45

In the next semester, I will be teaching for the first time a seminar course which is composed of about 25 students. I am not sure how to respond to prospective students who have asked me about the grade distribution of the course.

For courses with large enrollments, I have the impression from my colleagues that I have to follow the "typical grade distribution" at my university, which is something like 25% As, 50% Bs, 20% Cs, and some Ds and Fs if the students really did poorly.

However, for the new course that I am teaching, I am wondering if I will be given more leeway in the grade distribution because the course has a small enrollment. I would like to give grades based on the students' performance.

  • So if many of the students perform well, I would like to give 90% or even 100% of the students A grades (A-, A and A+).
  • Conversely, if most of the students perform poorly, I would like to be able to give a low proportion of A grades.

Every semester, a committee made up of members of my department which will meet to discuss the grade distributions of each and every course offered by our department. I am hesitant to recommend grades that deviate too far from the "typical grade distribution" at my university because I am afraid of receiving pushback from the committee members.

Questions:

  • Should I fight for the freedom to assign grades according to my academic judgment, without necessarily following the "typical grade distribution" at my university?
  • Is this a fight that I could win, and if so, how do I go about winning it?

Update

I had forgotten to include an important and relevant piece of information. The goal of the course is to teach students how to analyze and present business case studies. Consequently, by its nature, the evaluation of the students' performance in the course will tend to be more subjective, rather than objective (as it would be for a course on say, calculus).

The grading of the course is based purely on continuous assessment; we do not have a final exam for the course. (Given that the purpose of the course is to develop students' presentation skills, a written examination does not seem to be the right way to assess students' learning.)

  • 83
    If students ask about grade distributions in advance of a course, that points to a very unhealthy attitude towards this at the department level. Grade distributions are something that happen on their own, not something that are determined in advance. For large enough courses, one can assume that the student population approaches something like a representative sample and adjust grades accordingly based on exam performance (to adjust for the possibility of the exam being too hard or too easy), but for smaller courses this would be absurd. – Tobias Kildetoft Oct 11 '17 at 9:38
  • 114
    @louic "The purpose of an exam is to separate "good" from "bad" students." No it isn't. The purpose of an exam is for the student to show that he/she has successfully learnt the class material. – orlp Oct 11 '17 at 12:01
  • 7
    @orlp. I disagree, as do many others. An good exam tests intelligence, insight, and the ability to apply the theory a student has learned. A university is not a school. – louic Oct 11 '17 at 12:16
  • 7
    There is no one set "purpose of an exam". Personally I find the "separate good from bad students" purpose distasteful. It promotes unhealthy competition. Most likely your exam is not actually a valid instrument for making that distinction (how many good students will be identified as bad and bad students identified as good). It promotes stress - I have to cram for the exam b/c being identified as a bad student is a bad thing. – emory Oct 11 '17 at 13:24
  • 13
    @Iouic "the purpose is good exam design" is tautological. Of course, the exam should be designed well. Anyone who says the exam should be designed poorly is clearly wrong. It is not obvious what the exam should measure or what the level of a course means. – emory Oct 11 '17 at 13:54

12 Answers 12

75

You won't get into trouble if all your students earn grades of A. It has always been my hope that I'd have a class with all A grades, or even all A and B grades. It has never happened.

If you "give out" all grades of A, someone, perhaps your department chair or the committee you mention, is going to ask you to explain yourself.

Your course (presumably) has a set of learning outcomes. You should assess your students against those learning outcomes, at a level of difficulty suitable to the students' standing, e.g. a greater level of difficulty for seniors than for freshmen. When you've done that, the grades will take care of themselves, and if everyone earns an A, you've done an outstanding job with the course!

Edit: based on the update to the question: For the purpose of providing useful formative feedback, you need to assign a grade to each piece of work you assess. Doing so has the added benefit of covering your posterior. Such grades should be assigned according to a rubric that the students have seen before beginning work. There is a brief example of developing such a rubric here.

Even in a course such as you describe, grades can be assigned granularly and with objectivity. If you've done that, your students will thank you and your defense when questioned by chair or committee is that the students got the grades they earned and earned the grades they got.

  • 14
    If a class does get all As, you will most likely be asked to defend yourself on why. While grade distribution may not be enforced (learning outcome should be best measurement), it is often used as a soft measurement to make sure teachers aren't too easy or that the content of the class is in-depth enough. – Confuzing Oct 11 '17 at 18:56
  • 10
    "It has always been my hope that I'd have a class with all A grades, or even all A and B grades." Teach graduate classes. Seriously. – Pete L. Clark Oct 11 '17 at 19:21
  • 2
    I suppose technically only the higher-ups at the OP's department and/or university can promise that they won't get into trouble for giving out all A's. I could imagine a department having a policy that prohibits this, even though it would not be a very smart policy. – David Z Oct 11 '17 at 23:27
  • 4
    I am aware of one case where all students actually all got A grades and there was a long discussion with the examination office until they where finally accepted. So yes, you probably will have to explain yourself! – koalo Oct 12 '17 at 6:02
  • @PeteL.Clark I know. Sadly, I didn't get to do very much of that. – Bob Brown Oct 12 '17 at 11:45
29

There are three reasons your students are all getting straight A's:

  1. Your testing methods are not good enough and just about everyone can get an A, even if they didn't study or know the material well enough.
  2. You are just that good of a teacher.
  3. You have many, many brilliant students.

The first reason is the only one that could get you in trouble. The others are not a concern. You might be asked about it and asked for proof that the methods were adequate, and you should be prepared to present them.

Just keep note of what was given in class and what was given as homework. The exams should be a good indicator of whether it was easy or not to someone who is remotely relevant to the course.

  • 12
    This is a good answer. But I would modify 3. as follows: 3'. The generally agreed upon standard of student performance for an A grade is set at a place that the majority of students taking the course will achieve. This is not the same as having "brilliant students" unless your definition of brilliant is getting an A in this course. I have encountered university courses for which most of the students get an A. This is not so common in STEM fields, but in some other fields / some other specific courses, it does happen. – Pete L. Clark Oct 11 '17 at 19:18
  • 1
    @PeteL.Clark Grade inflation was so bad even 15 years ago at my conservative university that one STEM course I set the threshold for an A at 75% so the professor could keep grading standards high but not sink all of the scholarship students taking a difficult major. – chrylis -on strike- Oct 11 '17 at 21:19
  • 3
    A fourth reason plausible is that the class material is just that simple. – liori Oct 12 '17 at 0:01
  • 2
    @liori That's just a less flattering rephrasing of "You have many, many brilliant students". – Mateen Ulhaq Oct 12 '17 at 3:46
  • 2
    @MateenUlhaq, I don't think so. "many bright students" can change due to demographics (e.g. next year the class becomes less popular and bright students go elsewhere—something the faculty can't control). "material is simple" suggests something the faculty can control: curriculum can be changed, regardless of popularity of the class. The mechanics of the cause are different, even if their observable result might be similar. – liori Oct 12 '17 at 15:43
13

Should I fight for the freedom to assign grades according to my academic judgment, without necessarily following the "typical grade distribution" at my university?

Is this a fight that I could win, and if so, how do I go about winning it?

I'm approaching this question a bit differently, instead of pedagogy let's talk about office politics. I'd humbly suggest fine-tuning this mentality to steer away from being adversary. It's indeed true that most of us found these committees rubber-stamping and paper-pushing, some of them do work as a good quality control and even resources of advice.

However, for the new course that I am teaching, I am wondering if I will be given more leeway in the grade distribution because the course has a small enrollment. I would like to give grades based on the students' performance.

This is the main reason I suggested you to work with the committee. We can't decide if you can have the leeway, they can. I'm not sure what's the office climate is at your place, but generally, I'd suggest:

  1. Prepare your syllabus that has a fully developed assessment scheme. If this is the first time, it's better to have more assessments each with a lower contribution to the final grade. They'd provide more occasions to adjust your question/grading style and opportunities for students to get your style and better understand your requirements.

  2. Schedule a meeting with the chair of that committee, present the scheme. Ask for feedback and express concern about possible fluctuation in grades compared to the historic record; the chair may give you some expected range of change which would be useful. Listen to the chair's advice and concern. Try to reach a conclusion.

  3. After the meeting, send an e-mail follow up to summarize the decision, attach the scheme (or revised scheme.) Save this e-mail.

  4. When class starts, let students know of the expectations required in order to excel. Grading policy and rubrics may be useful to share at this point.

  5. Teach and assign grades. If anything seems awfully off, talk to the dean and/or the chair of the said committee.

6

From a quick glance through, I don't think any of the answers point out the following:

Grade distributions tend to vary a lot from course to course.

If you're being told about grade distributions from say large introductory courses, the point of this so you can see if you're giving grades in a way seems to be consistent with other classes. (Note grade distributions can vary quite a bit for different large classes, or even the same class in different semesters---e.g., students tend to do better in a Fall Calc I than a Spring Calc I at my school, for a few obvious reasons.)

If you're not being told about grade distributions for the kinds of classes you're teaching, then that likely means there are no set expectations for what kind of grades you should give out. Note the population of smaller, advanced classes is quite different than that for large introductory classes, which means that such classes tend to have quite different grade distributions. Also, when the grade distribution is a statistically much smaller sample, one-off grade distributions can't be used to conclude too much.

That said, I agree with Bob Brown's answer completely.

5

It probably depends on your department policy; however if there was indeed one that made you give away certain percentages of each grade it would be unfair to the students. A person should be assessed on the basis of that person's performance and the grade should not be dependent on external circumstances of other people's performance (unless we are talking about group projects). As others have pointed out, grades distribution should occur naturally. If you feel that you've given too many good grades that may indicate that your course is too easy and you should maybe consider increasing the difficulty next year.

In the beginning of courses that I teach I usually tell my students that they will earn points (homeworks and exams are the basis for scoring) and their grade will be based on the total number of points. In my country there are three passing grades and one failing and I communicate the thresholds that are connected with each grade (if one scores < 50% of points - that person fails the class; the other thresholds are set at 75% and 90%). I depart from the practice of not taking group performance into account when there are many failing grades - in that case that may mean that the course was too difficult and thresholds may lowered (that possibility is also communicated to the students at the beginning; however it makes me wonder if it doesn't affect student's discipline negatively). I would never raise them though if I gave away many highest grades - I would be happy if that once happened and attribute that to my exceptional teaching performance.

As for the case of getting into trouble after giving away too good degrees - that may, again, depend on your institution policy, but as my experiences go, I have never heard of anyone getting in trouble for that. I have known a number of lecturers that are in fact known to never fail a student and nobody is making problem with that. It's the other way around - one of colleagues had a talk with dean after he failed his whole class. If you are afraid of pushback from the faculty you may consider asking more senior and experienced members on the issue as well.

  • Glad someone brought up departmental policy - it really does vary. At my Uni some faculties enforce grading according to a distribution, while others do not. – beldaz Oct 12 '17 at 19:36
4

Is this a fight that I could win, and if so, how do I go about winning it?

This is a fight you could lose. See the example of Denis Rancourt who was suspended for giving A's to every student in his physics class.

Should I fight for the freedom to assign grades according to my academic judgment, without necessarily following the "typical grade distribution" at my university?

I would want to make sure my faculty association would support me. Is this something you have at your institution? I recommend checking in with them. In general, I definitely support "assigning grades according to (your) academic judgment". The bell curve simply cannot show up in each and every class, and forcing or coercing faculty to push every class under that curve is a mistake.

As a point of personal experience: I remember being told (as a graduate teaching assistant) that I was required to fail at least 40% of my college algebra class, or I would not be allowed to teach it again. That turned out to be hyperbole, as I only failed 20% and went about my business otherwise.

100% A's, on the other hand? I am positive I would have had a "talking to".

3

The main concern here as I see it is:

typical grade distribution

This requires further explanation as to how the "standard" is set up. There are several scenarios that can exist.

Fixed material: For example, if you are teaching specific course with specific requirements like Calculus I for example and at the end of your course 100% of students are able to show that they learned 90-100% How can you give some one a C? Your only option in making such course more difficult is by including content from Calculus II. This is exact area, you don't have to "fight" it, just present it as self evident fact that can be proven by looking at results of Exam papers.

Flexible material: This is for courses like Literature, Sociology, Politics, certain Core/Intro courses. (Although even these courses often have fixed departmental boundaries) However, here if all your students get A's that could mean two things. Either they were all smart, or you've made your course too easy by not including enough/complex materials. This is a grey area. If I'm the chairman and also a professor in the same department. And my students get much worse grades than yours, (or to put it in ego's words I'm a bad professor and you're a better one) I could just present it as a "fact" that you're being too lenient and too easy on your students. But here too, Exam papers are your proof of understood knowledge.

Whichever the scenario may be, make sure you conform to the pre-approved curriculum/syllabus and your Exams reflect that the students have actually grasped the materials you've set to cover. I would then ignore this "typical grade distribution" system and play it dumb asking (whoever challenges you) for advice on how to make things more "challenging".

Bottom line is, there is only so much material one can present in any given course - time being the limiting factor. There are some good professors/teachers and some bad. I've had both kinds, and everything in between. It's easy to tell which is which. The better the professor the less time students have to research the course-related information looking for answers elsewhere.

2

I would think this depends on what sort of institution you're at. There's a lot of variation. Once as a graduate teaching assistant, I was told that the several TAs leading discussion sections under one lecturer would be largely independent of each other, but that "If everyone in your section gets an 'A', there will be a row." This was not a serious possibility, unless it was.

Once a student newspaper and then a local regular newspaper published an article about a student threatening a lawsuit because she got an "A" where she deserved an "F". She was sure the grade was intended as a bribe to stop her from complaining about the professor. Early in the term she complained about lack of sufficient feedback on her performance. Then she dropped the course. But she didn't officially drop it because that was too much hassle; she was still registered. The professor decided everyone except one weak student should get an "A" and that one should get a "B". So he located that student's name on the list and wrote "B" and then entered "A" after all the other names, without looking at them. This student had fallen off his radar by not being there. After the matter got public attention, the department head intervened and changed the grade to an "F".

Three times I taught a course for (about a dozen) engineering graduate students in the joint program of MIT and WHOI, where I gave some feedback on homework but no scores on homework and no quizzes or exams and at the end certified that everyone passed. This wouldn't work at a community college nor even at a fairly well respected flagship state university.

2

Have previous test(s) results as a proof of fairness of your grading

First-hand experience: I went to a very selective college, they intentionally do not use "gaussian grading" on the behalf (which I fully share) that, since they are preparing students for the outside world/for higher academical achievements, it is not fair for individual scores to be influenced by how "good" the other classmates are, since a "class", for how big it may be, it is just not a representative sample of all the students in the field.

In a course full of talented students it is fair that there will be a lot of "A", in a less gifted/less previously prepared class it is absolutely ok that most of them will have positive but not so great grades.

In a course that I followed the group was less than excellent, as a result the negative results on finals (we have no mid-semester assignements or whatsoever) were over 70% of the total.

When challenged by a few students the professor simply showed to the board that the exams were really similar to those from past years, the material provided the same and the only real difference was, well, in the results. End of the thing.

What you need to do, if challenged, is the exact same thing (albeit with opposite data). This way you will show that you simply encountered a very good group of students and are giving them the high scores they deserve.

There is no point in lowering grades of very good students because they happened to be in the company of other great individuals, and the opposite applies too. It leads to questions like "Yes, X has good grades, but was he a good student or just happened that the other ~40 people he was in class with were just not that brilliant?"

EDIT: I feel adding my comment to another answer as part of this answer.

In my graduating group there was an abnormally large amount of great students who went on to purse phds or got into great companies, both with excellent results. There were a LOT of top marks in a lot of tests. It would have been very unfair if tests were made harder to have the grades fit a gaussian and thus have way less excellent students who happened to be in other "less gifted" groups get higher grades and get ahead of them when competing for doc.s, positions in r&d etc. The assumption that a group of ~50 people should fit the general normal distribution for whole field is a HUGE one

1

I had an incredible professor in college. He told us that, by definition, most people are average, and a C is by definition an average grade. That's how he would grade on a curve.

He expected to see a distribution like this:

 F       D      C      B      A
0-10%  0-10%  50-75%  5-10%  0-5%

If he saw too many A's he felt like he needed to increase the difficulty of the tests.

If he saw too many D's and F's then he needed to decrease the difficulty of the tests.

It was kind of interesting - there were a few of us who were paying for an education. We worked hard, studied, and got A's and B's on the tests, and I suspect we actually learned, compared to other students in the class.

He also ensured that there were extra credit opportunities - worth a full letter grade in the class.

And he also encouraged us to come by during office hours pre-tests to make sure that we had prepared enough, and he would happily help us improve where we lacked.


That contrasts with the "cheerleader stats" course that I took, where we could bring a full 8.5"x11" sheet of paper, one side totally full of notes to the test (coupled with my insane courseload that particular semester)... I got an A in that class, too, but not because I actually learned anything. I actually feel bad about that class because there were a lot of interesting things that I know I missed out on.


All of your students have different motivations for attending your class. Some want to be there because they want to learn. For others, they're just doing it because that was the next thing to do after high school, and do we really have to come to class every time? Uuuugh!

Decide what kind of teacher you want to be, and then tailor your tests and curriculum around that.

  • 12
    He might have been a great teacher, but his grading policy is nonsense. There is no reason to expect a normal distribution of student performance, nor a distribution with a slight skew toward better performance. The grades should be calibrated towards the learning goals, not toward some theoretical distribution. – henning Oct 11 '17 at 16:58
  • 2
    Well, as a philosophy professor, I think he had a philosophy that most students were interested in getting the maximum grade for the minimum effort, and he wasn't interested in teaching them. Though I could be wrong about that :) – Wayne Werner Oct 11 '17 at 17:19
  • 1
    In my graduating group there was an abnormally large amount of great students who went on to purse phds or got into great companies, both with excellent results. There were a LOT of top marks in a lot of tests. It would have been very unfair if tests were made harder to have the grades fit a gaussian and thus have way less excellent students who happened to be in other "less gifted" groups get higher grades and get ahead of them when competing for doc.s, positions in r&d etc. The assumption that a group of ~50 people should fit the general normal distribution for whole field is a HUGE one. – Caterpillaraoz Oct 13 '17 at 11:14
  • It is certainly not true that by definition, most people are average. – Michael Hardy Oct 26 '18 at 23:41
  • What exactly is "extra credit"? There are certain opportunities provided for students to do things that will improve their grades in a course. What is the effect of calling some of those opportunities "extra"? – Michael Hardy Oct 26 '18 at 23:43
1

As a student in Germany, I learned that professors seem to try to adapt grades in classes to the Normal distribution, So that "average" students are in the center and the rest move along with them.

The problem hits when trying to apply this to different classes and you got a very good one, a "normal" one and a bad one. now the curve is normal among these classes but not the good or the bad class. But is it good to change teaching in the good or bad classes to get the normal distribution inside of these? That's what most teachers tried to do during my time in school.

But I think it's just a thing of mileage, by getting older and having taught for a long time, it's easier to get it right. I don't think you can hit it in the beginning. Your best bet is to ask the students from time to time if it's good or not.

  • 1
    Actually, during my time working in Germany I was expressly encouraged not to adopt such a style, but instead find a minimum passing grade and an appropriate single "delta" for the transitions to the higher grade levels. That way there's minimal "I deserve a higher grade" ranting after the exam grades are posted. – aeismail Oct 13 '17 at 23:47
  • That's how it should be, but I got a lot of old and new teachers that tried to get this curve. – alsternerd Oct 30 '17 at 8:39
0

It depends on how "trouble" is defined.

If you are tenure track but pre-tenure, consistently giving out a very high percentage of A's might cause trouble for you that you never saw coming. A senior faculty member once told me that the problem with giving too many A's is that, when I go up for review, my reviewers will think my classes are "fluff". It is natural to think we need good teaching evaluations, but I get the sense that departments want good teaching evaluations together with evidence that you are challenging your students and holding them to the high standards of your institution (as well as being an effective teacher and other things...). Perhaps unfortunately, grades might be used as such evidence (despite how arbitrary the grades may actually be!)

This answer is not advocating giving artificial low grades, but rather highlighting a potential pitfall of designing courses in such a way that most students will get an A.

protected by StrongBad Oct 11 '17 at 21:21

Thank you for your interest in this question. Because it has attracted low-quality or spam answers that had to be removed, posting an answer now requires 10 reputation on this site (the association bonus does not count).

Would you like to answer one of these unanswered questions instead?

Not the answer you're looking for? Browse other questions tagged or ask your own question.