Maybe it has grown steadily since the 80s but the latest spike is related to the pandemic. Have a look at the chart in the article - there's a huge increase from 2019 to 2020.
> The sharp rise in top grades at A-level means that the proportion getting top A* and A grades has risen by almost 75% since the last time conventional exams were taken in 2019.
It seems to me that the article is suggesting that the last 2 years were exceptionally different.
What I'm wondering is what the factors truly are here and what to do with this data.
In my opinion these grades are there to differentiate the skill levels of the people, "short-term-memory" being the major skill here (another topic for another day).
But if nearly half of the people are the same elite... this system of skill-measuring is getting useless.
Exams were cancelled this year and grades were estimated based on students' classroom performance. So your argument doesn't really hold up in this case.
Grade deflation is pretty-much impossible in the UK due to pushy parents; I've heard talk on the news of changing the grades from alphabetic to numeric ...
This! As long as parents push students by comparing their grades to other (previous year) grades, we can't reverse course.
We should swap to a numeric system but shouldn't just swap an "A" for a "1"; we should show grades as percentiles. Being in the highest 10%/5% is a more meaningful comparison than saying "A" if the meaning of "A" changes over time.
There will be no extra doctors. Places on medical programs are strictly limited. Students in the UK don't automatically get in once they make the required grades, they still need to be offered a place.
Unfortunately, introducing new grading schemes is a slow and arduous process, since the current grades are used to determine entry requirements for Colleges and Universities.
Nitpick: Not of all the UK uses A-levels (the article is careful to say "England, Wales and Northern Ireland") - Scotland has a completely different system and exams.
Meta: I enjoy how HN has European stories and comments this time of day before the US wakes up. You naturally get sort of region categorised HN feeds based on time of day.
I got 110 out of 80 on an Algebra II test once. On the standard system of every ten percentage points being a letter grade, that'd be an AAAAA.
I assume what's going on there is that the teacher has to deal with the range of students in the class going from 8th graders who have no difficulties up to 12th graders who can barely understand the material, and the test's "full mark" is set at a level that allows the 12th graders to get passing scores.
The problem these is that there is too much noise in the signal. Given that so much of your grade is based on sitting a single exam, the variance in your performance on that day makes it pointless to differentiate between students who got for example 87% and 89% on the exam. All you can say is that both are "very good" at that subject, but you cannot meaningfully judge which is better.
On the other hand couldn't grade saturation be seen as a good thing, if it indicates that more and more students are successfully mastering all the material they are supposed to master.
I think grading on a curve means that student won't study in a collaborative manners: they don't discuss solution, sharing knowledge, etc... Which is the opposite of what we want.
But a nationwide curve brings back the incentive for collaboration - your group of friends or school will still push itself up compared to other schools.
I guess it fundamentally comes down to what you are looking for in a grade. To you want to identify the 'best' 5% of maths students in a given year or identify everybody who is at least 'this' good at maths.
For a bit of political context, last year there was a lot of controversy about people getting poor grades based on estimates, which reflected badly on the government. I suspect there was a lot of politics at play here to ensure that everyone was happy with their inflated grades.
It’s ridiculous that students couldn’t just sit their exams as usual this year.
Last year they took teacher assessments to rank students inside schools, and then mapped those ranks onto previous grades from the same school. So if previously only one person in the school got the top grade, last year only the top ranked student by teacher assessment could have got that grade, no matter how good the other students were. They’ve managed another shambles this year, by swinging to an opposite failing. The Education Secretary doesn’t seem very bright.
There was no level playing field between schools and students this year due to different students missing different amounts of school, having different capabilities to learn at home, and many other covid related factors. It does not seem fair to me that they should all sit the same exam as if they had all had similar normal schooling for the past two years.
I'm not saying I agree with the mitigations that were put in place this year and last year, but I certainly think normal exams would not have been good for students. Ideally we want people to fill the universities, the aim is not to maximize the number of people rejected.
I think the impact is that some students are going to get placements that they otherwise would not have gotten, and will struggle accordingly because the level of study required at Uni will be that much greater. Outside of gaining entry into a university, A-levels are fairly meaningless.
That greatly depends on your grades and grade history. If an applicant has high grades, meets the entry requirements, and shows a consistently high grade history, they won't be asked.
It really depends on how much information the Uni has and how contested each seat is. Some schools just make everyone take an entry exam anyway.
> Everyone who applies to study physics or physics and philosophy at Oxford, without exception, must take the Physics Aptitude Test (PAT), a two-hour test that evaluates a student’s ability in both physics and maths. We work in partnership with Cambridge Assessment Admissions Testing to administer the PAT.
get placements that they otherwise would not have gotten, and will struggle accordingly
It will be very interesting to see. Based on my own very limited observations, the correlation between high school grades and university performance isn't that strong.
I'd propose that a kid who worked their ass off to get a B+ will do much better at university than the lazy kid who got an A simply because they where smart and had knack for the subject.
My personal anecdote as someone who did "ok" in highschool (when I didn't care, 5 A-C GCSEs, an F in IT!) and scored in the 95th %ile at uni studying CS (when I did) relates to that too, but I'm not sure how accurate this is for most people.
I was raised that A and B were the only acceptable grades, C means you need to study more. But then I ended up dropping out of college so that wasn’t too helpful.
Isn't an education system where all students master all the material they are supposed to master, and thus everybody gets A's, a worthy goal to strive for?
I’ve been blessed in my life to attend an “elite” high school that produced IMO medalists on the regular. I’ve attended a sports academy as well from K to middle school, and briefly competed internationally in a junior age bracket.
I’ve lost any faith in the concept of “nurture” after watching two things right in front of my eyes:
ive watched some young kids steamroll adults who dedicated their entire life to the sport for a decade or two. it only took six months of training, and those prodigies weren’t particularly obsessed about their training either.
It was the same with academic studies. Some 1-in-million people were truly special. with either ridiculously deep insight where they could see patterns in problem sets nobody else could, or memory that allowed them to memorise 200 pages overnight. Not quite verbatim but quite precise, after only reading it once/twice.
True genius was nothing short of supernatural. There is a reason why we instinctively call it a “gift”.
Unfortunately, the other part of the spectrum exists as well, what we call special education, or the unfortunate kids with head injuries, bad infections, abused or just born that way. It’s quite sad.
What you are proposing is to create an educational system where both an IMO medalist and the special ed student both receive an A in math?
These are the extremes, but the gradient doesn’t just disappear between them. You’ll just teach at the slowest students pace, and eventually put stronger students on the “gifted” track, or super-specialize them based on their interests.
What you are proposing is to create an educational system where both an IMO medalist and the special ed student both receive an A in math?
If they have both mastered the material set out in the national curriculum and achieve the set number of right answers on the final exam, why shouldn't they both receive an A in math?
I majored in math at university and we had lots of students who had gotten an 'A' in mathematics, which included everything from a kid who represented his country at the IMO finals to lazy kids who had barely scraped over line. So even today their range of talent was massive and that 'A' grade in itself offered very little indication how well they would actually go on to do.
A High school grade in itself cannot be granular enough to identify those "1-in-million people", so it's pointless to even try to use it for that. It should be an indicator of how much of the curriculum you have mastered. If you want to further identify extraordinary talented and passionate future mathematicians (or historians or doctors) within that cohort, then you're going to need different tools.
This is a non-story because this year and last year, exams were cancelled and teachers effectively set the grades. These two years cannot be compared to previous ones.
Of course they can be compared. Direct comparisons may be hard if we are looking for a cause. But even though the grading mechanism is different, we can identify the overall trend. And that trend is bad for higher education.
I'd really like to have some system that can convert any grade or score to a "ranking in class".
Then I can see that someone was schooled in Romania, studied Maths, and did better than 95% of other Romanians in school. I don't need to know the details of how the grades work, the retakes, the 'easy years', etc.
It would make hiring much easier. Currently I pretty much throw out the education part of peoples CV's because comparing totally different exam systems is impossible.
I don't know why they don't switch to a percentile based system. All the arguments against it are nonsense because people implicitly do it in their heads anyway. Like when someone says they got an A now you're immediately going to think "in 2019? that's pretty good" or "in 2021? only in the top 50%!".
Here are some common arguments against percentiles:
1. You can't compare grades across years!
You can't anyway because of grade inflation.
2. What if one year is intrinsically better than the previous year? Giving them the same overall score would be unfair.
Subjects at A level are big enough that that is extremely unlikely. Even for small subjects there are ways of doing it fairly though, based on performance in previous years.
3. We don't want people concentrating on tiny differences in percentiles.
Because in a perfect educating system with perfect students and teachers everyone would get 100%, and a percentile system has the same problems as stack ranking. It punishes success.
I worked my fingers to the bone to get A grades in my highers. The grading was done based on a distribution where only a certain % of candidates could get each grade in a given exam. Teacher assesed grades are antithetical to the purpose of exams.
This has been an incredible act of political cowardice on behalf of the UK governments. The only mitigating factor I can see is the universities benefiting from a larger candidate pool in the UK since there has been massive drop-off in EU students applying.
For universities that are primarily concerned with their reputation as a scarce commodity over the scaling of a quality education, sure. The unfortunate thing is this has become nearly all universities. The social signaling is more important than the education. Let’s all remember that when deciding what to value in such a system.
> Teacher assesed grades are antithetical to the purpose of exams.
The point of an exam is to assess a students knowledge on the given subject not to rank students against their peers. It's especially unfair when the number of peers is low like in a school setting.
I don't see how this effects the size of the candidate pool -- Universities increase/decrease their entrance requirements each year, based on their current popularity + expected grades from students.
Of course, because you have to make all your offers, then see who makes their grades, it's annoying hard to predict.
> The grading was done based on a distribution where only a certain % of candidates could get each grade in a given exam
I don't think that's fair, either. That's suitable when you have a ranking and only the top n are accepted (like in admission exams) but IMHO not when the idea is to benchmark against a level of achievement: In which case it does not really matter how well others did, only how well each candidate did against the expectation.
So while I do think exams should be so that only a certain % of candidates usually get top grades, I think that this should be achieved by calibrating exam questions to the appropriate level of difficulty.
For instance, there should be a few stretch questions that are challenging enough that maybe only the top 10% on average are expected to get them right and the grades' thresholds should be set accordingly.
A-levels are ridiculous to the point that when it was becoming evident that too many As were handed a few years back the answer was to create a new A* grade for the "really good ones". But of course inflation did not stop so I fully expect someone to propose A* in a few years.
> but IMHO not when the idea is to benchmark against a level of achievement.
Yeah but the level of achievement required for starting university education has steadily dropped, and university enrollment has risen as a result.
I once read an article that claimed to have done the math for this, on math. It claimed that the absolute top grade in 1976 in high school math would have been insufficient to have gotten you into university just 15 years earlier. I believe they assumed you answered all the questions asked correctly and would leave any more advanced questions blank.
This tweet in a short thread from Sam Freedman hits the nail on the head for me:
"Incidentally I think grade inflation is the wrong terminology. It's not like the pound inflating. It's like we suddenly decided to switch to the yen. You simply cannot compare this year to previous ones in any meaningful way."
We compare pound prices to yen prices all the time. If Britain suddenly switched to the yen, it would be incredibly easy to do before-and-after price comparisons.
This feels like you’re almost deliberately missing the point. The metaphor isn’t perfect, but surely you get what he’s trying to express. If your quibble is with the metaphor, a more constructive contribution would be to give a better one.
The metaphor isn't that bad. The interpretation is bad. The claim that measuring inflation (1) between last year's prices in pounds and this year's prices in pounds, or (2) between last year's prices in pounds and this year's prices in yen, differ in any significant way, is a bit of nonsense that should discredit the author. The entire concept of inflation is that the value of the monetary unit differs from year to year, and you want to find out what the difference is.
And similarly, the notion that British grades from this year can't be compared to British grades from two years ago is also nonsense. They are fundamentally similar phenomena and they can be compared. For political reasons, much less data was collected this year. That doesn't mean this year's data is different in kind. It means this year's data is low-quality.
That was weird when some of my Norwegian friends went abroad for US one year for Uni. Could be ordinary C students here (which is good when it's a study only containing the top of the top), but in the US they all got straight As. Basically just doing the coursework and extra credits and it was impossible to fail. They even were told they wouldn't get to continue if they got more than a single B.
My uni have started to print the grade distribution for each subject on the report card, to show employers that a B is actually really good.
Grade inflation has always been a thing but the velocity of this will lead to unfair outcomes. For comparison, in 2019 25.2% of students achieved these grades [1].
It's believable that, in the future, candidate pools for certain masters, PhDs and employment positions will include both 2019 and 2021 high school graduates. These grades are almost always used as selection criteria. Can we really trust the process will be nuanced enough to account for the inflation, or will 2019 exam sitters be unfairly discriminated against by a surge of higher scoring students from 2020 and 2021?
Edit: The inflation is even more extreme at the very top: 2,664 students in 2018 received at least 3 subjects at the top grade; in 2021 that figure was 12,945 (+386%) (as someone that achieved this roughly a decade ago, I can't help but feel a bit bitter) [2]
That's not what 'A' means in this case. A means you scored over 85% (or whatever the number is) on your course work plus final exams. There is no reason why, in theory, every student sitting the exam couldn't get an A.
Based on the responses here can tell this is the crowd who like to recite CS101 algorithms on a whiteboard in interviews.
This approach is overall fairer to students by looking at the whole school year not one three hour exam based. UK exam setup has not really changed in 30 years plus and just because you did it is not a good reason to rethink how this works.
One of the things missed here also is a lot of exam boards historically let the teacher decide if the student is “not great” to sit the lower tier exam and cap there grade to a C no matter if you got 60 or 100% on the exam.
Schools and especially universities are viewed as products. Consumers want their product to be worth something. This is why parents put up such an uproar, they are defending their investment. It's bonkers.
This is such a big problem in society. We teach kids that every one of them is "exceptional" at everything they do and then, when they hit the real world, they shut down at the first sign of criticism or "failure."
72 comments
[ 5.2 ms ] story [ 127 ms ] threadhttps://www.alansmithers.com/reports/AL2014.pdf
> The sharp rise in top grades at A-level means that the proportion getting top A* and A grades has risen by almost 75% since the last time conventional exams were taken in 2019.
What I'm wondering is what the factors truly are here and what to do with this data.
In my opinion these grades are there to differentiate the skill levels of the people, "short-term-memory" being the major skill here (another topic for another day). But if nearly half of the people are the same elite... this system of skill-measuring is getting useless.
Maybe there should be a better alternative?
Grade deflation is pretty-much impossible in the UK due to pushy parents; I've heard talk on the news of changing the grades from alphabetic to numeric ...
We should swap to a numeric system but shouldn't just swap an "A" for a "1"; we should show grades as percentiles. Being in the highest 10%/5% is a more meaningful comparison than saying "A" if the meaning of "A" changes over time.
https://www.theguardian.com/education/2021/aug/05/grade-infl...
That they are in the top 45% of students might concern some ...
If everybody gets the same grade, other, potentially less “fair” methods will emerge.
I assume what's going on there is that the teacher has to deal with the range of students in the class going from 8th graders who have no difficulties up to 12th graders who can barely understand the material, and the test's "full mark" is set at a level that allows the 12th graders to get passing scores.
On the other hand couldn't grade saturation be seen as a good thing, if it indicates that more and more students are successfully mastering all the material they are supposed to master.
The SAT IIs were curved in the US and I saw no such effect.
It’s ridiculous that students couldn’t just sit their exams as usual this year.
I'm not saying I agree with the mitigations that were put in place this year and last year, but I certainly think normal exams would not have been good for students. Ideally we want people to fill the universities, the aim is not to maximize the number of people rejected.
At least I had to for Physics and Philosophy at Oxford, but I'm just an expendable American.
It really depends on how much information the Uni has and how contested each seat is. Some schools just make everyone take an entry exam anyway.
At least for Oxford, it seems universal.
It will be very interesting to see. Based on my own very limited observations, the correlation between high school grades and university performance isn't that strong.
I'd propose that a kid who worked their ass off to get a B+ will do much better at university than the lazy kid who got an A simply because they where smart and had knack for the subject.
I hope it is, I guess time will tell
I’ve lost any faith in the concept of “nurture” after watching two things right in front of my eyes:
ive watched some young kids steamroll adults who dedicated their entire life to the sport for a decade or two. it only took six months of training, and those prodigies weren’t particularly obsessed about their training either.
It was the same with academic studies. Some 1-in-million people were truly special. with either ridiculously deep insight where they could see patterns in problem sets nobody else could, or memory that allowed them to memorise 200 pages overnight. Not quite verbatim but quite precise, after only reading it once/twice.
True genius was nothing short of supernatural. There is a reason why we instinctively call it a “gift”.
Unfortunately, the other part of the spectrum exists as well, what we call special education, or the unfortunate kids with head injuries, bad infections, abused or just born that way. It’s quite sad.
What you are proposing is to create an educational system where both an IMO medalist and the special ed student both receive an A in math?
These are the extremes, but the gradient doesn’t just disappear between them. You’ll just teach at the slowest students pace, and eventually put stronger students on the “gifted” track, or super-specialize them based on their interests.
If they have both mastered the material set out in the national curriculum and achieve the set number of right answers on the final exam, why shouldn't they both receive an A in math?
I majored in math at university and we had lots of students who had gotten an 'A' in mathematics, which included everything from a kid who represented his country at the IMO finals to lazy kids who had barely scraped over line. So even today their range of talent was massive and that 'A' grade in itself offered very little indication how well they would actually go on to do.
A High school grade in itself cannot be granular enough to identify those "1-in-million people", so it's pointless to even try to use it for that. It should be an indicator of how much of the curriculum you have mastered. If you want to further identify extraordinary talented and passionate future mathematicians (or historians or doctors) within that cohort, then you're going to need different tools.
Then I can see that someone was schooled in Romania, studied Maths, and did better than 95% of other Romanians in school. I don't need to know the details of how the grades work, the retakes, the 'easy years', etc.
It would make hiring much easier. Currently I pretty much throw out the education part of peoples CV's because comparing totally different exam systems is impossible.
Here are some common arguments against percentiles:
1. You can't compare grades across years!
You can't anyway because of grade inflation.
2. What if one year is intrinsically better than the previous year? Giving them the same overall score would be unfair.
Subjects at A level are big enough that that is extremely unlikely. Even for small subjects there are ways of doing it fairly though, based on performance in previous years.
3. We don't want people concentrating on tiny differences in percentiles.
They would be quantised like grades are.
The point of an exam is to assess a students knowledge on the given subject not to rank students against their peers. It's especially unfair when the number of peers is low like in a school setting.
Of course, because you have to make all your offers, then see who makes their grades, it's annoying hard to predict.
I don't think that's fair, either. That's suitable when you have a ranking and only the top n are accepted (like in admission exams) but IMHO not when the idea is to benchmark against a level of achievement: In which case it does not really matter how well others did, only how well each candidate did against the expectation.
So while I do think exams should be so that only a certain % of candidates usually get top grades, I think that this should be achieved by calibrating exam questions to the appropriate level of difficulty.
For instance, there should be a few stretch questions that are challenging enough that maybe only the top 10% on average are expected to get them right and the grades' thresholds should be set accordingly.
A-levels are ridiculous to the point that when it was becoming evident that too many As were handed a few years back the answer was to create a new A* grade for the "really good ones". But of course inflation did not stop so I fully expect someone to propose A* in a few years.
Yeah but the level of achievement required for starting university education has steadily dropped, and university enrollment has risen as a result.
I once read an article that claimed to have done the math for this, on math. It claimed that the absolute top grade in 1976 in high school math would have been insufficient to have gotten you into university just 15 years earlier. I believe they assumed you answered all the questions asked correctly and would leave any more advanced questions blank.
"Incidentally I think grade inflation is the wrong terminology. It's not like the pound inflating. It's like we suddenly decided to switch to the yen. You simply cannot compare this year to previous ones in any meaningful way."
https://twitter.com/Samfr/status/1424806593187794944
And similarly, the notion that British grades from this year can't be compared to British grades from two years ago is also nonsense. They are fundamentally similar phenomena and they can be compared. For political reasons, much less data was collected this year. That doesn't mean this year's data is different in kind. It means this year's data is low-quality.
My uni have started to print the grade distribution for each subject on the report card, to show employers that a B is actually really good.
[1]: https://blogs.lse.ac.uk/impactofsocialsciences/2020/08/26/fk...
It's believable that, in the future, candidate pools for certain masters, PhDs and employment positions will include both 2019 and 2021 high school graduates. These grades are almost always used as selection criteria. Can we really trust the process will be nuanced enough to account for the inflation, or will 2019 exam sitters be unfairly discriminated against by a surge of higher scoring students from 2020 and 2021?
[1] https://www.gov.uk/government/news/guide-to-as-and-a-level-r...
Edit: The inflation is even more extreme at the very top: 2,664 students in 2018 received at least 3 subjects at the top grade; in 2021 that figure was 12,945 (+386%) (as someone that achieved this roughly a decade ago, I can't help but feel a bit bitter) [2]
[2] https://www.ft.com/content/c35e13f4-09cd-4700-9573-91fdfd012...
This approach is overall fairer to students by looking at the whole school year not one three hour exam based. UK exam setup has not really changed in 30 years plus and just because you did it is not a good reason to rethink how this works.
One of the things missed here also is a lot of exam boards historically let the teacher decide if the student is “not great” to sit the lower tier exam and cap there grade to a C no matter if you got 60 or 100% on the exam.