The goofiness of grading curves surfaced recently in Marquis v. University of Massachusetts. Brian Marquis, a student enrolled in the University of Massachusetts College of Social and Behavioral Sciences, sued the university, some of its officials, the chair of the Philosophy Department, and a teaching assistant, because the teaching assistant lowered Marquis' grade based on a grading curve. According to the complaint, Marquis enrolled in Philosophy 161, Problems in Social Thought, taught by Jeremy Cushing, a teaching assistant. On the first day of class, Cushing distributed a syllabus that explained the grading policy:
Each exam will be worth 25% of the final grade for a total of 75%; The response papers will be worth a total of 20% (or 5% each). Each paper will receive a number grade (from 0-5) in .5 point increment; The remaining 5% is for participation in class.Marquis finished the course with response paper scores of 5, 4, 4, and 4.5, for 17.5% out of 20%. His exams earned 23, 22.5 and 19.5 out of a possible 25%, for a total of 65% out of 75%. He earned 5% for class participation. His total score of 92.5% 'translat[ed], by universally accepted standards, into an A- letter grade." The complaint did not explain who set the "universally accepted standards" for converting percentages into letter grades.
When, after the semester, Marquis looked up his grade on the university's data system, he saw that a grade of C had been entered for the course. He contacted Cushing and asked for either an explanation or a re-evaluation. In his response Cushing wrote, "This brought your final grade to an 84 for the class. The numerical grades were on the high side .. but I thought your grade (of C) was a good reflection of your work. To make the grades more representative of student performances, I set a curve (or, more accurately, I drew up a new grade scale). There were two other students that had grades in this range, one with an 83.5 and one with an 84.5, both of these students also received a grade of C. In your case, the grade assigned by scale seemed to fit." When Marquis complained to the University Ombudsman, she replied, "I would urge you to accept this grade and continue on with your course work as these are no grounds for an academic grievance. For example, 84 points could range anywhere from a 'C' to possibility an 'A-,' at the extreme end."
Marquis alleged that Cushing's grading was arbitrary and capricious. He alleged that when he applied to graduate schools he would be disadvantaged, because C grades are viewed unfavorably by graduate programs, in part because they suggest the student is lazy and inattentive to studies. He alleged that the grade pretty much squashed his chances of getting into graduate school.
Marquis rested his case on several grounds. He alleged violations of the Massachusetts Consumer Protection Act. He alleged violation of his First Amendment rights because the ombudsman shut down the grievance procedure. He alleged violation of the Fifth Amendment because he was denied a chance to appeal and because the grading infringed on his rights by taking of liberty and property without due process. Marquis also alleged breach of contract and breach of special relationship. He added allegations of reckless action, conspiracy, deception, interference with civil rights, promissory estoppel, intentional infliction of emotional distress, and tortuous interference with economic advantage. He sought certification of the litigation as a class action.
In their Memorandum in Support of Motion to Dismiss, the defendants argued that Marquis had failed to state a claim upon which relief may be granted, had ignored existing judicial authorities rejecting many of his specific claims, had not followed specific procedural prerequisites for certain claims, had not demonstrated a contractual relationship with the individual defendants, and failed to allege facts supporting his tort and constitutional claims. The defendants argued that a course syllabus does not create binding contractual rights. They also alleged Marquis had not completed service of the complaint within the required time.
According to this Boston Globe story, District Court Judge Michael A. Ponsor dismissed the suit. Marquis has indicated he is considering an appeal.
On the law as it stands, the court was correct. Unfortunately, the law provides no useful remedy, in fact no remedy, for students harmed by the sort of situation Marquis encountered. It is unfortunate that "the law" leaves students at the mercy of educational institutions, aside from obvious instances of racial, gender, religious, or similar discrimination.
Of course, if educational institutions did things the way they should be done, these sorts of issues would not arise. The Marquis case is not an instance of a student who is unjustifiably unhappy with his or her grade. It involves a serious grading problem.
Let's treat the case as the court had to treat it, namely, we must assume that the facts alleged by the plaintiff are true. As best as I can tell, the defendants did not dispute any of the facts that bear on my criticism of what happened.
Cushing gave the students a partial explanation of how the course would be graded. To his credit, he did more than some teachers, especially college and university faculty, provide in this respect. What he omitted, of course, was the connection between the acquired percentages and the letter grades. One might assume that a 100% earns an A, but after reading what happened in the Marquis case, I'm unwilling to make that assumption. The response of the ombudsman to Marquis suggests that there is very little correlation between the level of achievement measured against a standard and the letter grade assigned by the grader. Professional educators tell us that a good grading system lets the students know what is expected and what grades can be earned when a student performs at a particular level. My students know that I have a benchmark, that roughly 75% or better is equivalent to an A, that 20% is the minimum to earn a passing grade, that roughly 50% is equivalent to a C. The law school has in place a policy that requires faculty to explain the bases on which grades in a course are earned. Thus, faculty must explain if the grade is based entirely on an examination, a paper, semester exercises, presentations, or some combination, and whether and how class participation affects the grade. Faculty are not required to explain how performance translates into a letter grade, though a few of us do provide that explanation.
Cushing then altered the grading system without notifying the students. What appeared to be a 92.5% became, somehow, and 84%. Professional educators agree that this is a terrible thing to do. To leave students thinking that their total scores will be based on a published explanation but to use a different grade scale without telling the students is appalling. Perhaps the law has no remedy, but ought not educational institutions, especially huge universities that have education departments, think twice about the adverse ramifications of faculty engaging in this sort of changing rules in the middle of the game?
Cushing claims he set a curve, or, more accurately, drew up a new grade scale. Nowhere in his explanation on the course syllabus does he mention a curve. Yes, he threw his students a curve ball when he introduced the curve. Was it really a curve? In a sense, yes, it was. Why did he not inform the students at the outset that he would be doing this? I have an idea, and I'll get to that in a moment.
Cushing gives as one reason for altering his grading his intent or attempt to "make the grades more representative of student performances." If an 84 or a 92.5 is a C, that leaves very, very fine lines between the C and C+, the C+ and the B-, the B- and the B, the B and the B+, the B+ and the A-, and the A- and the A. Even if the only grades are C, C+, B, B+ and A, that's a lot of line drawing to squeeze into a small numeric range.
The other reason is that "[t]he numerical grades were on the high side." So? High scores could mean that students learned more than was expected, or somehow figured out how to do a better job writing exams than did previous students. They could mean that a disproportionate number of very bright and high achieving students enrolled in the course for this particular semester. Cushing's explanation, though, suggests that what happened was that he over-valued responses and assigned more points than hindsight would admit. Well, that happens. The solution is to get better at designing exam questions and assignments, and to learn more about grading. The solution is not to change the benchmarks in the middle of the course, or after the exams and assignments are completed.
Students don't realize it, but designing examinations, semester exercises, assignments, and problems is a very challenging task. If the question inadvertently reaches too many issues, or if there are insufficient facts, or if the analysis requires students to know something beyond the scope of the course, the measurement of student achievement will be skewed. Good testing requires an understanding of what the course is designed to do, and requires that the tester have specific goals in mind with respect to each question or problem. This is something that professional educators learn in their education courses.
What Cushing did is typical of the tribulations afflicting first-time teachers who do not have education degrees (and perhaps some who do). Cushing is a teaching assistant. I wonder if anyone was mentoring him with respect to examination, testing, and grading. In many respects, the institution failed Cushing no less than, and probably more than, Cushing failed the institution and the students. Universities throw graduate students into teaching roles because they're inexpensive, and because their use frees up faculty for research and other activities not directly connected to teaching. So long as the law turns the other way when students suffer because of institutional inadequacies, the quality of higher education in this country will continue to suffer.
Grading curves hide a lot of errors in examination design, defects in questions, flaws in scoring, and shortcomings in converting scores to letter grades. Grading curves provide an excuse behind which faculty can hide. Grading curves distort the evaluation of student work. The problems in the Marquis case go way beyond grading curves, and yet grading curves took center stage because the grader resorted to one when the rest of the evaluation system broke down.
When speaking in opposition to grading curves, I use an analogy that curving advocates have yet to negate. If 30 people enter a race, and 10 of them run a mile in fewer than 4 minutes, so be it. If the benchmark for an A is a 4-minute mile, and if in past and future races, only 3 or 4 runners run that quickly, it ought not diminish what the 10 runners have accomplished. If it is necessary to select the fastest runner for a gold medal, that's fine, even though 9 other very fast runners won't take the gold. What is important is the achievement and not the rank. I'm told that those seeking to hire lawyers (and perhaps the same is true of those seeking to hire people with other degrees and majors) rely on rankings. That's dangerous. They should rely on grades, assuming, of course, that the grades reflect what the student has accomplished as against a benchmark. Being number one means only that one has higher grades than everyone else, but it doesn't necessarily mean a person is all that accomplished. Similarly, being number 50 out of 100 doesn't sound all that great, but if the those holding the top 49 spots are supergeniuses, one doesn't necessarily go wrong hiring number 50.
Grades matter. So, too, unfortunately, do ranks. Grading curves distort the evaluation process. Every excuse that I have been given for the use of grading curves comes across as a method to mask a flaw in testing design, scoring, and the setting of benchmarks. To quote someone who is a teacher and who teaches teachers, when she learned I had been appointed to a law faculty, "But you don't have an education degree." "No matter," I replied, "I don't think any of us do." Her response, something to the effect that this was a horrible or at least unsatisfactory situation, rings true to this day, perhaps even more loudly.
Some might argue that "the government" should require institutions of higher education to mandate education courses for their faculty, as is done for most K-12 teachers. I say, no, if institutions of higher learning can't figure that out for themselves and take the necessary steps on their own, they are setting themselves up for long-term failure, measured by the achievements of their graduates. But if the government is not to be the source of pressure, who is? The answer is simple. Those paying the tuition should demand that they get what they pay for, and that includes sensible grading standards, well-designed evaluative tools, and prohibition against changes in grading methods and policies after the end of the drop-add period. Some institutions do follow these best practices. It's time for the rest of them, no matter their rank or reputation, to get on board. The law may not demand it, but common sense, decency, fairness, and truth so require.