Problem is they've missed out as someone else now has the space and you can't suddenly add another 20 medical students in to a world of social distancing on small class sizes.
Eek twin b has been warned that classes may occur up to 9pm at night and between 10 to 4 on Saturdays and Sundays..
The Dems will get little boost from a virtual Convention. Indeed even real Conventions seem to produce relatively little in the way of bounce these days, stripped of all meaning and substance they have become rather empty affairs.
The upside is that Trump almost certainly needs these sort of events even more. He feeds off his crowds and hones his messages by how they respond. He looked utterly despondent when his last rally turned into a damp squib and for good reasons. He does not want to campaign by being interviewed by a largely hostile media day after day. He needs that direct link with his base and I am not sure how he is going to replace it.
His press conferences have pretty well tuned into rallies, if you listen to what he spews. But with smaller and tougher audiences.
He gets to jet off steam, and they do get widespread coverage.
Best wishes for the op.
Regarding the emptying hospitals, how much of that is now down to patients being reluctant to go in, and how much to the NHS ? Clearly there is a significant reduction in capacity owing to COVID measures, but it can’t just be that ?
My wife has an on-going small and apparently non-malignant dermatological problem. She had one growth removed early this year, privately. What with one thing and another, upwards of £1k, mainly the dermatologist's time and skill. Now it's recurred and the GP said she needs a further consultant review...... no, needn't pay that sort of money; I'll refer you. So we had a call from an appointment booking service..... you can go to.... naming four reasonably local centres. OK.... naming the most convenient. Not, sorry, nothing in the foreseeable. OK, another. Same. Turns out there's NO possibility of a dermatologist appointment this year in Mid or N Essex.
What your wife is experiencing with the NHS is endemic throughout it. Vast swathes of people are using Covid as a reason to not do their jobs. In this case what are the people who work in these dermatology centres doing all day if they are not seeing any patients. There are huge numbers of people in the NHS going to work each day and are doing nothing. My wife is one of them, she regularly comes home and says "yet another day when I have done nothing". She is a nurse on a ward.
The lunancy of shutting the NHS to protect it from Covid must be killing hundreds of people. Currently less than 1% of hospital beds are taken up with Covid yet we have shut the entire system beacuse of it.
This website provides an interesting snapshot of hospital data.
The most startling one for me was the number of NHS staff who were off sick on the 8th August 2020.
Have a guess.
Its 60,367.
TBH, while I agree about wondering what some of the staff in the dermatology unit are doing, the problem appears to be a shortage of senior medical staff. Nothing, or very little, to do with Covid-19.
The Dems will get little boost from a virtual Convention. Indeed even real Conventions seem to produce relatively little in the way of bounce these days, stripped of all meaning and substance they have become rather empty affairs.
The upside is that Trump almost certainly needs these sort of events even more. He feeds off his crowds and hones his messages by how they respond. He looked utterly despondent when his last rally turned into a damp squib and for good reasons. He does not want to campaign by being interviewed by a largely hostile media day after day. He needs that direct link with his base and I am not sure how he is going to replace it.
His press conferences have pretty well tuned into rallies, if you listen to what he spews. But with smaller and tougher audiences.
He gets to jet off steam, and they do get widespread coverage.
Best wishes for the op.
Regarding the emptying hospitals, how much of that is now down to patients being reluctant to go in, and how much to the NHS ? Clearly there is a significant reduction in capacity owing to COVID measures, but it can’t just be that ?
My wife has an on-going small and apparently non-malignant dermatological problem. She had one growth removed early this year, privately. What with one thing and another, upwards of £1k, mainly the dermatologist's time and skill. Now it's recurred and the GP said she needs a further consultant review...... no, needn't pay that sort of money; I'll refer you. So we had a call from an appointment booking service..... you can go to.... naming four reasonably local centres. OK.... naming the most convenient. Not, sorry, nothing in the foreseeable. OK, another. Same. Turns out there's NO possibility of a dermatologist appointment this year in Mid or N Essex.
What your wife is experiencing with the NHS is endemic throughout it. Vast swathes of people are using Covid as a reason to not do their jobs. In this case what are the people who work in these dermatology centres doing all day if they are not seeing any patients. There are huge numbers of people in the NHS going to work each day and are doing nothing. My wife is one of them, she regularly comes home and says "yet another day when I have done nothing". She is a nurse on a ward.
The lunancy of shutting the NHS to protect it from Covid must be killing hundreds of people. Currently less than 1% of hospital beds are taken up with Covid yet we have shut the entire system beacuse of it.
This website provides an interesting snapshot of hospital data.
The most startling one for me was the number of NHS staff who were off sick on the 8th August 2020.
Have a guess.
Its 60,367.
Can you give us an average figure for being off sick on a Saturday in early August? Although the number sounds high, with no reference value, it means nothing.
This links provides many years statistics of days lost to sickness.
Looking at it the lower your pay banding the more likely you are to go off sick.
In 2019/20 the sickness absence rate for Band 1 was 7.73% whereas for Band 9 it was 1.59%
Anyone with statistical chops got the time and inclination to wade through it ? (I suspect one reason it’s several hundred pages long is to discourage such an effort.)
I haven't read the whole thing myself but have seen the extract where they talk about their back testing and the back testing isn't right, they don't test the same thing as what they claim their model does. And they back test using as their test data a portion of their training data.
That is insane
One thing that seems really ridiculous is the situation with some pupils seemingly being randomly assigned "U"s or similar. Among all but the most useless or disinterested students I imagine a U (or very unexpectly poor mark) is often not remotely a sign of ability, but a consequence of random events on the day. The sort of event that can affect a normally straight A student as must as a borderline C/D student. Sometimes more so, because a straight A student is possibly more likely to panic when confronted with something they haven't properly been prepared for, and has more to lose.
No algorithm should be randomly giving people Us because it can have almost no basis for doing so. It can't rely on teacher's rankings because such outcomes in normal exam life won't easily follow ability like that. I imagine the algorithm has looked at schools, noted they normally get a percentage of Us every year, and then just given them out in proportion. Which is an absurd way to decide a pupil's future life. At least if they have had a panic on the day, they can rationalise the outcome (even if not truly reflective of their potential) and own it to some extent. But to be given a U by a computer!!!
I hope you are making this up. Any system that just randomly assigns "U" grades without good reason is a disgrace. Because some people think the whole procedure is a disgrace is not enough imply what you are claiming.
On what other basis can people be being given Us? By "random" i don't mean that it's completely random - it'll be following the algorithm.
But my point is that in any exam cohort there will be pupils who get unexpected Us (or very poor grades) on isolated papers, due to circumstances on the day. This is normally "random" in the sense that every teacher will know that there will probably be some pupils for whom this happens, but has almost no idea who it is going to happen to.
As the algorithm has tried to replicate past distribution of grades it has assigned "U" or very poor grades, to some pupils. But on what basis, who knows? Obviously it's going to be the exception rather than the rule, but across the entire country there will be unlucky kids. There was a case (obviously caveated as unverified) referenced on here year of a pupil who needed ABB for her university (which was presumably achievable or the Uni wouldn't have accepted her) who got ABU, and consequently couldn't apply anywhere!
"it has assigned "U" or very poor grades, to some pupils. But on what basis, who knows?"
- On the basis of the teacher-produced pupil ranking. If the algorithm says there should be a U in that subject at that particular school the lowest ranked pupil gets it.
Garbage In, Garbage Out.
Since when were teachers able to rank pupils 100% in the correct order with zero mistakes?
Well - since never - exactly - but I`d have a lot of faith that teachers will have a pretty good idea and that exam days generally produce few surprises to them. They`d never get it exactly right though. Too many variables.
If the teachers haven't got it 100% right with zero surprises then this algorithm is by definition unfair on those it got it wrong [on the down side] for since it hasn't given the benefit of the doubt to the pupils. Madness.
Seems to me that we have got to the point where Johnson needs to step in, over-rule Williamson, and do what Scotland did and award centre assessed grades where they are higher than the algorithm.
He would also need to sack him
Probably yes (mainly his fault or Ofqual`s?). Certainly politically he`d have to go.
The Dems will get little boost from a virtual Convention. Indeed even real Conventions seem to produce relatively little in the way of bounce these days, stripped of all meaning and substance they have become rather empty affairs.
The upside is that Trump almost certainly needs these sort of events even more. He feeds off his crowds and hones his messages by how they respond. He looked utterly despondent when his last rally turned into a damp squib and for good reasons. He does not want to campaign by being interviewed by a largely hostile media day after day. He needs that direct link with his base and I am not sure how he is going to replace it.
His press conferences have pretty well tuned into rallies, if you listen to what he spews. But with smaller and tougher audiences.
He gets to jet off steam, and they do get widespread coverage.
Best wishes for the op.
Regarding the emptying hospitals, how much of that is now down to patients being reluctant to go in, and how much to the NHS ? Clearly there is a significant reduction in capacity owing to COVID measures, but it can’t just be that ?
My wife has an on-going small and apparently non-malignant dermatological problem. She had one growth removed early this year, privately. What with one thing and another, upwards of £1k, mainly the dermatologist's time and skill. Now it's recurred and the GP said she needs a further consultant review...... no, needn't pay that sort of money; I'll refer you. So we had a call from an appointment booking service..... you can go to.... naming four reasonably local centres. OK.... naming the most convenient. Not, sorry, nothing in the foreseeable. OK, another. Same. Turns out there's NO possibility of a dermatologist appointment this year in Mid or N Essex.
What your wife is experiencing with the NHS is endemic throughout it. Vast swathes of people are using Covid as a reason to not do their jobs. In this case what are the people who work in these dermatology centres doing all day if they are not seeing any patients. There are huge numbers of people in the NHS going to work each day and are doing nothing. My wife is one of them, she regularly comes home and says "yet another day when I have done nothing". She is a nurse on a ward.
The lunancy of shutting the NHS to protect it from Covid must be killing hundreds of people. Currently less than 1% of hospital beds are taken up with Covid yet we have shut the entire system beacuse of it.
This website provides an interesting snapshot of hospital data.
The most startling one for me was the number of NHS staff who were off sick on the 8th August 2020.
Have a guess.
Its 60,367.
TBH, while I agree about wondering what some of the staff in the dermatology unit are doing, the problem appears to be a shortage of senior medical staff. Nothing, or very little, to do with Covid-19.
Why would there be a shortage of senior medical staff?
I have to say I've never had a remarkably high opinion of dermatology. In my experience patients rarely die..... although there are potentially fatal skin conditions. Equally, they take for ever to recover from whatever it is.
For an NHS positive, I've recently been discharged by the Prostate Cancer clinic; now down to six monthly blood tests, and the GP to review and refer if necessary. But the Registrar, when we had the phone consultation, said he didn't expect I'd be back.
It is also the fact that many GP surgeries are still refusing to see people, therefore are not referring people to hospital
Do you have a source for this "fact"? Most surgeries are I think triaging - initial consultation by phone or electronic, see them as needed. I had a minor issue last week - the GP dfiscussed it by phone and then asked me to look in.
Your posts seem simultaneously to argue that staff are lazily doing nothing and that it's somehow the fault of the NHS. There is pretty general caution among both staff and patients, but I very much doubt that the organisation of healthcare in Britain is a significant factor either way. Just as some people fetishise the NHS, others see it as the root of all problems, but a lot of issues are unrelated to the structure, and the psychoogical impact of COVID is one of them.
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.
On the NHS, I think the "Protect the NHS" slogan is wrong. It should be "Protect the Nation's Health", and the NHS should get back to work soon.
The Dems will get little boost from a virtual Convention. Indeed even real Conventions seem to produce relatively little in the way of bounce these days, stripped of all meaning and substance they have become rather empty affairs.
The upside is that Trump almost certainly needs these sort of events even more. He feeds off his crowds and hones his messages by how they respond. He looked utterly despondent when his last rally turned into a damp squib and for good reasons. He does not want to campaign by being interviewed by a largely hostile media day after day. He needs that direct link with his base and I am not sure how he is going to replace it.
His press conferences have pretty well tuned into rallies, if you listen to what he spews. But with smaller and tougher audiences.
He gets to jet off steam, and they do get widespread coverage.
Best wishes for the op.
Regarding the emptying hospitals, how much of that is now down to patients being reluctant to go in, and how much to the NHS ? Clearly there is a significant reduction in capacity owing to COVID measures, but it can’t just be that ?
My wife has an on-going small and apparently non-malignant dermatological problem. She had one growth removed early this year, privately. What with one thing and another, upwards of £1k, mainly the dermatologist's time and skill. Now it's recurred and the GP said she needs a further consultant review...... no, needn't pay that sort of money; I'll refer you. So we had a call from an appointment booking service..... you can go to.... naming four reasonably local centres. OK.... naming the most convenient. Not, sorry, nothing in the foreseeable. OK, another. Same. Turns out there's NO possibility of a dermatologist appointment this year in Mid or N Essex.
What your wife is experiencing with the NHS is endemic throughout it. Vast swathes of people are using Covid as a reason to not do their jobs. In this case what are the people who work in these dermatology centres doing all day if they are not seeing any patients. There are huge numbers of people in the NHS going to work each day and are doing nothing. My wife is one of them, she regularly comes home and says "yet another day when I have done nothing". She is a nurse on a ward.
The lunancy of shutting the NHS to protect it from Covid must be killing hundreds of people. Currently less than 1% of hospital beds are taken up with Covid yet we have shut the entire system beacuse of it.
This website provides an interesting snapshot of hospital data.
The most startling one for me was the number of NHS staff who were off sick on the 8th August 2020.
Have a guess.
Its 60,367.
TBH, while I agree about wondering what some of the staff in the dermatology unit are doing, the problem appears to be a shortage of senior medical staff. Nothing, or very little, to do with Covid-19.
Why would there be a shortage of senior medical staff?
Er....... AIUI the last dermatologist locally went back to his EU country.
The Dems will get little boost from a virtual Convention. Indeed even real Conventions seem to produce relatively little in the way of bounce these days, stripped of all meaning and substance they have become rather empty affairs.
The upside is that Trump almost certainly needs these sort of events even more. He feeds off his crowds and hones his messages by how they respond. He looked utterly despondent when his last rally turned into a damp squib and for good reasons. He does not want to campaign by being interviewed by a largely hostile media day after day. He needs that direct link with his base and I am not sure how he is going to replace it.
His press conferences have pretty well tuned into rallies, if you listen to what he spews. But with smaller and tougher audiences.
He gets to jet off steam, and they do get widespread coverage.
Best wishes for the op.
Regarding the emptying hospitals, how much of that is now down to patients being reluctant to go in, and how much to the NHS ? Clearly there is a significant reduction in capacity owing to COVID measures, but it can’t just be that ?
My wife has an on-going small and apparently non-malignant dermatological problem. She had one growth removed early this year, privately. What with one thing and another, upwards of £1k, mainly the dermatologist's time and skill. Now it's recurred and the GP said she needs a further consultant review...... no, needn't pay that sort of money; I'll refer you. So we had a call from an appointment booking service..... you can go to.... naming four reasonably local centres. OK.... naming the most convenient. Not, sorry, nothing in the foreseeable. OK, another. Same. Turns out there's NO possibility of a dermatologist appointment this year in Mid or N Essex.
What your wife is experiencing with the NHS is endemic throughout it. Vast swathes of people are using Covid as a reason to not do their jobs. In this case what are the people who work in these dermatology centres doing all day if they are not seeing any patients. There are huge numbers of people in the NHS going to work each day and are doing nothing. My wife is one of them, she regularly comes home and says "yet another day when I have done nothing". She is a nurse on a ward.
The lunancy of shutting the NHS to protect it from Covid must be killing hundreds of people. Currently less than 1% of hospital beds are taken up with Covid yet we have shut the entire system beacuse of it.
This website provides an interesting snapshot of hospital data.
The most startling one for me was the number of NHS staff who were off sick on the 8th August 2020.
Have a guess.
Its 60,367.
TBH, while I agree about wondering what some of the staff in the dermatology unit are doing, the problem appears to be a shortage of senior medical staff. Nothing, or very little, to do with Covid-19.
Why would there be a shortage of senior medical staff?
Er....... AIUI the last dermatologist locally went back to his EU country.
The Dems will get little boost from a virtual Convention. Indeed even real Conventions seem to produce relatively little in the way of bounce these days, stripped of all meaning and substance they have become rather empty affairs.
The upside is that Trump almost certainly needs these sort of events even more. He feeds off his crowds and hones his messages by how they respond. He looked utterly despondent when his last rally turned into a damp squib and for good reasons. He does not want to campaign by being interviewed by a largely hostile media day after day. He needs that direct link with his base and I am not sure how he is going to replace it.
His press conferences have pretty well tuned into rallies, if you listen to what he spews. But with smaller and tougher audiences.
He gets to jet off steam, and they do get widespread coverage.
Best wishes for the op.
Regarding the emptying hospitals, how much of that is now down to patients being reluctant to go in, and how much to the NHS ? Clearly there is a significant reduction in capacity owing to COVID measures, but it can’t just be that ?
My wife has an on-going small and apparently non-malignant dermatological problem. She had one growth removed early this year, privately. What with one thing and another, upwards of £1k, mainly the dermatologist's time and skill. Now it's recurred and the GP said she needs a further consultant review...... no, needn't pay that sort of money; I'll refer you. So we had a call from an appointment booking service..... you can go to.... naming four reasonably local centres. OK.... naming the most convenient. Not, sorry, nothing in the foreseeable. OK, another. Same. Turns out there's NO possibility of a dermatologist appointment this year in Mid or N Essex.
What your wife is experiencing with the NHS is endemic throughout it. Vast swathes of people are using Covid as a reason to not do their jobs. In this case what are the people who work in these dermatology centres doing all day if they are not seeing any patients. There are huge numbers of people in the NHS going to work each day and are doing nothing. My wife is one of them, she regularly comes home and says "yet another day when I have done nothing". She is a nurse on a ward.
The lunancy of shutting the NHS to protect it from Covid must be killing hundreds of people. Currently less than 1% of hospital beds are taken up with Covid yet we have shut the entire system beacuse of it.
This website provides an interesting snapshot of hospital data.
The most startling one for me was the number of NHS staff who were off sick on the 8th August 2020.
Have a guess.
Its 60,367.
Can you give us an average figure for being off sick on a Saturday in early August? Although the number sounds high, with no reference value, it means nothing.
Over a million staff, 60K sick. 6% ...
Yes, and that would include people isolating because of T and T.
Anyone with statistical chops got the time and inclination to wade through it ? (I suspect one reason it’s several hundred pages long is to discourage such an effort.)
I haven't read the whole thing myself but have seen the extract where they talk about their back testing and the back testing isn't right, they don't test the same thing as what they claim their model does. And they back test using as their test data a portion of their training data.
That is insane
One thing that seems really ridiculous is the situation with some pupils seemingly being randomly assigned "U"s or similar. Among all but the most useless or disinterested students I imagine a U (or very unexpectly poor mark) is often not remotely a sign of ability, but a consequence of random events on the day. The sort of event that can affect a normally straight A student as must as a borderline C/D student. Sometimes more so, because a straight A student is possibly more likely to panic when confronted with something they haven't properly been prepared for, and has more to lose.
No algorithm should be randomly giving people Us because it can have almost no basis for doing so. It can't rely on teacher's rankings because such outcomes in normal exam life won't easily follow ability like that. I imagine the algorithm has looked at schools, noted they normally get a percentage of Us every year, and then just given them out in proportion. Which is an absurd way to decide a pupil's future life. At least if they have had a panic on the day, they can rationalise the outcome (even if not truly reflective of their potential) and own it to some extent. But to be given a U by a computer!!!
I hope you are making this up. Any system that just randomly assigns "U" grades without good reason is a disgrace. Because some people think the whole procedure is a disgrace is not enough imply what you are claiming.
On what other basis can people be being given Us? By "random" i don't mean that it's completely random - it'll be following the algorithm.
But my point is that in any exam cohort there will be pupils who get unexpected Us (or very poor grades) on isolated papers, due to circumstances on the day. This is normally "random" in the sense that every teacher will know that there will probably be some pupils for whom this happens, but has almost no idea who it is going to happen to.
As the algorithm has tried to replicate past distribution of grades it has assigned "U" or very poor grades, to some pupils. But on what basis, who knows? Obviously it's going to be the exception rather than the rule, but across the entire country there will be unlucky kids. There was a case (obviously caveated as unverified) referenced on here year of a pupil who needed ABB for her university (which was presumably achievable or the Uni wouldn't have accepted her) who got ABU, and consequently couldn't apply anywhere!
"it has assigned "U" or very poor grades, to some pupils. But on what basis, who knows?"
- On the basis of the teacher-produced pupil ranking. If the algorithm says there should be a U in that subject at that particular school the lowest ranked pupil gets it.
Garbage In, Garbage Out.
Since when were teachers able to rank pupils 100% in the correct order with zero mistakes?
Well - since never - exactly - but I`d have a lot of faith that teachers will have a pretty good idea and that exam days generally produce few surprises to them. They`d never get it exactly right though. Too many variables.
If the teachers haven't got it 100% right with zero surprises then this algorithm is by definition unfair on those it got it wrong [on the down side] for since it hasn't given the benefit of the doubt to the pupils. Madness.
Yes, but if you are saying that the algorithm is shite and teacher rankings are unreliable (which you are) then both will throw up injusticies whether taken seperately or together. So what metric can be used in order to give the "benefit of doubt to the pupils". What can this be based on.
(By the way - thanks for calling them pupils. Calling pupils "students" is irritating.)
It is also the fact that many GP surgeries are still refusing to see people, therefore are not referring people to hospital
Do you have a source for this "fact"? Most surgeries are I think triaging - initial consultation by phone or electronic, see them as needed. I had a minor issue last week - the GP dfiscussed it by phone and then asked me to look in.
Your posts seem simultaneously to argue that staff are lazily doing nothing and that it's somehow the fault of the NHS. There is pretty general caution among both staff and patients, but I very much doubt that the organisation of healthcare in Britain is a significant factor either way. Just as some people fetishise the NHS, others see it as the root of all problems, but a lot of issues are unrelated to the structure, and the psychoogical impact of COVID is one of them.
Our experience with our GP is similar. An initial telephone consultation with a face to face follow up if required. It is very noticeable when you go to the surgery, however, that it is far quieter than it was before. From my observation I am not sure that the steady stream that used to get help from the Practice nurses are there. Not sure where they have gone.
Anyone with statistical chops got the time and inclination to wade through it ? (I suspect one reason it’s several hundred pages long is to discourage such an effort.)
I haven't read the whole thing myself but have seen the extract where they talk about their back testing and the back testing isn't right, they don't test the same thing as what they claim their model does. And they back test using as their test data a portion of their training data.
That is insane
One thing that seems really ridiculous is the situation with some pupils seemingly being randomly assigned "U"s or similar. Among all but the most useless or disinterested students I imagine a U (or very unexpectly poor mark) is often not remotely a sign of ability, but a consequence of random events on the day. The sort of event that can affect a normally straight A student as must as a borderline C/D student. Sometimes more so, because a straight A student is possibly more likely to panic when confronted with something they haven't properly been prepared for, and has more to lose.
No algorithm should be randomly giving people Us because it can have almost no basis for doing so. It can't rely on teacher's rankings because such outcomes in normal exam life won't easily follow ability like that. I imagine the algorithm has looked at schools, noted they normally get a percentage of Us every year, and then just given them out in proportion. Which is an absurd way to decide a pupil's future life. At least if they have had a panic on the day, they can rationalise the outcome (even if not truly reflective of their potential) and own it to some extent. But to be given a U by a computer!!!
I hope you are making this up. Any system that just randomly assigns "U" grades without good reason is a disgrace. Because some people think the whole procedure is a disgrace is not enough imply what you are claiming.
On what other basis can people be being given Us? By "random" i don't mean that it's completely random - it'll be following the algorithm.
But my point is that in any exam cohort there will be pupils who get unexpected Us (or very poor grades) on isolated papers, due to circumstances on the day. This is normally "random" in the sense that every teacher will know that there will probably be some pupils for whom this happens, but has almost no idea who it is going to happen to.
As the algorithm has tried to replicate past distribution of grades it has assigned "U" or very poor grades, to some pupils. But on what basis, who knows? Obviously it's going to be the exception rather than the rule, but across the entire country there will be unlucky kids. There was a case (obviously caveated as unverified) referenced on here year of a pupil who needed ABB for her university (which was presumably achievable or the Uni wouldn't have accepted her) who got ABU, and consequently couldn't apply anywhere!
Its the teacher rankings that is the issue. The algorithm has treated the teacher rankings as divinely accurate then provided the results based on that as to what would be accurate.
Which is patently absurd! The issue with algorithms is if you put garbage in, you get garbage out. Unless the teachers were able to with the Wisdom of Solomon and Mystic Meg combined get the ranking in 100% the right order then anyone ranked lower than they should have been will be punitively getting a terrible grade from the algorithm. The average comes out fine because for everyone getting a terrible grade below what they should have got there is someone else getting a great grade above what they should have got - but that doesn't make things better or acceptable for those that have suffered. If there is any doubt as to the accuracy of the results then being reasonable the benefit of the doubt should go to the student.
That is insane. It should never have happened. The results are not accurate, they are not fair, they don't meet what has historically happened and averaging out errors by insisting two wrongs make a right is madness.
They haven’t even done that. Teachers may well have felt that several student had a near equal chance if getting an A. But they were still required to rank those students in order.
It’s little wonder a large number of unfair grade allocations have been made. Add to that the evident statistical incompetence of Ofqal, which means that they have not even achieved the year on year equivalence that they gave priority to, then you have a complete car crash.
I think the only practical solution is to go with the (admittedly imperfect) teacher grades. The system simply cannot accommodate the vast number of appeals otherwise required.
And of course that would also require lifting the cap on university places,
If he was Boris Johnson, why would he have appointed Williamson as a cabinet minister in the first place. Not to mention it is highly unlikely he would be PM anyway.
The actual PM is not going to be a fan of resignations or sackings for incompetence for obvious reasons. Bluster your way through before u-turning and claiming world leading success is a blueprint that has worked wonders for him so far.
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.
On the NHS, I think the "Protect the NHS" slogan is wrong. It should be "Protect the Nation's Health", and the NHS should get back to work soon.
Congratulations you've just annoyed 100,000 students who should be doing what they want because their centre assessment grades met the university requirements but their actual grades didn't because the algorithms had been screwed by the A and A* awarded by small cohorts in private schools.
Seems to me that we have got to the point where Johnson needs to step in, over-rule Williamson, and do what Scotland did and award centre assessed grades where they are higher than the algorithm.
He would also need to sack him
Probably yes (mainly his fault or Ofqual`s?). Certainly politically he`d have to go.
As I understand it the law needs a change to do that and reports this morning confirm ofqual put out the process in a wide consultation and teachers unions and every interested party endorsed the system being used
I think the GCSE should have teacher assesments as Northern Ireland announced today but a fair appeal system should be agreed for the A levels in England and Wales
Seems to me that we have got to the point where Johnson needs to step in, over-rule Williamson, and do what Scotland did and award centre assessed grades where they are higher than the algorithm.
He would also need to sack him
Probably yes (mainly his fault or Ofqual`s?). Certainly politically he`d have to go.
As I understand it the law needs a change to do that and reports this morning confirm ofqual put out the process in a wide consultation and teachers unions and every interested party endorsed the system being used
I think the GCSE should have teacher assesments as Northern Ireland announced today but a fair appeal system should be agreed for the A levels in England and Wales
Oh - I didn`t know that - has NI followed Scotland??
The only reasons I can see for Trump's odds improving in the last days are his open attempts to cripple the postal service. Punters are rightly taking this seriously.
No, there's a strong CNN poll for him just out. Other recent polls have also been sort of ok for him, so the CNN figures may be reflecting a trend; it's not strong but I would have expected to see Biden pulling away by now and he's not.
Biden is on 51% in the 538 poll tracker, which is pretty much his high watermark. Trump has, however, improved off his lows.
Guessing this is Don't Knows opting for a candidate then since they're not excluded in US polls?
Historically do Don't Knows tend to go more towards the incumbent or the challenger when a POTUS seeks re-election? Or do they split relatively evenly?
Don't knows tend to break 60:40 for the incumbent, although there may be a shy Trump factor this time.
Still, if biden does get North of 51%, it's hard to see him losing. He's tracking about 4-5 points above where Ms Clinton was in 2016.
I'm not sure one CNN poll should move the betting that much. Currently Biden is 8.0% ahead nationally on the 538 average. A week ago he was 7.8% ahead. There's a bit of noise but his lead has been very stable for the last 3 weeks, while the betting has seen a bit of a rise in Trump's implied chances during this time.
I suspect the betting is moving based on concerns over the USPS and that Trump might pull out the stops to steal the election Belarus-style.
The terrible thing this year is we're not just betting on whom the electorate wants to vote for the most, but will they be able to do so freely. From the nation formerly known as the leader of the free world.
And how will the votes be counted. And what will the courts decide. Not who wins fair and square.
If he was Boris Johnson, why would he have appointed Williamson as a cabinet minister in the first place. Not to mention it is highly unlikely he would be PM anyway.
The actual PM is not going to be a fan of resignations or sackings for incompetence for obvious reasons. Bluster your way through before u-turning and claiming world leading success is a blueprint that has worked wonders for him so far.
isn't it better that williamson takes the hit for gcse mess and then is sacked?
Anyone with statistical chops got the time and inclination to wade through it ? (I suspect one reason it’s several hundred pages long is to discourage such an effort.)
I haven't read the whole thing myself but have seen the extract where they talk about their back testing and the back testing isn't right, they don't test the same thing as what they claim their model does. And they back test using as their test data a portion of their training data.
That is insane
One thing that seems really ridiculous is the situation with some pupils seemingly being randomly assigned "U"s or similar. Among all but the most useless or disinterested students I imagine a U (or very unexpectly poor mark) is often not remotely a sign of ability, but a consequence of random events on the day. The sort of event that can affect a normally straight A student as must as a borderline C/D student. Sometimes more so, because a straight A student is possibly more likely to panic when confronted with something they haven't properly been prepared for, and has more to lose.
No algorithm should be randomly giving people Us because it can have almost no basis for doing so. It can't rely on teacher's rankings because such outcomes in normal exam life won't easily follow ability like that. I imagine the algorithm has looked at schools, noted they normally get a percentage of Us every year, and then just given them out in proportion. Which is an absurd way to decide a pupil's future life. At least if they have had a panic on the day, they can rationalise the outcome (even if not truly reflective of their potential) and own it to some extent. But to be given a U by a computer!!!
I hope you are making this up. Any system that just randomly assigns "U" grades without good reason is a disgrace. Because some people think the whole procedure is a disgrace is not enough imply what you are claiming.
On what other basis can people be being given Us? By "random" i don't mean that it's completely random - it'll be following the algorithm.
But my point is that in any exam cohort there will be pupils who get unexpected Us (or very poor grades) on isolated papers, due to circumstances on the day. This is normally "random" in the sense that every teacher will know that there will probably be some pupils for whom this happens, but has almost no idea who it is going to happen to.
As the algorithm has tried to replicate past distribution of grades it has assigned "U" or very poor grades, to some pupils. But on what basis, who knows? Obviously it's going to be the exception rather than the rule, but across the entire country there will be unlucky kids. There was a case (obviously caveated as unverified) referenced on here year of a pupil who needed ABB for her university (which was presumably achievable or the Uni wouldn't have accepted her) who got ABU, and consequently couldn't apply anywhere!
"it has assigned "U" or very poor grades, to some pupils. But on what basis, who knows?"
- On the basis of the teacher-produced pupil ranking. If the algorithm says there should be a U in that subject at that particular school the lowest ranked pupil gets it.
Garbage In, Garbage Out.
Since when were teachers able to rank pupils 100% in the correct order with zero mistakes?
Well - since never - exactly - but I`d have a lot of faith that teachers will have a pretty good idea and that exam days generally produce few surprises to them. They`d never get it exactly right though. Too many variables.
If the teachers haven't got it 100% right with zero surprises then this algorithm is by definition unfair on those it got it wrong [on the down side] for since it hasn't given the benefit of the doubt to the pupils. Madness.
Yes, but if you are saying that the algorithm is shite and teacher rankings are unreliable (which you are) then both will throw up injusticies whether taken seperately or together. So what metric can be used in order to give the "benefit of doubt to the pupils". What can this be based on.
(By the way - thanks for calling them pupils. Calling pupils "students" is irritating.)
In the absence of any more reliable and more accurate alternative? The teacher's predicted grades will be the best alternative.
Alternatively if designing an algorithm from scratch then one that adjusts for any uncertainty and then rounds up at each stage of the calculation would have been more reasonable. It may have given better grades than the teacher's predictions though.
Any reasonable algorithm that had accounted for uncertainty and given the benefit of the doubt to the pupil (rounded up) would have seen a significant increase in pupils getting better grades overall. Or more importantly in the inverse a significant reduction in pupils getting bad grades. That doesn't seem to have happened.
The Dems will get little boost from a virtual Convention. Indeed even real Conventions seem to produce relatively little in the way of bounce these days, stripped of all meaning and substance they have become rather empty affairs.
The upside is that Trump almost certainly needs these sort of events even more. He feeds off his crowds and hones his messages by how they respond. He looked utterly despondent when his last rally turned into a damp squib and for good reasons. He does not want to campaign by being interviewed by a largely hostile media day after day. He needs that direct link with his base and I am not sure how he is going to replace it.
His press conferences have pretty well tuned into rallies, if you listen to what he spews. But with smaller and tougher audiences.
He gets to jet off steam, and they do get widespread coverage.
Best wishes for the op.
Regarding the emptying hospitals, how much of that is now down to patients being reluctant to go in, and how much to the NHS ? Clearly there is a significant reduction in capacity owing to COVID measures, but it can’t just be that ?
My wife has an on-going small and apparently non-malignant dermatological problem. She had one growth removed early this year, privately. What with one thing and another, upwards of £1k, mainly the dermatologist's time and skill. Now it's recurred and the GP said she needs a further consultant review...... no, needn't pay that sort of money; I'll refer you. So we had a call from an appointment booking service..... you can go to.... naming four reasonably local centres. OK.... naming the most convenient. Not, sorry, nothing in the foreseeable. OK, another. Same. Turns out there's NO possibility of a dermatologist appointment this year in Mid or N Essex.
What your wife is experiencing with the NHS is endemic throughout it. Vast swathes of people are using Covid as a reason to not do their jobs. In this case what are the people who work in these dermatology centres doing all day if they are not seeing any patients. There are huge numbers of people in the NHS going to work each day and are doing nothing. My wife is one of them, she regularly comes home and says "yet another day when I have done nothing". She is a nurse on a ward.
The lunancy of shutting the NHS to protect it from Covid must be killing hundreds of people. Currently less than 1% of hospital beds are taken up with Covid yet we have shut the entire system beacuse of it.
This website provides an interesting snapshot of hospital data.
The most startling one for me was the number of NHS staff who were off sick on the 8th August 2020.
Have a guess.
Its 60,367.
TBH, while I agree about wondering what some of the staff in the dermatology unit are doing, the problem appears to be a shortage of senior medical staff. Nothing, or very little, to do with Covid-19.
Why would there be a shortage of senior medical staff?
Er....... AIUI the last dermatologist locally went back to his EU country.
Yes, we have lost 3 senior EU staff since Jan, and another took early retirement as couldn't be doing with all the PPE malarkey.
Most have stayed, but the prospect of not seeing their families for a year has given some of them pause for thought.
Seems to me that we have got to the point where Johnson needs to step in, over-rule Williamson, and do what Scotland did and award centre assessed grades where they are higher than the algorithm.
He would also need to sack him
Probably yes (mainly his fault or Ofqual`s?). Certainly politically he`d have to go.
As I understand it the law needs a change to do that and reports this morning confirm ofqual put out the process in a wide consultation and teachers unions and every interested party endorsed the system being used
I think the GCSE should have teacher assesments as Northern Ireland announced today but a fair appeal system should be agreed for the A levels in England and Wales
Oh - I didn`t know that - has NI followed Scotland??
Not as far as I am aware, as I understand it this is for GCSE only
Also Stormant has been recalled over the A results adding credence to law changes
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
If he was Boris Johnson, why would he have appointed Williamson as a cabinet minister in the first place. Not to mention it is highly unlikely he would be PM anyway.
The actual PM is not going to be a fan of resignations or sackings for incompetence for obvious reasons. Bluster your way through before u-turning and claiming world leading success is a blueprint that has worked wonders for him so far.
isn't it better that williamson takes the hit for gcse mess and then is sacked?
Whether he gets sacked or not is secondary.
He should announce today that the GCSE mess and A-Level mess are being fixed by going with teacher's predictions. Like the Scots did beforehand and which he tried to fix with the last minute mocks fudge but that doesn't seem to have worked.
There is a clear problem and there is a clear solution. JFDI now.
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?
The actual PM is not going to be a fan of resignations or sackings for incompetence for obvious reasons. Bluster your way through before u-turning and claiming world leading success is a blueprint that has worked wonders for him so far.
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.
On the NHS, I think the "Protect the NHS" slogan is wrong. It should be "Protect the Nation's Health", and the NHS should get back to work soon.
I have been complaining about that slogan for weeks. The NHS is supposedly to protect the health of the nation not itself...
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.
Anyone with statistical chops got the time and inclination to wade through it ? (I suspect one reason it’s several hundred pages long is to discourage such an effort.)
I haven't read the whole thing myself but have seen the extract where they talk about their back testing and the back testing isn't right, they don't test the same thing as what they claim their model does. And they back test using as their test data a portion of their training data.
That is insane
One thing that seems really ridiculous is the situation with some pupils seemingly being randomly assigned "U"s or similar. Among all but the most useless or disinterested students I imagine a U (or very unexpectly poor mark) is often not remotely a sign of ability, but a consequence of random events on the day. The sort of event that can affect a normally straight A student as must as a borderline C/D student. Sometimes more so, because a straight A student is possibly more likely to panic when confronted with something they haven't properly been prepared for, and has more to lose.
No algorithm should be randomly giving people Us because it can have almost no basis for doing so. It can't rely on teacher's rankings because such outcomes in normal exam life won't easily follow ability like that. I imagine the algorithm has looked at schools, noted they normally get a percentage of Us every year, and then just given them out in proportion. Which is an absurd way to decide a pupil's future life. At least if they have had a panic on the day, they can rationalise the outcome (even if not truly reflective of their potential) and own it to some extent. But to be given a U by a computer!!!
I hope you are making this up. Any system that just randomly assigns "U" grades without good reason is a disgrace. Because some people think the whole procedure is a disgrace is not enough imply what you are claiming.
On what other basis can people be being given Us? By "random" i don't mean that it's completely random - it'll be following the algorithm.
But my point is that in any exam cohort there will be pupils who get unexpected Us (or very poor grades) on isolated papers, due to circumstances on the day. This is normally "random" in the sense that every teacher will know that there will probably be some pupils for whom this happens, but has almost no idea who it is going to happen to.
As the algorithm has tried to replicate past distribution of grades it has assigned "U" or very poor grades, to some pupils. But on what basis, who knows? Obviously it's going to be the exception rather than the rule, but across the entire country there will be unlucky kids. There was a case (obviously caveated as unverified) referenced on here year of a pupil who needed ABB for her university (which was presumably achievable or the Uni wouldn't have accepted her) who got ABU, and consequently couldn't apply anywhere!
"it has assigned "U" or very poor grades, to some pupils. But on what basis, who knows?"
- On the basis of the teacher-produced pupil ranking. If the algorithm says there should be a U in that subject at that particular school the lowest ranked pupil gets it.
Garbage In, Garbage Out.
Since when were teachers able to rank pupils 100% in the correct order with zero mistakes?
Well - since never - exactly - but I`d have a lot of faith that teachers will have a pretty good idea and that exam days generally produce few surprises to them. They`d never get it exactly right though. Too many variables.
If the teachers haven't got it 100% right with zero surprises then this algorithm is by definition unfair on those it got it wrong [on the down side] for since it hasn't given the benefit of the doubt to the pupils. Madness.
Yes, but if you are saying that the algorithm is shite and teacher rankings are unreliable (which you are) then both will throw up injusticies whether taken seperately or together. So what metric can be used in order to give the "benefit of doubt to the pupils". What can this be based on.
(By the way - thanks for calling them pupils. Calling pupils "students" is irritating.)
In the absence of any more reliable and more accurate alternative? The teacher's predicted grades will be the best alternative.
Alternatively if designing an algorithm from scratch then one that adjusts for any uncertainty and then rounds up at each stage of the calculation would have been more reasonable. It may have given better grades than the teacher's predictions though.
Any reasonable algorithm that had accounted for uncertainty and given the benefit of the doubt to the pupil (rounded up) would have seen a significant increase in pupils getting better grades overall. Or more importantly in the inverse a significant reduction in pupils getting bad grades. That doesn't seem to have happened.
Hold on - you pointed out that the teachers`rankings are unreliable but then say that the "teacher's predicted grades will be the best alternative" - they are the same thing. The predicted grades were translated into a rank (league style). If you are saying that the rank is wrong the grades must be wrong.
The actual PM is not going to be a fan of resignations or sackings for incompetence for obvious reasons. Bluster your way through before u-turning and claiming world leading success is a blueprint that has worked wonders for him so far.
Except that every claim in that Tweet is nonsense.
We don't have the worst per capita deaths, we do have an effective track and trace system and the virus is more under control now in this country than France, Spain, Sweden, Portugal, the USA and more.
It is also the fact that many GP surgeries are still refusing to see people, therefore are not referring people to hospital
Do you have a source for this "fact"? Most surgeries are I think triaging - initial consultation by phone or electronic, see them as needed. I had a minor issue last week - the GP dfiscussed it by phone and then asked me to look in.
Your posts seem simultaneously to argue that staff are lazily doing nothing and that it's somehow the fault of the NHS. There is pretty general caution among both staff and patients, but I very much doubt that the organisation of healthcare in Britain is a significant factor either way. Just as some people fetishise the NHS, others see it as the root of all problems, but a lot of issues are unrelated to the structure, and the psychoogical impact of COVID is one of them.
I only have anecdotal information as I am sure that the Government is not producing stats on how few people GP surgeries are seeing. When your wife and daughter are both nurses and lots of their friends are nurses the information I receive is pretty reliable. Some GP surgeries are now open almost normally, but most are still shut and are refusing to see people. Ask any pharmacist what they currently think of their local GP surgery.
I have been saying on here for months that hosptials are empty and its only when a poster goes to hospital or has experience of the NHS that they report the same. There is almost a disbelief that hospitals are quite places.
The obsession of the NHS with not being able to do anything due to Covid is costing thousand of lives.
0.6% of beds that are occupied have people in them that have tested positive for Covid. Out of total available beds just 0.3% are occupied with Covid positive. The NHS should be nearly back to normal now but it seems there is little interest in doing this. Just 56% of beds are occupied.
If he was Boris Johnson, why would he have appointed Williamson as a cabinet minister in the first place. Not to mention it is highly unlikely he would be PM anyway.
The actual PM is not going to be a fan of resignations or sackings for incompetence for obvious reasons. Bluster your way through before u-turning and claiming world leading success is a blueprint that has worked wonders for him so far.
isn't it better that williamson takes the hit for gcse mess and then is sacked?
Whether he gets sacked or not is secondary.
He should announce today that the GCSE mess and A-Level mess are being fixed by going with teacher's predictions. Like the Scots did beforehand and which he tried to fix with the last minute mocks fudge but that doesn't seem to have worked.
There is a clear problem and there is a clear solution. JFDI now.
It may need the recall of parliament to adjust legislation
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.
On the NHS, I think the "Protect the NHS" slogan is wrong. It should be "Protect the Nation's Health", and the NHS should get back to work soon.
Congratulations you've just annoyed 100,000 students who should be doing what they want because their centre assessment grades met the university requirements but their actual grades didn't because the algorithms had been screwed by the A and A* awarded by small cohorts in private schools.
The actual PM is not going to be a fan of resignations or sackings for incompetence for obvious reasons. Bluster your way through before u-turning and claiming world leading success is a blueprint that has worked wonders for him so far.
Except that every claim in that Tweet is nonsense.
We don't have the worst per capita deaths, we do have an effective track and trace system and the virus is more under control now in this country than France, Spain, Sweden, Portugal, the USA and more.
In any case - deaths shouldn`t be the metric, infections a slightly better guide, hospitalisations better still. A international comparision of covid hospitalisations per capita would be interesting.
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
Anyone with statistical chops got the time and inclination to wade through it ? (I suspect one reason it’s several hundred pages long is to discourage such an effort.)
I haven't read the whole thing myself but have seen the extract where they talk about their back testing and the back testing isn't right, they don't test the same thing as what they claim their model does. And they back test using as their test data a portion of their training data.
That is insane
One thing that seems really ridiculous is the situation with some pupils seemingly being randomly assigned "U"s or similar. Among all but the most useless or disinterested students I imagine a U (or very unexpectly poor mark) is often not remotely a sign of ability, but a consequence of random events on the day. The sort of event that can affect a normally straight A student as must as a borderline C/D student. Sometimes more so, because a straight A student is possibly more likely to panic when confronted with something they haven't properly been prepared for, and has more to lose.
No algorithm should be randomly giving people Us because it can have almost no basis for doing so. It can't rely on teacher's rankings because such outcomes in normal exam life won't easily follow ability like that. I imagine the algorithm has looked at schools, noted they normally get a percentage of Us every year, and then just given them out in proportion. Which is an absurd way to decide a pupil's future life. At least if they have had a panic on the day, they can rationalise the outcome (even if not truly reflective of their potential) and own it to some extent. But to be given a U by a computer!!!
I hope you are making this up. Any system that just randomly assigns "U" grades without good reason is a disgrace. Because some people think the whole procedure is a disgrace is not enough imply what you are claiming.
On what other basis can people be being given Us? By "random" i don't mean that it's completely random - it'll be following the algorithm.
But my point is that in any exam cohort there will be pupils who get unexpected Us (or very poor grades) on isolated papers, due to circumstances on the day. This is normally "random" in the sense that every teacher will know that there will probably be some pupils for whom this happens, but has almost no idea who it is going to happen to.
As the algorithm has tried to replicate past distribution of grades it has assigned "U" or very poor grades, to some pupils. But on what basis, who knows? Obviously it's going to be the exception rather than the rule, but across the entire country there will be unlucky kids. There was a case (obviously caveated as unverified) referenced on here year of a pupil who needed ABB for her university (which was presumably achievable or the Uni wouldn't have accepted her) who got ABU, and consequently couldn't apply anywhere!
"it has assigned "U" or very poor grades, to some pupils. But on what basis, who knows?"
- On the basis of the teacher-produced pupil ranking. If the algorithm says there should be a U in that subject at that particular school the lowest ranked pupil gets it.
Garbage In, Garbage Out.
Since when were teachers able to rank pupils 100% in the correct order with zero mistakes?
Well - since never - exactly - but I`d have a lot of faith that teachers will have a pretty good idea and that exam days generally produce few surprises to them. They`d never get it exactly right though. Too many variables.
If the teachers haven't got it 100% right with zero surprises then this algorithm is by definition unfair on those it got it wrong [on the down side] for since it hasn't given the benefit of the doubt to the pupils. Madness.
Yes, but if you are saying that the algorithm is shite and teacher rankings are unreliable (which you are) then both will throw up injusticies whether taken seperately or together. So what metric can be used in order to give the "benefit of doubt to the pupils". What can this be based on.
(By the way - thanks for calling them pupils. Calling pupils "students" is irritating.)
In the absence of any more reliable and more accurate alternative? The teacher's predicted grades will be the best alternative.
Alternatively if designing an algorithm from scratch then one that adjusts for any uncertainty and then rounds up at each stage of the calculation would have been more reasonable. It may have given better grades than the teacher's predictions though.
Any reasonable algorithm that had accounted for uncertainty and given the benefit of the doubt to the pupil (rounded up) would have seen a significant increase in pupils getting better grades overall. Or more importantly in the inverse a significant reduction in pupils getting bad grades. That doesn't seem to have happened.
Hold on - you pointed out that the teachers`rankings are unreliable but then say that the "teacher's predicted grades will be the best alternative" - they are the same thing. The predicted grades were translated into a rank (league style). If you are saying that the rank is wrong the grades must be wrong.
No because the teachers have done what the algorithm should have done and given the benefit of the doubt to the pupils.
EG if the teacher thinks that 3 pupils in their class are capable of an A and predicts each of them an A then I say they should have all gotten an A.
If historically randomly one of the three (lets say in this instance Pupil 2) may have fluffed their exam and gotten a B then under the algorithm Pupil 2 keeps their A but Pupil 3 gets a B. That is one correct grade and 2 incorrect grades, though the historical average of 2 As and a B is correct they've gone to the wrong people.
In that case the teacher's grades are more accurate on an individual basis as it gave the benefit of the doubt to the pupils so is getting the errors that are on the upside only. The algorithm has tried to keep the average right by doubling the errors, for every error on the upside there is one on the downside. And downside errors are far worse than upside ones.
Seems to me that we have got to the point where Johnson needs to step in, over-rule Williamson, and do what Scotland did and award centre assessed grades where they are higher than the algorithm.
He would also need to sack him
Probably yes (mainly his fault or Ofqual`s?). Certainly politically he`d have to go.
As I understand it the law needs a change to do that and reports this morning confirm ofqual put out the process in a wide consultation and teachers unions and every interested party endorsed the system being used
I think the GCSE should have teacher assesments as Northern Ireland announced today but a fair appeal system should be agreed for the A levels in England and Wales
Ofqal did not share the details of the algorithm it applied. It even refused to get outside help to validate it.
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.
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
If universities are doing more remote learning next year they can probably easier absorb the extra places without worrying about packed lecture halls.
The actual PM is not going to be a fan of resignations or sackings for incompetence for obvious reasons. Bluster your way through before u-turning and claiming world leading success is a blueprint that has worked wonders for him so far.
As I have pointed out many times the fact the test, track and trace fell under Hancock's remit when he was already overloaded with other equally important but immediate responsibilities was primarily a failure of his bosses (Johnson and Raab for an important spell) rather than Hancock.
If he was Boris Johnson, why would he have appointed Williamson as a cabinet minister in the first place. Not to mention it is highly unlikely he would be PM anyway.
The actual PM is not going to be a fan of resignations or sackings for incompetence for obvious reasons. Bluster your way through before u-turning and claiming world leading success is a blueprint that has worked wonders for him so far.
isn't it better that williamson takes the hit for gcse mess and then is sacked?
Whether he gets sacked or not is secondary.
He should announce today that the GCSE mess and A-Level mess are being fixed by going with teacher's predictions. Like the Scots did beforehand and which he tried to fix with the last minute mocks fudge but that doesn't seem to have worked.
There is a clear problem and there is a clear solution. JFDI now.
It may need the recall of parliament to adjust legislation
Anyone with statistical chops got the time and inclination to wade through it ? (I suspect one reason it’s several hundred pages long is to discourage such an effort.)
I haven't read the whole thing myself but have seen the extract where they talk about their back testing and the back testing isn't right, they don't test the same thing as what they claim their model does. And they back test using as their test data a portion of their training data.
That is insane
One thing that seems really ridiculous is the situation with some pupils seemingly being randomly assigned "U"s or similar. Among all but the most useless or disinterested students I imagine a U (or very unexpectly poor mark) is often not remotely a sign of ability, but a consequence of random events on the day. The sort of event that can affect a normally straight A student as must as a borderline C/D student. Sometimes more so, because a straight A student is possibly more likely to panic when confronted with something they haven't properly been prepared for, and has more to lose.
No algorithm should be randomly giving people Us because it can have almost no basis for doing so. It can't rely on teacher's rankings because such outcomes in normal exam life won't easily follow ability like that. I imagine the algorithm has looked at schools, noted they normally get a percentage of Us every year, and then just given them out in proportion. Which is an absurd way to decide a pupil's future life. At least if they have had a panic on the day, they can rationalise the outcome (even if not truly reflective of their potential) and own it to some extent. But to be given a U by a computer!!!
I hope you are making this up. Any system that just randomly assigns "U" grades without good reason is a disgrace. Because some people think the whole procedure is a disgrace is not enough imply what you are claiming.
On what other basis can people be being given Us? By "random" i don't mean that it's completely random - it'll be following the algorithm.
But my point is that in any exam cohort there will be pupils who get unexpected Us (or very poor grades) on isolated papers, due to circumstances on the day. This is normally "random" in the sense that every teacher will know that there will probably be some pupils for whom this happens, but has almost no idea who it is going to happen to.
As the algorithm has tried to replicate past distribution of grades it has assigned "U" or very poor grades, to some pupils. But on what basis, who knows? Obviously it's going to be the exception rather than the rule, but across the entire country there will be unlucky kids. There was a case (obviously caveated as unverified) referenced on here year of a pupil who needed ABB for her university (which was presumably achievable or the Uni wouldn't have accepted her) who got ABU, and consequently couldn't apply anywhere!
Its the teacher rankings that is the issue. The algorithm has treated the teacher rankings as divinely accurate then provided the results based on that as to what would be accurate.
Which is patently absurd! The issue with algorithms is if you put garbage in, you get garbage out. Unless the teachers were able to with the Wisdom of Solomon and Mystic Meg combined get the ranking in 100% the right order then anyone ranked lower than they should have been will be punitively getting a terrible grade from the algorithm. The average comes out fine because for everyone getting a terrible grade below what they should have got there is someone else getting a great grade above what they should have got - but that doesn't make things better or acceptable for those that have suffered. If there is any doubt as to the accuracy of the results then being reasonable the benefit of the doubt should go to the student.
That is insane. It should never have happened. The results are not accurate, they are not fair, they don't meet what has historically happened and averaging out errors by insisting two wrongs make a right is madness.
My point in the case of "U"s is that even if the teacher rankings weren't garbage, the algorithm would still fail. Because in normal times "U"s or unexpected poor individual results will strike randomly. It won't conveniently be the weakest pupil who will underperform. It might just as easily be the strongest pupil.
That's what makes the teacher rankings garbage . . .
That's not a failing of the teachers, its a failing of the system designed. Teachers can't be expected to 100% accurately foretell the ranking of their pupils.
Maybe the teacher assessments are more accurate than the exams?
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
If universities are doing more remote learning next year they can probably easier absorb the extra places without worrying about packed lecture halls.
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
If universities are doing more remote learning next year they can probably easier absorb the extra places without worrying about packed lecture halls.
I suspect Scotland's universities will see a big fall in foreign students due to covid providing more capacity, indeed I expect this to be an issue UK wide
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
If universities are doing more remote learning next year they can probably easier absorb the extra places without worrying about packed lecture halls.
Only if government lifts the cap on numbers.
JFDI.
That's my answer to all this. JFDI. There's a problem, it needs a solution. JFDI.
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.
How could they do that since they didn't have predicted rankings last year?
The issue is that they came up with a fudge to get the exams this year to match last years as an average. So as an average yes they would have passed your test, but the garbage is assuming you can accurately match individual pupils to the exact grades predicted that way.
Your test would have been hit as an average, but an average isn't good enough. It needs to match at least what everyone should have got individually not on average and it hasn't done that.
It is individuals that matter for their individual results, not averages. I always believe in individualism over collectivism, but grades is surely where individualism should be at the forefront. Saying that someone gets a worse grade than they should but that is OK because collectively someone else got a better one instead - that just adds salt to the wound.
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
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.
On the NHS, I think the "Protect the NHS" slogan is wrong. It should be "Protect the Nation's Health", and the NHS should get back to work soon.
Congratulations you've just annoyed 100,000 students who should be doing what they want because their centre assessment grades met the university requirements but their actual grades didn't because the algorithms had been screwed by the A and A* awarded by small cohorts in private schools.
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
If universities are doing more remote learning next year they can probably easier absorb the extra places without worrying about packed lecture halls.
Lecture halls can be more easily managed than laboratories.
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
Then recall Westminster and change the law. That is the government's prerogative not Ofqual's.
That MPs are on holiday is not a good enough reason not to recall Westminster. Especially when they have remote voting and a vote on this should go through pretty unanimously if whipped from all parties.
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.
On the NHS, I think the "Protect the NHS" slogan is wrong. It should be "Protect the Nation's Health", and the NHS should get back to work soon.
Congratulations you've just annoyed 100,000 students who should be doing what they want because their centre assessment grades met the university requirements but their actual grades didn't because the algorithms had been screwed by the A and A* awarded by small cohorts in private schools.
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.
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.
Anyone with statistical chops got the time and inclination to wade through it ? (I suspect one reason it’s several hundred pages long is to discourage such an effort.)
I haven't read the whole thing myself but have seen the extract where they talk about their back testing and the back testing isn't right, they don't test the same thing as what they claim their model does. And they back test using as their test data a portion of their training data.
That is insane
One thing that seems really ridiculous is the situation with some pupils seemingly being randomly assigned "U"s or similar. Among all but the most useless or disinterested students I imagine a U (or very unexpectly poor mark) is often not remotely a sign of ability, but a consequence of random events on the day. The sort of event that can affect a normally straight A student as must as a borderline C/D student. Sometimes more so, because a straight A student is possibly more likely to panic when confronted with something they haven't properly been prepared for, and has more to lose.
No algorithm should be randomly giving people Us because it can have almost no basis for doing so. It can't rely on teacher's rankings because such outcomes in normal exam life won't easily follow ability like that. I imagine the algorithm has looked at schools, noted they normally get a percentage of Us every year, and then just given them out in proportion. Which is an absurd way to decide a pupil's future life. At least if they have had a panic on the day, they can rationalise the outcome (even if not truly reflective of their potential) and own it to some extent. But to be given a U by a computer!!!
I hope you are making this up. Any system that just randomly assigns "U" grades without good reason is a disgrace. Because some people think the whole procedure is a disgrace is not enough imply what you are claiming.
On what other basis can people be being given Us? By "random" i don't mean that it's completely random - it'll be following the algorithm.
But my point is that in any exam cohort there will be pupils who get unexpected Us (or very poor grades) on isolated papers, due to circumstances on the day. This is normally "random" in the sense that every teacher will know that there will probably be some pupils for whom this happens, but has almost no idea who it is going to happen to.
As the algorithm has tried to replicate past distribution of grades it has assigned "U" or very poor grades, to some pupils. But on what basis, who knows? Obviously it's going to be the exception rather than the rule, but across the entire country there will be unlucky kids. There was a case (obviously caveated as unverified) referenced on here year of a pupil who needed ABB for her university (which was presumably achievable or the Uni wouldn't have accepted her) who got ABU, and consequently couldn't apply anywhere!
"it has assigned "U" or very poor grades, to some pupils. But on what basis, who knows?"
- On the basis of the teacher-produced pupil ranking. If the algorithm says there should be a U in that subject at that particular school the lowest ranked pupil gets it.
Garbage In, Garbage Out.
Since when were teachers able to rank pupils 100% in the correct order with zero mistakes?
Depends on the size of the group and how varied they are, but I would struggle once the numbers got much past ten.
I’m surprised just how badly this has gone, in part because I don’t have the faith that many of you seem to that the results they would have got had the exams run are a perfect representation of the ability and understanding of the candidates. Grades in normal years have a margin of error of at least one up or down (except perhaps for the top few who will get straight A*s or now 9s at GCSE) and all that I thought would happen this year is that the margin would increase. Certainly when I looked at the results of my own class their were no surprises: everyone seemed to have got a grade in line with my own assessment (which was NOT necessarily the same as the predicted grade sent of by the school).
However, it has become clear that in in a small percentage of cases, which is a large absolute number given how many people we are talking about, the algorithm used has done something stupid. It seems that there was no human in the loop at any stage with the ability to look at the outliers and say no. Ministers are not in control of the situation and, as someone put it earlier, OFQAL have proved even more useless than @ydoethur predicted they would be.
At this stage the only way out I can see is for ministers to bypass the exams completely and tell universities that there will be no penalty for accepting anyone they made an offer to even if the algorithm gave them grades which were too low.
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.
But how do you fix the other half of it. Universities have awarded places so how do you help the person who lost their university place, has been re-awarded it but the university no longer has accommodation..
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.
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.
The problem is predict accurately is different for a cohort of many pupils, and the individual pupil.
Experience and common sense would suggest, although must admit that I havent seen data for it, that students will tend to have much more unexpected downside risk than upside risk. A "B" student will sometimes get a D or E due to illness, stress, hangover, relationship breakup etc. By contrast it is very hard to see how they magic an A*.
A reasonable proportion of the lower grades will happen every year so the aggregate model will move the lowest "B"s into a C. In real life the lowest Bs would mostly be Bs still, and the Cs would come from across the board.
There will therefore be students who are odds on for a B, the correct prediction individually is B but an algorithm has to give them a C if it is required to balance to previous years.
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.
Whereas you seek to blame Johnson for the manifest errors of others.
Boris needs a reshuffle - and then a mighty kick up the arse to the new Cabinet. The six months since Covid hit have been largely woeful.
You know how if you buy clothes online you get a "Size Guide" where it tells you how many people thought the size was "as expected" or "bigger" or "smaller"?
Do the same with the grades - give them the teachers' predicted grades with a note saying in previous years x% of pupils at this school/teacher/ got the predicted grade y% got a better grade and z% got a worse grade. There's probably loads of things wrong with that idea, but might be fairer than anything else.
Anyone with statistical chops got the time and inclination to wade through it ? (I suspect one reason it’s several hundred pages long is to discourage such an effort.)
I haven't read the whole thing myself but have seen the extract where they talk about their back testing and the back testing isn't right, they don't test the same thing as what they claim their model does. And they back test using as their test data a portion of their training data.
That is insane
One thing that seems really ridiculous is the situation with some pupils seemingly being randomly assigned "U"s or similar. Among all but the most useless or disinterested students I imagine a U (or very unexpectly poor mark) is often not remotely a sign of ability, but a consequence of random events on the day. The sort of event that can affect a normally straight A student as must as a borderline C/D student. Sometimes more so, because a straight A student is possibly more likely to panic when confronted with something they haven't properly been prepared for, and has more to lose.
No algorithm should be randomly giving people Us because it can have almost no basis for doing so. It can't rely on teacher's rankings because such outcomes in normal exam life won't easily follow ability like that. I imagine the algorithm has looked at schools, noted they normally get a percentage of Us every year, and then just given them out in proportion. Which is an absurd way to decide a pupil's future life. At least if they have had a panic on the day, they can rationalise the outcome (even if not truly reflective of their potential) and own it to some extent. But to be given a U by a computer!!!
I hope you are making this up. Any system that just randomly assigns "U" grades without good reason is a disgrace. Because some people think the whole procedure is a disgrace is not enough imply what you are claiming.
On what other basis can people be being given Us? By "random" i don't mean that it's completely random - it'll be following the algorithm.
But my point is that in any exam cohort there will be pupils who get unexpected Us (or very poor grades) on isolated papers, due to circumstances on the day. This is normally "random" in the sense that every teacher will know that there will probably be some pupils for whom this happens, but has almost no idea who it is going to happen to.
As the algorithm has tried to replicate past distribution of grades it has assigned "U" or very poor grades, to some pupils. But on what basis, who knows? Obviously it's going to be the exception rather than the rule, but across the entire country there will be unlucky kids. There was a case (obviously caveated as unverified) referenced on here year of a pupil who needed ABB for her university (which was presumably achievable or the Uni wouldn't have accepted her) who got ABU, and consequently couldn't apply anywhere!
"it has assigned "U" or very poor grades, to some pupils. But on what basis, who knows?"
- On the basis of the teacher-produced pupil ranking. If the algorithm says there should be a U in that subject at that particular school the lowest ranked pupil gets it.
Garbage In, Garbage Out.
Since when were teachers able to rank pupils 100% in the correct order with zero mistakes?
Depends on the size of the group and how varied they are, but I would struggle once the numbers got much past ten.
I’m surprised just how badly this has gone, in part because I don’t have the faith that many of you seem to that the results they would have got had the exams run are a perfect representation of the ability and understanding of the candidates. Grades in normal years have a margin of error of at least one up or down (except perhaps for the top few who will get straight A*s or now 9s at GCSE) and all that I thought would happen this year is that the margin would increase. Certainly when I looked at the results of my own class their were no surprises: everyone seemed to have got a grade in line with my own assessment (which was NOT necessarily the same as the predicted grade sent of by the school).
However, it has become clear that in in a small percentage of cases, which is a large absolute number given how many people we are talking about, the algorithm used has done something stupid. It seems that there was no human in the loop at any stage with the ability to look at the outliers and say no. Ministers are not in control of the situation and, as someone put it earlier, OFQAL have proved even more useless than @ydoethur predicted they would be.
At this stage the only way out I can see is for ministers to bypass the exams completely and tell universities that there will be no penalty for accepting anyone they made an offer to even if the algorithm gave them grades which were too low.
Why not just do what the Scots have done and just overrule the algorithm altogether and reissue the grades? Its the least-bad solution that I can see.
Even if people get into university with a fudge they're still potentially going to have to say to employers years from now they got a bad grade worse than they should have got. That is unacceptable to me. Benefit of the doubt must go to the pupil above all else.
PS No doubt later this week someone will accuse me of always taking the government line when I back them on something else! Let there be no doubt I do not agree with this but will support a u-turn.
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.
Seems to me that we have got to the point where Johnson needs to step in, over-rule Williamson, and do what Scotland did and award centre assessed grades where they are higher than the algorithm.
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.
But how do you fix the other half of it. Universities have awarded places so how do you help the person who lost their university place, has been re-awarded it but the university no longer has accommodation..
Surely the problem of finding accommodation can be fixed? This is something where throwing a bit of money at the problem really does solve it, I'd have thought.
I'm loving the parody poster on here who is doing a pitch perfect impression of a self absorbed Tory. "My child's grades need to be raised otherwise their place in Oxbridge is threatened. But not the oiks, if you raise their grades that threatens the integrity of the system."
I'd be careful about a solution that involves accepting all teachers' predictions, as this will throw up a brand new set of inequities given that teachers, and their institutions, vary between the grimly realistic and the hopelessly optimistic.
The CEO of the Sixth Form Colleges' Association was making this point this morning. He was concerned that using just teachers' grades could throw up a whole new set of anomalies. As an alternative, he was arguing that Ofqual should sort out the bugs in the algorithm (e.g. the one that says subjects must have grade Us if there's been any Us in the past 3 years, and the small cohort bias), and run it again. He's saying this because he knows that sixth form colleges, given their size and nature, are in general much more accurate in their predictions than many schools with small sixth forms.
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.
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.
But how do you fix the other half of it. Universities have awarded places so how do you help the person who lost their university place, has been re-awarded it but the university no longer has accommodation..
I doubt that's a major problem. There's normally a lot of places going to clearing and to foreign students. Bumping up more domestic applicants up the list at this stage just means there'll be fewer spots available elsewhere - and considering the pandemic that's probably not a bad thing. The universities were warning they were facing a shortage of foreign students and the revenues they bring already.
And foreign students all need accommodation. Many domestic students still live at home.
Its the least-bad solution in a year when no exams have taken place. Just hurry up and do this already England.
Great they've now completely screwed up this year's A level results, you can't have one rule for one set of exams and another for the other set of exams.
Are you suggesting that we should screw up the GCSE results in order to maintain solidarity with the A Level results? Or achieve some sort of consistency?
Consistency is all important - problem is that by no doing this prior to the A level results it's now a bigger f*** up then it would have been on Thursday.
This Government really seems wants to play a game of "If you think that was a screw up, wait until you see how we do this..."
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.
On the NHS, I think the "Protect the NHS" slogan is wrong. It should be "Protect the Nation's Health", and the NHS should get back to work soon.
Congratulations you've just annoyed 100,000 students who should be doing what they want because their centre assessment grades met the university requirements but their actual grades didn't because the algorithms had been screwed by the A and A* awarded by small cohorts in private schools.
There’s also a much greater than zero chance that Ofqal’s statistical analyses were deeply flawed.
I honestly think it must be bust somehow. 40% of grades downgraded is just such a huge number.
It can't be the case that in a normal year, 40% of grades are less than what students were predicted. Surely?
Don't forget people sit 3 exams so it's highly likely people will drop a single grade or 2. How often do AAA predictions become ABB..
This is the statistic I would like to know. What proportion of pupils had a downgrade?
If 40% of A-levels are downgraded... and you treat the probabilities of a downgrade as independent (they probably aren't)... and assume people do 3 A levels... then the chances of any individual student *not* being downgraded are 21.6%.
So almost 80% of students drop a grade. And for a large proportion, that would affect a university offer presumably.
Seems to me that we have got to the point where Johnson needs to step in, over-rule Williamson, and do what Scotland did and award centre assessed grades where they are higher than the algorithm.
He would also need to sack him
Probably yes (mainly his fault or Ofqual`s?). Certainly politically he`d have to go.
As I understand it the law needs a change to do that and reports this morning confirm ofqual put out the process in a wide consultation and teachers unions and every interested party endorsed the system being used
I think the GCSE should have teacher assesments as Northern Ireland announced today but a fair appeal system should be agreed for the A levels in England and Wales
We weren’t told the truth. And even with the system as it was put forward, plenty of people who knew what they were talking about had serious reservations.
That said, I knew it would be a fiasco but I’m stunned at how bad it has been, far worse than I predicted, largely due to government lying, dithering and incompetence.
I'd be careful about a solution that involves accepting all teachers' predictions, as this will throw up a brand new set of inequities given that teachers, and their institutions, vary between the grimly realistic and the hopelessly optimistic.
The CEO of the Sixth Form Colleges' Association was making this point this morning. He was concerned that using just teachers' grades could throw up a whole new set of anomalies. As an alternative, he was arguing that Ofqual should sort out the bugs in the algorithm (e.g. the one that says subjects must have grade Us if there's been any Us in the past 3 years, and the small cohort bias), and run it again. He's saying this because he knows that sixth form colleges, given their size and nature, are in general much more accurate in their predictions than many schools with small sixth forms.
Too late.
Had the bugs been dealt with first time then fair enough, but its too little, too late to tinker with the edges now.
If you accept teachers grades then in general all the anomalies should be on the upside. That is better than anomalies on the downsides. It hurts anyone who may have outperformed the predictions but given the absence of exams I don't see any solution for that other than letting them take the exams later.
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. Also place numbers would be dictated by how many hospital slots in future years are available for the clinical training?
Its the least-bad solution in a year when no exams have taken place. Just hurry up and do this already England.
Great they've now completely screwed up this year's A level results, you can't have one rule for one set of exams and another for the other set of exams.
Are you suggesting that we should screw up the GCSE results in order to maintain solidarity with the A Level results? Or achieve some sort of consistency?
Consistency is all important - problem is that by no doing this prior to the A level results it's now a bigger f*** up then it would have been on Thursday.
This Government really seems wants to play a game of "If you think that was a screw up, wait until you see how we do this..."
Seems to me that we have got to the point where Johnson needs to step in, over-rule Williamson, and do what Scotland did and award centre assessed grades where they are higher than the algorithm.
He would also need to sack him
Probably yes (mainly his fault or Ofqual`s?). Certainly politically he`d have to go.
As I understand it the law needs a change to do that and reports this morning confirm ofqual put out the process in a wide consultation and teachers unions and every interested party endorsed the system being used
I think the GCSE should have teacher assesments as Northern Ireland announced today but a fair appeal system should be agreed for the A levels in England and Wales
We weren’t told the truth. And even with the system as it was put forward, plenty of people who knew what they were talking about had serious reservations.
That said, I knew it would be a fiasco but I’m stunned at how bad it has been, far worse than I predicted, largely due to government lying, dithering and incompetence.
It must bad bad then! Your words were, I believe, that that gov will "have its arse handed to it".
I`ve just read that NI is following Scotland on GCSEs. Surely Williamson has to back down? Grade inflation for a year is far preferable to this fiasco.
Comments
Eek twin b has been warned that classes may occur up to 9pm at night and between 10 to 4 on Saturdays and Sundays..
Nothing, or very little, to do with Covid-19.
Looking at it the lower your pay banding the more likely you are to go off sick.
In 2019/20 the sickness absence rate for Band 1 was 7.73% whereas for Band 9 it was 1.59%
https://digital.nhs.uk/data-and-information/publications/statistical/nhs-sickness-absence-rates/march-2020
Your posts seem simultaneously to argue that staff are lazily doing nothing and that it's somehow the fault of the NHS. There is pretty general caution among both staff and patients, but I very much doubt that the organisation of healthcare in Britain is a significant factor either way. Just as some people fetishise the NHS, others see it as the root of all problems, but a lot of issues are unrelated to the structure, and the psychoogical impact of COVID is one of them.
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.
On the NHS, I think the "Protect the NHS" slogan is wrong. It should be "Protect the Nation's Health", and the NHS should get back to work soon.
The rest - not so much as a sniff of democracy comes their way....
(By the way - thanks for calling them pupils. Calling pupils "students" is irritating.)
Handy little alibis, holidays!
Teachers may well have felt that several student had a near equal chance if getting an A. But they were still required to rank those students in order.
It’s little wonder a large number of unfair grade allocations have been made.
Add to that the evident statistical incompetence of Ofqal, which means that they have not even achieved the year on year equivalence that they gave priority to, then you have a complete car crash.
I think the only practical solution is to go with the (admittedly imperfect) teacher grades. The system simply cannot accommodate the vast number of appeals otherwise required.
And of course that would also require lifting the cap on university places,
The actual PM is not going to be a fan of resignations or sackings for incompetence for obvious reasons. Bluster your way through before u-turning and claiming world leading success is a blueprint that has worked wonders for him so far.
See https://twitter.com/branwenjeffreys/status/1295234429691015169 for confirmation as there is zero chance that is possible unless small cohorts impacted the grading.
I think the GCSE should have teacher assesments as Northern Ireland announced today but a fair appeal system should be agreed for the A levels in England and Wales
Thought not - Boris couldn't even time his holiday to provide suitable cover.
Alternatively if designing an algorithm from scratch then one that adjusts for any uncertainty and then rounds up at each stage of the calculation would have been more reasonable. It may have given better grades than the teacher's predictions though.
Any reasonable algorithm that had accounted for uncertainty and given the benefit of the doubt to the pupil (rounded up) would have seen a significant increase in pupils getting better grades overall. Or more importantly in the inverse a significant reduction in pupils getting bad grades. That doesn't seem to have happened.
Most have stayed, but the prospect of not seeing their families for a year has given some of them pause for thought.
Also Stormant has been recalled over the A results adding credence to law changes
He should announce today that the GCSE mess and A-Level mess are being fixed by going with teacher's predictions. Like the Scots did beforehand and which he tried to fix with the last minute mocks fudge but that doesn't seem to have worked.
There is a clear problem and there is a clear solution. JFDI now.
We don't have the worst per capita deaths, we do have an effective track and trace system and the virus is more under control now in this country than France, Spain, Sweden, Portugal, the USA and more.
Some GP surgeries are now open almost normally, but most are still shut and are refusing to see people. Ask any pharmacist what they currently think of their local GP surgery.
I have been saying on here for months that hosptials are empty and its only when a poster goes to hospital or has experience of the NHS that they report the same. There is almost a disbelief that hospitals are quite places.
The obsession of the NHS with not being able to do anything due to Covid is costing thousand of lives.
0.6% of beds that are occupied have people in them that have tested positive for Covid. Out of total available beds just 0.3% are occupied with Covid positive. The NHS should be nearly back to normal now but it seems there is little interest in doing this. Just 56% of beds are occupied.
EG if the teacher thinks that 3 pupils in their class are capable of an A and predicts each of them an A then I say they should have all gotten an A.
If historically randomly one of the three (lets say in this instance Pupil 2) may have fluffed their exam and gotten a B then under the algorithm Pupil 2 keeps their A but Pupil 3 gets a B. That is one correct grade and 2 incorrect grades, though the historical average of 2 As and a B is correct they've gone to the wrong people.
In that case the teacher's grades are more accurate on an individual basis as it gave the benefit of the doubt to the pupils so is getting the errors that are on the upside only. The algorithm has tried to keep the average right by doubling the errors, for every error on the upside there is one on the downside. And downside errors are far worse than upside ones.
https://twitter.com/jessphillips/status/1295264903335419904
That's my answer to all this. JFDI. There's a problem, it needs a solution. JFDI.
Williamson must go. The Tories have fucked up royally.
The issue is that they came up with a fudge to get the exams this year to match last years as an average. So as an average yes they would have passed your test, but the garbage is assuming you can accurately match individual pupils to the exact grades predicted that way.
Your test would have been hit as an average, but an average isn't good enough. It needs to match at least what everyone should have got individually not on average and it hasn't done that.
It is individuals that matter for their individual results, not averages. I always believe in individualism over collectivism, but grades is surely where individualism should be at the forefront. Saying that someone gets a worse grade than they should but that is OK because collectively someone else got a better one instead - that just adds salt to the wound.
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
40% of grades downgraded is just such a huge number.
It can't be the case that in a normal year, 40% of grades are less than what students were predicted. Surely?
That MPs are on holiday is not a good enough reason not to recall Westminster. Especially when they have remote voting and a vote on this should go through pretty unanimously if whipped from all parties.
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.
I’m surprised just how badly this has gone, in part because I don’t have the faith that many of you seem to that the results they would have got had the exams run are a perfect representation of the ability and understanding of the candidates. Grades in normal years have a margin of error of at least one up or down (except perhaps for the top few who will get straight A*s or now 9s at GCSE) and all that I thought would happen this year is that the margin would increase. Certainly when I looked at the results of my own class their were no surprises: everyone seemed to have got a grade in line with my own assessment (which was NOT necessarily the same as the predicted grade sent of by the school).
However, it has become clear that in in a small percentage of cases, which is a large absolute number given how many people we are talking about, the algorithm used has done something stupid. It seems that there was no human in the loop at any stage with the ability to look at the outliers and say no. Ministers are not in control of the situation and, as someone put it earlier, OFQAL have proved even more useless than @ydoethur predicted they would be.
At this stage the only way out I can see is for ministers to bypass the exams completely and tell universities that there will be no penalty for accepting anyone they made an offer to even if the algorithm gave them grades which were too low.
Experience and common sense would suggest, although must admit that I havent seen data for it, that students will tend to have much more unexpected downside risk than upside risk. A "B" student will sometimes get a D or E due to illness, stress, hangover, relationship breakup etc. By contrast it is very hard to see how they magic an A*.
A reasonable proportion of the lower grades will happen every year so the aggregate model will move the lowest "B"s into a C. In real life the lowest Bs would mostly be Bs still, and the Cs would come from across the board.
There will therefore be students who are odds on for a B, the correct prediction individually is B but an algorithm has to give them a C if it is required to balance to previous years.
Boris needs a reshuffle - and then a mighty kick up the arse to the new Cabinet. The six months since Covid hit have been largely woeful.
Do the same with the grades - give them the teachers' predicted grades with a note saying in previous years x% of pupils at this school/teacher/ got the predicted grade y% got a better grade and z% got a worse grade. There's probably loads of things wrong with that idea, but might be fairer than anything else.
Even if people get into university with a fudge they're still potentially going to have to say to employers years from now they got a bad grade worse than they should have got. That is unacceptable to me. Benefit of the doubt must go to the pupil above all else.
PS No doubt later this week someone will accuse me of always taking the government line when I back them on something else! Let there be no doubt I do not agree with this but will support a u-turn.
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.
https://twitter.com/keir_starmer/status/1295076611574784002?s=21
The CEO of the Sixth Form Colleges' Association was making this point this morning. He was concerned that using just teachers' grades could throw up a whole new set of anomalies. As an alternative, he was arguing that Ofqual should sort out the bugs in the algorithm (e.g. the one that says subjects must have grade Us if there's been any Us in the past 3 years, and the small cohort bias), and run it again. He's saying this because he knows that sixth form colleges, given their size and nature, are in general much more accurate in their predictions than many schools with small sixth forms.
And foreign students all need accommodation. Many domestic students still live at home.
https://twitter.com/HannaLiubakova/status/1295266199715094532
This Government really seems wants to play a game of "If you think that was a screw up, wait until you see how we do this..."
What proportion of pupils had a downgrade?
If 40% of A-levels are downgraded... and you treat the probabilities of a downgrade as independent (they probably aren't)... and assume people do 3 A levels... then the chances of any individual student *not* being downgraded are 21.6%.
So almost 80% of students drop a grade.
And for a large proportion, that would affect a university offer presumably.
That said, I knew it would be a fiasco but I’m stunned at how bad it has been, far worse than I predicted, largely due to government lying, dithering and incompetence.
Had the bugs been dealt with first time then fair enough, but its too little, too late to tinker with the edges now.
If you accept teachers grades then in general all the anomalies should be on the upside. That is better than anomalies on the downsides. It hurts anyone who may have outperformed the predictions but given the absence of exams I don't see any solution for that other than letting them take the exams later.
I`ve just read that NI is following Scotland on GCSEs. Surely Williamson has to back down? Grade inflation for a year is far preferable to this fiasco.