Why weren't teachers/schools more realistic with predicted grades?

[nervousnelly8]

Genuinely looking to understand how the A Level results seem to have gone so wrong. I don't work in education and DS is too young to be impacted, so hoping that those more informed might have some input!

Everywhere I see, people are calling for teacher assessed grades to be used. But if this happened, wouldn't the results be way out of line with history, rendering them useless as a form of comparison with other year groups? I understand that the model that has been used appears flawed when considering individuals, but does aggregate performance not also matter? Why wouldn't schools and teachers have been sensible in their predictions relative to previous cohorts so that their assessment could be used reliably?

Not really an AIBU I suppose, I'm sure IABU for seemingly missing the point completely!

Children often do unexpectedly badly because of nerves. I can think of four who were at school with me. Teachers will not have accounted for that.

Really good article about this here: www.spectator.co.uk/article/the-government-should-have-trusted-teachers-on-exams/amp?__twitter_impression=true&fbclid=IwAR1Y0uSy8d76Ay4QaGARLnS-wYfqCZHAaF9INJjblIvT9FonH6MIpNb1lfY

In our school we pretty much awarded the same number of grades (a-e) and (9-1) as in previous years with some tweaking on prior attainment. So no over egging.

I think it's an unprecedented situation.
I was talking to a family friend who's a teacher... For example she has one student who is GCSE age and pregnant. She very rarely attends school (like less than a handful of days a year) and told staff she wasn't gonna bother turning up to her exam. But the girl is actually clever and if she turned up she probably get around what was a "B" (sorry haven't got my head around the new system!). So do you give her a B as she is capable of it or a U as she wouldn't have turned up?
This is a situation no one has been in before and there seems to have been a lack of guidance... So there is bound to be issues.

My headteacher put me under quite a bit of pressure to over predict for my department (core subject). I resisted, but it didn't make me popular. I imagine other teachers decided to do as their HT wanted, fearing reprisals if they didn't.

Yes, you're missing the point.

Teachers can say quite confidently what a student should be capable of based on their prior attainment, effort, attitude etc. What we can't predict is what will happen once study leave starts - who will be bereaved, who gets ill, who doesn't bother to revise your subject, who's Dad gets arrested, who misunderstands the exam questions, who presses the wrong button on their calculator etc etc etc. So the 5 students you might think can get B grade might end up with ABBBC. How do you judge who will make a mistake? Do you arbitrarily decide that Joe will be pulled down and Sally pushed up, just because someone will have been? Or do you say, no, they are all capable of a B so I am going to predict a B for them all?

Teachers have said for a long time that basing our appraisals on student results isn't very fair. We can't control what's happening in their lives and what impact that will have on their results. We can only deliver our content in the best way possible and cross our fingers.

Morning @nervousnelly8

There was a thread about this a few days ago and a clear example was used then. Imagine if, as a teacher, you have five very similar students who are all on the B/C border. Based on the evidence you have, each of them have more than earnt a C and there is a reasonable possibility of each one earning a B.

In testing, not all of them would have been able to get that B. Perhaps 2 did, 2 stayed at a C and one found it overwhelming and ended up with a D. That sort of thing happens.

When producing their CAGs (centre-assessed grades), you have no choice but to give a B as that's what they each individually could have been capable of getting. Otherwise, you would have to choose which students would have been successful in crossing the boundary, which wouldn't and you'd also have to unfairly choose which student would fall at the final hurdle. All with no evidence as to why you chose that for each one.

Teacher CAGs were always going to be higher, that was known. What wasn't know was that Ofqual were going to create a model that was shown to only be accurate in around 60% of cases. In some subjects, it's below 50% accuracy. That's why there is all this outrage and drama around A levels and the coming GCSEs.

In most cases teachers will look favourably on their pupils and if in doubt, round up. The article linked to above confirms this. Minor changes in a few cases for one particular teacher - it soon adds up to massive grade inflation overall.

If 12% of students bomb an exam unexpectedly for a number of reasons - stress, illness, anxiety, break-ups - this wouldn’t be seen as massively unusual.

But how can teacher assessments add that to the mix? We can’t predict the ones who underachieve unexpectedly. That wouldn’t be fair. Teachers have predicted what the student would realistically get, given their ability, attitude, progress and effort, on a good day. Now that is fair. To mark anyone down from that is to assume they will underachieve without evidence.

Ah Darren’s mum had Covid, let’s put him down two grades as he would have been looking after his siblings and not revising all April.

I don’t think so.

@DoubleDeckerBusRideLover that is a really helpful article thank you. I have a teacher friend who said his department had an overall grade inflation of 2% compared to last year, which seems reasonable. But someone somewhere must have over-egged the predictions in order to get the aggregate inflation (he said it was around 30% across the board, but I don't know how he calculated that or where he saw it)

Is that you Gavin?

I'm not in the UK but our students' final year exams have been assessed only by their teachers and those results have been accepted by the govt.

Yeah, some results were higher than envisaged. Some were lower. So what? A generation of kids where I am (like in the UK) finished their 13 years of schooling, at home, on a computer, in the middle of a pandemic. That they come through this insanity without going bonkers is enough for them to deserve the results their teachers (= the people who scholastically know them and their abilities best) predicted for them. Or should have been.

The point is, in an unprecedented moment, unprecedented action has to be taken. You can't then morally turn round and say "nah, don't like how that turned out, let's roll the dice again".

Maybe it would be out of line with other year groups, but then it was always going to be different from other year groups because this year students weren’t allowed to take their exams.

Every other year students will have the human factors that affect them on the day that brings their grades down, or they may not have revised. It should have been expected that by giving every student the grade they were on track to achieve if nothing went wrong, that the results would end up a little higher than normal. But so what? Nothing else about this year is normal, and whatever grades they got they will always be tainted with being 2020 results anyway.

They were realistic. They spent hours and hours moderating and ranking and discussing students over Teams during lockdown. Predicted grades are what a student is capable of at their best, based on their work over the course.

It cannot predict the students who work extra hard up to the exam and somehow pull it out of the bag to go from a B to an A, or those who for whatever reason flunk it due to nerves or miss a question on the day or answer the wrong question etc and go from a potential B to a C or a D.

It’s not a perfect solution but it’s much fairer than this horrific lottery and is actually based on evidence from professionals who know each individual best.

This idea that teachers are artificially inflating grades across the board is frankly bollocks.

As an Exam Officer, our school predicted what we genuinely thought would have happened had the kids sat exams - broke my heart typing in U or 1 but it was the truth.

As it stands, very few of our A Level predictions were changed and our genuine A*s stood. Some went down but also some went up too. Out of 400 subject entries, there were approx 30 amendments.

I have to wonder if places who are shouting about gross discrepancies have clearly over-predicted

The thing is, going on the previous years’ results, some students, for example, would get a U. But it would be vanishingly rare to predict a student a U, because they get that by not answering questions in the exam or writing very little of use to the examiner. And if you thought your student was likely to do that you wouldn’t enter then for the exam on the grounds that it’s a waste of time and money.

So while U grades are possible, perhaps even likely, they’re inherently unpredictable. You know which students might get a U - and you certainly know which students almost definitely never will - but you can’t know until the day of exam who ^will*. Just from that alone, there will be an element of “overprediction”.

(The above is for A-level - U grades are more likely to be predicted at GCSE but still not common).

Teacher predictions are always over-optimistic, although with the best intentions. It would have been helpful if the algorithm had taken into account what each school had predicted in past years against what those students actually achieved.

@HoneysuckIejasmine and @Iamnotthe1 thank you for your explanations. I didn't mean to teacher bash in the slightest - I completely get that it was a nigh on impossible task. Your examples make sense. But in the case of your 5 students all predicted a B, but in the exam they might have had ABBBC, surely this would have led to the "right" average, whereas in fact the average was much higher?

I suppose it wouldn't matter really if results were higher this year, unless it meant that universities couldn't accept everyone they had made offers to?

@StationView

My headteacher put me under quite a bit of pressure to over predict for my department (core subject). I resisted, but it didn't make me popular. I imagine other teachers decided to do as their HT wanted, fearing reprisals if they didn't.

That appears to have happened in quite a few schools and colleges. There were some colleges where pretty much everyone was awarded A/A* by their tutors when in any other year there would be a normal distribution.

The trouble is that if they take the teacher predicted grades, those in schools who marked objectively would look bad in comparison to those at schools where grades were inflated.

Grade boundaries change each year. So what was a b last year could be a c the next. Some will be an A with 89% and other an A with 81% it's very tricky.

I agree with the others

Dd was downgraded for one but both her predicted Bs were ‘fair’ she probably could have managed As but we’ll never know

Ds1 who took his gcse a few years back was predicted 6 A and 3B, in the end because of his anxiety and the fact he tends to go to pieces in an exam it was the other way around

That would have appeared as if the teachers had over predicted his grades but he was and is more than capable of As

The big problem seems to be that children who were supposed to get all A and A* got dreadful grades in some cases...one child we know was predicted a U...he got a C!

It is really tricky too if you have a student on a grade boundary. If the boundary was say 50% and you had three students who averaged 50%, on the day one two or even three of them might come up against a harder question, maybe something you hadn't revised as recently and they drop down to 49%. Or maybe they get lucky and the topic they revised in the last session comes up and they go up to 54%. You couldn't predict which of those three would get 49% or how easy or difficult they would find the paper on the day so you adopt the approach that all three would get 50%. When magnified across the country all those benefit of the doubt decisions add up.

Predicted grades aren’t routinely reported, and no formal record is kept of predictions vs final grades, Omlette, so that wouldn’t be possible.

UCAS predictions are a different thing as they’re for a specific purpose and wouldn’t be the same as CAGs.

@TW2013

It is really tricky too if you have a student on a grade boundary. If the boundary was say 50% and you had three students who averaged 50%, on the day one two or even three of them might come up against a harder question, maybe something you hadn't revised as recently and they drop down to 49%. Or maybe they get lucky and the topic they revised in the last session comes up and they go up to 54%. You couldn't predict which of those three would get 49% or how easy or difficult they would find the paper on the day so you adopt the approach that all three would get 50%. When magnified across the country all those benefit of the doubt decisions add up.

Absolutely

It wasn't the teachers' fault, they ranked their pupils and tried to assess what they were capable of. They can't account for the fact some pupils every year underperform, they couldn't choose that could they!