Can someone explain why grade boundaries change

[Ceci03]

I'm still struggling to understand. Oning from Rep of Ireland where an A is always 85-100 and a B is always 70-84 I'm struggling to understand why the boundaries change for A level. And it seems for GCSE as well? Like how does a student know how they're doing when the boundaries are so fluid. I dont understand. Thanks !

Yeh but down below the poster said they don't use a quota system in England?

Try this.

Say 100 people sat a juggling exam in 2024. 10 got an A, 20 got B, 25 got a C, 20 got a D, 15 got E and 10 failed.

In 2025 the questions were a bit harder. If they kept the same boundaries, only 7 would get an A. So they move the grade boundaries down a bit so that the top 10 still get A.

In 2026, the questions were to be a bit easier. If they kept the same boundaries as 2025, 15 would have got an A. But as they only want the top 10 to get A, they raise the grade boundary.

Does that help?

(Actually, it’s a bit more complicated than that because they also consider ‘matched candidates’ and the academic profile of the cohort, but I didn’t want to over-complicate the explanation.)

Yes technically if you are a bright student in a bright year you may get a worse grade than a bright student in a poor year, but generally there are so many pupils taking exams across the whole country that the chance that an entire year is significantly brighter than another is quite low. The point is that on average each year will be roughly similar and so the top e.g. 10% get an A.

Generally A levels are used for university admissions, so same year being compared anyway. Its rare that a job past 18 is going to be awarded on A levels alone. The chance that 2 people of equal skill had such vast difference in brightness in their years that their A levels are miles apart is so low, and that's not going to be the only thing an employer looks at because everyone knows A level difficulty have changed over the years

The results have to be standardised somehow.

It's become more pronounced as we have moved from criteria to a norm system- before if you achieved certain mark or skills/knowledge, you would get a certain grade.

But as teaching to the exam improved, more and more were getting higher grades leading to complaints about grade inflation.

So we are now in a norm referenced exam system where it's all about having a similar amount of students getting a similar proportion of each grade each year.

Both systems have benefits and issues.

I don't think you do see what I mean tbh.

The ability of a child will not change. We are not saying a year is bright, we are saying that the exam in this yar may have been harder, in this year easier. That is actually assessed by how many people get the top grades - o they need to modify the boundaries?

So DD should have got a C. If the exam was harder and she was scoring a D, they will have amended the boundaries so she gained her C. I mean not literally for every child but that's the idea.

So the bright child in your example will do better than the less bright child in a different year. If the bright child does a tougher exam then they will need a lower mark for their A. Meanwhile the less able child next door who is working at around a B, will do an easier exam maybe but then will need a much higher grade for their B.

The thing is, saying "you're in a bright year" doesn't really apply for A levels, because so many students take the exams that it's really unlikely that one year group would be brighter than another year group. It can happen in a particular school - because in a smaller group it could be that there just happens to be lots of bright kids - but not across the whole A level cohort. It's just really unlikely that there would be lots of bright kids born in one year but not in the next. The average level of intelligence would be roughly the same for both year groups.

past papers are useful anyway as (as said already) the boundaries do not vary that much. A B in French in 2019 required 274/400; in 2018 you needed 2699/400. Pretty similar tbh. Clearly the paper in 2019 was ranked as a bit easier.

And you use the boundaries from that paper to assess what grade you would have got.

But the flaw in this is that it's artificial. What is the standard? If more pupils reach the standard some years than others is that not more fair to the good jugglers?

The ‘standard’ (in this hypothetical case) is the level that the top 10% of students who took the exam achieve in any one year.

This is a statistical calculation based on the assumption that over huge numbers of students, there isn’t any significant variation in the ability of a cohort from year to year. The top 10 students will be equally good at juggling exams in all three years, no better, no worse. The variable is how hard the exam paper was.

So this poster is incorrect and there are quotas then

So it's all about quotas?

No, it isn’t quotas. It’s a perfectly sensible assumption that the same percentage of students will have roughly the same ability each year, and deserve roughly the same grades as equally able students in other years.

if you are fed up with your own kids’ results or prospects, take it up with JCQ and Ofqual.

There's not quotas

Essentially they try and maintain the same percentage of students getting each grade but its not a direct 10% will get an A.

They hold a big meeting. They look at what the distribution of marks are, they compare this to previous years. They also look at the paper and again compare to previous years. They use previous data on the cohorts IQ and performance, and they use work of students on the grade boundary to account for changes in 'brightness of the year" and try and keep roughly fair. But generally given the numbers of people we are talking about there's not going to be masses of variation.

It's never going to be perfect. I'm not sure what you aren't getting about the difficulty of the paper varies year on year?

Essentially if you take 500000 people each year, and and measure their IQ and skill/ability then there's always going to be roughly the same distribution of "brightness".

There's not going to be wide variation. It's a reasonable assumption and it's not a blanket quota, there are many factors involved. The percentages do change slightly, but they have to standardise as the difficulties change year on year

It's much more reasonable to assume that 600000 16yr olds year on year will have the same distribution of ability than to assume we can write papers of exactly the same difficulty year on year.

You are meant to be doing the past papers to get used to the type of questions asked, to see the kind of answers the examiners are looking for, and to get used to writing an exam in a certain amount of time. Not to see what grade you’ll get.

OMG I would be wasting my time trying to get you to understand some of the complex, subtle and simple issues - you seem determined not to take on board and new perspectives or information so, you do you, can on thinking what you like - you've had some good answers reiterated many times over.

It’s a simple grading on a curve. It’s not quotas at all. If it were you would always have the same % of students at every grade, every year.

They calculate the bell curve, then the statistical deviations and that’s how they get the grade boundaries. Your grade depends on how well you do compared to how every other student also did.

Watch this video.
https://m.youtube.com/watch?v=uexhLXeAnbQ

No one decides which year is “hard” or “easy” that is just the conclusion drawn on a year where all the students score lower (hard year) than average of all years versus a year where all the students score higher (easy year) than average of all years.

No. It does not mean that. Your chance at an A is the same no matter what year you are in because you are competing against students the same age taking the same course and at the population level there is no such thing as there being more clever year of students than another year.

But that's because of the allowances made for covid. And it's not very helpful of the tutor to say because if she'd been in last year's cohort she'd likely have got a lower percentage on the paper than this year because of missed teaching, so still wouldn't have got a higher grade overall.

The Covid allowances were really the de facto lower average score of all students because the pandemic negatively affected their education. The grade boundaries were calculated the same way, there wasn’t anyone randomly bumping students up a grade.

Yeah, that was what I was trying to say really - that the OP's daughter wouldn't have got a better grade last year because she's in a post-covid cohort

If an employer say has 2 candidates and 1 sat A levels in 2020 and got AAA and the other one sat them in 2023 and got BBB but yet if you looked at the breakdown they both received the same mark.

Leaving aside the complications of the pandemic and the shambles of 2020, comparing marks isn't useful if one paper is harder than another. Someone who got 50% on a really difficult paper might be more able than someone who got 55% on a really easy paper. You can't compare people using their percentage scores unless they did the same test.

So they know roughly what proportion of the year should get each grade and then figure out where the cut offs are in terms of marks.

Say in English who is to say the paper is hard or easy. Is that not subjective.

They can tell this by how well the cohort performed.

her tutor / grind marked it and said if she had been marked last year she would have been a B yet she was awarded a D. Cos of the boundaries changing.

But they can't possibly say that, because she would have sat a different paper and achieved a different mark, likely lower. You can't compare grade boundaries from different years, they only apply to the one paper.

So I took on board @Octavia64 post below. Is she incorrect. I'm not trying to be obtuse I was genuinely confused . In the system I was brought up an A is always 85-100
And a B is 70-84 etc etc and if you learned and answered the questions you got the marks and the grade. There was no manipulation or massaging of the figures. It's just very different. Like it was a standard exam . You answered right you got the grade. I still feel students are at a disadvantage here as you never quite know what you need to get to achieve the grade you're looking for. Does anyone understand what im saying.