Tuesday, February 12, 2013

Vog3lfr3i Says: Relative Grading is Stupid. From InsideHigherEd.

Zack Budryk

Since he started teaching at Johns Hopkins University in 2005, Professor Peter Frölich has maintained a grading curve in which each class’s highest grade on the final counts as an A, with all other scores adjusted accordingly. So if a midterm is worth 40 points, and the highest actual score is 36 points, "that person gets 100 percent and everybody else gets a percentage relative to it,” said Frölich.

This approach, Frölich said, is the "most predictable and consistent way" of comparing students' work to their peers', and it worked well.

At least it did until the end of the fall term at Hopkins, that is.

As the semester ended in December, students in Frölich’s "Intermediate Programming", "Computer Science Fundamentals," and "Introduction to Programming for Scientists and Engineers" classes decided to test the limits of the policy, and collectively planned to boycott the final. Because they all did, a zero was the highest score in each of the three classes, which, by the rules of Frölich’s curve, meant every student received an A.

“The students refused to come into the room and take the exam, so we sat there for a while: me on the inside, they on the outside,” Frölich said. “After about 20-30 minutes I would give up.... Then we all left.” The students waited outside the rooms to make sure that others honored the boycott, and were poised to go in if someone had. No one did, though.



  1. Outstanding! Students themselves expose the absurdity of curving grades. Or of stating policies that can be "gamed". I've actually always thought exactly that: if you curve grades, what's to stop the class from collectively deciding not to do any of the work?

    In 20+ years of teaching college math, I have never "curved" grades. I used to state the no-curving policy on my syllabus, but more recently I realized this leads more people to drop the class, so I just do it (or rather, not do it) quietly.

    I just state "expected" grading cutoffs: 80% for an A, 70% for a B, 60% for a C. In the end I have to adjust these bands downwards by about 5 points anyway. Typically there are two or three As, very few Bs, and a number of "courtesy Cs". Most students are pleasantly surprised when they see their course grades.

  2. I have not, and will not, mark on a curve. I inform students that if they all get an A, then that's their grade. If they all get an F (more likely), then that's the mark.

    My question is, what kind of an idiot sets a curve so the highest grade is 100%? If a student gets 36 of 40 that is not, unless my math is really bad, 100%

  3. I have a certain amount of sympathy to upward adjustments, but I don't do it automatically. I have certain things I look for (like the sum of the high scores for all the exam questions, and questions with very low average scores, or that nobody earned full marks on). Then I decide whether the question was unfair, or the students unprepared.

    But I sure as hell don't tell students that I do that. When I bump scores up, it's a gift, not an entitlement.

  4. And now that I've read the whole thing... what a maroon.

    Hey, kids, unexcused absences count as zeros, not as 100%. That doesn't have to be in the syllabus because it's bleedin' obvious; no sane grade appeal will support you, and you lose.

  5. Allow myself to quote my syllabus:
    "You will not pass if you do not give every speech or miss an exam."

    Problem solved.

  6. A student occasionally asks if I curve the highest grade to 100% and then adjust all other grades accordingly. I say, "No, I am willing to lower the lowest grade to 0% and adjust all the other grades accordingly." That shuts them up for a while.

  7. When I was an undergrad (MANY years ago), I had a Hamster Psychology professor who graded on a curve... A bell curve! The guy posted the chart outside of his office (he wasn't joking). That was bad enough, but...

    We had the usual "freshman-as-experimental-subjects" requirements to fulfill, but were allowed to do extra credit (more experiments).

    And so... how did he count them? He piled in the extra credits INTO the regular grades in his bell curve. Thus, there really wasn't extra credit. If you didn't max out the extra credit, you didn't get the additional cumulative points (which would push you down in the curve).

    When he explained this to me, I looked at him and thought: "man, I'm just a punk undergraduate, but you are seriously math challenged".

  8. Where else but a classroom is anyone graded on a curve? Certainly not in the real world.

    In the end, you either have results or excuses.

  9. As a grad flake I taught for a Proffie who insisted that I force exam results into a bell curve. If too many people had Bs, then I had to go back through the exams and mark them down until the class scores fit that curve.

  10. I curve in my class. It's the only way for anyone to pass. My tests are too difficult, and the curve gives my students a chance.

    1. So the students are not demonstrating that they have achieved the course objectives, but they're still receiving As, Bs, and Cs? There are two other things you could try: changing the course objectives to ones that the students have a reasonable chance of learning, or changing the delivery of the material to a method that adequately prepares students to demonstrate that they've achieved those objectives. Curving a test because "it's too hard" seems like just giving up.

    2. Prof.Chiltepin, in an ideal world, I couldn't agree more. But in practice yes, I do end up giving Cs (not As or Bs) to students who have not really engaged with the material, and barely even gone through the motions.

      It's math, and I teach many required classes for non-majors. Their intrinsic ability for the subject and command of prerequisites do not fall into a continuum: thee is a small group at the top who can handle the material at an adequate level; then a large gap; then the majority of the class, who forgot all they had "learned" already, dislike the subject and struggle throughout. This has gotten worse over the years.

      Changing the course objectives? It's already at the lowest possible level if it is going to be useful for anything else (say, in engineering), not to mention this would trivialize it for the people who could actually learn something. Changing the delivery? That's not the problem. Ultimately, you can't teach anything new and difficult to someone who isn't willing to put in as much effort as it takes to learn it.

      My tests aren't deliberately difficult, but they assume the students have at least understood the solutions to the HW problems, a minimal requirement. There are so many who aren't even willing to do that, I end up more or less forced (tenure or not) to give some of them a courtesy pass. I still don't curve, but I make my grading scale unusually forgiving. It's not terribly honest, but addressing this would take systemic changes on a longer timescale than any individual's career, and for which there is no political will. We all know that, right?

  11. I think Professor Frölich should have given everybody a 0, not 100%. It stands to reason that the highest grade is obtained by writing the exam and that it cannot be a 0. If I were one of those students, I would have taken the exam even if I didn't think I would get an A on my own merits. The students are lucky that they have not been treated as if they just cheated, which they actually did.

  12. If my grades actually fit a bell curve, I become concerned. I'm not testing a random sample of students; I'm testing a group of students that I've carefully groomed to meet the learning goals of the course. They should, if I am doing my job well, skew to the right. This is why I think most talk of grade inflation is a load of shit.

    1. That, in itself, is a load of shit. If you don't have a reasonable grade distribution, you're not setting the course goals high enough. Getting an A should be a challenge that not everyone is up to meeting.

      Instead, it's the most common grade in America.

      Given the state of higher education oft bemoaned on this very blog, a set of challenging course goals will find a fair number of students who are too lazy to achieve them, and another significant number who don't belong in college and are thus incapable of achieving them well enough to earn an A or a B.

  13. "Skew to the right" does not mean the same as "I give everyone an A." In fact, there are a more than a few students who are, as you said, lazy and not up to the challenge, and they receive F's or D's. I am a bit surprised that you just assumed I give everyone an A. Is that the false dichotomy we're working under, here? Either you have a bell curve or you give everyone A's? It isn't possible for there to be any middle ground?

  14. Professor Chiltepin, if I understand your comment to mean that your students generally earn more A's, B's, and C's, I think you might mean their grades skew to the left. Also, I tend to agree that curving shouldn't need to occur if the tests or other various measures of assessment are realistically gauging student comprehension.

    I remember getting my Differential Equations midterm back from my T.A. with a big fat 55% on it. It turned out to be a B+. There was clearly a serious disconnect between the expectation of the professor and our ability to perform. That's a problem.

  15. I only consider a "curve" on individual tasks, and then only when I see this red flag: even the best, straight-A-til-now student can't manage the A. This happened once when I was a TA for a language course, and there was one section of a test that was weighted pretty heavily compared to the others, but we hadn't had the time to really go over the material in class like I'd wanted. I kept it in the exam just to see how it would go, but even the brightest student royally screwed it up (his classmates did even worse). So I dropped that part of the grade for everyone and graded the rest of it accordingly. Then it turned out pretty much as I would have predicted.

    Sometimes I still worry that my exams are too difficult, but I want to set up a challenge rather then just a hoop to jump through for the A. Snagging a B isn't all that difficult in my classes if you do what you're supposed to do and submit your shitte on time (which is, sadly, a lot rarer than it ought to be), but I make sure it takes going the extra mile (or having that extra talent) to achieve the A range. It takes active avoidance of responsibility to get a D or worse.


Note: Only a member of this blog may post a comment.