I am surprised that this “grading on a curve” BS is not banned yet. The students have my full support here.
For those that don’t realize what this is about - grading on a curve takes the best result of the exam (whatever that may be) and then distributes the marks to the rest of the class using either a bell curve or something approximating it working on the assumption that the exam results have to be normally distributed (in the sense of a Gaussian distribution).
Unfortunately, this is terribly flawed on two fronts. First, there is no reason why the exam results should be normally distributed - that would apply if you took a random sample of the population as your class and administered the exam there. However, the students in the class are not a random sample! (for all sorts of reasons - course choices, admission policies, gender/race balance, etc.). So this assumption doesn’t apply (statistics 101).
The second - and much worse - issue is that “grading on a curve” doesn’t evaluate the student ability at all. If you are in a class of 100 and 80 get a better score than you, you will end up with a terrible (possibly failing) mark because you “fit” into the last 20%. However, your performance was possibly only very slightly (e.g. by a point or two) worse than of those 80 in front of you. No matter, the curve grading will force you to fail the course because it forces the distribution of marks regardless of how the students have actually performed on the test - there are only so many good, average and failing marks “available”, so someone has to get the failing marks too. Basically it is impossible for everyone to pass such test from the get go, whether you pass or not doesn’t depend on what you know but on how many better performers you happen to sit in the exam with.
An even worse absurdity happens if everyone achieves the same score on the test - now the ordering is arbitrary and how do you distribute the grades using that curve? Who gets to pass with a distinction and who fails?
This is the same nonsense as the various “stack ranking” schemes some companies use - rank the employees and fire the bottom 10%. Again, it doesn’t evaluate any individual capability, only ranks/orders them according to some artificial criteria.
Exams and assessment in general are not supposed to create leaderboards but to evaluate what the individual students have learned and also to provide them with feedback on it. That’s why these “grading on a curve” schemes are so flawed - they completely miss this point, they only the students using some arbitrary criteria which have nothing to do with their actual ability.
(background - former university teacher with a course in university pedagogy under his belt …)