Author and Wharton Prof Adam Grant has an interesting and provocative piece in the SundayReview (NYT) about grading on a curve. His article is about the effect of curve grading on hyper-competitive B-school students, but I’m sure the concept would apply to many undergrad programs too.
He argues that a curved grade fosters competition rather than cooperation. If only 3 As can be awarded, regardless of the level of knowledge exhibited by each student, the students are pitted against each other. Of course, one solution is to do what Harvard and some other schools do - give just about everyone an A.
Grant experimented with grading techniques and testing options that lead to cooperation rather than competition, and had some good results.
For me, the best kind of grading system would be absolute, i.e., an “A” means a high level of mastery of the subject. In a given class, maybe lots of As would be handed out. In another one, nobody might get one. (Though in the latter case, one might take a hard look at the prof and curriculum to see why the level of mastery was so low.) Looking at a student’s transcript would give you a good indication of their actual competence in a subject vs. how well they fared compared to their peers. But, I don’t expect this to happen. Give a kid a C today, and he might lodge a formal complaint. Right before the parents helicopter in.
Couldn’t agree more. The author gets off on a tangent about student leaders being the ones who help others, but the comments on the disadvantages of the forced bell curve are relevant for discussion, I think.
Aside from standardized test scores (MCAT, LSAT,GRE, GMAT, etc.), if there is no “Bell Curve” grading, what other criteria can be used to evaluate and determine who gets into very limited space in prestigious medical schools or graduate schools.
@UCBUSCalum I’d say GPA is more meaningful to those grad/med schools when it’s a result of mastering material/applying concepts - grading the student against the material s/he is expected to learn - than when it’s a grade based on doing better than the others in one particular section of one class at one college.
@boneh3ad I am surprised at that statement because the really hard (for many students) classes like organic chemistry, P Chem, physics, and many of the technical engineering courses often have exams where the absolute grades range from 30-70 and are curved up. I have seen this in school after school, including the most prestigious ones where the students coming in were among the brightest. It is in many of the humanities courses where the grades don’t need to be curved because the grades on the exams and assignments are often quite high, especially those where the grades are based more on subjective essays rather than objective right/wrong answers.
@UCBUSCalum I believe for this system to work, everyone would have to adopt it so one was always comparing apples to apples, at least to some degree. Of course, there is no way to know if one profs test for Econ 101 at Cal was a lot easier than the one given by another prof at Wake Forest for the same course, but that is true now. So if this system produces more people that have A’s than the curve system, which I assume is what you are saying could be the issue, the schools would have to dig deeper into each student beyond just the GPA and MCAT/LSAT/… And that might not be such a bad thing, as there should be more to being a professional in these areas than just how you did on tests.
@fallenchemist If we are limiting the concept of “curving” to the more traditional process of fitting class grades into a normal distribution and assigning grades accordingly, then I really don’t know any professors who grade on a curve. I know that curve has also come to be used commonly to describe scaling exam results based on the overall grade range, e.g. making the top score a 100 and scaling the rest accordingly. I don’t know any professors who grade using the more traditional meaning of curve above. I do know a few, but not many, who use the second “curving” method to assign grades.
More common (in my experience) is setting a grading scale at the beginning of the semester (e.g. the 10-point scale) and saying anyone who meets those score ranges is guaranteed the grade assigned to it in the syllabus, while reserving the right to adjust the scale accordingly (usually to make it more generous) as needed. That is far and away the most common method used by most that I know, and it is typically the one I use. It is basically saying “if you, the student, demonstrate a certain level of mastery of the material, you will earn the appropriate corresponding grade regardless of how your classmates performed.” If, at the end of the semester, it seems like students really struggled overall, it merits me going back and looking at the exam problems I gave, and if it looks like they were too difficult for what I was trying to assess, then I always reserve the right to go back and make the grade ranges larger or to shift them to make the scale more forgiving.
Thanks much for the clarification. I was indeed using the term with imprecision. I agree I do not know any that use the classic curving of grades, the first method you describe. I know tons that use the second, where the top grade is either literally adjusted to be like a 100 and the other grades adjusted accordingly, or they just say that 63-70 was an A, 54-62 a B, etc. I see no meaningful difference whether they use a factor to normalize the grades to a 100 point scale or not, although I concede there can be minor differences. Certainly setting a minimum number that demonstrates excellent understanding, another for good understanding, etc. a priori is different. Personally I have seen less of the last, but certainly do see it reasonably often. I personally favor it, btw. I agree that there should be some standard that should be met regardless of the quality of that particular class.
The OP is discussing the type of curve where it is predetermined that only 10 As will be given, and 15 Bs and 20 Cs and so on. It’s used, apparently, to weed out students, or in situations where the desired outcome is only a certain # of students with each grade, and so students are graded against each other (in other words, if 20 students score in the typical A range of say 92-100, only the top 10 of those grades (maybe those over 96) will get an A, the rest will get A- or B or whatever).
It’s NOT referring to the type of curve where everyone is curved up.
Back at my business school they used a 10/80/10 grading system for every class. 10% received a “1” - the highest grade. 80% received a “2” - satisfactory, but middle of the pack. 10% received a “3” - obviously the lowest grade. Accumulate too many 3s and you were in danger of disciplinary action (got a couple of those myself).
In the end, only about 1-5% of students end up getting disciplined.
ETA: Accumulate enough 1s and you were designated a scholar at graduation - highly coveted. It definitely caused some hurtful competition.
In my long-ago engineering school days, the bell curve approach described by Grant was definitely used by many profs, particularly in intro courses intended, in part, to weed out weaker students. With the importance of retention metrics these days, I suspect the concept of weeding out the weakest third of the incoming class is over for good.
^^In my 20-years-ago engineering school days, the weed-out was accomplished through the material and pacing and the pressure to schedule many credits each semester in order to graduate on time. Never had a single course forced to a bell curve, though plenty of professors used the “uplift curve” if they felt the initial exam scores didn’t reflect actual student mastery of the material.
^ Same. I’m an engineering student at a hard school and the vast majority of courses aren’t curved. In theory, this means all students could get an A, but in reality, it means most fail. A 40 on an honors thermodynamics exam is still a 40. Rough but curves are dying.
In a STEM class the goal is not necessarily “mastery” of the material, but a thorough understanding of at least some portion of the professors vast expertise. Of course this applies more to upper level classes. If your professor is up there doing a brain dump for 4 months and you grasp 60-80% of it well enough to answer test questions that is probably pretty good. I would rather be exposed to everything a top level research PhD knows than have him/her cap what he puts out there in an effort to make sure a high percentage of the class reaches “mastery” level.
When I taught a law school class for a few years, I was specifically directed to give no more than a certain percentage of As and a certain percentage of Bs. I was not required to flunk anyone, but there definitely was a firm curve in the sense of “you are competing directly against each other for the limited number of good grades.”
I didn’t like doing it, but it wasn’t my call. It definitely made the students unwilling to cooperate with each other in a course where cooperation would have been beneficial for all to learn the material better.
Traditional bell curve is used in my daughter’s engineering classes, also physics classes. It is a very competitive environment. Interestingly my daughter has found the older students, 30’s & 40’s, are more willing to collaborate than those her age, in those kinds of classes.
@fallenchemist : Actually, hat you describe in STEM is kind of a typical curve honestly. For difficult courses, the professor or department has a target average or median course GPA in mind (I usually see B-/B in elite schools so like 2.6-3.0 is normal) and since exam and assignment means are on the low end, the grades are fit to some distribution leading to that. By default since means in such courses typically range from 50-70, the 90 is going to get an A grade (more precise teachers will do it based on standard deviation). The problem with these grading curves is:http://digitalcommons.georgiasouthern.edu/cgi/viewcontent.cgi?article=1101&context=ij-sotl
Performance will often be determined by luck and raw talent in the particular subject perhaps even more so than hard work. However, I notice that studying for a difficult STEM (or any course giving exams and assignments on the higher ends of cognitive complexity) course usually requires different study habits that are efficient and conducive to higher levels of thinking. Many students do not get this and think that studying longer to simply memorize more is sufficient (like you would for an extremely content heavy course that requires little deep level problem solving or analysis) so waste lots of time doing things that do not develop skills to tackle difficult problems.
Business school curves are stupid IMHO in the sense that they often do cause scores to be adjusted downward versus a normal scale because many courses consistently have high averages. They are full of non-quiz and exam assessments and unlike a difficult STEM course, the quiz and exam assessments are usually not the types where students must truly “guess” and derive answers for stuff they have not been directly exposed to before. They are typically straight forward factual recall, basic understanding, and algorithmic problem solving. The times where you will see major distributions occur is when exams require a decent level of mathematical reasoning and logic (and maybe long reading passages will challenge students like they do on many harder STEM exams). Instead of putting grades on such curves, they should maybe adjust the level of the course so it automatically fits the desired distribution (which on most b-school and econ. curves at top schools is like B/B+ in a core class and consistently B+ in a non) or courses that are by nature just bound to be simple should not be graded on the curve. Easier courses should not warrant competition especially when group work is involve. Then the positive dynamics of a group are hampered in such an environment.
*Don’t ask why I’m up now lol. Gotta go grade lots of papers soon
@fallenchemist that’s interesting, because I repeatedly heard about the curve you mentioned about STEM classes, but none of my college’s teachers-Ochem, ochem lab, cell bio, thermodynamics, ChemE intro, etc-use bell curve at all AFAIK. They all use absolute grading system. They “curve” the score by replacing worst midterm grade with final test grade, but I doubt it’s the same thing as bell curve.
Some times I wish professors use bell curve but at the same time I m relieved that my scores are direct reflections of my effort. I wonder if it’s also frequent for professors NOT to use bell curve?