLike
gabrielrossman
January 2, 2011 at 9:15 pm
A couple of quick thoughts. First, yes, grubbing and inflation are closely tied; and yes, Mikaila’s point that incoming first-years, particularly at flagship campuses, are used to getting A’s, is true too.
Second: I think there are actually three related but distinct phenomena to worry about: grade inflation (the overall rising average); grade compression (the rising average combined with a constant ceiling reduces the “space” available to document variation in performance); and systematic grade inequality (a significant portion of each student’s GPA is a function of the discipline/instructor mix s/he encounters, not of her performance at all). Of these, inflation is IMHO the *least* important, for many reasons including the one Thorfinn outlines above. The most important is inequality, because that threatens the fairness of overall performance measures.
Third: differences in grading style are not just contingency or personal preference, though that’s certainly part of it. To a large extent they reflect legitimate intellectual and pedagogical differences. Disciplines like chemistry and cultural studies are sufficiently different in their very cores that imposing any uniform grading practice will (a) backfire; and (b) never be endorsed by any faculty governance council anyway. So whatever policy approaches we pursue, they have to be flexible enough to encompass wide variations in grading philosophies.
LikeLike
andrewperrin
January 3, 2011 at 3:43 am
When these questions emerge at my institution, a friend who is an undergraduate advisor reminds me to determine how well each student (as an individual) met the learning objectives set forth in the syllabus. Evaluation should be based on whether the student expressed an understanding/ability of the learning objectives (and not just that they “put forth effort”—this undermines the goal of “learning”).
Further, the grade of a D should not be accompanied by comments on what the student *didn’t do* (with the assumption that he/she “started” with an A…). Instead, rubrics can be developed for qualitative work that describes four levels of achievement from “minimal” to “excellent” (D-A), each one building on the “D” rather than being “subtracted” from the “A”.
Hypothetically, if you have one class of students who are “better” collectively than another class, the “better” students are penalized by any cap on high grades (whether it is a curve or absolute number) just by being in the “better” class. Is it possible that every student can excel? We should allow for this possibility, even though it is unlikely. But if all of our students are easily earning “A”s, then perhaps our courses do not challenge them enough?
LikeLike
Rebecca
January 5, 2011 at 9:42 pm