When I was going through school, grading on the curve was more the rule than any other system.
I question the reasons for grading on a curve…
Why must what students learn be be distributed using something that describes a random process.
If a teacher sets levels of accomplishment, and then teaches effectively where students learn, why shouldn’t the entire class get high grades.
Am I sounding too much like Wildest Bill?? 
For several reasons, I’d guess.[list=a][]Because, in the vast majority of human endeavors, the ability of people seems to fall into a normal curve.[]Because, IIRC, the SAT (by far the most-widely-taken test) seems to end up that way.[]It helps fight grade inflation.[]To piss off high school students.[/list=a]
Not meant as a defense of the system, but one thing that grading on a curve will do is take out possible test issues such as ‘it was too easy’ or ‘question 5 was very ambiguous’.
Well, the curve wouldn’t be necessary if the teacher taught at optimal efficiency, and the test perfectly measured the effectiveness of the teacher’s teaching. Both of these factors are out of the control of the student, but they can easily affect the absolute performance of the student on the test.
So the goal of the curve is to remove these 2 factors, and therefore give an evaluation of the student on some kind of absolute scale.
Also, most teachers (at least in my experience) curve up to fit the normal, but never curve down.
The main problem i see with the system is that a strict curve takes away a great deal of a students control over his/her grade. It bases their grade on the grades of other students. This isnt fair because a student might do fairly good work and test fairly well yet still fail, or get a low grade simply because there was a great deal of competition. Also, different students with comparible proformances in different classes (same subject/course) could get vastly different grades based on who their classmates are and how well they proform. Meaning student x’s A might indicate the same level of proformance as student y’s B, or C even.
Curve, smurve. Teach the stuff, test the kids, and get out of the way. Curves are a waste of time. If the entire class fails then maybe we could learn why they’re failing and how to teach better, instead of covering up the mistake by passing them all because the class dweeb got a C Minus.
grading on a curve forces the students to compete with each other in any way the teacher decides. if the book contained all the relevant material then theoretically the student could never come to class and still get an A, but if the teacher says things in class that are not in the book and test for that then the student is forced to come to class. if the teacer designs the test, s/he can make it a memorization of trivial details test.
in subjects like math and physics we should create a standard set of 1000 questions with 10 levels of difficulty. a computer randomly selects 2 questions from each level. no 2 students get the same test, use barcodes to specify answers so the tests can be checked by computer. the teacher just teaches and doesn’t test. the results of the tests show how good the teacher is. everyone can get an A by a dificulty set nationally. use MIT as the standard.
don’t change books year after year. physics doesn’t change below bachelors degree level.
Dal Timgar
I agree if were dealing with ability, but learning in some subjects isn’t necessarily an ability. Through grades K-6, the ability of the kids are about the same, and the material being taught isn’t so difficult that we should see a curve in ability.
The SAT is an acceptance test, not a test of what students learned. It is used to mesure the level of what someone learned, which across the nation is random because of different teaching requirements and styles. It is entirely different than say a midterm or a final in which we are measuring what students learned with all the variables approximately the same, and all in the control of the teacher.
Why is grade inflation a problem?
Depending on the subject, some students are hoping for a curve, and then they get pissed off at one kid, that actually studied, and busts the curve.
Another beef
Can I hijaak my own thread?
Why don’t teachers teach, instead of covering material. Seems that they tend to cover the material, and leave it up to the student to learn. Doesn’t the teacher bear some responsibility to ‘cause’ learning, instead of just covering the material.
It would seem to me that the curve reflects real life a bit better. It results in competition, which is generally a good thing. .
The idea of having to ‘cover’ a certain amount of material regardless of whether it is learned or not is repulsive, and possibly a major flaw in the educational system.
However, how does one “cause” learning? If a student doesn’t want to or isn’t interested in learning, how can it be caused by someone else? It certainly can’t be forced.
As MilTan mentioned, I’ve very rarely in my 19 years of schooling* seen an honest-to-God curve. (I have seen it once or twice, though.) What usually happens is more of a “shift”–either the grades themselves get a given amount added to them, or the grade cutoffs are moved.
This is a good practice, for two reasons I can think of.
1.) It allows for the possiblity that the test will be too hard. If a teacher fails to get a certain point across, or if he misjudges the level of detail the students were able to grasp, then a “curve” allows the teacher to adjust the grades accordingly after the fact.
2.) You can make a much more accurate test if you make it “too hard”. I had an organic chemistry professor (three times, for the same class!) who wanted his tests to have an average of around 50. He said that this allowed for the best spread of grades, and the best assessment of understanding. As nerve-wracking as it was at the time, I agree. He did set definite points–if you got an 80, it was a definite A, a 70 was a definite B, etc., but reserved the right to move those cutoffs down. (He usually did, by a great deal.)
As to an actual “curve”–no, I can’t defend that. It should be theoretically possible for the entire class to fail or to get an A. If that happens, though, it’s probably the teacher’s fault more than the class.
Dr. J
I’ve seen this only in a movie and I’ve always wondered if it were true…
In a medical school, where it seems like they encourage the weak ones to fail, the professor said, for an exam, “there will be 1 “A” given, 2 “B’s”, 2 “C’s”, and the rest fail.” (there were about 10 students).
In otherwords, the students competed against each other for the medal grades. I actually thought this would be the perfect way to ensure that only the best doctors will graduate, since society would only want the best.
But it also seemed unfair, because if you were a real good student who happened to be in a class with a zealous over-achiever, then you were destined to fail.
I used to teach at the university and community college levels. I always reserved the right to “curve” any particular exam, and add a final curve at the end of the semester if necessary. As other posters have noted, the point was not to get a bell-type distribution. In fact, the “scaling” of grades could only help the student.
I took the approach to challenge my students. Of course, they tended to whine a lot (“that test was hard!”). If a particular test had an average of, say, 62, with a class high score of 89, I would probably add 8 points to everyone’s test. Class size, subject, and level; as well as grade distribution, were always considered, so the actual scale would be situation-specific. The main thing is to be internally consistent within any given class.
At the end of the semester I would analyze the final averages and look for natural breaks in the distribution. Any additional curve could only work to a student’s advantage. That is, I always used a 10-point scale. At times I would maybe group the 87s-89s in with the As; but I would never ‘scale up’ to make the lowest ‘A’ a 93 or something.
So, the students are not really competing against each other. They are still competing against the 10-point scale.
I grade on a curve because if I, a grad student, graded my freshmen against my own idea of a well-written paper, there wouldn’t be any As and precious few Bs.
These students were placed in my class because they started off at roughly the same level – otherwise they’d be in remedial comp, or advanced comp, or they’d place out of comp entirely. So, given that they all came to the class with approximately the same level of ability, it seems fair to grade them on what they do with that ability.
Of course, I teach a subject where there are few hard-and-fast answers; in science or math I imagine it’s another story altogether.
Why grade at all? After all, grades (like curves) are artificial constructs that may or may not reflect upon what the student actually learned. I believe that what the student learns is more important than how that student fits into some standard.
That said, in my experience, curves have been reserved for when a group of students does exeptionally bad on a test, and the teacher figures that the test was flawed.
No, grades are a good idea. I spent a year at Hampshire College, an overpriced “progressive” school with no grading system. What it amounted to was that you had no way of being sure of your standing in the class, you didn’t have much to show for your efforts in the end, and 3/4 of the time your written evaluations from your professors just echoed what you wrote on your self-evaluation, because, realistically, a professor can’t be perfectly attuned to a class to each of the 100 or so students they may deal with. Now I’m at state school and loving my 3.75. Personally, curving “up” tends to irritate me; it allows too many slackers to slide by. Sure, I don’t mind the occasional boost, but I don’t like to see people who should just drop out being coddled.
Nav, you’re tackling some of the biggest issues in current pedagogical theory.
I’m currently earning my teaching certification at California State University of San Marcos in a program that is extremely reform-minded. One of the first things they started yelling about in the pre-requisite classes was the skewed focus in assessment (“assessment” = pedagogical jargon for “testing”).
The ideal the faculty and program teach to is that teachers have an obligation and an opportunity to see that all students succeed in meeting educational objectives. Therefore, an expectation of a bell curve distribution in grades is not just inappropriate; it’s downright harmful. We shouldn’t expect a portion of the students to fail. We should do everything necessary to see that each student succeeds.
As to “curving” a test - which I always took to mean “take the highest score and add the difference between it and a perfect score to every student’s test score” - I think it’s too easy for a teacher to continually curve test scores instead of looking at why the students aren’t succeeding or why the assessment does not accurately reflect what is being taught (or why what’s being taught doesn’t match the stated objectives). [And, Fretful, this is in absolutely NO way a slam against you. I’m sure that you are a fantastic teacher.]
One of the big issues we discuss frequently is the idea of “authentic assessment”. It’s verging on becoming a buzzword in educational circles, it’s discussed so much. The basic argument goes “there are many ways to find out whether or not a student has achieved the stated learning objectives. By using only standard testing formats, we are limiting ourselves and our students.” In California and many other states, there is a growing movement towards more authentic forms of assessment. In the school where I student taught this semester, all graduating senior must make a portfolio presentation to a panel of teachers, administrators, and community leaders. Putting together the portfolio and presenting have become a very powerful tool in measuring student progress.
Another big topic for debate are standardized tests. Standardized tests - like the SAT I - have their uses, but those uses are much more narrow than most people will admit. The largest problems with standardized tests are:
they are norm-referenced, forcing students to compete against each other to get that bell-curve distribution.
the test is rarely aligned with the goals and objectives the state, school district, school, and teachers have decided upon.
Space Vampire, the problem with doing away with grades is when they don’t replace it with some other formalized method of assessment and evaluation. For instance, a college could conceivably do away with grades if the students received a formal written evaluation by the instructor or went through some sort of proving test necessary to receive credit. Just doing away with grades, though, won’t do it.
Now, anyone want to ask me about student-centered learning, multiple intelligence theory, or multicultural education?
[QUOTE]
*Originally posted by Wrath *
**I’ve seen this only in a movie and I’ve always wondered if it were true…
In a medical school, where it seems like they encourage the weak ones to fail, the professor said, for an exam, “there will be 1 “A” given, 2 “B’s”, 2 “C’s”, and the rest fail.” (there were about 10 students).
**
I believe the movie was Flatliners, although I can’t justify knowing that nugget of information.
At my medical school, this is not the way it works. The top 10% got honors, the next 70% passed, and 10% marginal passed, and 10% failed. If more than 10% got above 90/100, then usually more than 10% honored. Granted, we weren’t on an ABCD system of grading. Medical schools have a very high retention rate and it is not their aim to fail students, in my experience. It is very difficult to do well in medical school, but not very hard to pass.
I have just finished grading the exam for the class that I TA’ed (graduate school Genetics B). We made the test difficult and lengthy. The top grades in the class were around 93, and the average was around 75. I don’t know yet if we will need a curve. A good test is one in which nobody gets 100 and nobody gets 0, the peak is around 75, and it fits a bell curve. Since we only really have A, B or Fail in graduate school, we will move the B/Fail cutoff to 1 standard deviation below the mean.
The justification of a curve is as follows, IMHO. You have to define the term “failure” very carefully. Do you want to make someone repeat a class because they only knew less than 70% of the information? Probably. Can we write a test to accurately determine this? Probably not. At what point do you want to insist that a person needs to repeat a class? How can you determine from your test where that cutoff is? All of these things can be determined by looking at the test numbers, determining a mean, and determining a standard deviation.
So, the mean of the class (at least in graduate school where we don’t worry that the entire class is underacheiving or stupid) should be somewhere above passing, to control for teaching. You could pass if you were within a standard deviation of the mean. You could get an A if you are greater than a standard deviation (or two, we have to crunch the numbers) above the mean.