GCSEs and A-Levels 2020: Warning that 40% of GCSE grades may change

[Iris2212]

TES has just posted this article - www.tes.com/news/exclusive-warning-40-gcse-grades-may-change

It isn't looking good for GCSEs and A-Levels this August. I am fearing a massive backlash, especially after IB disaster. Ofqual and the exam boards better be prepared.

It's not very surprising.

Imagine you're a teacher. You've got some kids you are pretty sure will get 7, some you are pretty sure will get a 6, and some who you're almost certain will get 6 or 7 but it could go either way on the day. I reckon when I was teaching it would probably be half where I was reasonably sure of the grade they'd get and half who would be on a borderline. Add in that you don't know how well they'll use the last couple of months of study time, and it's less predictable than usual.

So, if your kid is one of those 6/7 borderliners, what would you like the teacher to put down? 7, obviously; 6 would make it unlikely that they would get a 7. The teacher knows that some of those 6/7s will come out as 6s when they do the statistical adjustments, but not how many. So they might as well put them all down as 7. They will probably have agonised about the ranking, knowing that the bottom few are more or less certain to lose out. I expect many have gone for a formula (eg mock score with adjustment for work seen since then) to keep it objective - that will inevitably work better for some than others.

So it's entirely plausible that over half of pupils are in that "borderline" position where the teacher has put the higher of two possible grades, and you might reasonably expect at least half of them to get the lower one.

In the end of the day, there will be a bit of randomness at the borderlines - but then there always is, because everyone knows that there's a bit of luck of the draw in the questions on the exam.

I suspect most grades will not move more than one notch, so all those 6/7 borderliners will get a 6 or a 7.

Yes, it's going to be frustrating, particularly for those on the more crucial borderlines.

I am fearing a massive backlash

From who? The ‘change’ is irrelevant if students are presented with a set of grades that are proportionally in line with previous years. If they argue with the ‘change’ from what the teacher assessed, they’re arguing with a statistical model. If they argue with what the teacher assessed they’ll be pointed towards the Autumn exams.

I would guess most of the children taking the autumn exams will be the ones whose parents believed them at the post-mocks parents evening when the child swore faithfully that they were just about to being a world-beating revision programme under their own steam. Never seen a single child who promised that actually do it.

I think using teacher assessment, which will have to be based on data, is probably the fairest way as it will reward consistency of effort. There will always be parents bemoaning the grade is low because Mrs ABC doesn't like little Johnny, but if the data is there then it's there.

Well if anyone is aggrieved they can sit the Autumn exams.

If they argue with the ‘change’ from what the teacher assessed,

They shouldn't actually know what the teacher assessed...

I thought the schools had to "rank" the pupils in order as well as suggesting grades. So, the kids in your example who maybe get a 6 or 7 will be ranked, i.e. the most likely to get 7 will be ranked higher than those least likely. When the grade boundaries are reviewed (standardisation), some of those with lower ranked 7s will be moved down to 6s if too many people have been suggested a 7.

I’m not worried about the teacher predictions, I’m worried about the downgrading. This affects high achieving students in low performing schools or low performing subjects. If the student is predicted A’s by the teachers (and got 8’s or 9’s in gcse), but the school usually gets C grades in those subjects, then that student is very likely to be downgraded.

It’s all well and good to say you can sit the exams in autumn. But two things - firstly many students haven’t even finished being taught the whole syllabus, and they have had zero teaching since March. And secondly, the autumn exams are not the same. They are 100% written. My daughter takes music and theatre studies - so she would not be assessed at all in performance and composition (in which she got top marks at gcse) and the written exam would be 100% instead of 40%. Same with theatre studies - no performing or devising, or coursework. So she is unlikely to get an A or A* in the autumn exams, as the written is her weakest part, and she hasn’t finished the syllabus.

The problem we have noble is that our cohort this year is hugely different due to border changes. Previously we have had lower grades due to the area. Five years ago when boundaries changed we became a completely different school. This is the first year where we would have exceeded targets by quite a long way as our kids are so different - we are now oversubscribed etc.
These kids will suffer because of this. It's a real shame.

If your cohort is different, then their SATs profile will be different?

I agree that schools that have made rapid improvement will potentially be screwed - I would think those schools would be making plans for the Autumn?

They shouldn't actually know what the teacher assessed...

They won’t have been told, no. But they can ask to be told. I think it would be much better if they couldn’t.

We have always been screwed by one school with high SATs results. Which don't correlate with the kids we had. Said teacher was rumbled and fired - so this years SATs are similar but with kids who were actually working at that level - so in previous years we have been in minus figures this year we would have been in plus.
Our previous kids came from a couple of really deprived estates/areas but a new school was built so local kids came back.

The number grades have only been in for a couple of years too, haven't they? I remember last year the teachers at my nieces school made it clear that any predictions were extremely subjective as they weren't confident how grade boundaries would fall.

I don't envy teachers sorting this lot out - and I suspect had it been my year I'd have been one of the 'losers' in this situation as I totally over performed in the exams compared with my work prior.

Having said that, what's the alternative?

The prior cohort data correlates to the actual cohort so those boundary changes mentioned above would be factored in.

I’m not worried about the teacher predictions, I’m worried about the downgrading. This affects high achieving students in low performing schools or low performing subjects. If the student is predicted A’s by the teachers (and got 8’s or 9’s in gcse), but the school usually gets C grades in those subjects, then that student is very likely to be downgraded.

If the whole cohort is better, then that will be evident from the GCSE results (I presume we're talking A-level here).
If the school's prior results and any differences in cohort suggest that most results will be B-D, and the school predicts that but with a couple of A/A*s, then hopefully the exam board will trust them on the outliers. The problem would be if the school predicts loads of higher results.

I don't envy them the job of doing the statistical wrangling for smaller subjects and smaller schools at A-level.