SERIOUSLY! (to what I bolded) So, if a good teacher has smart students, and EARN an ‘A’ or ‘B’, if too many are that smart, they get a grade less than they earned?
At U.S. law schools, it’s very common for the school to tell a professor how many of each grade they are permitted to give out. I don’t think they force them to assign Ds or Fs, but they do require a minimum number of Cs and C-minuses.
It was known in the legal community that my law school had a tendency to give out lower grades than other comparable law schools.
And it’s very common for a grade in a law school course to be based on just one final exam.
The purpose of the class is for the student to learn the material. How well this is achieved is independent of the grading system.
The purpose of the assigned grade can be any of several things, with two standing out:
(1) to quantify how well the student has learned the material on some absolute scale
(2) to quantify how well the student has learned the material relative to the other students
Both are useful, but option (2) has the huge practical advantage that it is self-calibrating. If you get a ‘B’, it doesn’t indicate what mastery of the material you have. It only indicates that 10% (say) of students have a higher mastery.
Option (1) often leads to useless (to a consumer of these grades) answers, since setting the cutoffs ahead of time will most likely end up with a high- or low-skewed distribution.
So, someone who gets an ‘F’ in a curve-graded course should not interpret that as meaning “I failed” but rather as “I did worse than 90% (say) of the class”, which is exactly the information someone might want to know, since setting an absolute scale on a course-by-course basis that is consistent from year-to-year is basically impossible.
The absolute scale comes from knowing the general level of student at the school. If I see a transcript for a student from Caltech or M.I.T., a ‘B+’ is notable, but perhaps not terribly off-putting. If I see a transcript for a student from Midwest Podunk College, a ‘B’ is very worrisome, as even at that school he/she wasn’t able to excel relative to the other students. (The pool of other students, thus, establishes the long-term, consistent reference against which to interpret the grades).
I suppose every situation is different, but the only classes I’ve ever had that were purposely graded on a bell were based on the assumption that no one would get a perfect score. The bell was based on the highest score in the class. Then the top 20% got the A, the next 20% got the B and so on. However, the bottom 20% only failed if they actually failed. If the passing grade was 70% and the range of scores was 70-85, then the 85 score got the A, but the 70 scores still got their D.
I recall one test where the top score in the class was 55%. No one could argue that they flunked because of the curve.
If I only have room for 30 majors in my program, and I have 150 people in the introductory class, then I need some way to pick out the top 30. What else would you suggest?
Depending on the field of study, there are two types of intro classes: 1 for majors, 1 so people can fill distribution (general ed) requirements. Why would you (IMO artificially) cap how many students can have a major.
Most of the large-class courses I’ve taken explicitly stated in their syllabus that grades are curved, but you won’t fail if you get at least, say 60%. It could be that the FML quote in the OP is about someone who got a final score of 47%, but the first draft of the grades passed him; then the prof’s superiors tell her to reassign grades with new criteria, so the student ends up getting the failing grade he actually deserved.
I haven’t encountered any too-easy courses where everyone scored super-high anyway. Even in courses I found particularly easy had a wide distribution of scores. It’s been much more common to say “All hail the curve!” before or after a particularly brutal exam; you know that even if you score 68% on the final, you can still get an A if the rest of the class did even worse.
Use a more fine-grained scale than the discrete five-point A through F system.
Lots of reasons. The major is a competitive/elite program and the school doesn’t want to expand it precisely because it’s competitive and elite.
A science course with a limited amount of laboratory or other facilities.
A field where a limited amount of academic staff are available.
An area which the school is gradually de-emphasizing.
A school with a practice of having senior professors teach both undergraduate and graduate courses, and keeping both groups small.
Part of a university system where the curators have decided that another campus will be the “focus school” for that particular major.
How do you define “earn”? If the established system is such that the top 10% get As and the next 20% get Bs, then to “earn” an A or a B, you’d better out-perform 70% of the students. Anything short of that, and you didn’t earn an A or B.
Another point: you may feel that it is unfair for someone in the “middle” of the pack if the class has more smart students than normal. This sort of thing is way less of a random factor than the year-to-year changes in instruction and testing. And, this is exactly why a curve is useful.
In other words, the distribution of student skill level is fairly consistent from year to year in a given course. To be sure, small courses can be affected heavily by one or two geniuses or clowns, but small courses aren’t usually graded on curves anyway. (It’s the bigger courses that are.) What does change dramatically from year to year is the instruction and the testing. Certainly the instructor changing has a big impact on the raw numerical score distribution. Further, unless the instructor is being lazy, the midterm and final exams will use brand new questions – certainly not the same ones as in the previous couple of years – and so the numerical scores cannot be directly compared to those in previous years. What is consistent year-to-year is the aggregate student body. So, knowing how well a student performed relative to others is a much more useful measure than knowing how well they performed on one particular, uncalibrated version of the course.
my definition of “earning an ‘A’”-You mastered at least 90% of the material in a class.
That’s just as arbitrary as any other definition.
The problem is that that is impossible to measure or interpret in any consistent way, per my latest post.
Here’s a brief blurb from the SAT folks about grading on a curve. It’s a very large-scale version of a course-by-course curve. It is very useful knowing that 720 for a student one year means the same as 720 for a student the next year, even though their initial raw scores might differ (since they took different tests).
that would be the part where I said the test should be reviewed. I had a school teacher who taught very little in class. He let the better students teach segments of it which forced them to prepare more for the chapter they taught. He also paired good students with struggling ones in study sessions and projects. I guarantee you his students did better than other equivalent groups. Should they be punished for it?
Thank you for the replies everyone. For some reason I go through bubbles of no notification emails. Then I get inundated with 4 days worth. It’s very annoying. I thought this thread had died.
I understand what you are all saying. I guess it hasn’t come up in my >30 (frequently >25) size classes in Community College. One more thing to look forward to when I transfer.
I always thought that grading on the curve caused a class of all lower scores to get higher grades, but not if someone skewed the curve. I never thought that high scores would get reduced.
It depends on the curve. For example, some course grades at my high school were graded on the curve y = 10√x. With such a curve, you need a pre-curve grade of only 25% to pass, nobody can score more than 100%, and no grade will ever be lowered.
supposedly, my HS crush’s Chem class HATED her for wrecking the curve.
We’ve had threads in the past on what “grading on a curve” means, and how it works. If you need more information, check out threads like
Grading on a curve
Define: Grading On The Curve?
In my experience as an instructor, that has been my practice. I would never drop someone’s grade to fit into some curve. Maybe it’s because my tests are invariably on the difficult side so there aren’t too many high grades anyway. As to limiting the number of high grades or guaranteeing some failures in each class, that doesn’t happen with me either. Yes, there are emails from the administration about average grades and the fight against grade inflation, but I’ve never heard of any instructor who got dinged for his/her grade assignments. The fact of the matter is that instructors have very wide latitude in their grading. If an instructor is particularly harsh in his grading, it’s likely because that’s the way he does it, not because he has to.