64

In my undergraduate Physics courses, I have heard from my Professor last semester that he got in trouble from the physics department for passing too many students.

This semester, the TA in the lab portion of my class (who grades our lab reports) came flapping around a memo he got, also from the Physics department, stating that the average grade needed to be 75% and he will now be grading accordingly, after a few weeks of normal grading.

I typically spend up to 4-5 hours on these reports, typing up formulas and doing analysis, calculations, etc. I have gotten a 100 on every lab report in my first semester and up to that point in the second. Now, after the ultimatum, I'm losing points for things I didn't before, and being asked for additional analysis that has never been required. I already spend as much time as I can, and this harsh grading feels out of nowhere and undeserved.

Questions:

  1. Is this normal that teachers have quotas?

  2. Should it affect students who actually put in the work?

  3. Does this seem like an ethical issue of arbitrary grading that I should bring up to the department or my professor?

Nat
  • 6,227
  • 3
  • 23
  • 45
dattebane
  • 611
  • 1
  • 5
  • 6
  • 4
    Is it certain that the professors have been given actual quotas, or is the department asking only that their average grades come out to a C? There are several ways of doing this, which include grading more harshly, giving more difficult assignments, giving more work, and indeed quotas. If no particular method has been prescribed, your professor has some leeway in how they choose to meet that target. – Alex Reinking Mar 16 '18 at 07:14
  • 1
    Related question (linking to my related answer...): https://academia.stackexchange.com/questions/97177/would-i-get-into-trouble-if-i-give-out-too-many-a-grades/97300#97300 Hope this helps. – Caterpillaraoz Mar 16 '18 at 08:42
  • 61
    This is called curve grading. It's widely considered an educational anti-pattern (at least my text book in Higher Education Pedagogy calls it that), which of course does not stop some administrators from asking for it because it leads to "nice" grade distributions by construction, independently of how sucky tests are designed. – xLeitix Mar 16 '18 at 09:08
  • 6
    In the UK this is the norm. Often as a TA, I graded everything, and then they took my grades and made sure that the semi-Gaussian distribution that arises had specific mean and std. – Ander Biguri Mar 16 '18 at 09:58
  • 11
    @AnderBiguri - It's certainly not the norm everywhere in the UK. I've never heard of it being done in schools (and many of my family are teachers), and it wasn't done in university where I attended/taught. – Guy G Mar 16 '18 at 14:16
  • @GuyG Interesting. Even if it is not the norm, it is a commonly widespread thing in Universities. I am not a Brit, so I only know it in Uni environment, but I've seen this a lot. – Ander Biguri Mar 16 '18 at 15:00
  • 2
    If you got 100% on EVERY lab for 1 and half semesters, and especially if that h append to other students, that is a problem which needed to be fixed. It meant that either the lab was too easy, or the grading is too lenient. Based on your complaint, the grading style was too generous before, and they fixed that. – Nick S Mar 16 '18 at 16:45
  • @NickS I can't say I agree without more information, given we have only one example from the class in question. If he's been at the top of his class for two semesters, that seems perfectly reasonable. It's also unclear whether these labs are best compared to tests/homework/projects, as each category tends to have it's own distribution. – TemporalWolf Mar 16 '18 at 22:27
  • 9
    Normal? Completely. Ethical? Debatable. Do I think it's the right/best thing to do? Absolutely not. – Lightness Races in Orbit Mar 16 '18 at 22:44
  • @GuyG: In my experience, while it may of course not be applied "everywhere", it certainly is the default approach across the UK, both in schools and in Universities. Or, at the very least, every one that I've ever been enrolled in, or had friends enrolled in who discussed the topic, and so forth. – Lightness Races in Orbit Mar 16 '18 at 22:45
  • 5
    Everyone seems to have answered the "normal" part of the question but not the "ethical" part. Regardless of whether it's normal, it's highly unethical. – R.. GitHub STOP HELPING ICE Mar 16 '18 at 23:22
  • I won't speak to undergraduate matters, which is the topic of the question here, but grading on a curve is the norm in U.S. law schools. – TimothyAWiseman Mar 19 '18 at 01:32
  • At my alma mater (US elite private school) virtually every course was curved. However, it was exceptionally rare for the curve to lower your grade in the way that you’re describing. Instead, it was used to allow professors to assign very difficult assignments without worry about destroying students’ grades. Several professors claimed this was desirable because more difficult exams make it easier to tell the difference between the people who are good at the subject and the people who are really brilliant. – Stella Biderman Mar 19 '18 at 14:25
  • 1
    The title of this question gives a good illustration of why I find the "slash construction" so frustrating. – Spiny Mar 19 '18 at 14:58
  • A follow-up to my previous comment: the kind of curving you’re describing is unheard of in my experience. – Stella Biderman Mar 19 '18 at 18:39

7 Answers7

66

Unfortunately, the universities and colleges are run by managers and not only by academics. In ideal world, all students' work should be marked according to its merit; for example, if all mathematical problems are solved correctly, then the paper deserves 100%. Academics usually understand and share this view (apart of a few strongly believing that "no work ever deserves 100%").

Unfortunately, people who manage universities, are not always guided by common sense or mathematical reasoning. They seem to invent some obscure metrics and force them on the rest of academic staff. Academics do not think it is correct, but they have to play by the rules which are set by non-academics.

Unfortunately, students become a collateral damage in a clash of two cultures: academic and management. If you want to query this, best ask not your professor, but a Dean (anonymously or as a part of larger group). The misguided policies arrive from the very top; your professor has not much power against it.

Dmitry Savostyanov
  • 55,416
  • 14
  • 140
  • 202
  • 23
    Where in the question does it say that a non-academic made this decision? It could be that the department chair (an academic) was unimpressed with the professor giving nothing less than an A in a term. – Alex Reinking Mar 16 '18 at 07:15
  • 13
    Believing that "all students' work should be mark[ed] [according] to its merit" is not incompatible with the disbelief that the work of every student in a large lecture merits an A. The particular goal (average of 75%) might be satisfied by simply making the course as difficult as other existing courses in the department. – Alex Reinking Mar 16 '18 at 07:20
  • 17
    I think the distinction depends on what one considers to be the "point" of grades. If criteria are given based on understanding of the course content (A students should understand 90% of the course objectives, B students 70%, etc.), then a good lecturer teaching a well-designed course might expect to give 90%+ of students As. If, instead, grades are intended to allow one to differentiate between students in a cohort, one would expect grades (on average, perhaps over multiple years) to follow a normal distribution. Either way, this should be consistent across the university. – georgewatson Mar 16 '18 at 07:35
  • 5
    @AlexReinking You can adjust the difficulty to make expected average 75%, but you can't always make real papers to meet your expectation. Problem is, managers kind of force exactly this. – Dmitry Savostyanov Mar 16 '18 at 07:54
  • @georgewatson - Yes, that makes perfect sense. I personally fall into the "understanding of the course content" camp rather than the "differentiate between students in a cohort" one, but regardless I agree that it should be consistent across the university. – Alex Reinking Mar 16 '18 at 07:56
  • 10
    A very important point is to go to the very top as a group and protest this. It works wonders, especially in US. –  Mar 16 '18 at 09:46
  • 1
    @gergewatson Across the university? Maybe across the faculty, but even that will be hard. I have no idea how they teach and mark in medicine or theology, or classical philology. – Vladimir F Героям слава Mar 16 '18 at 13:01
  • 1
    In my experience it is extremely rare that students can get 100% on every (every is emphasized by the student) lab for one and half semester. It can happen, but students are human and humans make small mistakes from time to time. Also, from my experience, if a professor receives a memo that the class average needs to be brought down to 75%, that means that the average is WAY over 75%, and that is typically a problem with the Lab or grading. – Nick S Mar 16 '18 at 16:54
  • 1
    @georgewatson I’ve always wondered why a “normal” (Gaussian) distribution is assumed. When a particular university draws students from the general population by merit (getting into the university vs not), then you’re already sampling a normal distribution and limiting the downside. That means the new distribution is probably not normal (excuse the pun). – Peter K. Mar 18 '18 at 14:44
  • @PeterK: Central limit theorem... the features that were selected for during admission (and are now strongly non-normal) are only a subset of the features determining performance in any particular course. – Ben Voigt Mar 18 '18 at 18:27
  • @BenVoigt Possibly, but I’ve not seen any justification for the assumption. I might do some digging, because I can’t see it’s justified on ethical or pedagogical grounds, so I wonder if it’s statistically justifiable. And besides... CLT requires larger numbers than the class sizes that grading to a curve is sometimes applied in (fewer than 10 students a). – Peter K. Mar 18 '18 at 23:40
  • @BenVoigt One can argue that across a cohort there are much more similar factors (e.g. the same lecturer, the same entry criteria) than factors able to differentiate the students. – Dmitry Savostyanov Mar 19 '18 at 18:38
52

Yes, this is (unfortunately) fairly normal in the sense that it is done at a number of locations.

The places I've heard it done, the main concern is at a department level, with courses of many sections, in which instructor difficulty has high variability; some instructors are "hard" and others "easy". Admittedly, this causes some initial level of unfairness in the luck of the draw as regards who each student gets for an instructor. The fixed-statistic doctrine forces the harsh instructors to scale up grades to look more like other sections, and so forth (this reduces student complaints to the dean/department). The resulting counter-unfairness is that if lots of legitimately strong students all get in the same section at once, they will be effectively penalized... however this becomes somewhat masked because the grade-data is now mangled, and all you have left are subjective student complaints that are likely ignored. I know that I've had multiple sections of the same course in a semester, taught identically, with wildly varying outcomes (40% passing in one section and 80% in the other).

I think the gold-standard way of handling this would be to have joint tests that are team-graded (i.e., same one or two professors grading each problem and verifying each others' judgement). However, that is logistically expensive and rarely done by tenured academics, I think.

My father had a similar down-grading in a college class, for similar reasons, circa 1966 and he hasn't stopped complaining about it yet.

Daniel R. Collins
  • 36,531
  • 13
  • 99
  • 143
  • 3
    I am in total agreement that this could be a "sampling problem." In other words take 30 students randomly from all physics students to make a class. There will be a distribution of how well all the students will do in the class, and another distribution of how well a particular class does when compared to all classes. The more students in the class the more the sample will agree with the norm. Also all the physics majors might be forced into one particular section of two being offered since they also need some other course scheduled at the same time as the other section. – MaxW Mar 17 '18 at 17:38
  • 2
    I chose this as the answer because it makes the most sense in my situation. Certainly it has to do with the variance in instructor quality that I see in my school. Thanks for your analysis! – dattebane Mar 18 '18 at 04:32
  • 4
    "My father had a similar down-grading in a college class, for similar reasons, circa 1966 and he hasn't stopped complaining about it yet." Ha, ha. Brilliant! Wish I could up-vote this more than +1. – Ben Mar 18 '18 at 23:18
25

While it sounds like your professor may have originally been grading too easy, increasing the difficulty of assignments and grading on a curve are different.

If it is on a curve, on the topic of ethics I'd make the following argument against grading on a curve, which may give you some traction if you try to get this policy changed:

Grading on a curve turns grades into a zero sum game, which penalizes group study and makes the class competitive rather than cooperative: the better you do, the worse off I will be. It's now in my best interest not to help you (and to actively harm if I'm so inclined: peer reviews being a prime candidate), because that will maximize my ranking on the grading scale.

This actively harms all student learning, as teaching others is one of the best ways of learning material and this disincentivizes students from helping/teaching one another: Those who would be willing to teach should, if working from their best interests grade-wise, not. Those who need additional help are then less likely to receive it.

One of the articles I most agree with around the web on this topic is Why We Should Stop Grading Students on a Curve which covers this topic with more depth, to include that this idea of life being a zero-sum game is ultimately to the student's detriment.

TemporalWolf
  • 361
  • 3
  • 7
8

I've been out of circulation for a while, but I can tell you for example that in a chemistry department at a prominent US university, the target for all exams was 50%, because this would give a symmetric bell-curve distribution. Then letter grades were given based upon standard deviation.

The more well-spread the grades are, the fairer it is to good students.

DavePhD
  • 818
  • 5
  • 12
  • 1
    That justification doesn't make sense; many natural grading distributions are skewed or bimodal (in fact, I've come to expect the latter). People and particles don't follow the same behavior. – Daniel R. Collins Mar 17 '18 at 12:25
  • @DanielR.Collins I don't mean it that literally. Really I'm just saying that 50% was thought of as ideal for spreading out the grades. The students (say in general chemistry with a class of a few hundred) who got near the average, whether that be a grade of 40% or 50% or 60%, would get something like a B- . The other grades were based on standard deviation. – DavePhD Mar 17 '18 at 13:02
  • 1
    I agree with the final sentence but not with the means of implementing it. If the issue is with too many students getting high grades then just make the exam harder. Then only the really good students will get the high grades. – bon Mar 17 '18 at 14:14
  • 5
    @bon maybe I'm not explaining well. The 50% was only targeted based on the difficulty/easiness of the test. Grading was very objective. Grading was not modified to target the 50%. – DavePhD Mar 17 '18 at 14:48
1

It's probably not unethical. (Note: Don't infer this answer to mean it's good. Or bad.)

At the risk of pointing out the obvious that no one else desires to, grades serve 2 roles:

  1. To evaluate how well the student has already mastered the material.

  2. To predict how well the student is likely to learn and apply such material in the future.

This means:

  • Giving everyone an A indicates that every student learned what the instructor was trying to teach, but does not at all indicate whether two students from that class will perform similarly if faced with different (but related) problems in the future.

  • Similarly, forcing the lowest grade to an F and the highest grade to an A fixes the second problem, but does not at all indicate whether the students in fact learned the material that was taught.

So, basically, what we have here is a "multi-objective optimization problem", which generally implies there is a trade-off and no single mathematically optimal answer.

(You'd think we could give 2 separate numbers for these 2 things rather than just 1 number, but the real world probably isn't very open to this idea.)

This means it is at the instructor's/department's/school's discretion to figure out how to combine these 2 measures into 1 measure, and that is something students must fundamentally accept. So your notion of being "ethical" must hinge on more than merely you disliking where the professor struck the balance between these two objectives.

So now we get to the particulars of your situation, which boil down to this quote:

I'm losing points for things I didn't before, and being asked for additional analysis that has never been required.

If I understand this correctly, this means one of the following:

  1. You were marked down for not performing analysis that were not required.

  2. You did analysis that were not required, and were marked down for mistakes in those.

Unless the analyses were blatantly off-topic (such as analyzing Shakespeare's plays after solving your linear circuits, which I assume they were not), in the first scenario, what has happened is that you have done less thorough and/or less correct work compared to your peers, and your professor has deemed it necessary for this to be reflected in your grade. Presumably this has happened either because he thinks it is a likely indicator that you didn't understand the material as well (reason #1), or that it is a likely indicator that you would not be as correct or as thorough about such material in the future (reason #2).

In any case, I hope you can see that there is nothing unethical about this behavior. It is clearly a judgment call, but it was quite clearly done in good faith and completely within his discretion and responsibilities as an instructor. You're welcome to claim that it is a poor judgment call and complain based on that, but I see no evidence for the claim that it is unethical.

user541686
  • 4,710
  • 3
  • 23
  • 26
  • 2
    I cannot agree. We either believe that merit-based grades predict the future abilities of the student to apply course concepts to real problems (in which case a merit-based grade is enough), or we believe otherwise (in which case no grading system achieves the second goal). – svavil Mar 17 '18 at 06:51
  • @svavil: I don't follow where you got the notion that that's impossible. There are times when you can clearly see that one student is likelier to be stronger than another in the future, but is not currently. (Say, maybe someone already knew the material before coming in and got 90% on every exam by virtue of that, but doesn't really show any indication of learning. Whereas someone had a deficient background but improved their score from 20% to 85%.) To me it makes perfect sense to say the second student seems likelier to perform better in the future, but doesn't currently. – user541686 Mar 17 '18 at 07:12
  • I don't follow either. Does your comment suggest to give student a course grade based on your subjective evaluation of how good he is going to become? Or, for a better phrasing, how this is different from giving a subjective grade? – svavil Mar 17 '18 at 07:15
  • @svavil: No, I'm replying to your assertion that "no grading scheme achieves the second goal". In the situation I just told you you can clearly quantify both of these by giving 2 grades instead of 1. You could do it subjectively or objectively; that's beside the point. If you want to do it objectively, an easy objective way would be to show the score and the improvement, e.g. first one (90, 0) and second one (85, 55). First one is about how well they currently did, second one is about how well they are likely to do in the future. – user541686 Mar 17 '18 at 07:21
  • 2
    How do you come up with the second figure? If I understand correctly, it's not based on current student performance. If you base it on how much the student improved during the course, students who had previous experience are penalized. More, I don't think that students who knew the material of the course beforehand are less likely to apply the material than students who have recently learned it. – svavil Mar 17 '18 at 07:36
  • @svavil: I come up with it by messing up when I subtract 20 from 85. – user541686 Mar 17 '18 at 07:49
  • My concern is more on the second figure in (90, 0) pair. Effectively, you are saying that a student who knew 90% of the course before starting and demonstrated 90% knowledge on the course exam is getting a fail grade indicating his inability to apply the course concepts in the future. I argue that this student will be better at it due to prolonged exposure to course concepts. – svavil Mar 17 '18 at 07:53
  • @svavil: Whether (90, 0) is a pass or a fail is something you decided, not me. – user541686 Mar 17 '18 at 07:56
  • 1
    I don't understand how "forcing the lowest grade to an F and the highest grade to an A" helps to "predict how well the student is likely to learn and apply such material in the future". It helps to see which of the students is better/worse than the other students (of the same course), but that is not the same thing. – Paŭlo Ebermann Mar 18 '18 at 00:21
  • @PaŭloEbermann: It just depends how pedantically you want to read the sentence. It certainly doesn't help show which student is better, since whoever looks at the grades can see which student is currently better than which one regardless of whether the grades are pre-saturated to A-F or not. So the instructor has this degree of freedom to encode something besides past or current performance. I suppose he doesn't have to use it to encode anything, but we can assume he'd do the sensible thing, which is to use it to encode predicted future performance. – user541686 Mar 18 '18 at 00:36
  • So maybe you are encoding which of the students has a predicted better/worse future performance than other students in the same course, but you can't encode the overall predicted performance. For example, if the course better succeeds to prepare the students for their future (or you just get more better students), I would expect overall better grades, but this is impossible when "grading on a curve". – Paŭlo Ebermann Mar 18 '18 at 01:18
  • @PaŭloEbermann: Indeed, that's exactly what I said: when I said "the second problem" I was referring to addressing the second part of the preceding bullet point: "...but does not at all indicate whether two students from that class will perform similarly if faced with different (but related) problems in the future." – user541686 Mar 18 '18 at 07:41
1

I cannot answer as to whether or not this is common, however I can say that in my first year at university a 'pass' for physics depended on one (4hour) exam at the end of the year. We were told before the exam that regardless of marks, only 50% of the students would be given a 'pass'. It was a long time ago but with dim recollection I think that for my year you needed to get at least 75% right to scrape through.

Fair? No. The system was set up purely to only pass as many students as they could admit to second year.

Alastair
  • 11
  • 1
-1

Let me take the possible situation you're in to an extreme. Imagine your TA had a marking rubric that consisted of "5 marks out of 10 for putting your name and assignment number on page 1, and 1 mark for each page submitted." 5 pages, good format: 10/10. Not only would you get 100% on every lab, so would the rest of the class, except for a few who only submitted 4 page reports. The department would not feel that your TA and prof were doing a good job preparing you for future courses or for work, or evaluating your competence, if that was the actual marking rubric. Furthermore, it's pretty easy to argue that this imaginary rubric is unethical.

So in a department where someone is doing that, it's not unethical to tell them "you have enough students in your class that your distributions should match such-and-such a pattern, and your average should be X." While it's possible that much higher or lower averages in a class are caused by happening to get a much smarter class, or the prof being an amazing instructor, it's far more likely to be caused by easy marking. Easy marking feels fun while it's happening. Everyone likes getting 100%. Even when you make mistakes (leave things out or do things wrong, you're not clear) you like getting 100% anyway. Where it's not so much fun is next year, when you get a hard marker. Or when you realize you don't actually know how to do certain things that everyone assumes you learned how to do in this course.

Having quotas is a way that departments try to rein in both easy markers (who you like, but are actually treating you unethically because they are not teaching you everything you need to learn) and hard markers (who crush your spirit with their nitpicking and unfair expectation.) It is not normal for everyone who "puts in the work" to get 100% on everything. And marks are not assigned as a way to be fair and recognize the time you put in.

Kate Gregory
  • 5,746
  • 2
  • 24
  • 28