The students who are really going to suffer, and who will probably make up the bulk of the autumn exam sitting, are those who have suffered from downgrading but are trying to do something where the universities aren't flexible about grades, even in clearing. These tend to be the "professionally accredited" courses such as Medicine and 4 year Engineering. I can see a lot of competition for 2021 places.
We know someone who is a senior lecturer at a Russell Group university in Computer Science. They were expecting to be between 200 and 300% over-subscribed on people who had met their (very high) offers. The same was true of Mathematics, Business, and Economics, but most other departments were expecting spaces, as there is a trend towards certain degree subjects being chosen for employability, and I think that the private schools probably push this a bit more.
I believe most universities have set up a system for letting them know about grade appeals for those who have missed out or are in clearing.
Be careful using the term ‘predicted grades’.
The teacher assessment grades sent to exam boards based on continuous assessment are a one off (hopefully).
Students usually have predicted grades on their university application that are renowned to be highly inaccurate & depend more on how much they beg & cajole their teachers & tutors & how realistic their choices are in relation to their abilities & efforts.
(I work with A-Level students)
The figure I saw quoted was 75% of predicted grade were over generous.
The problem is that they are trying to synthesize data where none exists, this is a policy failure rather than something that can be fixed with an algorithm.
On the radio, I heard someone explaining that the adjustment wasn't applied to small cohorts because it wasn't worth it. Independent schools are often over represented in this group - where only a small number each year take each subject. Meanwhile, one small independent school was publicising online that, this year, they had their highest results on record.
On the radio, I heard someone explaining that the adjustment wasn't applied to small cohorts because it wasn't worth it.
One other possible side effect of this has been a much greater proportion of students than normal getting higher grades in Further Maths than they do in Maths.
So if you just allow the results to stand, then more students will go to the better universities this year than in normal years, but those universities should have the flexibility to allow this.
That would be fine, if it weren’t for the fact that the government, in its wisdom, decided to impose a student number control (i.e. a maximum number of students we can recruit) on us this year. In theory this was to protect the worse universities from potential failure due to all the better universities over-recruiting UK students to make up for their predicted drop in overseas ones.
Maybe universities should have addressed this too. Worked out how they could get the best students without exam grades. They knew this was all coming anyway.
Neither UCAS nor the government would allow us to make offers that weren’t based on grades. And we’ve been kind of busy working out how the hell to balance next years budget without most of (the income from) our overseas students. While everyone has been working from home (often with kids under our feet) because all the campuses are closed due to Covid.
Besides which, six months is nowhere near enough time to develop an entirely new admissions system that’s both without a nationally-standardised examination system and fair to all applicants from the whole country.
The only practical solution for universities and colleges would have been to have good discussions with teachers (more than just the 'predicted grades' they supplied) and based on those offer unconditional places.
The government banned us from making unconditional offers this year as well, ostensibly for the same reason they imposed the student number control.
Besides, do you know how many individual applicants a given university gets in a given year? Are you suggesting that for each one the Admissions department should be having a chat with their teachers to see how good they are? Even if so many conversations were possible, how on earth would you assess a teacher from Oldham’s opinion of their student against that of another teacher from Bognor Regis? And would those teachers have the time to have six such conversations (the maximum number of universities a student can apply to through UCAS) for each of their students?
“Practical” is not the word I would use for such a suggestion.
One other possible side effect of this has been a much greater proportion of students than normal getting higher grades in Further Maths than they do in Maths.
Wait - that actually happens? There are people who take Further Maths who can't get an A**!, or whatever the top grade is these days, in Maths with one hand tied behind their back?
Back in my day, my school wouldn't let anyone who wasn't a cast-iron guaranteed A at Maths A-level take the Further Maths course.
Back in my day, my school wouldn't let anyone who wasn't a cast-iron guaranteed A at Maths A-level take the Further Maths course.
No (well I'm sure events happen), but this year children have been (dis)credited as though it happened (the proportion of grades is the same).
One possibility is that Maths as a large subject they've corrected for the teachers being wrong, while for Further Maths they've trusted the teachers. [See Chorister's post]
I don't know if there's a way to muddle up distributions, for another explaination.
Let’s assume for a moment that someone somehow held a gun to the governments head that meant they couldn’t defer universities a year but fund them anyway - so therefore some fudge was going to be done, knowing it would be wrong.
Why are ‘we’ more worried about predicting too high than too low ? Why not have the teachers predictions and accept the year results will be anomalous ?
As Marvin said it’s impractical & would do some students a huge mis-service by relying on teacher opinion.
This reliance hasn't been removed by using an algorithm -- which in this case still depends on teacher's ability to perfectly rank order their students.
Letting high predictions unchallenged does favour 'crooked' teachers, so i can see why you want to avoid it (the gov paper does talk about them failing high).
But if you are going to correct you need to do it consistently. Which they didn't. Hence somehow Independent Schools ended up getting 5% more A's than usual.
They
a) Should have looked at this ages ago. Granted the Education department itself should have been looking at getting schools to work, but the exam boards had free time, they don't have to actually mark the papers.
b) If/As you can't get something that replicates normal A-levels, they shouldn't have tried passing them off as them. Present them in an exceptional way matching the exceptional circumstance (perhaps replace with a nominal normal grade in 5 years time, [when resits could have happened, or they've graduated, got work experience]).
c) Now they shouldn't try to pretend that this is robust or dependable.
d) And part of fixing it means allowing the universities to do their bit.
The only practical solution for universities and colleges would have been to have good discussions with teachers (more than just the 'predicted grades' they supplied) and based on those offer unconditional places.
The government banned us from making unconditional offers this year as well, ostensibly for the same reason they imposed the student number control.
Besides, do you know how many individual applicants a given university gets in a given year? Are you suggesting that for each one the Admissions department should be having a chat with their teachers to see how good they are?
When looking to fill a job position, no one interviews every applicant, there's a process of whittling down potentially hundreds of applicants to the four you ask for interview. So, you don't need to talk to all those teachers - you can do a preliminary screen on GCSE results (no question about them) and the teacher predictions. Do universities still invite those who apply to interview and see the department, university and city? I was interviewed by 4 of the 5 universities I applied to (and had been to open days at two more). The problem being that while with a job interview it's very unlikely that someone interviewed will say no there's a much higher chance of an applicant student to decide another university is better for them. It's certainly more work, but doesn't seem impossible - except that the government decided against that to bias the process away from hard working and intelligent kids for deprived backgrounds so that dumb posh kids could have a better chance.
Neither UCAS nor the government would allow us to make offers that weren’t based on grades. And we’ve been kind of busy working out how the hell to balance next years budget without most of (the income from) our overseas students. While everyone has been working from home (often with kids under our feet) because all the campuses are closed due to Covid.
Besides which, six months is nowhere near enough time to develop an entirely new admissions system that’s both without a nationally-standardised examination system and fair to all applicants from the whole country.
Besides, do you know how many individual applicants a given university gets in a given year? Are you suggesting that for each one the Admissions department should be having a chat with their teachers to see how good they are? Even if so many conversations were possible, how on earth would you assess a teacher from Oldham’s opinion of their student against that of another teacher from Bognor Regis? And would those teachers have the time to have six such conversations (the maximum number of universities a student can apply to through UCAS) for each of their students?
“Practical” is not the word I would use for such a suggestion.
Totally agree. No-one on this thread so far has been able to suggest a workable alternative. It would be impossible to have the sort of conversations to which Alan Cresswell refers in a way that eliminated personal biases - apart from the time such a procedure would take. Exams are far from perfect but what is?
The only practical solution for universities and colleges would have been to have good discussions with teachers (more than just the 'predicted grades' they supplied) and based on those offer unconditional places.
The government banned us from making unconditional offers this year as well, ostensibly for the same reason they imposed the student number control.
Besides, do you know how many individual applicants a given university gets in a given year? Are you suggesting that for each one the Admissions department should be having a chat with their teachers to see how good they are?
When looking to fill a job position, no one interviews every applicant, there's a process of whittling down potentially hundreds of applicants to the four you ask for interview. So, you don't need to talk to all those teachers - you can do a preliminary screen on GCSE results (no question about them) and the teacher predictions.
When, exactly? The UCAS process is year-long, and was already well underway by the time lockdown started. And remember, back then they were saying it would only be for 12 weeks tops.
If we’d known in March that we’d be where we are now then maybe something could have been cobbled together, but we didn’t. And even if we had, there would have been some imperfection in the jury-rigged system that we put together at the last minute that people would have been having a go at. Do you think it’s easy doing this job?
Do universities still invite those who apply to interview and see the department, university and city?
In a word, no. For some of the more selective courses, sure, but for the majority it’s all done by UCAS.
the government decided against that to bias the process away from hard working and intelligent kids for deprived backgrounds so that dumb posh kids could have a better chance.
I seriously doubt that was in any way the reason for their choices. It’s far more likely that it was just an attempt to replicate the normal exams without actually having exams, and somewhere in the mess of algorithms and statistical analyses it ended up going wrong.
Neither UCAS nor the government would allow us to make offers that weren’t based on grades. And we’ve been kind of busy working out how the hell to balance next years budget without most of (the income from) our overseas students. While everyone has been working from home (often with kids under our feet) because all the campuses are closed due to Covid.
Besides which, six months is nowhere near enough time to develop an entirely new admissions system that’s both without a nationally-standardised examination system and fair to all applicants from the whole country.
Besides, do you know how many individual applicants a given university gets in a given year? Are you suggesting that for each one the Admissions department should be having a chat with their teachers to see how good they are? Even if so many conversations were possible, how on earth would you assess a teacher from Oldham’s opinion of their student against that of another teacher from Bognor Regis? And would those teachers have the time to have six such conversations (the maximum number of universities a student can apply to through UCAS) for each of their students?
“Practical” is not the word I would use for such a suggestion.
Totally agree. No-one on this thread so far has been able to suggest a workable alternative. It would be impossible to have the sort of conversations to which Alan Cresswell refers in a way that eliminated personal biases - apart from the time such a procedure would take. Exams are far from perfect but what is?
A workable alternative would have been repeating the academic year and taking the exams in summer 2021.
Be careful using the term ‘predicted grades’.
The teacher assessment grades sent to exam boards based on continuous assessment are a one off (hopefully).
Students usually have predicted grades on their university application that are renowned to be highly inaccurate & depend more on how much they beg & cajole their teachers & tutors & how realistic their choices are in relation to their abilities & efforts.
(I work with A-Level students)
The figure I saw quoted was 75% of predicted grade were over generous.
The problem is that they are trying to synthesize data where none exists, this is a policy failure rather than something that can be fixed with an algorithm.
Last year it seems that 75% of predictions were over generous vs the real grade. Why are we surprised at a 40% mark down?
Not so much the system (although the latter is risible) but can we ask the question why the teachers are getting their predictions wrong year after year? After all, they know and teach the pupils.
We had that discussion a while back, and I remain unconvinced that what you suggest is actually workable, but as neither of us convinced the other before I can’t see any point in re-hashing the argument.
A lot of that argument seemed to consist of, kids have worked really hard for their qualifications this year we’ll demotivate them if we hold them back - I am suggesting it would be less demotivating than just under a half the kids being screwed over in this manner.
A lot of that argument seemed to consist of, kids have worked really hard for their qualifications this year we’ll demotivate them if we hold them back - I am suggesting it would be less demotivating than just under a half the kids being screwed over in this manner.
I think the bigger arguments were logistical. Voluntary re-takes with a "no detriment" requirement for universities considering applications from those sitting them would suffice.
A workable alternative would have been repeating the academic year and taking the exams in summer 2021.
Certainly an alternative, but how workable would it have been? You'd have had double the usual number of students next year (assuming that the crisis has died away by the start of the next school year) and it's pretty unlikely that extra staff and accommodation could have been available in time. Then there'd be a similar doubling up at tertiary institutions at the start of the 2121 academic year.
Be careful using the term ‘predicted grades’.
The teacher assessment grades sent to exam boards based on continuous assessment are a one off (hopefully).
Students usually have predicted grades on their university application that are renowned to be highly inaccurate & depend more on how much they beg & cajole their teachers & tutors & how realistic their choices are in relation to their abilities & efforts.
(I work with A-Level students)
The figure I saw quoted was 75% of predicted grade were over generous.
The problem is that they are trying to synthesize data where none exists, this is a policy failure rather than something that can be fixed with an algorithm.
Last year it seems that 75% of predictions were over generous vs the real grade. Why are we surprised at a 40% mark down?
Not so much the system (although the latter is risible) but can we ask the question why the teachers are getting their predictions wrong year after year? After all, they know and teach the pupils.
You can either assume that teachers predict potential rather than results or that teachers predictions are inherently flawed -- in which case the current algorithm still assumes they can perfectly rank order their classes.
@Marvin the Martian - I didn't realise that government controls were so restrictive. And by "The Universities" I mean also the government, so I meant the entire sector. I didn;t mean to dismiss the incredible with that you and so many others have been doing these last few months.
I suppose I mean that this was clearly coming. And it seems that nobody - i.e. Gavin WIlliamson - just ignored it.
Be careful using the term ‘predicted grades’.
The teacher assessment grades sent to exam boards based on continuous assessment are a one off (hopefully).
Students usually have predicted grades on their university application that are renowned to be highly inaccurate & depend more on how much they beg & cajole their teachers & tutors & how realistic their choices are in relation to their abilities & efforts.
(I work with A-Level students)
The figure I saw quoted was 75% of predicted grade were over generous.
The problem is that they are trying to synthesize data where none exists, this is a policy failure rather than something that can be fixed with an algorithm.
Last year it seems that 75% of predictions were over generous vs the real grade. Why are we surprised at a 40% mark down?
Not so much the system (although the latter is risible) but can we ask the question why the teachers are getting their predictions wrong year after year? After all, they know and teach the pupils.
I've seen that the teacher grade is partially a capability\potential assessment (so assumes they don't have a horrendous breakup, the wrong exam questions).
The mark down is ugly but I can see why it was seen as necessary.
Realistically it's not surprising students don't perform at their best, but there isn't a nice way to say it when your the one deciding the reality.
That also includes the beginnings of correction for teachers bias. It needed handling nicer, and the students need support, but is not bad in itself.
The recreation of systemamic inequalities (school weightings) is even more ugly. I really don't like it, but don't really have a better idea.
The way the system then was selectively applied to let the small classes off that grade correction, giving a 5% increase in independent a grades on the other hand. That is something that should have been caught, and needs fixing asap. I gather there are some rounding things that bring the corrected values even lower so the students at large colleges are being tripply screwed.
The mark down is ugly but I can see why it was seen as necessary.
Realistically it's not surprising students don't perform at their best, but there isn't a nice way to say it when your the one deciding the reality.
That also includes the beginnings of correction for teachers bias. It needed handling nicer, and the students need support, but is not bad in itself.
The problem is that the adjustments don't take account of exceptional individuals, nor of the fact that while teachers may have over-predicted on average, a lot of that is because if you've got 10 candidates who are capable of a B you've little way of knowing which of those 10 will achieve it and which will fall short because it's essentially down to a combination of what happens between prediction and exam and sheer chance of which questions come up, the vagaries of marking (particularly in essay-based subjects) and who is suffering badly from a high pollen count or period cramps on the day of the exam. And that assumes the downgrading was proportionate and consistent, which it doesn't appear to have been.
Maybe certificates should simply be awarded with a teacher-assessed grade as well as the calculated grade, so employers and universities can be clear about what happened, if the government aren't prepared to accept a one-off spike in attainment.
The rounding down error is particular stupid as it pushes all grades down. And the giving out of Us is abhorrent, that is a grade that is given out for not turning up for the exam so shouldn’t be given out according to supposed ability as that isn’t relevant.
I don’t understand how or why they didn’t pick up on these issues. My university faculty was lucky in that our modules have graded essays throughout so when our end of year assessments were cancelled we had moderated coursework to rely on. But, never the less, we went to a great effort to consider the situation of the students who fell below grade boundaries, allow for extenuating circumstances and recognise that the students didn’t know all their weighting was going to be on the coursework; being seen to be fair was considered important. This does not appear to have happened with A levels.
From this Guardian article.
So the first problem is that the decision has been made to round averages down. Say it looks at the school's performance over the last five years. If one person in the school failed in the last five years then the algorithm will make somebody fail this year (even it appears if that person was an outlier). Meanwhile, if a school had one A* level in four of those years, but none in one year, nobody in the school will get an A*.
But as a second problem, according to the Guardian there are other "corrections" at work: so that the rounding based on past attainments has meant that one school with a 12.5% historical average A* has been given a 5.7% rating, which is then rounded down again as in the previous paragraph. So that a school that has been getting three to four A* per year on average and no Us is getting one of each.
So it appears that a school's past performance is being treated as a cap not as a target, and likewise the school's predictions for its students are being treated as caps not targets. That seems as if the algorithm is looking for any excuse to get you.
Looking into the link from the Guardian article cited, it appears that the downgrading to allow for over-optimistic teacher assessments is based on the nationwide trend. So that schools that have striven not to be over-optimistic are being penalised for other schools' optimism and schools that are more optimistic than average are being rewarded.
(My political prejudices are that independent schools and academies are going to be more motivated to be over-optimistic.)
From this Guardian article.
So the first problem is that the decision has been made to round averages down. Say it looks at the school's performance over the last five years. If one person in the school failed in the last five years then the algorithm will make somebody fail this year (even it appears if that person was an outlier). Meanwhile, if a school had one A* level in four of those years, but none in one year, nobody in the school will get an A*.
But as a second problem, according to the Guardian there are other "corrections" at work: so that the rounding based on past attainments has meant that one school with a 12.5% historical average A* has been given a 5.7% rating, which is then rounded down again as in the previous paragraph. So that a school that has been getting three to four A* per year on average and no Us is getting one of each.
So it appears that a school's past performance is being treated as a cap not as a target, and likewise the school's predictions for its students are being treated as caps not targets. That seems as if the algorithm is looking for any excuse to get you.
It seems like they have over-adjusted where they did have statistical evidence in order to counteract the rise in grades at centres where they did not (because of new courses or small cohorts). This is why private schools have done well out of it and large 6th form colleges been utterly screwed.
The rounding down error is particular stupid as it pushes all grades down. And the giving out of Us is abhorrent, that is a grade that is given out for not turning up for the exam so shouldn’t be given out according to supposed ability as that isn’t relevant.
Not true.
If you don’t turn up it’s an X. Lots of students achieve a U in an A-Level each year - even if the proportion is relatively small. & indeed some of those appealing having received D’s & E’s would have actually achieved a U had they sat the exams & may well do so if they sit the Autumn exam series.
And that assumes the downgrading was proportionate and consistent, which it doesn't appear to have been.
Is is my belief also that it wasn't consistent (hopefully you got them from the 2nd and 3rd parts). Wrt that, my only questions are deciding if it were too much or far too much, if it were negligence or willful and (relevant to the board title) will justice be done in this world or the next. (technically I do, also leave the possibility of being surprised by new evidence).
Maybe certificates should simply be awarded with a teacher-assessed grade as well as the calculated grade, so employers and universities can be clear about what happened, if the government aren't prepared to accept a one-off spike in attainment.
I think something of the sort should have happened, don't pretend it's comparable (related yes) to the normal assessment of A-Levels. Call them Red, Orange... or something.
(and give the option of taking the exams in future years)
The problem is that the adjustments don't take account of exceptional individuals, ...
I'm going to say that's not a problem caused by the downgrading itself. But it is related to why I think it ugly.
Given these problems and the loss of agency that's already occured (although hayfever, isn't exactly a choice, and I hadn't thought about discriminatory issues, that the traditional exam system brings up), you need to be very sensitive.
And there is a lot to think about about that section as issues of it's own.
@Marvin the Martian - I didn't realise that government controls were so restrictive
They usually aren’t. But the first strategy the top universities thought of to make up for their predicted loss of overseas students was to massively increase their home recruitment. Which would probably have resulted in some of the lowest-ranked universities folding due to lack of students. Despite all its bluster about creating a market in HE, the government wasn’t prepared to allow that outcome so it told all universities that they couldn’t recruit more students than they’d already submitted in their financial forecasts.
They’d been trying to get rid of unconditional or non-results-based offers for a while, Covid just gave them the perfect excuse.
If you don’t turn up it’s an X. Lots of students achieve a U in an A-Level each year - even if the proportion is relatively small. & indeed some of those appealing having received D’s & E’s would have actually achieved a U had they sat the exams & may well do so if they sit the Autumn exam series.
In any individual case the odds are against it (if it were likely they wouldn't be appealing).
There will be some students who were predicted a C's who would get a U (I got a U in a general studies AS exam part, and 100%(scaled) in the matching A2, i'm not sure what my prediction was), but the algorithm has no way of picking up who that will be.
It depends why you don't turn up. I missed one of my GCSE Latin papers completely due to a badly timed attack of Norovirus. I was therefore marked on just the one I had done. I should have a # next to the grade on my certificate, but they cocked up the first result and sent it as an E which was challenged straight away by my school, and the resulting correction upwards missed off the #.
As to interviews, apart from Oxford and Cambridge, and Medicine type courses, they tend to be limited to practical subjects that want to see a portfolio. One of the local universities has a very strong Art and Design offering, and in the spring it's not uncommon to see prospective students around there clutching their folders of work. I think for the university, what they can actually produce is more important than grades.
It's strange now to think that my experience of applying is now out of date, and that the four interviews I had from 6 applications (Cambridge, UCL, Royal Holloway & Lancaster interviewed, Edinburgh and Durham did not) for Maths/Physics courses are no longer the norm.
It seems like they have over-adjusted where they did have statistical evidence in order to counteract the rise in grades at centres where they did not
Yeah, and a moments thought shows how ridiculous this is. Incidentally, the report itself is illuminating in other ways; at GCSE level - and once you get past maths and sciences = there's a less than 70% chance that different markers would award the same grade to the same paper/body of work.
As I said above; this is a policy issue, and one that has been made worse by successive Education Ministers (most notably Gove) having deprecated coursework in favour of a final set of exams.
Ah yes, Interviews. I suppose the sheer number applying to university leads to a large degree of anonymity. As well as getting good grades (for those days), I had to attend an interview and take an essay-based entrance exam. And that was for Teacher Training.
A workable alternative would have been repeating the academic year and taking the exams in summer 2021.
Certainly an alternative, but how workable would it have been? You'd have had double the usual number of students next year (assuming that the crisis has died away by the start of the next school year) and it's pretty unlikely that extra staff and accommodation could have been available in time. Then there'd be a similar doubling up at tertiary institutions at the start of the 2121 academic year.
This is not quite right. I don't know if delaying everything for a year is the right approach but it is certainly practicable.
Imagine a 3 year degree in Shipoffoolsology* at Uxbridge University. Every year has approximately 50 students so the undergraduate body is 150. If we had no 2020 intake, then for 20/21 there would be 100 students in total. A loss of income for the uni that needs covering has to be considered. In 2021, the intake is 100 students instead of the usual 50 but because there is an empty 2nd year, this is manageable- ultimately instead of running SOF20x courses, you run two streams of SOF10x course if needed. So for 21/22, the undergraduate body is back to 150. In 22/23 it's the second year that poses the same problem but there is no 3rd year. 2023/24 is the year where you end up with a larger student body at 200 undergraduates. So you have 3 years to plan and prepare.
As I said, I don't know if it's the right approach but it certainly is possible.
Medicine would have an issue as the annual new crop of baby doctors are important to the workforce. But with 5 years to plan for it, I would not think that a deal breaker.
I am struggling to find detailed information on the algorithm but using the school average was always going to be desperately unfair.
We already have an education system that entrenched wealth inequality. This seems to be a system designed to make that even worse.
It stinks.
AFZ
*I will not speculate here on whether a degree in Shipoffoolsology would be a BSc or a BA... although I reckon we might have fun choosing the faculty. Presumably lectures would take place in the Aligator Lecture Theatre.
If you don’t turn up it’s an X. Lots of students achieve a U in an A-Level each year - even if the proportion is relatively small. & indeed some of those appealing having received D’s & E’s would have actually achieved a U had they sat the exams & may well do so if they sit the Autumn exam series.
My apologies, thank you.
But it still isn’t a grade that should be handed out so easily, especially to people predicted a C. And some schools have had students receive a U when the school has never had a U before.
I am struggling to find detailed information on the algorithm but using the school average was always going to be desperately unfair.
Just assigning each pupil the average that their school got would be desperately unfair, but that's clearly not what they've done. Assigning pupils at a school grades based on the past average and distribution might be a little unfair on the outliers at the school, but would probably be fair to most pupils. But that's not what they've done either.
I too am struggling to find detailed information on the algorithm, but school past performance seems to have little effect on the final outcome. (See the example in the Guardian of a school which has had a 12.5% average A* grade for one subject in previous years, given only one pupil i.e. 3.7% with that grade.) That is, schools' past performance is being downgraded based on the "prior attainment" of the pupils. How and why that is done is it seems completely opaque. What does seem to be the case is that, among other effects, schools with a history of good teaching, that help their pupils achieve better than they previously did, are having that good teaching factored out. The system is set up to favour schools with a selective intake, above and beyond those schools' past performance.
I suppose they could have just left the predicted results, without adjusting them down. Then someone with an A* might have been told to 'get in the queue with all those other feckers'.
I suppose Universities who were really bothered, might have viva-ed applicants to sort them out.
It fucks me off a bit, that we are - oh, a few weeks now - past the point where we all wondered if we might die, and now we're right back into a big 'It's Not Fair' trip. It wasn't very fair on those bus drivers, doctors and nurses, the disabled, the elderly, and all those fat people, black people, and men. It still isn't fair for all those Brazilians, Mexicans and Texans, Floridians and Californians; we have brushed the New Yorkers under the carpet and the dead residents of Bergamo are probably just about all buried by now. It isn't fair that Mancunian lives are restricted by the antics of a small number of Oldhamers. It isn't fair that my neighbours don't give a shit and are having another party tonight. They're fat and black and some are men; well, they probably won't get ill and that won't be fair either.
There is a difference between unfairness caused by deliberate decision making and unfairness caused by accident. There is also a difference between unfairness that can be rectified and that which can't.
Imagine a 3 year degree in Shipoffoolsology* at Uxbridge University....
*I will not speculate here on whether a degree in Shipoffoolsology would be a BSc or a BA... although I reckon we might have fun choosing the faculty. Presumably lectures would take place in the Aligator Lecture Theatre.
I think the degree in Chaplaincy Studies dreamt up by my friends when I was at university was individual choice as to BSc or BA, though we were quite Maths biased.
[/tangent]
This is also an act of organizational unfairness that is going to change some people's lives forever by wrecking career chances: if you've just been rejected as a student Vet say, it would be hard to go and pick that up in the future - for starters there are very few spaces and a lot of competition, even for graduate entry courses.
One of the Oxford colleges has said that it will just accept all the people it made offers to, as that's the fairest option, and I have seen an alumni campaign circulating for Cambridge to do the same.
This is also an act of organizational unfairness that is going to change some people's lives forever by wrecking career chances: if you've just been rejected as a student Vet say, it would be hard to go and pick that up in the future - for starters there are very few spaces and a lot of competition, even for graduate entry courses.
In many ways it is about to be worse than that when the same method is used for GCSE results in a few weeks time and a bunch of kids don't get 'C's in Maths/English (the latter of which is only marked consistently 50% of the time - see figures above).
This is also an act of organizational unfairness that is going to change some people's lives forever by wrecking career chances: if you've just been rejected as a student Vet say, it would be hard to go and pick that up in the future - for starters there are very few spaces and a lot of competition, even for graduate entry courses.
In many ways it is about to be worse than that when the same method is used for GCSE results in a few weeks time and a bunch of kids don't get 'C's in Maths/English (the latter of which is only marked consistently 50% of the time - see figures above).
Very true. I see a lot of resits for those in the next year. If the government actually tries to publish any kind of league table, especially based on "student progress", I can see them being thoroughly derided.
The borderline 3/4 kids will be on tenterhooks this week. Having a grade ‘C’ equivalent in English, and to a lesser extent, maths has a big impact on options available. If under 19 they are expected to continue to work towards this otherwise their education provider aren’t funded for them to do their main programme of study.
Very true. I see a lot of resits for those in the next year. If the government actually tries to publish any kind of league table, especially based on "student progress", I can see them being thoroughly derided.
Only by the people that care, not by the people that make the decisions.
I had a look to see how the U grades changed, and on whatever that search picked up it's was a bit depressing the wording for places celebrating their 'fantastic/excellent/unprecedented' results (though not sure what I'd actually write). The unprecedented one, unsurprisingly, seems to have about 20 students a year.
Very true. I see a lot of resits for those in the next year. If the government actually tries to publish any kind of league table, especially based on "student progress", I can see them being thoroughly derided.
Only by the people that care, not by the people that make the decisions.
I had a look to see how the U grades changed, and on whatever that search picked up it's was a bit depressing the wording for places celebrating their 'fantastic/excellent/unprecedented' results (though not sure what I'd actually write). The unprecedented one, unsurprisingly, seems to have about 20 students a year.
Well, I suppose those forgotten 'Red Wall' places will get what they voted for, and their young people will be less able to leave for work/college/university.
Comments
We know someone who is a senior lecturer at a Russell Group university in Computer Science. They were expecting to be between 200 and 300% over-subscribed on people who had met their (very high) offers. The same was true of Mathematics, Business, and Economics, but most other departments were expecting spaces, as there is a trend towards certain degree subjects being chosen for employability, and I think that the private schools probably push this a bit more.
I believe most universities have set up a system for letting them know about grade appeals for those who have missed out or are in clearing.
Ofqual tested the model they used on the 2019 data, and even where the model worked best were predicting 1/3 grades incorrectly.
The problem is that they are trying to synthesize data where none exists, this is a policy failure rather than something that can be fixed with an algorithm.
One other possible side effect of this has been a much greater proportion of students than normal getting higher grades in Further Maths than they do in Maths.
Very true. Very true indeed.
That would be fine, if it weren’t for the fact that the government, in its wisdom, decided to impose a student number control (i.e. a maximum number of students we can recruit) on us this year. In theory this was to protect the worse universities from potential failure due to all the better universities over-recruiting UK students to make up for their predicted drop in overseas ones.
Neither UCAS nor the government would allow us to make offers that weren’t based on grades. And we’ve been kind of busy working out how the hell to balance next years budget without most of (the income from) our overseas students. While everyone has been working from home (often with kids under our feet) because all the campuses are closed due to Covid.
Besides which, six months is nowhere near enough time to develop an entirely new admissions system that’s both without a nationally-standardised examination system and fair to all applicants from the whole country.
The government banned us from making unconditional offers this year as well, ostensibly for the same reason they imposed the student number control.
Besides, do you know how many individual applicants a given university gets in a given year? Are you suggesting that for each one the Admissions department should be having a chat with their teachers to see how good they are? Even if so many conversations were possible, how on earth would you assess a teacher from Oldham’s opinion of their student against that of another teacher from Bognor Regis? And would those teachers have the time to have six such conversations (the maximum number of universities a student can apply to through UCAS) for each of their students?
“Practical” is not the word I would use for such a suggestion.
Wait - that actually happens? There are people who take Further Maths who can't get an A**!, or whatever the top grade is these days, in Maths with one hand tied behind their back?
Back in my day, my school wouldn't let anyone who wasn't a cast-iron guaranteed A at Maths A-level take the Further Maths course.
No (well I'm sure events happen), but this year children have been (dis)credited as though it happened (the proportion of grades is the same).
One possibility is that Maths as a large subject they've corrected for the teachers being wrong, while for Further Maths they've trusted the teachers. [See Chorister's post]
I don't know if there's a way to muddle up distributions, for another explaination.
Why are ‘we’ more worried about predicting too high than too low ? Why not have the teachers predictions and accept the year results will be anomalous ?
At least gov is looking to fund students who want to resit their second year.
This reliance hasn't been removed by using an algorithm -- which in this case still depends on teacher's ability to perfectly rank order their students.
But if you are going to correct you need to do it consistently. Which they didn't. Hence somehow Independent Schools ended up getting 5% more A's than usual.
They
a) Should have looked at this ages ago. Granted the Education department itself should have been looking at getting schools to work, but the exam boards had free time, they don't have to actually mark the papers.
b) If/As you can't get something that replicates normal A-levels, they shouldn't have tried passing them off as them. Present them in an exceptional way matching the exceptional circumstance (perhaps replace with a nominal normal grade in 5 years time, [when resits could have happened, or they've graduated, got work experience]).
c) Now they shouldn't try to pretend that this is robust or dependable.
d) And part of fixing it means allowing the universities to do their bit.
Totally agree. No-one on this thread so far has been able to suggest a workable alternative. It would be impossible to have the sort of conversations to which Alan Cresswell refers in a way that eliminated personal biases - apart from the time such a procedure would take. Exams are far from perfect but what is?
When, exactly? The UCAS process is year-long, and was already well underway by the time lockdown started. And remember, back then they were saying it would only be for 12 weeks tops.
If we’d known in March that we’d be where we are now then maybe something could have been cobbled together, but we didn’t. And even if we had, there would have been some imperfection in the jury-rigged system that we put together at the last minute that people would have been having a go at. Do you think it’s easy doing this job?
In a word, no. For some of the more selective courses, sure, but for the majority it’s all done by UCAS.
I seriously doubt that was in any way the reason for their choices. It’s far more likely that it was just an attempt to replicate the normal exams without actually having exams, and somewhere in the mess of algorithms and statistical analyses it ended up going wrong.
A workable alternative would have been repeating the academic year and taking the exams in summer 2021.
Last year it seems that 75% of predictions were over generous vs the real grade. Why are we surprised at a 40% mark down?
Not so much the system (although the latter is risible) but can we ask the question why the teachers are getting their predictions wrong year after year? After all, they know and teach the pupils.
I think the bigger arguments were logistical. Voluntary re-takes with a "no detriment" requirement for universities considering applications from those sitting them would suffice.
Certainly an alternative, but how workable would it have been? You'd have had double the usual number of students next year (assuming that the crisis has died away by the start of the next school year) and it's pretty unlikely that extra staff and accommodation could have been available in time. Then there'd be a similar doubling up at tertiary institutions at the start of the 2121 academic year.
You can either assume that teachers predict potential rather than results or that teachers predictions are inherently flawed -- in which case the current algorithm still assumes they can perfectly rank order their classes.
I suppose I mean that this was clearly coming. And it seems that nobody - i.e. Gavin WIlliamson - just ignored it.
The mark down is ugly but I can see why it was seen as necessary.
Realistically it's not surprising students don't perform at their best, but there isn't a nice way to say it when your the one deciding the reality.
That also includes the beginnings of correction for teachers bias. It needed handling nicer, and the students need support, but is not bad in itself.
The recreation of systemamic inequalities (school weightings) is even more ugly. I really don't like it, but don't really have a better idea.
The way the system then was selectively applied to let the small classes off that grade correction, giving a 5% increase in independent a grades on the other hand. That is something that should have been caught, and needs fixing asap. I gather there are some rounding things that bring the corrected values even lower so the students at large colleges are being tripply screwed.
The problem is that the adjustments don't take account of exceptional individuals, nor of the fact that while teachers may have over-predicted on average, a lot of that is because if you've got 10 candidates who are capable of a B you've little way of knowing which of those 10 will achieve it and which will fall short because it's essentially down to a combination of what happens between prediction and exam and sheer chance of which questions come up, the vagaries of marking (particularly in essay-based subjects) and who is suffering badly from a high pollen count or period cramps on the day of the exam. And that assumes the downgrading was proportionate and consistent, which it doesn't appear to have been.
Maybe certificates should simply be awarded with a teacher-assessed grade as well as the calculated grade, so employers and universities can be clear about what happened, if the government aren't prepared to accept a one-off spike in attainment.
I don’t understand how or why they didn’t pick up on these issues. My university faculty was lucky in that our modules have graded essays throughout so when our end of year assessments were cancelled we had moderated coursework to rely on. But, never the less, we went to a great effort to consider the situation of the students who fell below grade boundaries, allow for extenuating circumstances and recognise that the students didn’t know all their weighting was going to be on the coursework; being seen to be fair was considered important. This does not appear to have happened with A levels.
So the first problem is that the decision has been made to round averages down. Say it looks at the school's performance over the last five years. If one person in the school failed in the last five years then the algorithm will make somebody fail this year (even it appears if that person was an outlier). Meanwhile, if a school had one A* level in four of those years, but none in one year, nobody in the school will get an A*.
But as a second problem, according to the Guardian there are other "corrections" at work: so that the rounding based on past attainments has meant that one school with a 12.5% historical average A* has been given a 5.7% rating, which is then rounded down again as in the previous paragraph. So that a school that has been getting three to four A* per year on average and no Us is getting one of each.
So it appears that a school's past performance is being treated as a cap not as a target, and likewise the school's predictions for its students are being treated as caps not targets. That seems as if the algorithm is looking for any excuse to get you.
(My political prejudices are that independent schools and academies are going to be more motivated to be over-optimistic.)
It seems like they have over-adjusted where they did have statistical evidence in order to counteract the rise in grades at centres where they did not (because of new courses or small cohorts). This is why private schools have done well out of it and large 6th form colleges been utterly screwed.
Not true.
If you don’t turn up it’s an X. Lots of students achieve a U in an A-Level each year - even if the proportion is relatively small. & indeed some of those appealing having received D’s & E’s would have actually achieved a U had they sat the exams & may well do so if they sit the Autumn exam series.
I think something of the sort should have happened, don't pretend it's comparable (related yes) to the normal assessment of A-Levels. Call them Red, Orange... or something.
(and give the option of taking the exams in future years)
I'm going to say that's not a problem caused by the downgrading itself. But it is related to why I think it ugly.
Given these problems and the loss of agency that's already occured (although hayfever, isn't exactly a choice, and I hadn't thought about discriminatory issues, that the traditional exam system brings up), you need to be very sensitive.
And there is a lot to think about about that section as issues of it's own.
They usually aren’t. But the first strategy the top universities thought of to make up for their predicted loss of overseas students was to massively increase their home recruitment. Which would probably have resulted in some of the lowest-ranked universities folding due to lack of students. Despite all its bluster about creating a market in HE, the government wasn’t prepared to allow that outcome so it told all universities that they couldn’t recruit more students than they’d already submitted in their financial forecasts.
They’d been trying to get rid of unconditional or non-results-based offers for a while, Covid just gave them the perfect excuse.
There will be some students who were predicted a C's who would get a U (I got a U in a general studies AS exam part, and 100%(scaled) in the matching A2, i'm not sure what my prediction was), but the algorithm has no way of picking up who that will be.
As to interviews, apart from Oxford and Cambridge, and Medicine type courses, they tend to be limited to practical subjects that want to see a portfolio. One of the local universities has a very strong Art and Design offering, and in the spring it's not uncommon to see prospective students around there clutching their folders of work. I think for the university, what they can actually produce is more important than grades.
Yeah, and a moments thought shows how ridiculous this is. Incidentally, the report itself is illuminating in other ways; at GCSE level - and once you get past maths and sciences = there's a less than 70% chance that different markers would award the same grade to the same paper/body of work.
As I said above; this is a policy issue, and one that has been made worse by successive Education Ministers (most notably Gove) having deprecated coursework in favour of a final set of exams.
This is not quite right. I don't know if delaying everything for a year is the right approach but it is certainly practicable.
Imagine a 3 year degree in Shipoffoolsology* at Uxbridge University. Every year has approximately 50 students so the undergraduate body is 150. If we had no 2020 intake, then for 20/21 there would be 100 students in total. A loss of income for the uni that needs covering has to be considered. In 2021, the intake is 100 students instead of the usual 50 but because there is an empty 2nd year, this is manageable- ultimately instead of running SOF20x courses, you run two streams of SOF10x course if needed. So for 21/22, the undergraduate body is back to 150. In 22/23 it's the second year that poses the same problem but there is no 3rd year. 2023/24 is the year where you end up with a larger student body at 200 undergraduates. So you have 3 years to plan and prepare.
As I said, I don't know if it's the right approach but it certainly is possible.
Medicine would have an issue as the annual new crop of baby doctors are important to the workforce. But with 5 years to plan for it, I would not think that a deal breaker.
I am struggling to find detailed information on the algorithm but using the school average was always going to be desperately unfair.
We already have an education system that entrenched wealth inequality. This seems to be a system designed to make that even worse.
It stinks.
AFZ
*I will not speculate here on whether a degree in Shipoffoolsology would be a BSc or a BA... although I reckon we might have fun choosing the faculty. Presumably lectures would take place in the Aligator Lecture Theatre.
I too am struggling to find detailed information on the algorithm, but school past performance seems to have little effect on the final outcome. (See the example in the Guardian of a school which has had a 12.5% average A* grade for one subject in previous years, given only one pupil i.e. 3.7% with that grade.) That is, schools' past performance is being downgraded based on the "prior attainment" of the pupils. How and why that is done is it seems completely opaque. What does seem to be the case is that, among other effects, schools with a history of good teaching, that help their pupils achieve better than they previously did, are having that good teaching factored out. The system is set up to favour schools with a selective intake, above and beyond those schools' past performance.
I suppose Universities who were really bothered, might have viva-ed applicants to sort them out.
It fucks me off a bit, that we are - oh, a few weeks now - past the point where we all wondered if we might die, and now we're right back into a big 'It's Not Fair' trip. It wasn't very fair on those bus drivers, doctors and nurses, the disabled, the elderly, and all those fat people, black people, and men. It still isn't fair for all those Brazilians, Mexicans and Texans, Floridians and Californians; we have brushed the New Yorkers under the carpet and the dead residents of Bergamo are probably just about all buried by now. It isn't fair that Mancunian lives are restricted by the antics of a small number of Oldhamers. It isn't fair that my neighbours don't give a shit and are having another party tonight. They're fat and black and some are men; well, they probably won't get ill and that won't be fair either.
Private Eye is not innocent in this regard.
[/tangent]
This is also an act of organizational unfairness that is going to change some people's lives forever by wrecking career chances: if you've just been rejected as a student Vet say, it would be hard to go and pick that up in the future - for starters there are very few spaces and a lot of competition, even for graduate entry courses.
One of the Oxford colleges has said that it will just accept all the people it made offers to, as that's the fairest option, and I have seen an alumni campaign circulating for Cambridge to do the same.
In many ways it is about to be worse than that when the same method is used for GCSE results in a few weeks time and a bunch of kids don't get 'C's in Maths/English (the latter of which is only marked consistently 50% of the time - see figures above).
Very true. I see a lot of resits for those in the next year. If the government actually tries to publish any kind of league table, especially based on "student progress", I can see them being thoroughly derided.
I had a look to see how the U grades changed, and on whatever that search picked up it's was a bit depressing the wording for places celebrating their 'fantastic/excellent/unprecedented' results (though not sure what I'd actually write). The unprecedented one, unsurprisingly, seems to have about 20 students a year.
Well, I suppose those forgotten 'Red Wall' places will get what they voted for, and their young people will be less able to leave for work/college/university.