At my university, most of the large-class courses are curved, with the exception that anyone who gets over at least, say, 60%, are guaranteed to pass if even if they’re the bottom of the barrel (they usually aren’t). Profs will also sometimes adjust the letter grades if there’s a cluster of scores (e.g. if the original plan was to give the top three students an A+ and the top six scores are 98.9%, 98.6%, 98.6%, 98.5%, 94.2% and 94.1%, they’ll bump the fourth student up to an A+).
This for me exactly. Just did semester one of 1L. Waiting to see where on that curve I fall…
Related to the thread: What is the purpose of grading on a curve?
I would think that the purpose of a class is to teach a topic up to a given level and certify that anyone who passed the class had that minimum level.
If the purpose is to determine who is best at it, then that can be accomplished by using averages over the course of the whole degree.
Grading on a curve seems to open the possibility of failing people who do well because they didn’t do as well as the rest (when nearly all students master the topic) while also making people who suck but not as much as the rest look good (when nearly all student don’t master the topic).
Me: Never graded on a curve in philosophy, anthropology, history, management, economics or law. I’m Canadian.
When I was an undergraduate in engineering it was widely believed that most classes were not only graded on a curve but that this was done in order to “weed out” some predetermined number of aspiring engineers.
A few years later when I went back to the same university as a grad student, I was assigned to teach (and grade) several of these “weed-out” courses. The department said nothing about any grading system or any need to fail or pass some number of students. In fact, all they said was, here’s the book and syllabus - go to it. I graded on a straight scale except a couple of times where I gave test questions that turned out to be harder than I’d intended. A few students objected to the grades I gave them, so I told them to go to the department chair if they wanted to complain. I never got any pressure from anyone to give any particular grade.
An interesting question would be whether there has ever been an undergraduate course in which there weren’t at least a few students who just couldn’t or wouldn’t do passing work.
I was told that I aced a high school History class final (an exam prepared by the Ministry of Education, not a school exam) but that due to an error in translation from the original French to the English exam, the English one was considered after the fact to be easier than the French one, so the English exam was curved down.
I ended up with a 94%.
I passed Statistics and Methods part I in Sociology as a grad student with a 50% test average. No one who ended up failing on the curve would have passed if it hadn’t been curved.
Yes.
My Introduction to Physical Anthropology teacher was a right cunt. He graded on a hard curve–the highest 8 scores (in a class of 80) received an A, the next 16 a B, and so on. I once earned an A with an 85% and once earned a B with a 92%.
There are a few professors I respect more, but not many.
I can recall verbatim one exam question from my undergrad years, posed by the above-mentioned professor: “Inasmuch as baboons are terrestrial creatures, they have Ischial callosities.”
The answer is “false.” I remember the howl that went up from the class when he explained that baboons do have Ischial callosities, are terrestrial creatures, and that the two facts are unrelated.
To the best of my recollection, he said "I expect college students to know that ‘inasmuch’ means ‘because’, and thereby connotes causality. If you missed the question, blame your high school English teacher, not me’.
Yes, I missed the question. I’ve never forgotten the meaning of inasmuch, though.
I am certain there are many of them. There are plenty of crappy colleges among the 3000 or so in the U.S. and any of them can have students that are going through anything from having a pressure breakdown to just not believing they had to do the work to pass. I went to Tulane as an undergraduate and professors there were more than happy to fail people even at the prices most parents were paying to send their snowflakes there. Many got kicked out of engineering and had to transfer to the liberal arts and some didn’t even make it there. Pre-med classes in general were brutal just to slap down physician dreams before they really took hold. That sucked if you just wanted to take Organic Chemistry or Biochemistry just because you were casually interested in it but didn’t want to spend 20 hours a week on one course.
I went to grad school at Dartmouth. It was a different story there. When I first served as a teaching assistant and had to do grading of my own, I was told that students were to be divided into two piles: A and B. I asked about assigning grades lower than that if the situation warranted it and the reply was that would not and could not happen. The students were all smart but some had some fairly liberal attitudes about class attendance and turning in work. I wondered how a few managed to get in at all even though they were smarter than average. I would have failed some of them if they were judged against their peers but grade inflation was policy then and part of an unwritten contract between the parents, students, and school.
It was very difficult to get less than a B at Harvard during the same time (1990’s and earlier) for the same reasons. They revised their grading scale to be much more in line with a curve (but not a true one) when it was publicized that the majority of students were graduating with honors. Keep that in mind when you read about politicians from JFK to John Kerry to George W Bush getting less than stellar grades at the Ivy’s when they attended. That was a remarkable achievement but not a positive one.
I think some of my undergrad courses were graded on a curve. There’d always be a few dumbasses, so the problem of a halfway-decent student still getting an F didn’t arise.
To be honest, i think that one of the few really good justifications for it is if you have multiple sections or lectures of the same class, taught by different professors. In such cases, agreeing upon a curve over all sections of the class can help to overcome differences in grading styles and harshness.
If you assume that, on average, each teacher should have roughly the same distribution of excellent, good, decent, mediocre, and shitty students, then it also makes sense to ensure that you don’t end up with a situation where one class has 18 A’s, 26 B’s, 18 C’s, 16 D’s and 9 F’s while another has 28 A’s, 38 B’s, 16 C’s, 5 D’s and no F’s.
Obviously, this works better as the class size gets bigger, because chance variations in overall student quality between classes become less likely as the classes get bigger. In two classes of 10 students each, you might, by pure chance, end up with almost all the “A” students in one class. That is far less likely in two classes of 100 students each.
And this can also work over time, from year to year, as well as between two classes being taught simultaneously in the same semester. A curve can ensure (assuming that your year-to-year student intake is of similar overall quality) that you don’t end up with a situation where the average grade for the same course is a C+ one year and an A- the next.
There is, in my own discipline (history) and in the humanities more generally, another way to deal with differences in grading style, at least when teaching the same class at the same time, and that is for the teachers to actually consult with one another. In grad school i TAed a survey class with two other grad students. When we sat down to grade papers and exams, we would spend some time swapping papers or exams with one another so that we could make sure that we were being fair.
I would select an A, B, C, and D paper (in my opinion) from my pile, and let the other two TAs read them. They would do the same. Then we would all talk about areas where we might disagree, and about the sort of things we were looking for in the work. If it turned out that i was grading my students easier or harder than the other two, we would make adjustments accordingly. This sort of consultation is especially useful in disciplines where there are not simply right and wrong answers, and where students write essays and other long pieces of work for their assessment tasks.
I don’t claim to be an expert in the pedagogy on the natural sciences or engineering, but i think that if my students, even students i knew were intelligent and capable, were consistently getting grades of below 50% on exams, i would question the quality of my syllabus and my exam writing rather than simply curving the class so that more people got a passing grade. There may be some pedagogical justification for making exams so hard that even good students “fail” them, but it doesn’t really seem like a very productive exercise to me.
Summer Econ 101 class. The teacher passed out 2 A’s, 2 B’s, 30 or so C’s, and several D’s and F’s. Hardest damn A I ever earned.
This is the reason given by my law school. All first year students take the same classes, but different professors are assigned to different sections - and you don’t choose professors, you just get a schedule and that’s that. The hard curve addresses disparity between sections, not disparity within sections.
Small classes (15 or less I believe) and “skills” classes (where you are practicing a task like negotiation, not just learning principles) are exempt from the hard curve.
Not in class, but our evaluation scores at work are done on a bell curve with the proportions handed down by corporate HQ. 80% are to fall in the “meets expectations” catagory, with a minimal raise. 10% are to fall in the “below expectations” catagory and recieve no raise. 10% are to fall into the “exceeds expectations” catagory and receive a higher pittance than the “meets expectations” crew. They even draw it out in bell curve format to show all employees where the score ranges put you. So basically, the managers have to fiddle with the evaluation scores until they get the correct numbers to meet corporate expectations.
StG
Most of the schools I went to graded on curves, with the class average of a test far more significant than the actual distribution of scores. I TAed an assembly language class which was the second CS class for most people, and which we did use to weed out students who thought they were whizzes because they could write Fortran code. Our first assignment was quite hard. We didn’t fail many people since the school had a late drop date. One thing to remember - if students do drop, assign them to the bottom of the curve and don’t renormalize it. That isn’t fair to those who stick around.
When the two methods meet there can be problems. I taught data structures at another school, who used straight grading. I wanted to push them so I gave a hard test, not really thinking about straight grades. My students freaked out when the class average was 60 or so, and I had to reassure them that they were not all going to fail. It did help me to identify the one guy who really got the material.
When I was teaching, I often did this after looking at the general distribution of the test results. My theory was that if everyone got, say, less than 80% right, then it’s possible that (1) something was wrong with the test or (2) I hadn’t taught the lessons as well as I thought.
My high school PE teacher graded on a true C-centered curve, around ability. Effort and improvement didn’t matter, only how good you were at the end of the section on that sport compared to the other people of your gender in your class. For each sport, the top student of each gender got an A, the bottom student of each got Fs, and the majority of us ended up with Cs.
Almost all of my undergrad was curved. After my first test in college, one of my friends walked out and said “I didn’t know whether to hand my test to the TA or just drop it in garbage.” Class average was a 26. Luckily for almost everyone, the prof blamed himself because with scores that uniformly awful, he assumed that there was something wrong with either his teaching or the test itself. They usually didn’t use a true C-centered bell curve, average would be a B. Averages and ranges were announced after every test so that you could figure out what your test score actually meant.
I had a structures design teacher who believed that if someone got all the questions on a test right he’d not made it hard enough. I saw test averages in the 30’s in that class.
Yes, my college Calc I class. I came out of it with an A at 63%. But theoretically, if everybody had scored 90% or above, 90% would have been an F and 100% would have been an A. That just never, *ever *happens. They deliberately style the exams so the average person gets less than 50%. That’s actually why I changed majors (from Physics to Computer Tech). I don’t like the mindset that nobody is “supposed” to master all of the material.
Actuarial exams are graded on a “sort of” curve. The Society of Actuaries vigorously denies that there is a curve, and insists that they grade based on the required knowledge. But if you pay attention to the way they describe their grading system, it’s ultimately a curve.
Essentially, they set the pass mark based on how difficult the exam is, and they decide how difficult the exam is by how people did on the exam as compared to other exams. So in effect, it’s a curve.
Which makes a difference sometimes. The Society has occasionally divided one exam into two parts, and at this point the curve has an impact. Because you could theoretically just decide to take both of the new exams, which are the equivalent material that was previously covered on the one exam. But it’s not the same thing. Because if most of the exam takers are going to be taking only one exam and you’re taking two, then you’ll be at a huge disadvantage. If there was no curve, then you would just have to do as well as the old pass mark. As a result of the splitting of the exam, the curve will shift higher.
Exams should be hard enough that you can reliably separate the A students from the B students (or even the A students from the A+ students, if you’re the kind of professor who believes in giving A+'s). From the point of view of exam design, the “best” exam is one where the class average is 50, the best student gets a 99%, and the worst student gets a 1% - this means that you have “spread out” the class along the curve in a way that lets you actually assess their abilities accurately.
I had a professor my freshman year who explained this to us, and then explained that he found that the best balance between spreading out his students so he could rank them and not being deluged with “it’s not faiiiiir” whines was at a class average of about 75%, so that was what he aimed for. He also pointed out that students really should be hoping for exams with as low a class average as possible. If the average score on a test is 10%, then you can never fall more than 10% below the mean, while you could have a really good day and be 90% above it. If the average score on a test is 90%, then you better be sure you’re not having a bad day, or you could get a C for a couple of careless errors that would normally be pretty inconsequential.
That professor also had a policy of figuring out grades two different ways: 60% final, 30% midterm, 10% labs, or 30% final, 30% midterm, 20% labs, 20% homework. He said that we could skip all the homework and take the risk of having the final count heavily if we wanted, but that his experience was that it was almost impossible to do well on the exams unless you had done most of the homework. Nevertheless, the choice was up to us. (Again, he said something like, “It always seems to make my class evaluation scores go up when I do this, even though it usually only affects two or three students of the 200 of you. It’s just one more column in my Excel spreadsheet for me, so why not?”)