The government’s handling of the exam results situation is not seen as particularly successful with only 17% of voters in England approving and 42% disapproving.
Wonderfully dry understatement there CHB.
‘The war situation has developed, not altogether to Japan’s advantage.’
Perhaps the government should award the teachers' grades for GCSEs but stick with the algorithm grades for A-levels. Appeals against GCSE grades could overload the system.
However a big increase in A-level grades would disrupt the universities as some would have too many students meeting their offers, resulting in others having too few.
Does anyone know how many extra places would be needed and where if all students with offers were given places? Might be the only way out at this point
Would that many extra places be needed? Or would it just be fewer for clearing?
No idea, would be worth finding out. Then the actual A level grades could be sorted out later.
I may be wrong but as far as I know Scotland has created 10,000 new places
But this is complete tosh. The Scottish government pays Universities the princely sum of £1850 per student. Scottish Universities survive because this is cross subsidised by English students pay £9250 and non EU students paying £20k+. Scottish Universities face a financial crisis and simply cannot afford to take on another 10k additional Scottish students that cost them money.
I am correct though on the 10,000 places am I not
The Scottish Government has asked for them, do we have any evidence that the universities actually created the places and are admitting students?
Yes.
I don't see anyone in Scotland still appearing on the media saying they've been denied a place despite getting the grades now. In which case absence of evidence is itself evidence of absence.
The only place those stories would appear would be in very local papers, the news agenda moved on once SQA changed their results.
Surely the first test of algorithm would have been to predict accurately last years exam results. I wonder if they did that.
Williamson must go. The Tories have fucked up royally.
There is no appropriate dataset from the last year to work from (exam boards didn't previously ask for predicted grades).
What they needed to do would be to identify statistical anomalies (the AAB that becomes AAU) and then either override the system to fix those or rewrite the algorithm until those issues disappeared.
There have been lots of reports on this UCAS constantly tack this. Grades predicted versus those achieved have been running at 40% overestimation for ages.
What people are getting excited about isnt HMG marking, but teachers overforecasting and setting pupil expectations at the wrong level.
Really this years pupils have nothing to moan about, every year behind them was the same pattern.
Really being awarded a U because the computer decided you were the student who walked out 10 minutes into the exam is not a reason to moan?
There have been daft anomalies for ever in exam marking. But the big pattern has always been predicted grades are higher than those achieved. My own kids never hit all the grades they were forecast to hit. The root cause of this isnt the algorithm but the base data is duff in the first place.
No doubt there are lots of reasons why, but until that is corrected the over prediction will still stay and lots of kids will be disappointed on the day.
The UK's Quangocracy really are having a shocker of a year.
Presumably you include Cummings and the Boris bus in that?
Boris has been elected within the past year and has a popular mandate.
The rest - not so much as a sniff of democracy comes their way....
Whereas technically you are correct, PM Johnson and Mr Cummings do seem to consider these semi-autonomous government bodies as handy shields behind which they can hide any errors.
That's not at all new - it has been a major part of the point of quangos (and criticism of them) for decades.
Not to be totally negative about it - it's not the worst thing for industries to be subject to the relative predictability of bureaucratic bodies rather than constant, fickle political intervention by headline obsessed politicians.
However, it only works to a degree. You can avoid commenting on last night's controversial TV show ("that's for Ofcom to say") or a gas billing snafu that crops up somewhere or other ("I've asked Ofgem to look into it urgently"). But, as Williamson is finding, you can't dodge the really big questions. A-levels and GCSEs are such a massive part of the school system, and this situation (deciding grades given inability to run exams) has been brewing for many months - as Education Secretary, you can't say "not really my remit".
Frankly I couldn't care less if they're a shield so long as the right thing is done now. Even if the answer is "Ofqual have called this wrong, I am stepping in now to overrule them and we will change the law to permit this" then I would be happy to accept that and move on.
It should never be too late to do the right thing. Ofqual were given a chance to call this right, they've failed, its time to move on.
I agree with that.
The point I was making is that quangos are a good way to avoid the need to give a running commentary on relatively small but somewhat complex day to day stuff about spectrum licensing, electricity and water markets, exam boards and so on.
But I agree that this isn't small stuff - it affects large numbers and is absolutely an area where the call for Government to step in isn't some kind of faddish or fringe thing.
The daft thing is this should have been obvious as soon as it became clear the exams couldn't go ahead. The Education Secretary should've been all over it - good politicians see the political weather coming.
Thinking about the 'need' to avoid exceptionally good results in any one year, how did this cohort of students do at GCSE level. Was 2018 good, bad or around the mid-point f the expected range? Particularly for those pupils who were in the higher result brings, and, presumably, went on to A levels.
My son is back to school today. One of his pieces of "homework" over the summer was an entry into an economics essay competition comparing the effects of the Black Death and Covid.
Although there are some surprising similarities the major difference is the scale. When I was a lad the general presumption was that 1/3 of the world (ie Europe) died as a result of the Black Death. The view from historians now seems to be that this was based on serious under estimates of where the population stood pre-plague and it was in fact more like 50-60% of the population who died in the various waves.
Which does rather put the 1-2% of Covid into perspective, doesn't it?
Especially when you consider the vast majority of those killed by covid would never had made it in life long enough to be killed by covid.
They would have died due to low life expectancy rates or from lack of treatment for the c0-morbidities they have.
Even by your quite low standards, that doesn’t seem to make sense.
Incidentally, did you know that statistically the most dangerous human activity is breathing? Everyone who breathes, dies.
It really is quite amazing that after such a long time so many people are so ignorant of what COVID is and who it affects.
You said the majority of those killed by Covid would not have lived long enough to die of it.
Which is an effect of this virus I will admit I was unaware of.
No it is one of the first things he's said that makes sense.
Life expectancy in the dark and middle ages were such that there wouldn't have been a pool of vulnerable elderly around to be killed off by a virus that targets them like COVID19.
Which would almost certainly be negated by the fact that many of them would have had underlying respiratory problems, e.g. TB, or gastro-intestinal problems, e.g. worms.
Perhaps the government should award the teachers' grades for GCSEs but stick with the algorithm grades for A-levels. Appeals against GCSE grades could overload the system.
However a big increase in A-level grades would disrupt the universities as some would have too many students meeting their offers, resulting in others having too few.
Does anyone know how many extra places would be needed and where if all students with offers were given places? Might be the only way out at this point
Would that many extra places be needed? Or would it just be fewer for clearing?
No idea, would be worth finding out. Then the actual A level grades could be sorted out later.
I may be wrong but as far as I know Scotland has created 10,000 new places
But this is complete tosh. The Scottish government pays Universities the princely sum of £1850 per student. Scottish Universities survive because this is cross subsidised by English students pay £9250 and non EU students paying £20k+. Scottish Universities face a financial crisis and simply cannot afford to take on another 10k additional Scottish students that cost them money.
I am correct though on the 10,000 places am I not
The Scottish Government has asked for them, do we have any evidence that the universities actually created the places and are admitting students?
Yes.
I don't see anyone in Scotland still appearing on the media saying they've been denied a place despite getting the grades now. In which case absence of evidence is itself evidence of absence.
The only place those stories would appear would be in very local papers, the news agenda moved on once SQA changed their results.
Local papers? What is this, the twentieth century?
If that was happening then it wouldn't take very long for something like that to go viral on Twitter and then be shared here by someone like CarlottaVance as evidence the SNP have screwed up.
My son is back to school today. One of his pieces of "homework" over the summer was an entry into an economics essay competition comparing the effects of the Black Death and Covid.
Although there are some surprising similarities the major difference is the scale. When I was a lad the general presumption was that 1/3 of the world (ie Europe) died as a result of the Black Death. The view from historians now seems to be that this was based on serious under estimates of where the population stood pre-plague and it was in fact more like 50-60% of the population who died in the various waves.
Which does rather put the 1-2% of Covid into perspective, doesn't it?
Especially when you consider the vast majority of those killed by covid would never had made it in life long enough to be killed by covid.
They would have died due to low life expectancy rates or from lack of treatment for the c0-morbidities they have.
Even by your quite low standards, that doesn’t seem to make sense.
Incidentally, did you know that statistically the most dangerous human activity is breathing? Everyone who breathes, dies.
It really is quite amazing that after such a long time so many people are so ignorant of what COVID is and who it affects.
You said the majority of those killed by Covid would not have lived long enough to die of it.
Which is an effect of this virus I will admit I was unaware of.
Its absolutely true. The numbers say yu have got to be pretty ill and old to die from COVID essentially. Over 80 with at least one co-morbodity.
In the middle ages, in case you were wondering, the was no such thing as managing illnesses like hypertension, heart disease and diabetes. Chaucer strangely doesn't refer to transplant surgery in the Canterbury tales.
Thinking about the 'need' to avoid exceptionally good results in any one year, how did this cohort of students do at GCSE level. Was 2018 good, bad or around the mid-point f the expected range? Particularly for those pupils who were in the higher result brings, and, presumably, went on to A levels.
Difficult to say, because it was the first year of new exams and OFQUAL fucked them up.
Perhaps the government should award the teachers' grades for GCSEs but stick with the algorithm grades for A-levels. Appeals against GCSE grades could overload the system.
However a big increase in A-level grades would disrupt the universities as some would have too many students meeting their offers, resulting in others having too few.
Does anyone know how many extra places would be needed and where if all students with offers were given places? Might be the only way out at this point
Would that many extra places be needed? Or would it just be fewer for clearing?
It's too late now. I would have thought all the spare places that were created in medicine/vet/ etc will have been filled by now.
What evidence do you have for that claim?
Especially if the government offers money to expand places while also considering the Scots fixed it after the fact without that being an issue and the absence of foreign students?
It's not a claim, it's a talking point, perhaps I should have added a Question mark as well?
Most universities (mine included) would happily bump up intake of UK students to help with the expected overseas student shortfall (albeit there's not much profit in home students compared to overseas). My uni's mid-point estimate is that overseas students are will be down 70%, so there will be spare capacity. However, as I understand it (can't find source now, may be something I was told at work) there's an agreement between universities not to increase home student intake over 5% above normal levels, to avoid completely shafting the lower ranked universities by taking all their students (universities further down could still be in a lot of trouble if the top 20 all take 5% extra, obviously, as that's one average sized university worth of students gone).
Thought that too. But the cap wasn't an entirely bad idea. Virtually everyone who wants to go to University can go now, so expansion of more prestigious courses and Unis will leave less prestigious ones in deep trouble.
Trouble is that the admissions process depends on the grades being reasonably predictable. The combination of ups and downs in the Ofqual algorithm was bad enough, the subsequent squirming is making things far worse.
Maybe I'm being a bit stupid here, but why is the expansion of more prestigious courses for this year's intake a problem? Just give the less prestigious institutions the money for the students they have lost, and everyone's happy, or am I missing something? Sure, it will cost extra, but surely not much in the grand scheme of coronavirus expenditure.
The UK's Quangocracy really are having a shocker of a year.
Presumably you include Cummings and the Boris bus in that?
Boris has been elected within the past year and has a popular mandate.
The rest - not so much as a sniff of democracy comes their way....
Whereas technically you are correct, PM Johnson and Mr Cummings do seem to consider these semi-autonomous government bodies as handy shields behind which they can hide any errors.
That's not at all new - it has been a major part of the point of quangos (and criticism of them) for decades.
Not to be totally negative about it - it's not the worst thing for industries to be subject to the relative predictability of bureaucratic bodies rather than constant, fickle political intervention by headline obsessed politicians.
However, it only works to a degree. You can avoid commenting on last night's controversial TV show ("that's for Ofcom to say") or a gas billing snafu that crops up somewhere or other ("I've asked Ofgem to look into it urgently"). But, as Williamson is finding, you can't dodge the really big questions. A-levels and GCSEs are such a massive part of the school system, and this situation (deciding grades given inability to run exams) has been brewing for many months - as Education Secretary, you can't say "not really my remit".
Frankly I couldn't care less if they're a shield so long as the right thing is done now. Even if the answer is "Ofqual have called this wrong, I am stepping in now to overrule them and we will change the law to permit this" then I would be happy to accept that and move on.
It should never be too late to do the right thing. Ofqual were given a chance to call this right, they've failed, its time to move on.
I agree with that.
The point I was making is that quangos are a good way to avoid the need to give a running commentary on relatively small but somewhat complex day to day stuff about spectrum licensing, electricity and water markets, exam boards and so on.
But I agree that this isn't small stuff - it affects large numbers and is absolutely an area where the call for Government to step in isn't some kind of faddish or fringe thing.
The daft thing is this should have been obvious as soon as it became clear the exams couldn't go ahead. The Education Secretary should've been all over it - good politicians see the political weather coming.
Yes agreed.
There should probably be a reshuffle next year once COVID is fading into history and we look to move on and rebuild. I would be quite OK with seeing Williamson go then, but for now he needs to do the right thing and deal with this.
On the topic of the exam fiasco: having skim-read OFQAL's methodological paper, it seems to me that the major unforced error in their work is their ad hoc small subject entry adjustment and how that interacts with the state/private school mix. Their own analysis shows that their algo boosts the number of A grades awarded to private school pupils by 4.7 % points compared to last year, far outstripping the improvement for any other type of school, and yet at no point do they consider this to be problematic or even worthy of note. Only in a country where it had become an ingrained norm for the wealthy to game the system would this be possible. Other than that, my sense is they adopted a reasonable methodology but that whatever they chose the result was likely to be politically untenable.
Using the percentage point increase makes it look bad. Using the percentage increase and it was about 10% in both private schools and in comprehensives (in other words for every 10 As that each type of school got last year they got 11 this year).
Independent schools are often academically selective, particularly at A-level and spend more per pupil. It is unsurprising that their results are better.
Give me the same time per pupil as my independent school colleagues get and I would expect to do just as well.
Surely the first test of algorithm would have been to predict accurately last years exam results. I wonder if they did that.
Williamson must go. The Tories have fucked up royally.
There is no appropriate dataset from the last year to work from (exam boards didn't previously ask for predicted grades).
What they needed to do would be to identify statistical anomalies (the AAB that becomes AAU) and then either override the system to fix those or rewrite the algorithm until those issues disappeared.
There have been lots of reports on this UCAS constantly tack this. Grades predicted versus those achieved have been running at 40% overestimation for ages.
What people are getting excited about isnt HMG marking, but teachers overforecasting and setting pupil expectations at the wrong level.
Really this years pupils have nothing to moan about, every year behind them was the same pattern.
Really being awarded a U because the computer decided you were the student who walked out 10 minutes into the exam is not a reason to moan?
There have been daft anomalies for ever in exam marking. But the big pattern has always been predicted grades are higher than those achieved. My own kids never hit all the grades they were forecast to hit. The root cause of this isnt the algorithm but the base data is duff in the first place.
No doubt there are lots of reasons why, but until that is corrected the over prediction will still stay and lots of kids will be disappointed on the day.
It only matters because we do UCAS applications before students get their results not after.
In reality the only people who should know where they are going before A level result days are those people heading off to do vocational subjects where interviews, tests or auditions are required.
Suggests that the vast majority of students are overpredicted in a normal year -> 75%.
Yes, that sounds about right. But this paper is about UCAS predicted grades. 75% overpredictions partly because teachers want to give their students the best chance of getting in to the university of their choice, even though they know the predictions are ambitious. The question is whether the predictions given to Ofqual were the same as the UCAS predictions. They should have been lower. UCAS predictions are what you hope they will get; Ofqual predictions should be what you think they will get.
But I suspect that in many institutions this distinction was not made.
I'd be interested in @ydoethur 's comments on this, as from memory, the exercise was conducted on the basis of realism rather than optimism by his school ?
In any event, outsize aviations between schools in the optimism of their predictions ought to have been something comparatively simple to control for, statistically.
The exercise which Ofsted undertook was a bizarre method of applying an overall population judgment to individuals, with the only individual input the class ranking orders (which were themselves artificially forced by banning equal rankings).
One of the legal actions against Ofqal argues that they exceed their statutory remit. I'm not clear on the details, as I just heard a snippet on the radio this morning, but will be very interesting to follow.
To be honest, I don’t know enough to comment. Certainly there has always been an issue with UCAS predicted grades not matching final grades, but as @Fysics_Teacher notes, these are often marginal calls. My own A-levels were ABC, but I was two marks in two subjects off AAB.
At A-level, the commonest predictions I made this year were jointly A and C, but as I noted I had a very able cohort. I predicted Ds as well, and got them. Nobody got less than a D, but then in the six years I have been teaching at that school only two students ever got Es, so that seems to me to be fair enough. I can’t remember that any were shockingly out of line with the UCAS predictions, and indeed some of them may have been an improvement on them. But as noted above, it is a very inexact science.
For GCSE I would rather not comment yet.
Incidentally, you mean OFQUAL not OFSTED.
There is a difference in predicted grdes between normal years and this year though. Normally a teacher deliberately bumping up the predicted grade two levels (per subject) will be counterproductive for the student, when they underperform in the actual A-Levels, do not make their offer and are left scrambling through clearing to find a last minute place.
This year the teachers had to submit predicted grades knowing that their predictions would not be compared against actual exam results. As teachers in England are trying to get the highest grades for their students there is a tension between "doing the right thing" and being very generous to their students.
Presumably in normal year the predicted A-Level grades are only assigned in the Autumn term for students who apply to uni, whereas this year the teachers have had to predict for all students in the Summer term.
My son is back to school today. One of his pieces of "homework" over the summer was an entry into an economics essay competition comparing the effects of the Black Death and Covid.
Although there are some surprising similarities the major difference is the scale. When I was a lad the general presumption was that 1/3 of the world (ie Europe) died as a result of the Black Death. The view from historians now seems to be that this was based on serious under estimates of where the population stood pre-plague and it was in fact more like 50-60% of the population who died in the various waves.
Which does rather put the 1-2% of Covid into perspective, doesn't it?
Especially when you consider the vast majority of those killed by covid would never had made it in life long enough to be killed by covid.
They would have died due to low life expectancy rates or from lack of treatment for the c0-morbidities they have.
Even by your quite low standards, that doesn’t seem to make sense.
Incidentally, did you know that statistically the most dangerous human activity is breathing? Everyone who breathes, dies.
It really is quite amazing that after such a long time so many people are so ignorant of what COVID is and who it affects.
You said the majority of those killed by Covid would not have lived long enough to die of it.
Which is an effect of this virus I will admit I was unaware of.
Its absolutely true. The numbers say yu have got to be pretty ill and old to die from COVID essentially. Over 80 with at least one co-morbodity.
In the middle ages, in case you were wondering, the was no such thing as managing illnesses like hypertension, heart disease and diabetes. Chaucer strangely doesn't refer to transplant surgery in the Canterbury tales.
Correct.
So the mortality rate would have been higher, because many of them would have underlying untreated conditions.
The fatality rate of bubonic plague today is around 10-14%. We don’t know the exact figure for Covid-19 which makes comparison difficult.
However as bubonic plague is not infectious unless it gets into the lungs, Covid-19 would have spread much more rapidly.
My son is back to school today. One of his pieces of "homework" over the summer was an entry into an economics essay competition comparing the effects of the Black Death and Covid.
Although there are some surprising similarities the major difference is the scale. When I was a lad the general presumption was that 1/3 of the world (ie Europe) died as a result of the Black Death. The view from historians now seems to be that this was based on serious under estimates of where the population stood pre-plague and it was in fact more like 50-60% of the population who died in the various waves.
Which does rather put the 1-2% of Covid into perspective, doesn't it?
Especially when you consider the vast majority of those killed by covid would never had made it in life long enough to be killed by covid.
They would have died due to low life expectancy rates or from lack of treatment for the c0-morbidities they have.
Even by your quite low standards, that doesn’t seem to make sense.
Incidentally, did you know that statistically the most dangerous human activity is breathing? Everyone who breathes, dies.
It really is quite amazing that after such a long time so many people are so ignorant of what COVID is and who it affects.
You said the majority of those killed by Covid would not have lived long enough to die of it.
Which is an effect of this virus I will admit I was unaware of.
Its absolutely true. The numbers say yu have got to be pretty ill and old to die from COVID essentially. Over 80 with at least one co-morbodity.
In the middle ages, in case you were wondering, the was no such thing as managing illnesses like hypertension, heart disease and diabetes. Chaucer strangely doesn't refer to transplant surgery in the Canterbury tales.
No the numbers don't say that.
The numbers say that with our healthcare, and with our treatments you are more likely to be pretty old and ill to die. But younger people especially those with co-morbidities are possible to die too even with our healthcare looking after them - and @ydoethur is right comorbidities and ill health were rife then.
With the absence of any antibiotics or medicine then young people with TB (a major issue then) or some other comorbidities could have been slaughtered in vast numbers then.
Here’s something else to throw into the mix about Teacher predictions. Isn’t it the case at A-Level now that the A* grade was introduced that was specifically determined not by performance in the exam, but (top) performance relative to the entire year’s exam cohort? How were teachers supposed to predict something which by definition is beyond their capability to predict?
Thinking about the 'need' to avoid exceptionally good results in any one year, how did this cohort of students do at GCSE level. Was 2018 good, bad or around the mid-point f the expected range? Particularly for those pupils who were in the higher result brings, and, presumably, went on to A levels.
Difficult to say, because it was the first year of new exams and OFQUAL fucked them up.
Even more reason for teacher assessment I would say. The Class of 2018 and 202 had a raw deal!
On the topic of the exam fiasco: having skim-read OFQAL's methodological paper, it seems to me that the major unforced error in their work is their ad hoc small subject entry adjustment and how that interacts with the state/private school mix. Their own analysis shows that their algo boosts the number of A grades awarded to private school pupils by 4.7 % points compared to last year, far outstripping the improvement for any other type of school, and yet at no point do they consider this to be problematic or even worthy of note. Only in a country where it had become an ingrained norm for the wealthy to game the system would this be possible. Other than that, my sense is they adopted a reasonable methodology but that whatever they chose the result was likely to be politically untenable.
Yes- with a few tweaks, the Ofqual model was non-ideal but probably tolerable in a crisis.
But the failure to do anything about small cohorts was a massive problem. Remember @TSE's (?) dictum about statistics revealing interesting things and hiding really interesting things? The 4.7% uplift for private schools covers large schools like Eton (whose results were put through the mincer) and tiny schools, whose increases must have been stratospheric.
There's also a weird effect at the bottom of each class; kids who miss out on a C don't get nudged down to a D, but sometimes tumble all the way to a U. Worse than that, the model seems to assume that the ability profile of classes in schools is stable. True at national level, probably true for selective schools, really not true for some schools and colleges.
All of that could have been fixed, given time. Probably using a month or so of the last five for a pre-appeal to bring teacher grades and algorithm grades together would have been enough. But that didn't happen, presumably because of massive hubris about the output of a computer model.
Suggests that the vast majority of students are overpredicted in a normal year -> 75%.
Yes, that sounds about right. But this paper is about UCAS predicted grades. 75% overpredictions partly because teachers want to give their students the best chance of getting in to the university of their choice, even though they know the predictions are ambitious. The question is whether the predictions given to Ofqual were the same as the UCAS predictions. They should have been lower. UCAS predictions are what you hope they will get; Ofqual predictions should be what you think they will get.
But I suspect that in many institutions this distinction was not made.
I'd be interested in @ydoethur 's comments on this, as from memory, the exercise was conducted on the basis of realism rather than optimism by his school ?
In any event, outsize aviations between schools in the optimism of their predictions ought to have been something comparatively simple to control for, statistically.
The exercise which Ofsted undertook was a bizarre method of applying an overall population judgment to individuals, with the only individual input the class ranking orders (which were themselves artificially forced by banning equal rankings).
One of the legal actions against Ofqal argues that they exceed their statutory remit. I'm not clear on the details, as I just heard a snippet on the radio this morning, but will be very interesting to follow.
To be honest, I don’t know enough to comment. Certainly there has always been an issue with UCAS predicted grades not matching final grades, but as @Fysics_Teacher notes, these are often marginal calls. My own A-levels were ABC, but I was two marks in two subjects off AAB.
At A-level, the commonest predictions I made this year were jointly A and C, but as I noted I had a very able cohort. I predicted Ds as well, and got them. Nobody got less than a D, but then in the six years I have been teaching at that school only two students ever got Es, so that seems to me to be fair enough. I can’t remember that any were shockingly out of line with the UCAS predictions, and indeed some of them may have been an improvement on them. But as noted above, it is a very inexact science.
For GCSE I would rather not comment yet.
Incidentally, you mean OFQUAL not OFSTED.
There is a difference in predicted grdes between normal years and this year though. Normally a teacher deliberately bumping up the predicted grade two levels (per subject) will be counterproductive for the student, when they underperform in the actual A-Levels, do not make their offer and are left scrambling through clearing to find a last minute place.
This year the teachers had to submit predicted grades knowing that their predictions would not be compared against actual exam results. As teachers in England are trying to get the highest grades for their students there is a tension between "doing the right thing" and being very generous to their students.
Presumably in normal year the predicted A-Level grades are only assigned in the Autumn term for students who apply to uni, whereas this year the teachers have had to predict for all students in the Summer term.
Sorry, eristdoof but that is the wrong way round. For UCAS teachers tend to predict the highest grade they think their students can get. For this exercise we were predicting what they were realistically capable of getting. That said, there was no external quality control so I can’t answer for every school.
Surely the first test of algorithm would have been to predict accurately last years exam results. I wonder if they did that.
Williamson must go. The Tories have fucked up royally.
There is no appropriate dataset from the last year to work from (exam boards didn't previously ask for predicted grades).
What they needed to do would be to identify statistical anomalies (the AAB that becomes AAU) and then either override the system to fix those or rewrite the algorithm until those issues disappeared.
There have been lots of reports on this UCAS constantly tack this. Grades predicted versus those achieved have been running at 40% overestimation for ages.
What people are getting excited about isnt HMG marking, but teachers overforecasting and setting pupil expectations at the wrong level.
Really this years pupils have nothing to moan about, every year behind them was the same pattern.
Really being awarded a U because the computer decided you were the student who walked out 10 minutes into the exam is not a reason to moan?
There have been daft anomalies for ever in exam marking. But the big pattern has always been predicted grades are higher than those achieved. My own kids never hit all the grades they were forecast to hit. The root cause of this isnt the algorithm but the base data is duff in the first place.
No doubt there are lots of reasons why, but until that is corrected the over prediction will still stay and lots of kids will be disappointed on the day.
It only matters because we do UCAS applications before students get their results not after.
In reality the only people who should know where they are going before A level result days are those people heading off to do vocational subjects where interviews, tests or auditions are required.
and around 40% of those applications grades will never be attained.
Thinking about the 'need' to avoid exceptionally good results in any one year, how did this cohort of students do at GCSE level. Was 2018 good, bad or around the mid-point f the expected range? Particularly for those pupils who were in the higher result brings, and, presumably, went on to A levels.
As @ydoethur says, 2018 was the first year of a new version of the GCSE using numbered grades rather than letters. As the grade bands don’t line up comparisons with previous years are problematic.
Suggests that the vast majority of students are overpredicted in a normal year -> 75%.
Yes, that sounds about right. But this paper is about UCAS predicted grades. 75% overpredictions partly because teachers want to give their students the best chance of getting in to the university of their choice, even though they know the predictions are ambitious. The question is whether the predictions given to Ofqual were the same as the UCAS predictions. They should have been lower. UCAS predictions are what you hope they will get; Ofqual predictions should be what you think they will get.
But I suspect that in many institutions this distinction was not made.
I'd be interested in @ydoethur 's comments on this, as from memory, the exercise was conducted on the basis of realism rather than optimism by his school ?
In any event, outsize aviations between schools in the optimism of their predictions ought to have been something comparatively simple to control for, statistically.
The exercise which Ofsted undertook was a bizarre method of applying an overall population judgment to individuals, with the only individual input the class ranking orders (which were themselves artificially forced by banning equal rankings).
One of the legal actions against Ofqal argues that they exceed their statutory remit. I'm not clear on the details, as I just heard a snippet on the radio this morning, but will be very interesting to follow.
To be honest, I don’t know enough to comment. Certainly there has always been an issue with UCAS predicted grades not matching final grades, but as @Fysics_Teacher notes, these are often marginal calls. My own A-levels were ABC, but I was two marks in two subjects off AAB.
At A-level, the commonest predictions I made this year were jointly A and C, but as I noted I had a very able cohort. I predicted Ds as well, and got them. Nobody got less than a D, but then in the six years I have been teaching at that school only two students ever got Es, so that seems to me to be fair enough. I can’t remember that any were shockingly out of line with the UCAS predictions, and indeed some of them may have been an improvement on them. But as noted above, it is a very inexact science.
For GCSE I would rather not comment yet.
Incidentally, you mean OFQUAL not OFSTED.
There is a difference in predicted grdes between normal years and this year though. Normally a teacher deliberately bumping up the predicted grade two levels (per subject) will be counterproductive for the student, when they underperform in the actual A-Levels, do not make their offer and are left scrambling through clearing to find a last minute place.
This year the teachers had to submit predicted grades knowing that their predictions would not be compared against actual exam results. As teachers in England are trying to get the highest grades for their students there is a tension between "doing the right thing" and being very generous to their students.
Presumably in normal year the predicted A-Level grades are only assigned in the Autumn term for students who apply to uni, whereas this year the teachers have had to predict for all students in the Summer term.
It's a brave and fool(hardy) person to talk about what teachers would do when a teacher has answered the question.
I take it you follow Johnson, Cumming and Gove's viewpoint of experts and expertise.
Perhaps the government should award the teachers' grades for GCSEs but stick with the algorithm grades for A-levels. Appeals against GCSE grades could overload the system.
However a big increase in A-level grades would disrupt the universities as some would have too many students meeting their offers, resulting in others having too few.
Does anyone know how many extra places would be needed and where if all students with offers were given places? Might be the only way out at this point
Would that many extra places be needed? Or would it just be fewer for clearing?
It's too late now. I would have thought all the spare places that were created in medicine/vet/ etc will have been filled by now.
What evidence do you have for that claim?
Especially if the government offers money to expand places while also considering the Scots fixed it after the fact without that being an issue and the absence of foreign students?
It's not a claim, it's a talking point, perhaps I should have added a Question mark as well?
Most universities (mine included) would happily bump up intake of UK students to help with the expected overseas student shortfall (albeit there's not much profit in home students compared to overseas). My uni's mid-point estimate is that overseas students are will be down 70%, so there will be spare capacity. However, as I understand it (can't find source now, may be something I was told at work) there's an agreement between universities not to increase home student intake over 5% above normal levels, to avoid completely shafting the lower ranked universities by taking all their students (universities further down could still be in a lot of trouble if the top 20 all take 5% extra, obviously, as that's one average sized university worth of students gone).
Thought that too. But the cap wasn't an entirely bad idea. Virtually everyone who wants to go to University can go now, so expansion of more prestigious courses and Unis will leave less prestigious ones in deep trouble.
Trouble is that the admissions process depends on the grades being reasonably predictable. The combination of ups and downs in the Ofqual algorithm was bad enough, the subsequent squirming is making things far worse.
Maybe I'm being a bit stupid here, but why is the expansion of more prestigious courses for this year's intake a problem? Just give the less prestigious institutions the money for the students they have lost, and everyone's happy, or am I missing something? Sure, it will cost extra, but surely not much in the grand scheme of coronavirus expenditure.
There are several reasons, but the main one is that the vast majority of courses at good unis already are against physical limits such as size of lecture room and teaching rooms, first year student accommodation.
Also in Oxbridge the tutorials will have to be for four students not three
Perhaps the government should award the teachers' grades for GCSEs but stick with the algorithm grades for A-levels. Appeals against GCSE grades could overload the system.
However a big increase in A-level grades would disrupt the universities as some would have too many students meeting their offers, resulting in others having too few.
Does anyone know how many extra places would be needed and where if all students with offers were given places? Might be the only way out at this point
Would that many extra places be needed? Or would it just be fewer for clearing?
It's too late now. I would have thought all the spare places that were created in medicine/vet/ etc will have been filled by now.
What evidence do you have for that claim?
Especially if the government offers money to expand places while also considering the Scots fixed it after the fact without that being an issue and the absence of foreign students?
It's not a claim, it's a talking point, perhaps I should have added a Question mark as well?
Most universities (mine included) would happily bump up intake of UK students to help with the expected overseas student shortfall (albeit there's not much profit in home students compared to overseas). My uni's mid-point estimate is that overseas students are will be down 70%, so there will be spare capacity. However, as I understand it (can't find source now, may be something I was told at work) there's an agreement between universities not to increase home student intake over 5% above normal levels, to avoid completely shafting the lower ranked universities by taking all their students (universities further down could still be in a lot of trouble if the top 20 all take 5% extra, obviously, as that's one average sized university worth of students gone).
Ah, there it is then - and not altruistic from the top universities afterall, as it was sold to me. Still, if those levels were exceeded then the lower ranked universities (which generally have lower research income and therefore more dependent on student income) will really be in trouble.
Here’s something else to throw into the mix about Teacher predictions. Isn’t it the case at A-Level now that the A* grade was introduced that was specifically determined not by performance in the exam, but (top) performance relative to the entire year’s exam cohort? How were teachers supposed to predict something which by definition is beyond their capability to predict?
Someone I know received 4 A*s. He got 9 nines for his GCSEs so it wasn't really a stretch. I expect most A* grades awarded are to similiarly obviously bright pupils.
On the topic of the exam fiasco: having skim-read OFQAL's methodological paper, it seems to me that the major unforced error in their work is their ad hoc small subject entry adjustment and how that interacts with the state/private school mix. Their own analysis shows that their algo boosts the number of A grades awarded to private school pupils by 4.7 % points compared to last year, far outstripping the improvement for any other type of school, and yet at no point do they consider this to be problematic or even worthy of note. Only in a country where it had become an ingrained norm for the wealthy to game the system would this be possible. Other than that, my sense is they adopted a reasonable methodology but that whatever they chose the result was likely to be politically untenable.
Using the percentage point increase makes it look bad. Using the percentage increase and it was about 10% in both private schools and in comprehensives (in other words for every 10 As that each type of school got last year they got 11 this year).
Independent schools are often academically selective, particularly at A-level and spend more per pupil. It is unsurprising that their results are better.
Give me the same time per pupil as my independent school colleagues get and I would expect to do just as well.
As I explained to isam last night, 10% may seem consistent when you look at those who have received good grades but it is very inconsistent when you look at those who have not received the grades which is where the problem lies. Comprehensive pupils were far more likely to miss out than independent ones, the percentage difference there is massive.
That's true but the issue isn't the percentages its the individual cases.
The thing is, every year some people mess up their exams and don't get what they thought they would. Surely when predicting the grades whoever is doing so has to take that into account?
Only if they can accurately predict who would mess up their exams without making any errors.
If you can't then you're simply punishing people who would have done well in place of those who would have messed up their exams.
Its not complex: Two wrongs don't make a right.
OK. But why is it perceived to be the well off that have had it off due to the way that it has been worked out, at the expense of the poorest kids, when Comps have gone up as much as Private Schools, and Academies up quite a bit too?
Because you're looking at the wrong percentage. What matters is not those that have done well, but those who haven't done well but should have. So you need to look at not the percentage of passes, but the percentage of failures.
Independent students NOT to get an A have gone down by 8.4% Comprehensive students NOT to get an A have gone down by 2.5%
Independent students NOT to get a C or above have gone down by 18.5% Comprehensive students NOT to get a C or above have gone down by 8.9%
That's worse for the rich isn't it?
No. You want the proportion NOT to get a good grade to go down.
The proportion NOT to get an A going down by 8.4% is miles better than going down by 2.5% The proportion NOT to get a C or above going down by 18.5% is miles better than going down by 8.9%
You've lost me.
I have had an extremely long and stressful weekend, and don't really care about this subject much anyway. I don't feel the way to unwind is to get involved in a row about it with you, so farewell
Lets say that you were worried about being in a car crash - would you rather have your risk of a car crash go down by 8.4% or go down by 2.5% ?
People getting a better grade than they should have isn't the problem. People getting a worse grade than they should have is the problem. That's much more likely in the Comprehensive than Independent sector.
Righto. I only saw the figures for the grade inflation I think
The inflation isn't the problem though, the problem is people missing out. For that you need to look at the inverse which makes the percentages much, much worse.
For not getting As: Independent 51.4 didn't get an A in 2020, 56.1 in 2019. -4.7 or 8.4% Comprehensive 78.2 didn't get an A in 2020, 80.2 in 2019. -2.0 or 2.5%
So yes the Independent sectors haven't had much inflation relative to the proportion who were already doing well . . . but they have had a much greater reduction in the number of people who weren't doing well.
My son is back to school today. One of his pieces of "homework" over the summer was an entry into an economics essay competition comparing the effects of the Black Death and Covid.
Although there are some surprising similarities the major difference is the scale. When I was a lad the general presumption was that 1/3 of the world (ie Europe) died as a result of the Black Death. The view from historians now seems to be that this was based on serious under estimates of where the population stood pre-plague and it was in fact more like 50-60% of the population who died in the various waves.
Which does rather put the 1-2% of Covid into perspective, doesn't it?
Especially when you consider the vast majority of those killed by covid would never had made it in life long enough to be killed by covid.
They would have died due to low life expectancy rates or from lack of treatment for the c0-morbidities they have.
Even by your quite low standards, that doesn’t seem to make sense.
Incidentally, did you know that statistically the most dangerous human activity is breathing? Everyone who breathes, dies.
It really is quite amazing that after such a long time so many people are so ignorant of what COVID is and who it affects.
You said the majority of those killed by Covid would not have lived long enough to die of it.
Which is an effect of this virus I will admit I was unaware of.
Its absolutely true. The numbers say yu have got to be pretty ill and old to die from COVID essentially. Over 80 with at least one co-morbodity.
Surely the first test of algorithm would have been to predict accurately last years exam results. I wonder if they did that.
Williamson must go. The Tories have fucked up royally.
There is no appropriate dataset from the last year to work from (exam boards didn't previously ask for predicted grades).
What they needed to do would be to identify statistical anomalies (the AAB that becomes AAU) and then either override the system to fix those or rewrite the algorithm until those issues disappeared.
There have been lots of reports on this UCAS constantly tack this. Grades predicted versus those achieved have been running at 40% overestimation for ages.
What people are getting excited about isnt HMG marking, but teachers overforecasting and setting pupil expectations at the wrong level.
Really this years pupils have nothing to moan about, every year behind them was the same pattern.
Really being awarded a U because the computer decided you were the student who walked out 10 minutes into the exam is not a reason to moan?
Don’t think it’s only this year it happened: getting a U because someone forgot to put the marks for one of your papers in to their system is equally galling, particularly if by the time it is sorted your university place has gone to someone else.
On the topic of the exam fiasco: having skim-read OFQAL's methodological paper, it seems to me that the major unforced error in their work is their ad hoc small subject entry adjustment and how that interacts with the state/private school mix. Their own analysis shows that their algo boosts the number of A grades awarded to private school pupils by 4.7 % points compared to last year, far outstripping the improvement for any other type of school, and yet at no point do they consider this to be problematic or even worthy of note. Only in a country where it had become an ingrained norm for the wealthy to game the system would this be possible. Other than that, my sense is they adopted a reasonable methodology but that whatever they chose the result was likely to be politically untenable.
Yes- with a few tweaks, the Ofqual model was non-ideal but probably tolerable in a crisis.
But the failure to do anything about small cohorts was a massive problem. Remember @TSE's (?) dictum about statistics revealing interesting things and hiding really interesting things? The 4.7% uplift for private schools covers large schools like Eton (whose results were put through the mincer) and tiny schools, whose increases must have been stratospheric.
There's also a weird effect at the bottom of each class; kids who miss out on a C don't get nudged down to a D, but sometimes tumble all the way to a U. Worse than that, the model seems to assume that the ability profile of classes in schools is stable. True at national level, probably true for selective schools, really not true for some schools and colleges.
All of that could have been fixed, given time. Probably using a month or so of the last five for a pre-appeal to bring teacher grades and algorithm grades together would have been enough. But that didn't happen, presumably because of massive hubris about the output of a computer model.
Yes, I imagine there is far more variation year on year (both due to random strength of year groups and in school improvement/decline) in state schools (especially non selective ones) than private schools that specifically control their intakes to ensure high performance as a default. And where parents pay for this. Generally private schools are probably pretty consistent in their placing in league tables, state institutions will move up and down like yoyos
Surely the first test of algorithm would have been to predict accurately last years exam results. I wonder if they did that.
Williamson must go. The Tories have fucked up royally.
There is no appropriate dataset from the last year to work from (exam boards didn't previously ask for predicted grades).
What they needed to do would be to identify statistical anomalies (the AAB that becomes AAU) and then either override the system to fix those or rewrite the algorithm until those issues disappeared.
There have been lots of reports on this UCAS constantly tack this. Grades predicted versus those achieved have been running at 40% overestimation for ages.
What people are getting excited about isnt HMG marking, but teachers overforecasting and setting pupil expectations at the wrong level.
Really this years pupils have nothing to moan about, every year behind them was the same pattern.
Really being awarded a U because the computer decided you were the student who walked out 10 minutes into the exam is not a reason to moan?
Don’t think it’s only this year it happened: getting a U because someone forgot to put the marks for one of your papers in to their system is equally galling, particularly if by the time it is sorted your university place has gone to someone else.
yup, happened one of my kids their marks were added up wrong, fortunately at GCSE so no uni implications.
What I have found strange is that all the exam result opening videos I have watched everyone seems really happy with their results.
Because obviously, somebody who burst into tears, tore up the paper or shouted 'I'm going to kill you motherf***ers' is going to give permission for the TV to use that footage.
Suggests that the vast majority of students are overpredicted in a normal year -> 75%.
Yes, that sounds about right. But this paper is about UCAS predicted grades. 75% overpredictions partly because teachers want to give their students the best chance of getting in to the university of their choice, even though they know the predictions are ambitious. The question is whether the predictions given to Ofqual were the same as the UCAS predictions. They should have been lower. UCAS predictions are what you hope they will get; Ofqual predictions should be what you think they will get.
But I suspect that in many institutions this distinction was not made.
I'd be interested in @ydoethur 's comments on this, as from memory, the exercise was conducted on the basis of realism rather than optimism by his school ?
In any event, outsize aviations between schools in the optimism of their predictions ought to have been something comparatively simple to control for, statistically.
The exercise which Ofsted undertook was a bizarre method of applying an overall population judgment to individuals, with the only individual input the class ranking orders (which were themselves artificially forced by banning equal rankings).
One of the legal actions against Ofqal argues that they exceed their statutory remit. I'm not clear on the details, as I just heard a snippet on the radio this morning, but will be very interesting to follow.
To be honest, I don’t know enough to comment. Certainly there has always been an issue with UCAS predicted grades not matching final grades, but as @Fysics_Teacher notes, these are often marginal calls. My own A-levels were ABC, but I was two marks in two subjects off AAB.
At A-level, the commonest predictions I made this year were jointly A and C, but as I noted I had a very able cohort. I predicted Ds as well, and got them. Nobody got less than a D, but then in the six years I have been teaching at that school only two students ever got Es, so that seems to me to be fair enough. I can’t remember that any were shockingly out of line with the UCAS predictions, and indeed some of them may have been an improvement on them. But as noted above, it is a very inexact science.
For GCSE I would rather not comment yet.
Incidentally, you mean OFQUAL not OFSTED.
There is a difference in predicted grdes between normal years and this year though. Normally a teacher deliberately bumping up the predicted grade two levels (per subject) will be counterproductive for the student, when they underperform in the actual A-Levels, do not make their offer and are left scrambling through clearing to find a last minute place.
This year the teachers had to submit predicted grades knowing that their predictions would not be compared against actual exam results. As teachers in England are trying to get the highest grades for their students there is a tension between "doing the right thing" and being very generous to their students.
Presumably in normal year the predicted A-Level grades are only assigned in the Autumn term for students who apply to uni, whereas this year the teachers have had to predict for all students in the Summer term.
Sorry, eristdoof but that is the wrong way round. For UCAS teachers tend to predict the highest grade they think their students can get. For this exercise we were predicting what they were realistically capable of getting. That said, there was no external quality control so I can’t answer for every school.
One thing I was wondering that you may have a (partial) answer to. UCAS predictions are routinely shared with students/parents. But were the Ofqual predictions similarly shared? If so, this would put pressure on teachers/schools/colleges. I suspect it wouldn't go down well with student/parents if their Ofqual predicted grades were lower than their UCAS predicted grades.
Everyone seems desperate to throw OFQUAL under the bus, but I think the Gov't should stick to it's guns. If pupils want to get a better grade they can sit an autumn exam.
My son is back to school today. One of his pieces of "homework" over the summer was an entry into an economics essay competition comparing the effects of the Black Death and Covid.
Although there are some surprising similarities the major difference is the scale. When I was a lad the general presumption was that 1/3 of the world (ie Europe) died as a result of the Black Death. The view from historians now seems to be that this was based on serious under estimates of where the population stood pre-plague and it was in fact more like 50-60% of the population who died in the various waves.
Which does rather put the 1-2% of Covid into perspective, doesn't it?
Yes, but that was in a time without medicine when the science of germs and bacteria was at best imperfectly understood.
If Covid had rampaged through the world instead of Spanish flu, what would the consequences have been?
The Black Death is an interesting one economically though because it led to a short lived economic boom. This was followed, over the next eighty years, by a decreasing trade balance and a shortage of silver, culminating in the Great Slump of the mid fifteenth century (which was, among other things, a cause of the Wars of the Roses).
What some of this more recent research has shown is that although there was an increase in nominal wages there was also a sharp burst of inflation and the apparent increase in wages believed to have occurred as a result of the shortage of labour was mainly illusory. In many ways it is remarkable that society survived at all given the level of carnage.
Curiously, the responses included face masks, social distancing and quarantine all of which seem somewhat familiar. The pocket full of posies from the line was an attempt to remove the infection perceived to be in the air by perfume. Given the lack of knowledge of viruses, bacteria or modes of transmission it is interesting that they ended up in similar places.
Here’s something else to throw into the mix about Teacher predictions. Isn’t it the case at A-Level now that the A* grade was introduced that was specifically determined not by performance in the exam, but (top) performance relative to the entire year’s exam cohort? How were teachers supposed to predict something which by definition is beyond their capability to predict?
Exam boards publish exact grade boundaries for all grades including A*, so I don’t think that’s right.
Suggests that the vast majority of students are overpredicted in a normal year -> 75%.
Yes, that sounds about right. But this paper is about UCAS predicted grades. 75% overpredictions partly because teachers want to give their students the best chance of getting in to the university of their choice, even though they know the predictions are ambitious. The question is whether the predictions given to Ofqual were the same as the UCAS predictions. They should have been lower. UCAS predictions are what you hope they will get; Ofqual predictions should be what you think they will get.
But I suspect that in many institutions this distinction was not made.
I'd be interested in @ydoethur 's comments on this, as from memory, the exercise was conducted on the basis of realism rather than optimism by his school ?
In any event, outsize aviations between schools in the optimism of their predictions ought to have been something comparatively simple to control for, statistically.
The exercise which Ofsted undertook was a bizarre method of applying an overall population judgment to individuals, with the only individual input the class ranking orders (which were themselves artificially forced by banning equal rankings).
One of the legal actions against Ofqal argues that they exceed their statutory remit. I'm not clear on the details, as I just heard a snippet on the radio this morning, but will be very interesting to follow.
To be honest, I don’t know enough to comment. Certainly there has always been an issue with UCAS predicted grades not matching final grades, but as @Fysics_Teacher notes, these are often marginal calls. My own A-levels were ABC, but I was two marks in two subjects off AAB.
At A-level, the commonest predictions I made this year were jointly A and C, but as I noted I had a very able cohort. I predicted Ds as well, and got them. Nobody got less than a D, but then in the six years I have been teaching at that school only two students ever got Es, so that seems to me to be fair enough. I can’t remember that any were shockingly out of line with the UCAS predictions, and indeed some of them may have been an improvement on them. But as noted above, it is a very inexact science.
For GCSE I would rather not comment yet.
Incidentally, you mean OFQUAL not OFSTED.
There is a difference in predicted grdes between normal years and this year though. Normally a teacher deliberately bumping up the predicted grade two levels (per subject) will be counterproductive for the student, when they underperform in the actual A-Levels, do not make their offer and are left scrambling through clearing to find a last minute place.
This year the teachers had to submit predicted grades knowing that their predictions would not be compared against actual exam results. As teachers in England are trying to get the highest grades for their students there is a tension between "doing the right thing" and being very generous to their students.
Presumably in normal year the predicted A-Level grades are only assigned in the Autumn term for students who apply to uni, whereas this year the teachers have had to predict for all students in the Summer term.
Sorry, eristdoof but that is the wrong way round. For UCAS teachers tend to predict the highest grade they think their students can get. For this exercise we were predicting what they were realistically capable of getting. That said, there was no external quality control so I can’t answer for every school.
One thing I was wondering that you may have a (partial) answer to. UCAS predictions are routinely shared with students/parents. But were the Ofqual predictions similarly shared? If so, this would put pressure on teachers/schools/colleges. I suspect it wouldn't go down well with student/parents if their Ofqual predicted grades were lower than their UCAS predicted grades.
It was a disciplinary offence, punishable by being struck off from teaching, to share these grades in advance. OFQUAL seemed obsessed about it well past the point of reason.
Now we know why, of course. Although how they think it will have helped I have no idea.
You will notice I am still being very careful what I say about GCSE grades.
Everyone seems desperate to throw OFQUAL under the bus, but I think the Gov't should stick to it's guns. If pupils want to get a better grade they can sit an autumn exam.
Are you offering to pay their salaries for the lost working year?
What is odd is that no ministers spotted the huge political elephant trap -- another sign of Boris's inexperienced Cabinet?
My son is back to school today. One of his pieces of "homework" over the summer was an entry into an economics essay competition comparing the effects of the Black Death and Covid.
Although there are some surprising similarities the major difference is the scale. When I was a lad the general presumption was that 1/3 of the world (ie Europe) died as a result of the Black Death. The view from historians now seems to be that this was based on serious under estimates of where the population stood pre-plague and it was in fact more like 50-60% of the population who died in the various waves.
Which does rather put the 1-2% of Covid into perspective, doesn't it?
Especially when you consider the vast majority of those killed by covid would never had made it in life long enough to be killed by covid.
They would have died due to low life expectancy rates or from lack of treatment for the c0-morbidities they have.
Even by your quite low standards, that doesn’t seem to make sense.
Incidentally, did you know that statistically the most dangerous human activity is breathing? Everyone who breathes, dies.
It really is quite amazing that after such a long time so many people are so ignorant of what COVID is and who it affects.
You said the majority of those killed by Covid would not have lived long enough to die of it.
Which is an effect of this virus I will admit I was unaware of.
Its absolutely true. The numbers say yu have got to be pretty ill and old to die from COVID essentially. Over 80 with at least one co-morbodity.
In the middle ages, in case you were wondering, the was no such thing as managing illnesses like hypertension, heart disease and diabetes. Chaucer strangely doesn't refer to transplant surgery in the Canterbury tales.
No the numbers don't say that.
The numbers say that with our healthcare, and with our treatments you are more likely to be pretty old and ill to die. But younger people especially those with co-morbidities are possible to die too even with our healthcare looking after them - and @ydoethur is right comorbidities and ill health were rife then.
With the absence of any antibiotics or medicine then young people with TB (a major issue then) or some other comorbidities could have been slaughtered in vast numbers then.
There’s plenty of places in the world with only a pretty Middle Ages standard of healthcare available to most people and with young populations. We don’t have to guess. We are simply not seeing mass casualties of the young in these places. This is a nasty pandemic but that’s all.
It is currently quite a fringe view to question the wisdom of lockdowns, partly because governments are great at communicating fear, partly because people have until now been shielded by the economic effects. It won’t stay this way. If Keir was smart, he’d be pitching his flag firmly in this territory, skating to where the puck is going, not where it is. That he has not makes me doubt his wisdom as a politician, even if he’d know doubt prove a more competent administrator than we have right now.
Surely the first test of algorithm would have been to predict accurately last years exam results. I wonder if they did that.
Williamson must go. The Tories have fucked up royally.
It is ofqual who have caused this crisis and must provide an appeal process today
Williamson is hopeless and must go but ultimately any overruling of ofqual may well require a change in the law as seen in Scotland and Stormant being recalled in NI
The buck stops for this utter debacle with the elected ministers. No ifs, no buts. No fig leaves to hide behind. Take responsibility and go.
Take responsibility and fix it. If they do that there's no need to go. Ofqual dropped the ball but they're fixing it.
If they don't fix it urgently though, then yes they should go. U-turning should not lead to sackings if its the right thing to do.
Only the diehards have any confidence in Williamson to put things right. He needs to go. Should have already gone. Utterly useless.
He is but right now he has to put it right
He can’t fix it because he has lost confidence. The sooner he goes the sooner thing can improve.
He can fix it by following the path the Scots laid out. It would have been better if he'd done it on Wednesday in hindsight but better late than never.
Did his Scottish equivalent resign? They did the same thing.
The Scottish Government is not high on my list of favourite organisations, but they are not short on basic political cunning. They will generally let the UK Government take the tough decisions, and either stand back and avoid, or implement something similar but with little twiddly differences that they can claim are improvements.
This was an opportunity for the UK Government, conversely, to see the Scottish Government fall flat on its face, and neatly avoid doing so itself. I really don't get why this hasn't happened. All I can come up with is that educational rigour is so close to Cumming's (and perhaps Gove's) heart that he's forced Williamson to adopt this approach. It isn't Boris - he's not a 'path of most resistance' kind of guy.
It's going to teach the Government a lot of great lessons, and that has to be a good thing.
Step 1) Mandate that all conditional offers from universities and technical colleges for British citizens and residents will be made unconditional. Solves the issue for anyone staying in (British) education. Any duds that slip through can come out in the wash of First Year exams.
Step 2) If necessary bung some extra cash the universities way. They are going to need a bail out without the foreign (Chinese) students anyway.
Step 3) Declare 2020 null and void for A Level unless you have taken the exam. Offer an exam in both autumn and spring, with a generous one off educational grant to anyone who defers going on the dole to instead complete their studies. Special tuition for those who want it.
Step 4) Said exams if necessary to be held in requisitioned conference facilities with the million person volunteer army to help oversea on the day.
Step 5) GCSEs cancelled this year, just like when Voldemort killed that kid in Harry Potter. In the grand scheme, on well.
Step 6) Hysteria to be dialled down a notch. Cv-19 is thankfully not the plague. School to be mandatory rather than optional from here on in.
Personally I see this as one of the more easily solvable problems caused by the lockdown. Money has been spaffed about to help all of pensioners (qe to support stock market bubble), property owners (stamp holidays), small business owners (no questions asked grants and govt backed loans) and workers (furlough). What about some love for the poor sods who aren’t at all risk from this disease and have had their life chances stunted in a final spiteful act by the over 60s?
A shame we don’t have another few Rishi’s knocking about the make up the numbers in the rest of Cabinet.
Suggests that the vast majority of students are overpredicted in a normal year -> 75%.
Yes, that sounds about right. But this paper is about UCAS predicted grades. 75% overpredictions partly because teachers want to give their students the best chance of getting in to the university of their choice, even though they know the predictions are ambitious. The question is whether the predictions given to Ofqual were the same as the UCAS predictions. They should have been lower. UCAS predictions are what you hope they will get; Ofqual predictions should be what you think they will get.
But I suspect that in many institutions this distinction was not made.
I'd be interested in @ydoethur 's comments on this, as from memory, the exercise was conducted on the basis of realism rather than optimism by his school ?
In any event, outsize aviations between schools in the optimism of their predictions ought to have been something comparatively simple to control for, statistically.
The exercise which Ofsted undertook was a bizarre method of applying an overall population judgment to individuals, with the only individual input the class ranking orders (which were themselves artificially forced by banning equal rankings).
One of the legal actions against Ofqal argues that they exceed their statutory remit. I'm not clear on the details, as I just heard a snippet on the radio this morning, but will be very interesting to follow.
To be honest, I don’t know enough to comment. Certainly there has always been an issue with UCAS predicted grades not matching final grades, but as @Fysics_Teacher notes, these are often marginal calls. My own A-levels were ABC, but I was two marks in two subjects off AAB.
At A-level, the commonest predictions I made this year were jointly A and C, but as I noted I had a very able cohort. I predicted Ds as well, and got them. Nobody got less than a D, but then in the six years I have been teaching at that school only two students ever got Es, so that seems to me to be fair enough. I can’t remember that any were shockingly out of line with the UCAS predictions, and indeed some of them may have been an improvement on them. But as noted above, it is a very inexact science.
For GCSE I would rather not comment yet.
Incidentally, you mean OFQUAL not OFSTED.
There is a difference in predicted grdes between normal years and this year though. Normally a teacher deliberately bumping up the predicted grade two levels (per subject) will be counterproductive for the student, when they underperform in the actual A-Levels, do not make their offer and are left scrambling through clearing to find a last minute place.
This year the teachers had to submit predicted grades knowing that their predictions would not be compared against actual exam results. As teachers in England are trying to get the highest grades for their students there is a tension between "doing the right thing" and being very generous to their students.
Presumably in normal year the predicted A-Level grades are only assigned in the Autumn term for students who apply to uni, whereas this year the teachers have had to predict for all students in the Summer term.
Sorry, eristdoof but that is the wrong way round. For UCAS teachers tend to predict the highest grade they think their students can get. For this exercise we were predicting what they were realistically capable of getting. That said, there was no external quality control so I can’t answer for every school.
OK, I stand corredted, but I am surprised. I guess then that the benefit of predicting up UCAS candidates, is that the students get considered for uni at all.
On the other hand you have made it clear that you here strong on the side of -what I call- "doing the right thing" by grade predictions and implied that this was not the case in all schools.
Obviously not a great poll for Biden, but as ever with HYUFD, check the other polls. Yesterday 2 other national polls were released: Yougov Biden 10 ahead NBC Biden 9 ahead
The last CNN national poll seems to have been back in the beginning of June and showed Biden 14 ahead.
Yougov are releasing a lot more regular polls (about 2 a week nationally), recent polls since the beginning of July have Biden ahead between 6 and 10 points. So a pretty good poll for Biden yesterday. It was also a "likely voter" poll and showed no change in Biden's lead compared to the "registered voter" poll released by Yougov last week. This is also good news for Biden, as often the switch to "likely voter" polling is worth a couple of points to the Republican candidate.
Biden's lead in the NBC poll (which seem to be 1 a month) is 2 down on a month ago, same as 2 months ago, and 2 up on 3 months ago.
Overall Biden's national lead seems to be pretty stable, currently 8.0% down from the peak of 9.6% in the 538 average on June 24th, but up a bit from a low of 7.6% on August 6th. State polling is very similar, with the exception of Minnesota where Biden's lead is lower than it was. I can't see any polling reasons for the recent improvement in Trump's betting odds, except Minnesota.
Suggests that the vast majority of students are overpredicted in a normal year -> 75%.
Yes, that sounds about right. But this paper is about UCAS predicted grades. 75% overpredictions partly because teachers want to give their students the best chance of getting in to the university of their choice, even though they know the predictions are ambitious. The question is whether the predictions given to Ofqual were the same as the UCAS predictions. They should have been lower. UCAS predictions are what you hope they will get; Ofqual predictions should be what you think they will get.
But I suspect that in many institutions this distinction was not made.
I'd be interested in @ydoethur 's comments on this, as from memory, the exercise was conducted on the basis of realism rather than optimism by his school ?
In any event, outsize aviations between schools in the optimism of their predictions ought to have been something comparatively simple to control for, statistically.
The exercise which Ofsted undertook was a bizarre method of applying an overall population judgment to individuals, with the only individual input the class ranking orders (which were themselves artificially forced by banning equal rankings).
One of the legal actions against Ofqal argues that they exceed their statutory remit. I'm not clear on the details, as I just heard a snippet on the radio this morning, but will be very interesting to follow.
To be honest, I don’t know enough to comment. Certainly there has always been an issue with UCAS predicted grades not matching final grades, but as @Fysics_Teacher notes, these are often marginal calls. My own A-levels were ABC, but I was two marks in two subjects off AAB.
At A-level, the commonest predictions I made this year were jointly A and C, but as I noted I had a very able cohort. I predicted Ds as well, and got them. Nobody got less than a D, but then in the six years I have been teaching at that school only two students ever got Es, so that seems to me to be fair enough. I can’t remember that any were shockingly out of line with the UCAS predictions, and indeed some of them may have been an improvement on them. But as noted above, it is a very inexact science.
For GCSE I would rather not comment yet.
Incidentally, you mean OFQUAL not OFSTED.
There is a difference in predicted grdes between normal years and this year though. Normally a teacher deliberately bumping up the predicted grade two levels (per subject) will be counterproductive for the student, when they underperform in the actual A-Levels, do not make their offer and are left scrambling through clearing to find a last minute place.
This year the teachers had to submit predicted grades knowing that their predictions would not be compared against actual exam results. As teachers in England are trying to get the highest grades for their students there is a tension between "doing the right thing" and being very generous to their students.
Presumably in normal year the predicted A-Level grades are only assigned in the Autumn term for students who apply to uni, whereas this year the teachers have had to predict for all students in the Summer term.
Sorry, eristdoof but that is the wrong way round. For UCAS teachers tend to predict the highest grade they think their students can get. For this exercise we were predicting what they were realistically capable of getting. That said, there was no external quality control so I can’t answer for every school.
One thing I was wondering that you may have a (partial) answer to. UCAS predictions are routinely shared with students/parents. But were the Ofqual predictions similarly shared? If so, this would put pressure on teachers/schools/colleges. I suspect it wouldn't go down well with student/parents if their Ofqual predicted grades were lower than their UCAS predicted grades.
It was a disciplinary offence, punishable by being struck off from teaching, to share these grades in advance. OFQUAL seemed obsessed about it well past the point of reason.
Now we know why, of course. Although how they think it will have helped I have no idea.
You will notice I am still being very careful what I say about GCSE grades.
Wow! So it really wasn't transparent at all, when it should have been. Students had no idea what their schools/colleges submitted to Ofqual. This means presumably that some students who are fuming actually did receive their centre assessed grade, but possibly not their UCAS grade.
This is a complete shambles; heads really should roll.
Suggests that the vast majority of students are overpredicted in a normal year -> 75%.
Yes, that sounds about right. But this paper is about UCAS predicted grades. 75% overpredictions partly because teachers want to give their students the best chance of getting in to the university of their choice, even though they know the predictions are ambitious. The question is whether the predictions given to Ofqual were the same as the UCAS predictions. They should have been lower. UCAS predictions are what you hope they will get; Ofqual predictions should be what you think they will get.
But I suspect that in many institutions this distinction was not made.
I'd be interested in @ydoethur 's comments on this, as from memory, the exercise was conducted on the basis of realism rather than optimism by his school ?
In any event, outsize aviations between schools in the optimism of their predictions ought to have been something comparatively simple to control for, statistically.
The exercise which Ofsted undertook was a bizarre method of applying an overall population judgment to individuals, with the only individual input the class ranking orders (which were themselves artificially forced by banning equal rankings).
One of the legal actions against Ofqal argues that they exceed their statutory remit. I'm not clear on the details, as I just heard a snippet on the radio this morning, but will be very interesting to follow.
To be honest, I don’t know enough to comment. Certainly there has always been an issue with UCAS predicted grades not matching final grades, but as @Fysics_Teacher notes, these are often marginal calls. My own A-levels were ABC, but I was two marks in two subjects off AAB.
At A-level, the commonest predictions I made this year were jointly A and C, but as I noted I had a very able cohort. I predicted Ds as well, and got them. Nobody got less than a D, but then in the six years I have been teaching at that school only two students ever got Es, so that seems to me to be fair enough. I can’t remember that any were shockingly out of line with the UCAS predictions, and indeed some of them may have been an improvement on them. But as noted above, it is a very inexact science.
For GCSE I would rather not comment yet.
Incidentally, you mean OFQUAL not OFSTED.
There is a difference in predicted grdes between normal years and this year though. Normally a teacher deliberately bumping up the predicted grade two levels (per subject) will be counterproductive for the student, when they underperform in the actual A-Levels, do not make their offer and are left scrambling through clearing to find a last minute place.
This year the teachers had to submit predicted grades knowing that their predictions would not be compared against actual exam results. As teachers in England are trying to get the highest grades for their students there is a tension between "doing the right thing" and being very generous to their students.
Presumably in normal year the predicted A-Level grades are only assigned in the Autumn term for students who apply to uni, whereas this year the teachers have had to predict for all students in the Summer term.
Sorry, eristdoof but that is the wrong way round. For UCAS teachers tend to predict the highest grade they think their students can get. For this exercise we were predicting what they were realistically capable of getting. That said, there was no external quality control so I can’t answer for every school.
OK, I stand corredted, but I am surprised. I guess then that the benefit of predicting up UCAS candidates, is that the students get considered for uni at all.
On the other hand you have made it clear that you here strong on the side of -what I call- "doing the right thing" by grade predictions and implied that this was not the case in all schools.
1) Yes. And it has got worse with the scrapping of AS levels.
2) How can I put this? I jumped through many hoops to get my A-level grades right. And it seems the exam boards are happy that I did. I cannot say that was the same in all schools from my personal knowledge. What I can say is that where I do have knowledge, the process seems to have been fairly rigorous.
Suggests that the vast majority of students are overpredicted in a normal year -> 75%.
Yes, that sounds about right. But this paper is about UCAS predicted grades. 75% overpredictions partly because teachers want to give their students the best chance of getting in to the university of their choice, even though they know the predictions are ambitious. The question is whether the predictions given to Ofqual were the same as the UCAS predictions. They should have been lower. UCAS predictions are what you hope they will get; Ofqual predictions should be what you think they will get.
But I suspect that in many institutions this distinction was not made.
I'd be interested in @ydoethur 's comments on this, as from memory, the exercise was conducted on the basis of realism rather than optimism by his school ?
In any event, outsize aviations between schools in the optimism of their predictions ought to have been something comparatively simple to control for, statistically.
The exercise which Ofsted undertook was a bizarre method of applying an overall population judgment to individuals, with the only individual input the class ranking orders (which were themselves artificially forced by banning equal rankings).
One of the legal actions against Ofqal argues that they exceed their statutory remit. I'm not clear on the details, as I just heard a snippet on the radio this morning, but will be very interesting to follow.
To be honest, I don’t know enough to comment. Certainly there has always been an issue with UCAS predicted grades not matching final grades, but as @Fysics_Teacher notes, these are often marginal calls. My own A-levels were ABC, but I was two marks in two subjects off AAB.
At A-level, the commonest predictions I made this year were jointly A and C, but as I noted I had a very able cohort. I predicted Ds as well, and got them. Nobody got less than a D, but then in the six years I have been teaching at that school only two students ever got Es, so that seems to me to be fair enough. I can’t remember that any were shockingly out of line with the UCAS predictions, and indeed some of them may have been an improvement on them. But as noted above, it is a very inexact science.
For GCSE I would rather not comment yet.
Incidentally, you mean OFQUAL not OFSTED.
There is a difference in predicted grdes between normal years and this year though. Normally a teacher deliberately bumping up the predicted grade two levels (per subject) will be counterproductive for the student, when they underperform in the actual A-Levels, do not make their offer and are left scrambling through clearing to find a last minute place.
This year the teachers had to submit predicted grades knowing that their predictions would not be compared against actual exam results. As teachers in England are trying to get the highest grades for their students there is a tension between "doing the right thing" and being very generous to their students.
Presumably in normal year the predicted A-Level grades are only assigned in the Autumn term for students who apply to uni, whereas this year the teachers have had to predict for all students in the Summer term.
Sorry, eristdoof but that is the wrong way round. For UCAS teachers tend to predict the highest grade they think their students can get. For this exercise we were predicting what they were realistically capable of getting. That said, there was no external quality control so I can’t answer for every school.
One thing I was wondering that you may have a (partial) answer to. UCAS predictions are routinely shared with students/parents. But were the Ofqual predictions similarly shared? If so, this would put pressure on teachers/schools/colleges. I suspect it wouldn't go down well with student/parents if their Ofqual predicted grades were lower than their UCAS predicted grades.
It was a disciplinary offence, punishable by being struck off from teaching, to share these grades in advance. OFQUAL seemed obsessed about it well past the point of reason.
Now we know why, of course. Although how they think it will have helped I have no idea.
You will notice I am still being very careful what I say about GCSE grades.
Wow! So it really wasn't transparent at all, when it should have been. Students had no idea what their schools/colleges submitted to Ofqual. This means presumably that some students who are fuming actually did receive their centre assessed grade, but possibly not their UCAS grade.
This is a complete shambles; heads really should roll.
Surely the first test of algorithm would have been to predict accurately last years exam results. I wonder if they did that.
Williamson must go. The Tories have fucked up royally.
There is no appropriate dataset from the last year to work from (exam boards didn't previously ask for predicted grades).
What they needed to do would be to identify statistical anomalies (the AAB that becomes AAU) and then either override the system to fix those or rewrite the algorithm until those issues disappeared.
There have been lots of reports on this UCAS constantly tack this. Grades predicted versus those achieved have been running at 40% overestimation for ages.
What people are getting excited about isnt HMG marking, but teachers overforecasting and setting pupil expectations at the wrong level.
Really this years pupils have nothing to moan about, every year behind them was the same pattern.
Really being awarded a U because the computer decided you were the student who walked out 10 minutes into the exam is not a reason to moan?
Don’t think it’s only this year it happened: getting a U because someone forgot to put the marks for one of your papers in to their system is equally galling, particularly if by the time it is sorted your university place has gone to someone else.
yup, happened one of my kids their marks were added up wrong, fortunately at GCSE so no uni implications.
My first ever summer job as a student was at the JMB in Manchester. Adding up the marks for each page on O Level papers and totalling them to check their accuracy. A shocking number were wrong. Sometimes ludicrously so.
Everyone seems desperate to throw OFQUAL under the bus, but I think the Gov't should stick to it's guns. If pupils want to get a better grade they can sit an autumn exam.
Are you offering to pay their salaries for the lost working year?
What is odd is that no ministers spotted the huge political elephant trap -- another sign of Boris's inexperienced Cabinet?
Everyone seems to want airy fairy artificially inflated grades. The young man I refferred to in my previous post went to a standard comp not some tiny cohort private school - my guess is actually the algorithm is rather better than most people like to think it is.
Perhaps the government should award the teachers' grades for GCSEs but stick with the algorithm grades for A-levels. Appeals against GCSE grades could overload the system.
However a big increase in A-level grades would disrupt the universities as some would have too many students meeting their offers, resulting in others having too few.
Does anyone know how many extra places would be needed and where if all students with offers were given places? Might be the only way out at this point
Would that many extra places be needed? Or would it just be fewer for clearing?
It's too late now. I would have thought all the spare places that were created in medicine/vet/ etc will have been filled by now.
What evidence do you have for that claim?
Especially if the government offers money to expand places while also considering the Scots fixed it after the fact without that being an issue and the absence of foreign students?
It's not a claim, it's a talking point, perhaps I should have added a Question mark as well?
Most universities (mine included) would happily bump up intake of UK students to help with the expected overseas student shortfall (albeit there's not much profit in home students compared to overseas). My uni's mid-point estimate is that overseas students are will be down 70%, so there will be spare capacity. However, as I understand it (can't find source now, may be something I was told at work) there's an agreement between universities not to increase home student intake over 5% above normal levels, to avoid completely shafting the lower ranked universities by taking all their students (universities further down could still be in a lot of trouble if the top 20 all take 5% extra, obviously, as that's one average sized university worth of students gone).
Thought that too. But the cap wasn't an entirely bad idea. Virtually everyone who wants to go to University can go now, so expansion of more prestigious courses and Unis will leave less prestigious ones in deep trouble.
Trouble is that the admissions process depends on the grades being reasonably predictable. The combination of ups and downs in the Ofqual algorithm was bad enough, the subsequent squirming is making things far worse.
Maybe I'm being a bit stupid here, but why is the expansion of more prestigious courses for this year's intake a problem? Just give the less prestigious institutions the money for the students they have lost, and everyone's happy, or am I missing something? Sure, it will cost extra, but surely not much in the grand scheme of coronavirus expenditure.
There are several reasons, but the main one is that the vast majority of courses at good unis already are against physical limits such as size of lecture room and teaching rooms, first year student accommodation.
Also in Oxbridge the tutorials will have to be for four students not three
That may be true, but I intended to ask why the problem for lesser unis losing students can't be simply fixed with giving them the money anyway (rather than problems caused for the more popular unis having too many students, which posters above seem to think wouldn't be a problem - indeed those unis would apparently welcome the extra students).
Here’s something else to throw into the mix about Teacher predictions. Isn’t it the case at A-Level now that the A* grade was introduced that was specifically determined not by performance in the exam, but (top) performance relative to the entire year’s exam cohort? How were teachers supposed to predict something which by definition is beyond their capability to predict?
Exam boards publish exact grade boundaries for all grades including A*, so I don’t think that’s right.
Agreed, but on the older modular schemes A* was awarded on the basis of performance in A2 units only at grade A standard, so it had no boundary score.
On the topic of the exam fiasco: having skim-read OFQAL's methodological paper, it seems to me that the major unforced error in their work is their ad hoc small subject entry adjustment and how that interacts with the state/private school mix. Their own analysis shows that their algo boosts the number of A grades awarded to private school pupils by 4.7 % points compared to last year, far outstripping the improvement for any other type of school, and yet at no point do they consider this to be problematic or even worthy of note. Only in a country where it had become an ingrained norm for the wealthy to game the system would this be possible. Other than that, my sense is they adopted a reasonable methodology but that whatever they chose the result was likely to be politically untenable.
Using the percentage point increase makes it look bad. Using the percentage increase and it was about 10% in both private schools and in comprehensives (in other words for every 10 As that each type of school got last year they got 11 this year).
Independent schools are often academically selective, particularly at A-level and spend more per pupil. It is unsurprising that their results are better.
Give me the same time per pupil as my independent school colleagues get and I would expect to do just as well.
As I explained to isam last night, 10% may seem consistent when you look at those who have received good grades but it is very inconsistent when you look at those who have not received the grades which is where the problem lies. Comprehensive pupils were far more likely to miss out than independent ones, the percentage difference there is massive.
That's true but the issue isn't the percentages its the individual cases.
The thing is, every year some people mess up their exams and don't get what they thought they would. Surely when predicting the grades whoever is doing so has to take that into account?
Only if they can accurately predict who would mess up their exams without making any errors.
If you can't then you're simply punishing people who would have done well in place of those who would have messed up their exams.
Its not complex: Two wrongs don't make a right.
OK. But why is it perceived to be the well off that have had it off due to the way that it has been worked out, at the expense of the poorest kids, when Comps have gone up as much as Private Schools, and Academies up quite a bit too?
Because you're looking at the wrong percentage. What matters is not those that have done well, but those who haven't done well but should have. So you need to look at not the percentage of passes, but the percentage of failures.
Independent students NOT to get an A have gone down by 8.4% Comprehensive students NOT to get an A have gone down by 2.5%
Independent students NOT to get a C or above have gone down by 18.5% Comprehensive students NOT to get a C or above have gone down by 8.9%
That's worse for the rich isn't it?
No. You want the proportion NOT to get a good grade to go down.
The proportion NOT to get an A going down by 8.4% is miles better than going down by 2.5% The proportion NOT to get a C or above going down by 18.5% is miles better than going down by 8.9%
You've lost me.
I have had an extremely long and stressful weekend, and don't really care about this subject much anyway. I don't feel the way to unwind is to get involved in a row about it with you, so farewell
Lets say that you were worried about being in a car crash - would you rather have your risk of a car crash go down by 8.4% or go down by 2.5% ?
People getting a better grade than they should have isn't the problem. People getting a worse grade than they should have is the problem. That's much more likely in the Comprehensive than Independent sector.
Righto. I only saw the figures for the grade inflation I think
The inflation isn't the problem though, the problem is people missing out. For that you need to look at the inverse which makes the percentages much, much worse.
For not getting As: Independent 51.4 didn't get an A in 2020, 56.1 in 2019. -4.7 or 8.4% Comprehensive 78.2 didn't get an A in 2020, 80.2 in 2019. -2.0 or 2.5%
So yes the Independent sectors haven't had much inflation relative to the proportion who were already doing well . . . but they have had a much greater reduction in the number of people who weren't doing well.
Make sense?
Yup, good point. Also if you look at state selective schools, whose 2019 outcomes were closest to private schools (36% at A vs 44%), the increase was 1.2%-pts vs 4.7%-pts. and sixth form colleges (21% at A in 2019) saw a 0.7%-pt increase. Let's think of it the other way round, if private schools had been disadvantaged relative to state schools, does anyone think it would have been allowed to stand? A few calls from some top Tory donors whose children or grandchildren had been affected and you would have seen an immediate U-turn and the OFQUAL leadership out on their ear.
Surely the first test of algorithm would have been to predict accurately last years exam results. I wonder if they did that.
Williamson must go. The Tories have fucked up royally.
There is no appropriate dataset from the last year to work from (exam boards didn't previously ask for predicted grades).
What they needed to do would be to identify statistical anomalies (the AAB that becomes AAU) and then either override the system to fix those or rewrite the algorithm until those issues disappeared.
There have been lots of reports on this UCAS constantly tack this. Grades predicted versus those achieved have been running at 40% overestimation for ages.
What people are getting excited about isnt HMG marking, but teachers overforecasting and setting pupil expectations at the wrong level.
Really this years pupils have nothing to moan about, every year behind them was the same pattern.
Really being awarded a U because the computer decided you were the student who walked out 10 minutes into the exam is not a reason to moan?
Don’t think it’s only this year it happened: getting a U because someone forgot to put the marks for one of your papers in to their system is equally galling, particularly if by the time it is sorted your university place has gone to someone else.
yup, happened one of my kids their marks were added up wrong, fortunately at GCSE so no uni implications.
My first ever summer job as a student was at the JMB in Manchester. Adding up the marks for each page on O Level papers and totalling them to check their accuracy. A shocking number were wrong. Sometimes ludicrously so.
They are now of course done by computer.
This can in itself lead to embarrassment of course. A computer flagged up a question I had marked as out by 80%. I'd actually given it the same mark as the examiner (this being before AQA abandoned standardisation) but like a twat I had put the mark in the wrong box, so I had given them 20/25 for a question they hadn't attempted.
My son is back to school today. One of his pieces of "homework" over the summer was an entry into an economics essay competition comparing the effects of the Black Death and Covid.
Although there are some surprising similarities the major difference is the scale. When I was a lad the general presumption was that 1/3 of the world (ie Europe) died as a result of the Black Death. The view from historians now seems to be that this was based on serious under estimates of where the population stood pre-plague and it was in fact more like 50-60% of the population who died in the various waves.
Which does rather put the 1-2% of Covid into perspective, doesn't it?
Especially when you consider the vast majority of those killed by covid would never had made it in life long enough to be killed by covid.
They would have died due to low life expectancy rates or from lack of treatment for the c0-morbidities they have.
Even by your quite low standards, that doesn’t seem to make sense.
Incidentally, did you know that statistically the most dangerous human activity is breathing? Everyone who breathes, dies.
It really is quite amazing that after such a long time so many people are so ignorant of what COVID is and who it affects.
You said the majority of those killed by Covid would not have lived long enough to die of it.
Which is an effect of this virus I will admit I was unaware of.
Its absolutely true. The numbers say yu have got to be pretty ill and old to die from COVID essentially. Over 80 with at least one co-morbodity.
In the middle ages, in case you were wondering, the was no such thing as managing illnesses like hypertension, heart disease and diabetes. Chaucer strangely doesn't refer to transplant surgery in the Canterbury tales.
No the numbers don't say that.
The numbers say that with our healthcare, and with our treatments you are more likely to be pretty old and ill to die. But younger people especially those with co-morbidities are possible to die too even with our healthcare looking after them - and @ydoethur is right comorbidities and ill health were rife then.
With the absence of any antibiotics or medicine then young people with TB (a major issue then) or some other comorbidities could have been slaughtered in vast numbers then.
There’s plenty of places in the world with only a pretty Middle Ages standard of healthcare available to most people and with young populations. .
I'm sorry this is just rubbish. You obviously have no idea of medicine in the middle ages.
There are certainy many places in the world which have no money for good medicines and equipment, but decent knowledge of medicine is almost everywhere. Even if rural developing areas are only using early 20th century medicine practices (which I doubt) they are still centuries ahead of middle ages medicine.
Anyway, this is the government that will be in charge of a No Deal Brexit post 1/1/2021.
The algorithms determining which firms are allowed to import and export, who gets medicines and how food will be distributed should be an absolute hoot. I can hardly wait.
A couple of observations on the A levels grading (insert suitable term here).
1. In a normal year thousands get lower results than expected and some appeal. It is natural that this year anybody in that situation and their parents will blame the authorities.
2. Some reported cases seem to contradict the algorithm as summarised in the press (I haven't time to delve). What is the betting on coding errors compounding the situation?
I think the suggestion of modifying and re-running would be best approach: " Bill Watkin, of the Sixth Form Colleges Association, said that correcting and re-running the algorithm would be a 'national, institution level, automated single appeal'".
No more than one grade drop, eliminate U grades and no grades lower on second run than first would rescue situation along with appeals and sitting in October.
Government response very poor in crisis management terms!
Everyone seems desperate to throw OFQUAL under the bus, but I think the Gov't should stick to it's guns. If pupils want to get a better grade they can sit an autumn exam.
Are you offering to pay their salaries for the lost working year?
What is odd is that no ministers spotted the huge political elephant trap -- another sign of Boris's inexperienced Cabinet?
Everyone seems to want airy fairy artificially inflated grades. The young man I refferred to in my previous post went to a standard comp not some tiny cohort private school - my guess is actually the algorithm is rather better than most people like to think it is.
I don't want artificially inflated grades. I want every student to at least get what they should have got.
The algorithm is good as an average, but average isn't good enough. If we know that one person in a class of 20 is a murderer and then we without a trial semi at random pick one of the 20 and imprison them for murder without a trial or evidence of their actual guilt then would that be OK?
Every single person should get what they should have got. The only way to achieve that is to give them the benefit of the doubt, which means having some inflation. Some inflation is better than major deflation for those who miss out on what they should have got.
Perhaps the government should award the teachers' grades for GCSEs but stick with the algorithm grades for A-levels. Appeals against GCSE grades could overload the system.
However a big increase in A-level grades would disrupt the universities as some would have too many students meeting their offers, resulting in others having too few.
Does anyone know how many extra places would be needed and where if all students with offers were given places? Might be the only way out at this point
Would that many extra places be needed? Or would it just be fewer for clearing?
It's too late now. I would have thought all the spare places that were created in medicine/vet/ etc will have been filled by now.
What evidence do you have for that claim?
Especially if the government offers money to expand places while also considering the Scots fixed it after the fact without that being an issue and the absence of foreign students?
It's not a claim, it's a talking point, perhaps I should have added a Question mark as well?
Most universities (mine included) would happily bump up intake of UK students to help with the expected overseas student shortfall (albeit there's not much profit in home students compared to overseas). My uni's mid-point estimate is that overseas students are will be down 70%, so there will be spare capacity. However, as I understand it (can't find source now, may be something I was told at work) there's an agreement between universities not to increase home student intake over 5% above normal levels, to avoid completely shafting the lower ranked universities by taking all their students (universities further down could still be in a lot of trouble if the top 20 all take 5% extra, obviously, as that's one average sized university worth of students gone).
Thought that too. But the cap wasn't an entirely bad idea. Virtually everyone who wants to go to University can go now, so expansion of more prestigious courses and Unis will leave less prestigious ones in deep trouble.
Trouble is that the admissions process depends on the grades being reasonably predictable. The combination of ups and downs in the Ofqual algorithm was bad enough, the subsequent squirming is making things far worse.
Maybe I'm being a bit stupid here, but why is the expansion of more prestigious courses for this year's intake a problem? Just give the less prestigious institutions the money for the students they have lost, and everyone's happy, or am I missing something? Sure, it will cost extra, but surely not much in the grand scheme of coronavirus expenditure.
There are several reasons, but the main one is that the vast majority of courses at good unis already are against physical limits such as size of lecture room and teaching rooms, first year student accommodation.
Also in Oxbridge the tutorials will have to be for four students not three
That may be true, but I intended to ask why the problem for lesser unis losing students can't be simply fixed with giving them the money anyway (rather than problems caused for the more popular unis having too many students, which posters above seem to think wouldn't be a problem - indeed those unis would apparently welcome the extra students).
Partly because they've been banned from taking extra students, with financial penalties attached.
But also because they've sacked all their teaching staff, so the infrastructure isn't there.
Suggests that the vast majority of students are overpredicted in a normal year -> 75%.
Yes, that sounds about right. But this paper is about UCAS predicted grades. 75% overpredictions partly because teachers want to give their students the best chance of getting in to the university of their choice, even though they know the predictions are ambitious. The question is whether the predictions given to Ofqual were the same as the UCAS predictions. They should have been lower. UCAS predictions are what you hope they will get; Ofqual predictions should be what you think they will get.
But I suspect that in many institutions this distinction was not made.
I'd be interested in @ydoethur 's comments on this, as from memory, the exercise was conducted on the basis of realism rather than optimism by his school ?
In any event, outsize aviations between schools in the optimism of their predictions ought to have been something comparatively simple to control for, statistically.
The exercise which Ofsted undertook was a bizarre method of applying an overall population judgment to individuals, with the only individual input the class ranking orders (which were themselves artificially forced by banning equal rankings).
One of the legal actions against Ofqal argues that they exceed their statutory remit. I'm not clear on the details, as I just heard a snippet on the radio this morning, but will be very interesting to follow.
To be honest, I don’t know enough to comment. Certainly there has always been an issue with UCAS predicted grades not matching final grades, but as @Fysics_Teacher notes, these are often marginal calls. My own A-levels were ABC, but I was two marks in two subjects off AAB.
At A-level, the commonest predictions I made this year were jointly A and C, but as I noted I had a very able cohort. I predicted Ds as well, and got them. Nobody got less than a D, but then in the six years I have been teaching at that school only two students ever got Es, so that seems to me to be fair enough. I can’t remember that any were shockingly out of line with the UCAS predictions, and indeed some of them may have been an improvement on them. But as noted above, it is a very inexact science.
For GCSE I would rather not comment yet.
Incidentally, you mean OFQUAL not OFSTED.
There is a difference in predicted grdes between normal years and this year though. Normally a teacher deliberately bumping up the predicted grade two levels (per subject) will be counterproductive for the student, when they underperform in the actual A-Levels, do not make their offer and are left scrambling through clearing to find a last minute place.
This year the teachers had to submit predicted grades knowing that their predictions would not be compared against actual exam results. As teachers in England are trying to get the highest grades for their students there is a tension between "doing the right thing" and being very generous to their students.
Presumably in normal year the predicted A-Level grades are only assigned in the Autumn term for students who apply to uni, whereas this year the teachers have had to predict for all students in the Summer term.
Sorry, eristdoof but that is the wrong way round. For UCAS teachers tend to predict the highest grade they think their students can get. For this exercise we were predicting what they were realistically capable of getting. That said, there was no external quality control so I can’t answer for every school.
OK, I stand corredted, but I am surprised. I guess then that the benefit of predicting up UCAS candidates, is that the students get considered for uni at all.
On the other hand you have made it clear that you here strong on the side of -what I call- "doing the right thing" by grade predictions and implied that this was not the case in all schools.
Processes evolve over time. The University application approach seems to be- you have 30 places to fill so you select 35-45 candidates you want next year based on target grades. 5-15 of them will fail to get the grades and you get the class of 30..
The only way to fix that is to move to post exam applications but the timescales don't exist and it's impossible to create a suitable timescale to do it without fundamental changes to either school or University timetables.
Step 1) Mandate that all conditional offers from universities and technical colleges for British citizens and residents will be made unconditional. Solves the issue for anyone staying in (British) education. Any duds that slip through can come out in the wash of First Year exams.
Step 2) If necessary bung some extra cash the universities way. They are going to need a bail out without the foreign (Chinese) students anyway.
Step 3) Declare 2020 null and void for A Level unless you have taken the exam. Offer an exam in both autumn and spring, with a generous one off educational grant to anyone who defers going on the dole to instead complete their studies. Special tuition for those who want it.
Step 4) Said exams if necessary to be held in requisitioned conference facilities with the million person volunteer army to help oversea on the day.
Step 5) GCSEs cancelled this year, just like when Voldemort killed that kid in Harry Potter. In the grand scheme, on well.
Step 6) Hysteria to be dialled down a notch. Cv-19 is thankfully not the plague. School to be mandatory rather than optional from here on in.
Personally I see this as one of the more easily solvable problems caused by the lockdown. Money has been spaffed about to help all of pensioners (qe to support stock market bubble), property owners (stamp holidays), small business owners (no questions asked grants and govt backed loans) and workers (furlough). What about some love for the poor sods who aren’t at all risk from this disease and have had their life chances stunted in a final spiteful act by the over 60s?
A shame we don’t have another few Rishi’s knocking about the make up the numbers in the rest of Cabinet.
What is surreal is the apparent determination of both the SQA and now OFQAL to award those marks with as little contact with actual output of the students as possible. I am obviously overly simplistic but is the logical response to the exams not being sat not to look to find some other assessment based upon the student's work such as mocks and class work?
Good news in the cricket. As they're not going to finish the Test, Ofqual are going to give us a result using their algorithm.
Joe Root got a duck, Ben Stokes surprisingly rattled off a half-century despite not playing, but it was all in vain as both Pakistan and England got downgraded and the Test was awarded to India because they're currently top of the rankings.
Everyone seems desperate to throw OFQUAL under the bus, but I think the Gov't should stick to it's guns. If pupils want to get a better grade they can sit an autumn exam.
Are you offering to pay their salaries for the lost working year?
What is odd is that no ministers spotted the huge political elephant trap -- another sign of Boris's inexperienced Cabinet?
Everyone seems to want airy fairy artificially inflated grades. The young man I refferred to in my previous post went to a standard comp not some tiny cohort private school - my guess is actually the algorithm is rather better than most people like to think it is.
I don't want artificially inflated grades. I want every student to at least get what they should have got.
The algorithm is good as an average, but average isn't good enough. If we know that one person in a class of 20 is a murderer and then we without a trial semi at random pick one of the 20 and imprison them for murder without a trial or evidence of their actual guilt then would that be OK?
Every single person should get what they should have got. The only way to achieve that is to give them the benefit of the doubt, which means having some inflation. Some inflation is better than major deflation for those who miss out on what they should have got.
I think everyone should get what they deserve.
I'm struggling to think of exactly what OFQUAL deserve for this mess, though most of them seem to end with various forms of slow death.
Anyway, this is the government that will be in charge of a No Deal Brexit post 1/1/2021.
The algorithms determining which firms are allowed to import and export, who gets medicines and how food will be distributed should be an absolute hoot. I can hardly wait.
It's going to be fun. I'm waiting for more obvious things like what export forms look like so we can being computerising them..
Everyone seems desperate to throw OFQUAL under the bus, but I think the Gov't should stick to it's guns. If pupils want to get a better grade they can sit an autumn exam.
Are you offering to pay their salaries for the lost working year?
What is odd is that no ministers spotted the huge political elephant trap -- another sign of Boris's inexperienced Cabinet?
Everyone seems to want airy fairy artificially inflated grades. The young man I refferred to in my previous post went to a standard comp not some tiny cohort private school - my guess is actually the algorithm is rather better than most people like to think it is.
These days A-levels exist mainly as a gateway to university -- the number of university places is the important thing, and if I were Boris I'd pressure the universities to speed up the acceptance process. Almost no-one cares about A-levels beyond that. And what on earth is the point of GCSEs now (almost) no-one can leave school at 16 any more?
And that means grade inflation in A-levels for one year really does not matter.
The big scandal is algorithmic U grades -- they should be revoked and replaced immediately. Again, I am surprised ministers did not spot this beforehand.
It was striking that Trump's price shortened abruptly on news of his brother dying. A humanizing event and one that he (unusually) reacted to in a way that was not especially crass.
He will have clocked the impact, I'm sure. If he doesn't follow the betting somebody in his camp will be and will have told him. "Mr President, it looks like Robert passing away has given you a boost for November."
I wonder what this could mean for other family members now. One doesn't like to contemplate such things but it's clear that there is very little he won't do to avoid defeat on 3/11.
Surely the first test of algorithm would have been to predict accurately last years exam results. I wonder if they did that.
Williamson must go. The Tories have fucked up royally.
It is ofqual who have caused this crisis and must provide an appeal process today
Williamson is hopeless and must go but ultimately any overruling of ofqual may well require a change in the law as seen in Scotland and Stormant being recalled in NI
The buck stops for this utter debacle with the elected ministers. No ifs, no buts. No fig leaves to hide behind. Take responsibility and go.
Take responsibility and fix it. If they do that there's no need to go. Ofqual dropped the ball but they're fixing it.
If they don't fix it urgently though, then yes they should go. U-turning should not lead to sackings if its the right thing to do.
Only the diehards have any confidence in Williamson to put things right. He needs to go. Should have already gone. Utterly useless.
He is but right now he has to put it right
He can’t fix it because he has lost confidence. The sooner he goes the sooner thing can improve.
He can fix it by following the path the Scots laid out. It would have been better if he'd done it on Wednesday in hindsight but better late than never.
Did his Scottish equivalent resign? They did the same thing.
The Scottish Government is not high on my list of favourite organisations, but they are not short on basic political cunning. They will generally let the UK Government take the tough decisions, and either stand back and avoid, or implement something similar but with little twiddly differences that they can claim are improvements.
This was an opportunity for the UK Government, conversely, to see the Scottish Government fall flat on its face, and neatly avoid doing so itself. I really don't get why this hasn't happened. All I can come up with is that educational rigour is so close to Cumming's (and perhaps Gove's) heart that he's forced Williamson to adopt this approach. It isn't Boris - he's not a 'path of most resistance' kind of guy.
It's going to teach the Government a lot of great lessons, and that has to be a good thing.
It's generals fighting the last war. Grade inflation was the last war, and may indeed be a future war, but this year that was not the issue as any fule kno. Being a creature of statute Ofqual does not think it needs political antennae, so its toppling will a piquant nemesis.
Comments
‘The war situation has developed, not altogether to Japan’s advantage.’
No doubt there are lots of reasons why, but until that is corrected the over prediction will still stay and lots of kids will be disappointed on the day.
The point I was making is that quangos are a good way to avoid the need to give a running commentary on relatively small but somewhat complex day to day stuff about spectrum licensing, electricity and water markets, exam boards and so on.
But I agree that this isn't small stuff - it affects large numbers and is absolutely an area where the call for Government to step in isn't some kind of faddish or fringe thing.
The daft thing is this should have been obvious as soon as it became clear the exams couldn't go ahead. The Education Secretary should've been all over it - good politicians see the political weather coming.
If that was happening then it wouldn't take very long for something like that to go viral on Twitter and then be shared here by someone like CarlottaVance as evidence the SNP have screwed up.
In the middle ages, in case you were wondering, the was no such thing as managing illnesses like hypertension, heart disease and diabetes. Chaucer strangely doesn't refer to transplant surgery in the Canterbury tales.
There should probably be a reshuffle next year once COVID is fading into history and we look to move on and rebuild. I would be quite OK with seeing Williamson go then, but for now he needs to do the right thing and deal with this.
Independent schools are often academically selective, particularly at A-level and spend more per pupil. It is unsurprising that their results are better.
Give me the same time per pupil as my independent school colleagues get and I would expect to do just as well.
In reality the only people who should know where they are going before A level result days are those people heading off to do vocational subjects where interviews, tests or auditions are required.
If ABBA had taken their A levels this year, they would be known as CDDC
Normally a teacher deliberately bumping up the predicted grade two levels (per subject) will be counterproductive for the student, when they underperform in the actual A-Levels, do not make their offer and are left scrambling through clearing to find a last minute place.
This year the teachers had to submit predicted grades knowing that their predictions would not be compared against actual exam results. As teachers in England are trying to get the highest grades for their students there is a tension between "doing the right thing" and being very generous to their students.
Presumably in normal year the predicted A-Level grades are only assigned in the Autumn term for students who apply to uni, whereas this year the teachers have had to predict for all students in the Summer term.
So the mortality rate would have been higher, because many of them would have underlying untreated conditions.
The fatality rate of bubonic plague today is around 10-14%. We don’t know the exact figure for Covid-19 which makes comparison difficult.
However as bubonic plague is not infectious unless it gets into the lungs, Covid-19 would have spread much more rapidly.
The numbers say that with our healthcare, and with our treatments you are more likely to be pretty old and ill to die. But younger people especially those with co-morbidities are possible to die too even with our healthcare looking after them - and @ydoethur is right comorbidities and ill health were rife then.
With the absence of any antibiotics or medicine then young people with TB (a major issue then) or some other comorbidities could have been slaughtered in vast numbers then.
But the failure to do anything about small cohorts was a massive problem. Remember @TSE's (?) dictum about statistics revealing interesting things and hiding really interesting things? The 4.7% uplift for private schools covers large schools like Eton (whose results were put through the mincer) and tiny schools, whose increases must have been stratospheric.
There's also a weird effect at the bottom of each class; kids who miss out on a C don't get nudged down to a D, but sometimes tumble all the way to a U. Worse than that, the model seems to assume that the ability profile of classes in schools is stable. True at national level, probably true for selective schools, really not true for some schools and colleges.
All of that could have been fixed, given time. Probably using a month or so of the last five for a pre-appeal to bring teacher grades and algorithm grades together would have been enough. But that didn't happen, presumably because of massive hubris about the output of a computer model.
I take it you follow Johnson, Cumming and Gove's viewpoint of experts and expertise.
Also in Oxbridge the tutorials will have to be for four students not three
I expect most A* grades awarded are to similiarly obviously bright pupils.
Cons 325
Lab 240
LDs 4
SNP 58
Plaid 4
NI 18
https://www.electoralcalculus.co.uk/fcgi-bin/usercode.py?scotcontrol=Y&CON=42&LAB=39&LIB=5&Brexit=2&Green=3&UKIP=&TVCON=&TVLAB=&TVLIB=&TVBrexit=&TVGreen=&TVUKIP=&SCOTCON=21&SCOTLAB=19&SCOTLIB=6&SCOTBrexit=0&SCOTGreen=0&SCOTUKIP=&SCOTNAT=53&display=AllChanged&regorseat=(none)&boundary=2019
That's like saying people only share selfies they're happy with so nobody has issues with how they look otherwise.
Curiously, the responses included face masks, social distancing and quarantine all of which seem somewhat familiar. The pocket full of posies from the line was an attempt to remove the infection perceived to be in the air by perfume. Given the lack of knowledge of viruses, bacteria or modes of transmission it is interesting that they ended up in similar places.
Now we know why, of course. Although how they think it will have helped I have no idea.
You will notice I am still being very careful what I say about GCSE grades.
What is odd is that no ministers spotted the huge political elephant trap -- another sign of Boris's inexperienced Cabinet?
It is currently quite a fringe view to question the wisdom of lockdowns, partly because governments are great at communicating fear, partly because people have until now been shielded by the economic effects. It won’t stay this way. If Keir was smart, he’d be pitching his flag firmly in this territory, skating to where the puck is going, not where it is. That he has not makes me doubt his wisdom as a politician, even if he’d know doubt prove a more competent administrator than we have right now.
When the government was under pressure over taxes on pasties.
This was an opportunity for the UK Government, conversely, to see the Scottish Government fall flat on its face, and neatly avoid doing so itself. I really don't get why this hasn't happened. All I can come up with is that educational rigour is so close to Cumming's (and perhaps Gove's) heart that he's forced Williamson to adopt this approach. It isn't Boris - he's not a 'path of most resistance' kind of guy.
It's going to teach the Government a lot of great lessons, and that has to be a good thing.
On the other hand you have made it clear that you here strong on the side of -what I call- "doing the right thing" by grade predictions and implied that this was not the case in all schools.
Yougov Biden 10 ahead
NBC Biden 9 ahead
The last CNN national poll seems to have been back in the beginning of June and showed Biden 14 ahead.
Yougov are releasing a lot more regular polls (about 2 a week nationally), recent polls since the beginning of July have Biden ahead between 6 and 10 points. So a pretty good poll for Biden yesterday. It was also a "likely voter" poll and showed no change in Biden's lead compared to the "registered voter" poll released by Yougov last week. This is also good news for Biden, as often the switch to "likely voter" polling is worth a couple of points to the Republican candidate.
Biden's lead in the NBC poll (which seem to be 1 a month) is 2 down on a month ago, same as 2 months ago, and 2 up on 3 months ago.
Overall Biden's national lead seems to be pretty stable, currently 8.0% down from the peak of 9.6% in the 538 average on June 24th, but up a bit from a low of 7.6% on August 6th. State polling is very similar, with the exception of Minnesota where Biden's lead is lower than it was. I can't see any polling reasons for the recent improvement in Trump's betting odds, except Minnesota.
This line has obviously been focus grouped to death
This is a complete shambles; heads really should roll.
2) How can I put this? I jumped through many hoops to get my A-level grades right. And it seems the exam boards are happy that I did. I cannot say that was the same in all schools from my personal knowledge. What I can say is that where I do have knowledge, the process seems to have been fairly rigorous.
A shocking number were wrong. Sometimes ludicrously so.
Let's think of it the other way round, if private schools had been disadvantaged relative to state schools, does anyone think it would have been allowed to stand? A few calls from some top Tory donors whose children or grandchildren had been affected and you would have seen an immediate U-turn and the OFQUAL leadership out on their ear.
This can in itself lead to embarrassment of course. A computer flagged up a question I had marked as out by 80%. I'd actually given it the same mark as the examiner (this being before AQA abandoned standardisation) but like a twat I had put the mark in the wrong box, so I had given them 20/25 for a question they hadn't attempted.
In my defence, it was 2am...
There are certainy many places in the world which have no money for good medicines and equipment, but decent knowledge of medicine is almost everywhere. Even if rural developing areas are only using early 20th century medicine practices (which I doubt) they are still centuries ahead of middle ages medicine.
The algorithms determining which firms are allowed to import and export, who gets medicines and how food will be distributed should be an absolute hoot. I can hardly wait.
1. In a normal year thousands get lower results than expected and some appeal. It is natural that this year anybody in that situation and their parents will blame the authorities.
2. Some reported cases seem to contradict the algorithm as summarised in the press (I haven't time to delve). What is the betting on coding errors compounding the situation?
I think the suggestion of modifying and re-running would be best approach: " Bill Watkin, of the Sixth Form Colleges Association, said that correcting and re-running the algorithm would be a 'national, institution level, automated single appeal'".
No more than one grade drop, eliminate U grades and no grades lower on second run than first would rescue situation along with appeals and sitting in October.
Government response very poor in crisis management terms!
The algorithm is good as an average, but average isn't good enough. If we know that one person in a class of 20 is a murderer and then we without a trial semi at random pick one of the 20 and imprison them for murder without a trial or evidence of their actual guilt then would that be OK?
Every single person should get what they should have got. The only way to achieve that is to give them the benefit of the doubt, which means having some inflation. Some inflation is better than major deflation for those who miss out on what they should have got.
But also because they've sacked all their teaching staff, so the infrastructure isn't there.
The only way to fix that is to move to post exam applications but the timescales don't exist and it's impossible to create a suitable timescale to do it without fundamental changes to either school or University timetables.
As they're not going to finish the Test, Ofqual are going to give us a result using their algorithm.
Joe Root got a duck, Ben Stokes surprisingly rattled off a half-century despite not playing, but it was all in vain as both Pakistan and England got downgraded and the Test was awarded to India because they're currently top of the rankings.
I'm struggling to think of exactly what OFQUAL deserve for this mess, though most of them seem to end with various forms of slow death.
And that means grade inflation in A-levels for one year really does not matter.
The big scandal is algorithmic U grades -- they should be revoked and replaced immediately. Again, I am surprised ministers did not spot this beforehand.
There are plenty of parts of the world where it is pretty bad.
It was striking that Trump's price shortened abruptly on news of his brother dying. A humanizing event and one that he (unusually) reacted to in a way that was not especially crass.
He will have clocked the impact, I'm sure. If he doesn't follow the betting somebody in his camp will be and will have told him. "Mr President, it looks like Robert passing away has given you a boost for November."
I wonder what this could mean for other family members now. One doesn't like to contemplate such things but it's clear that there is very little he won't do to avoid defeat on 3/11.
Being a creature of statute Ofqual does not think it needs political antennae, so its toppling will a piquant nemesis.