Harvard capping high marks; should we?

[poetryandwine]

Harvard University academics have voted to cap the percentage of ‘A’ grades awarded in undergraduate course modules to 20% of enrolled students, plus or minus four students. There is no cap on other grades, including the grade A- ( ‘A minus’). This will be from Autumn 2027. The change will be assessed after three years.

About 70% voted in favour of the cap.

The grade of A at Harvard is supposed to be reserved for work of exceptional merit. The last time the percentage of A grades was as low as 20% was 2005. In 2024-25, about 60% of awarded module grades were A’s.

Students are not happy. There is a fundamental difference between what students want from assessment and what academics and employers want from assessment.

UK First Class degree awards have undergone a similar but less dramatic change over the last few decades, especially since tuition fees increased. We know employers don’t find undergraduate marks and degree classifications as useful as they used to, and why they now set such high bars.

Do MumsNetters think that capping the percentage of First Class Marks in each undergraduate course module, which can easily be achieved by rescaling marks similarly to what we already do, would be a good idea, or not?

Capping awards of A grades is surely coming at it from the wrong end. The number of A grades awarded should depend on the number of students who have done well enough to achieve an A grade. As you say, they could raise the grade boundary to make it harder to achieve an A but capping the number of A grades awarded surely just means that some students will be prevented from being awarded As simply because - in the university’s view - too many other students have already been awarded As. How can that be right?

It sounds more like a signal than an effective approach to teaching, but maybe they felt the need to make that signal.

Cue endless complaints from students who claim to be in the top 20%?

Don't agree with this - I hate the idea that students can’t achieve because they happen to be in a really good cohort, but if they’d been in the year before or after they’d have done better. I much prefer a pass mark.

That increase also suggests that students are getting better at reaching the expected standard probably not because they’re fundamentally better but because they understand what’s expected of them and can use examples etc to make sure they get the marks (or however it’s done). Is knowing what you have to do and doing it unacceptable?

instead Harvard could change their assessment (more work for them).

I also think school students here should have their % reported, perhaps alongside an indicative average to accommodate for subject differences , rather than a grade though so that’s where I’m coming from. I don’t think it’s fair that our system builds in failure.

It depends what the mark is meant to represent and that's up to the institution of course. Is it meant to be the top slice of a particular cohort or is it meant to indicate a certain level of competence/ knowledge.

There's pretty good evidence that students genuinely are better at studying and/or try harder than that used to. So in some ways the top 20% thing makes it harder to compare people from different cohorts.

I was looking at doing a graduate certificate at Harvard but you have to get all A or B grades to complete the certificate. I would rather spend my money somewhere that gives me a fighting chance of finishing the course.

How does this work? How do you look at say, 27 out of 100 students who all got the same grades, but say only 24 of you can get a good grade?

You rank everyone and award the A to the top 20%.

Lazy, inaccurate and unfair way to assess. Some years they will get more capable cohorts that others.

If so many students are reaching A grade then the fair way is make the course and exams more challenging so students still have the opportunity to influence their results.

Completely unfair. If you have two identical grades that fall on the boundary, one gets it and one doesn’t?

I mean if elite unis here did the same and some employers are uni blind, then who in their right mind is going to be going to an elite uni.
Unless it is to cement the elite uni brand and stop employers being uni blind in the long run by creating the impression that there is no grade inflation at certain select unis, whereas there is elsewhere.

In the US, I think the name of the uni you attend like Harvard still matters for jobs and connections for life. So an A from Harvard if they do this is really only for the top of the top. And I guess the US does not mind that sort of thing as much as we seem to these days.

We already kind of do this at my institution in the sense that we have to report all our module stats to the external examiner and exam board. If our module has too high/low an average or too many firsts/fails in proportion to other modules then we essentially need to explain ourselves and adapt the module content next year to ensure we bring it back in line. So no cap but a process to stop grade de/inflation.

I do have a fair few American students though and they get VERY upset at what we would consider top marks. If my top mark is 85 on my module I will get the American students complaining they didn't get 98/99. I always think of clueless when cher argued her way up the grade card and wonder if they've been taught to do this.

I may be an outlier here but I'm completely in favour of bell curve marking. I worked in a school exam board in the late 70s and 80s when this was the model and it meant that within a cohort you knew exactly where people were graded against their peers. I don't believe for one moment that yearly cohorts vary that much, so a set of A grades at A level then was a real achievement and said something about the person who had got them. Equally first class degrees were vanishingly rare and only awarded to exceptional students.
When grade inflation at all levels mean that a set of 9s at GCSE or a first class honours degree is commonplace, they are no longer useful tools to measure anything as you have no way of judging between the truly exceptional, the very good, the average, or the rest.

You change the boundary....

That's what happens in the UK.

With a score of 90/100 in an exam, One year you might get an 8, another year a 7, or a 9.... It all just depends.

Yes but that’s not what Harvard would do. You would only be able to give a certain number of students an A grade. So even if you have two identical grades, only one would be allowed an A. It’s a horrible system.

A set of alm 9s at GCSE is absolutely not common place! 0.19% of all students achieved all 9s....

https://www.gov.uk/government/publications/infographic-gcse-results-2024/infographics-for-gcse-results-2024-accessible

Of course they would.

How else would they cap if more than 20% for the required grade? Pick from a hat?

Thy will have to change the boundaries.

Ot make it much harder to achieve the top marks, by making assesment more difficult...

but it equates to the same thing..

That sounds nuts but then I don’t understand the grading system anyway. My independent school graded us on the percentage where 80% was an A. Or maybe it was 85%. This was a primary school though where we had exams each summer. I can’t get my head around grade boundaries changing all the time and marks in the 60s getting you an A or equivalent. 65% was a C. To me it makes the grades meaningless. I must be missing something though because my daughter swears it makes sense.

Would be fair to raise the grade boundary or the criteria for getting an A be higher . What happens when more than 20 % meet the demands of the A? How do they choose if it’s capped ?

They change the boundaries again...
Make the assesment harder in the first place.

It does not equate to the same thing .

If you only want 20% of people achieving top grades, then you either change the grade boundaries and/or don't give such high marks

Let's say 90% is the A grade.

If 25% if people are getting 90+% the assesment is too easy.... So you make the assesment harder so fewer people achieve the 90%.

And then if it's still more than 20% achieving 90%+, you just look at the marks awarded and then find the boundaries, and going "oh actually this year anyone with 97%+ will get the A grade"

It's the same thing in the end.

Up to 20% of people achieving the A grade.

This. Something needs to be done. Employers can’t currently differentiate the extraordinarily able from the quite able as both receive the top grade.

The only issue to me is whether there is significant variation in results between years meaning a clever child could miss out one year through being in a very able cohort.
(This is probably under the control of the institutions in terms of the level of offers they make anyway.)

The other issue is the difficulty of subjects. We all know some subjects are much harder than others.

A few years ago my DD did a LSE summer school course which was popular with US Ivy students. She got the impression that most of the Ivy students were grade obsessed eg tears as they hadn’t got over 90%, asking everyone their marks as they needed to know their class position. She and the other UK students found it most odd. More focus on getting high marks and gaming the system to get these marks than learning itself. Although my DD would admit she gamed her final year choices to get a first (but would agree she is no better than her BF who didn’t and got a high 2:1).

Don’t agree with caps though.

Most UK graduate schemes have their own testing so degree class (or at least 1st/2:1) not so important. My DD home uni gave you a percentage and some stats (median/range) for the module. Little discussion of degree class until final year.

My DD applied for lots of internships/placements and just had to tick box that she was on course for at least a 2:1 and had covered certain modules (no interest in grades).

It’s called grading on the curve. You see it in many work performance management schemes linked to bonuses. Usually called forced distribution. Although most of the major banks have moved away from and now do an ‘expected’ distribution.

Not sure really when it comes to grades, but I can appreciate if the whole class is getting an A plus, there will be varying levels of quality in the work. Even if it is A plus quality, when compared to another A plus it’s only 80% as good, then maybe it should be a lower grade.

@poetryandwine As someone whose children have recently been through the entire US school system from pre-k thru to the end of 12th grade, I'm aware that "grade inflation" has been a huge problem in the past 10-15 yrs at the very least.

I would say it is most likely that Harvard has been responding to the general expectations of students over this time and inflating grades as well, to keep everything "aligned"

It is a massive problem in many US schools, especially in more affluent areas where the parents who donate to the schools (state schools as well as private) exert tangible influence over such matters.

I genuinely don't think the situation is comparable to anything in the UK universities.

I do think that making A levels "easier" has led to the introduction of the A star grade. I think that's obvious to everyone.

As to capping the number of 1sts that go out - sure, why not.