AIBU that we must accept many teachers do not have the appropriate professional judgment regarding what students need to achieve for A-Levels?

[darkwader]

There is no reason to suggest that nationally this year’s students are different significantly to previous years – certainly not as demonstrated by GCSE results.

Unclear why, but exam boards have been generous in this years results in all categories, showing higher results than last year, but needing to downgrade almost 40% of teacher assessed grade to remotely be a normal year.

Despite what teachers are claiming, it must be the case that 40% of grades were inflated by teachers – even if the individual students who had these inflated grades are hard to determine. The number of A/A*’s would not jump by 10%.

If every teacher had correctly provided grades, then the national mix would match previous years and no downgrading would have occurred. – so although maybe not the teacher who is specifically involved with a set of students; overall teachers are responsible for the disappointment because of poor grade assessment in the first place in aggregate.

Given that teachers have been predicting grades for university entrance for years and marking coursework in some cases – this shows the unfairness of such a system, as they are incapable of doing so to any degree of accuracy or potentially without bias towards those they know.

Students across this country are now being affected by this incompetence – even if not the students own teacher, the professional standards are to blame.

AIBU to now understand that this professional judgment does not exists for many, many teachers and they need to be evaluated each year before being allowed to be involved in marking and grading?

If AIBU - what am I missing?

@Piggywaspushed

As you broke etiquette to look up previous posts and make assertions about what my role would entail, I simply wrote in the correct terms for the role.

The same 'Jargon' is being used by teachers.

It all comes down to the same thing - either estimates are accurate or not. And if not, why are they used or not adjusted based on being inaccurate previously.

What? Ermmm... that wasn't me sunshine.

We get it, you are cleverer and more competent than us. Well done. Come over to the dark side and join us.

@Piggywaspushed

Understood - so unlike other colleagues who have said they can predict accurately, not every teacher can.

The CAGs were not predictions for students, but were simply inputs to a model - and hence, no student this year should be disappointed, as they were not expecting a particular grade - because the CAG certainly wasn't a prediction.

It doesn't seem that students think this way though.

Although checking detail and names and reading may not be your strong suit.

They are 17 and 18 year olds with emotions.

The algorithm has absolutely been mechanically and bluntly applied to my class.

If it help you, I haven't told a single student a CAG today.

But they had university offers and lives and futures to sort out. Some of them have been crushed by an algorithm that says 'thou shalt not have an A*. No one form your school has had one in 3 years. How could you be different? Have a B.'

I wonder how you would have felt if you got say BBC instead of your expected (because you would know how good you were, in reality, give or take AAA, or whatever)?

We deal with humans.

@Piggywaspushed

Not at all, and sorry if I misattributed you.

I just don't understand how so many students are anticipating grades that they could never have achieved - and so far, everyone is saying they either can or cannot predict grades - but for some reason this is a fiasco.

If grades can be predicted - then clearly they were wrong, as would have achieved massive grade inflation.

If they can't be predicted, then no student would have been anticipating getting any specific grade. So the model is not a problem.

Sorry - I realise I see it simply, but sometimes an outside view doesn't need to take into account all the internal noise.

I think the vast majority of teachers have the best interests of their pupils at heart and worked very hard as a team to make the most accurate predictions possible. They knew exactly what was riding on them and more than usual as usually the students do do exams and there is something for unis to work with when offering places.

I do think there is a tiny minority who would deliberately mark students down but would hope that the school moderation process would weed those out.

And I think the bias in the algorithm that was used to change results, as well as the ridiculous bell curve has been far more damaging than any inaccuracy in the school predictions.

You’re also missing the point that many teachers are very aware of the difficulties in predicting grades, dislike doing so and would be quite happy not to predict any grades for students ever but instead just get on with teaching.

If teachers as a profession are so bad at it, rather than the system of grade prediction being inherently flawed, what’s the solution? Bring in people who are excellent grade predictors but may be terrible teachers? If we over predict grades it’s because we see the best in students and often stay hopeful in the face of lots of evidence telling us not to be. Traits that are actually pretty useful in teachers.

I think you'd probably get on better with your analysis if you think of this as an example of game theory rather than an exercise in forecasting. I mean, my CAGs were accurate in the sense they were based on mocks marked to last year's grade boundaries, moderated by the previous achievement of students doing my subject at my school, but I also had to take into account that Ofqual/the exam board would probably be looking to moderate down, so I didn't want to be too pessimistic. I also do not know if senior management put forward my CAGs or if they moderated them before submission. Trying to second-guess what several other parties will do is hardly straightforward forecasting. UCAS is even more gamified, because whether students get an offer is dependent on what grades they're predicted, and quite often they get in anyway even if they miss one (or they may get an unconditional offer).

Little about this is straightforward.

@darkwader

It's not that they are anticipating grades they could never have achieved. It's that they would have achieved them but for that to be true others would have not achieved.

Take school A, weaker cohort than previously, however inputs realistic CAGs of B, B, B. Students in 2017/18/19 got A, A, A*. In all likelihood the 2020 algorithm would spit out at least BBB, possibly AAB. Including an A or two in this schools results means that the quota of As countrywide is reduced.

Now to school B. Predicted B, B, B. However 2017/8/9 no one achieved above a C. You'll be moderated down to CCC as a maximum. If they had sat the exam, the results may have been ABB, ABC. Someone from school A may have scored lower on the exam, meaning there is space for an A in the quota now. There was no exam, so the students in school A are assumed to score above those in school B because of results in 2017/18/19. Over small subject groups this is incredibly unfair.

@GravityFalls

Just teach and leave the grades to exams, and the first sight is exam results day? Unless very accurate, simply don't do it.

This year, as shown, the predictions were not really needed - just the ranking. And ranking is much easier to do - as it only ever needs a comparison of two people ate a time.

This is going to blow your mind OP. Did you know we all also get given target grades /indicators , worked out by algorithms, national and school data , likely outcomes and GCSE performance : sometimes these are called ALPs? I am now probably going to be judged on my students not attaining these because Ofqual said, in fact, they were wrong.

Marvellous : do we send those ranking to UCAS then instead?

@Hercwasonaroll

I don't think that's really how the algorithm worked, as presumably the weaker cohort would have lower GCSE results - otherwise why would they be weaker. The algorithm works on a significant value add basis of GCSE results from what I quickly scanned.

It is unfair for small subject groups - and that's why the algorithm doesn't apply (the paper says so) - it basically uses the predicted grades much more strongly, even though they are known to be too generous. But it is very valid for larger groups.

I think the government said today no teachers would be assessed based on this years grades.

I find it fascinating that the grades went up in some independent schools. How very predictable.

The GCSE results of the individual students weren't included. They were banded into higher attainers, middle, and lowest at a school level.

Small groups is less than 5.

Groups size 6-15ish have been screwed over. Statistically small enough to fluctuate massively.

Eg 2017 BBBCDD
2018 AAABBE
2018 BCCCCD

Huge variety of results there. You could be legitimately expecting AAABEE this year and your top end would be capped at an A at best.

@Hercwasonaroll

Why - I read that three years are being taken into account historically; and that between 5-15 they are getting more influence of CAG, so will be upgraded rather than left behind.

It is the larger classes that will be normalised to mean the most - the smaller the class, the greater the use of teacher inflated grades rather than statistical modelling.

If students were expecting the grades predicted for their UCAS application. These are often ‘aspirational’ and IME dictated by demanding parents.
If students were expecting their ALPS predicted grade, these are often split grades. So if a student has a B/C ALPS prediction, statistically the best 25% of students with this target would typically get that grade. This means that a teacher predicting a B in this case is likely to be regarded as over-inflation, even though that is in line with the B/C ALPS prediction.

yes, I won't be judged in the same way : but do you not get how frustrating it is to have indicator grades (based on prior performance) saying my class should have got AA BBBBBC (of course knowing they wouldn't all do that) and then for me to CAG A A BBCCCD and be given ABCCCCDD by Ofqual's stuff?

What's your skin in this game OP?

We don't use split ALPs.

Although your point is correct, of course.

What you read is not what happened in quite a lot of schools with classes of between 5 -15.

Ok, so I'll jump in.
OP, are you a teacher?
If you aren't, I don't think you have any idea of how much teachers are under pressure to try and predict grades accurately - every year, not just this year.
We are judged on it, and are held responsible if they are not relatively accurate (given that students can over or under perform in an exam situation).
Many schools, including mine, make students sit mock exams under full exam conditions (including having external invigilators) for the very reason that we can then have a good idea of what progression the average student will make compared to mocks. Obviously there will always be the few students who don't actually do as expected, but that is entirely unpredictable.
We hold this data going back years, so we can keep track of how good our predictions are, and we keep refining year on year to try and make them as accurate as possible.
In my school, across the current cohort, and all of the different exams offered in my subject (four), we are only going to appeal two results because we believe the rest are fair.
What would you have us do instead I wonder?