Politics isn’t a battle between good and evil. It entails largely well-intentioned people making choices between imperfect options. Last week’s A-Level fiasco was a case in point. Gavin Williamson’s task throughout this pandemic has been near-impossible. Most Education Secretaries deal with angry unions, failing schools and controversial curriculum changes. He has had to juggle a global pandemic with providing England’s children with the education they deserve. Whatever he did about schools and exams, there would have been public outcry. He had the near-impossible task of choosing how to disadvantage the fewest pupils both now and in the future. Undoubtedly, he has made mistakes. But the situation before his u-turn was far from as bad as it was made out to be – and the situation since it is far worse. He should stay and get England’s schools reopened and clean up the mess from a u-turn he shouldn’t have been forced into.
Arguing for his resignation is easy. Across England last week, thousands of Year 13s were angry to see their grades downgraded from their predicted ones by Ofqual’s algorithm. It provided their A Level results in lieu of exams. What made this mathematical model egregious was that this year’s cohort saw their grades moderated according to previous years’. This flattened individual performances seen in exams and meant bright pupils in underperforming state schools were more likely to have grades marked down than average pupils in private schools.
Consequently, it was, in Timandra Karkness’ phrase, a “prejudice engine” – bright students at schools which don’t usually get exceptional results or send students to top universities were said to have been deprived of chances to show their individual brilliance in exams by Williamson. To many, it made a mockery of the government’s pledge to “level up” the life chances of those across the country.
Presiding over this, Williamson has made mistakes. Even after Scotland had the same problem the previous week, he failed to prepare pupils and parents for what was coming. Once results were announced, he resolutely refused to u-turn. He eventually he responded to outrage by saying mock exams would be accepted over results if they were higher. That then descended into farce when Ofqual both published and retracted their criteria for using mocks within a few hours. Finally, he bowed to the inevitable and copied the Scottish policy by replacing results with predicted grades. He looks a fool. Ahead of schools reopening in September, shouldn’t he be doing the decent thing and make way for someone more competent?
However, his resignation would be no panacea. In fact, those mistakes identified have been exaggerated and others have been ignored. Williamson’s original sin was months ago when he shut schools. The Swedish example shows open schools see the virus spread no more than elsewhere. If opening schools was impossible, then socially distanced exams should have been hosted. They have been with great success in Germany. Children should have been doing lessons and taking exams. Because they weren’t, state school pupils were disadvantaged twice over. Firstly, they were far less likely to be provided with lessons during lockdown than their privately-educated counterparts. Secondly, above-average pupils at schools with poorer results were unable to show their talents in exams. Whilst I show below that this was less problematic than commonly thought, this doesn’t detract that the biggest mistake was made a lot longer ago than last week.
But Williamson isn’t solely to blame for schools being shut. After all, he wanted schools reopened before the summer holidays. That was stopped by parents who didn’t want to send their children to school and teachers who didn’t want to teach. Williamson was surprised by the public’s unwillingness to have children in school even when they had a right to be there. The government had expected 20% of pupils to be in school during lockdown – the children of key workers. The reality was 2%. England’s teaching unions were also uninterested in teaching children, neither consenting in many cases to reopen schools or provide alternative education. Williamson therefore had to choose between reopening schools with few pupils or teachers or keep them shut and cancel exams. It isn’t a surprise he did so.
What has also been missed in last week’s rancour is that the results produced by the algorithm were actually quite good. Before the u-turn, this year’s results were already better than last year’s. More students received A*/As. 88% of students got into their first-choice university. More 18 year olds than ever before were set to go to university – and the number of those from disadvantaged backgrounds was at a record high. Had this been a normal year, Williamson would have been feted for overseeing an excellent set of results, rather than seeing his position under threat.
The outcry was from those who felt the system gave them worse results than they deserved. This is especially the case for those students whose results saw significant drops on their predicted grades. But from the best available evidence, predicted grades have never been good guides to exam outcomes- and give false hopes to many students. Though the process of moderation saw some students’ results significantly drop from those predicted, this reflected as much the historical inability of teachers to correctly predict grades as it the algorithm’s callousness. Between 2013 and 2015 around 80% of grades were predicted incorrectly;[ER1] this is disproportionately the case in state-schools and disadvantaged areas. The likelihood is that many of those who underperformed their predicted grades would have done so anyway. [ER2] The disappointment they felt at their computerised grades would normally be directed at over-optimistic teachers.
The failure of students usually to meet their grades is not only the fault of the algorithm, or the tendency of many teachers to over-predict, but the long-standing differences in quality between Britain’s state and private schools. Unfortunately, more state school pupils were penalised by the algorithm because [ER3] they were receiving a poorer general standard of education than those at private schools – a difference I have been able to experience myself. We are all familiar with the litany: 7% of UK pupils attend private schools, make up 18% of those doing A-levels, 34% of Oxbridge applications and 42% of eventual Oxbridge students. Of course, Oxbridge isn’t the sum of all British education, and there are as many excellent state schools out there as private ones. But it is a good marker for the advantages private education confers – and of which I am a lucky product.
Being educated privately is the sort of advantage baked into the algorithm that won’t be removed until England’s state schools are as successful as England’s private ones.[ER4] This is not the fault of Williamson, who has been in his role just over a year. Nor is it specifically the fault of the Conservatives. In fact, they deserve credit that since 2010, state school results have improved against private ones.
In August 2010, 4.8 million or two-thirds of children were learning in a ‘Good’ or ‘Outstanding’ school (as defined by Ofsted, which investigates all state, and some independent, schools). In 2018, that figure had grown to 6.7 million, or 86% of the school population. Rapid improvements have also been seen amongst those from disadvantaged backgrounds. In 2010, a child eligible for free school meals was half as likely to achieve 5 or more grade A*- Cs at GCSE (including English and Maths) than a child from a wealthier background. According to the Education Endowment Foundation that had been halved to 25% by 2019.
Moreover, the number of children from disadvantaged backgrounds applying to university rose from 15% in 2009 to 23% in 2018. So the idea the algorithm was a product of some Tory plot to help the privileged over the disadvantaged is nonsense. The story since 2010 is of a rapid improvement in prospects for those from the poorest backgrounds. Though inequalities exist, the gap is closing fast, and this year’s results – with more children from disadvantaged backgrounds going to university anyway, before the u-turn – was a reflection of an improving situation and a narrowing in educational inequality that happened under ten years of Conservative government.
As such, whilst privately educated pupils did better out of the algorithm than state-educated ones, this does not mean the algorithm actively sought to disadvantage pupils in the state-sector. They are historically more likely to under-perform over-optimistic predicted grades than their privately-educated counterparts.[ER5] This may seem callous, but to ignore this fact would be to ignore the (happily, narrowing) divide in our education system. For example, private schools have been more likely to provide something approaching usual levels of teaching during lockdown than state ones. Had exams gone ahead, this divergence in teaching would have been reflected in results – as it is every year state schools do worse on average than their private equivalents. Whilst it may be tragic that children in this country are still victims of a parental and postcode lottery, this shouldn’t detract from the accuracy of the algorithm in producing results in line with would have usually been the case, even if it was not what many wanted or thought they deserved.
Obviously, the argument against saying the collective outcome was fair is that the algorithm obscured individuals who lost out opportunities to prove their ability in exams. In doing so, they may have outperformed their schools’ average, or surprised compared to prior performance. Though the results before the u-turn were better than last year’s, I will concede that individuals who may have exceeded expectations lost out. But I maintain that the algorithm was still a better policy than the u-turn, even if schools should have stayed open and exams should have gone ahead. That’s because the unfortunate but predictable consequences of the government’s u-turn are becoming more obvious every day.
Suddenly reversing the grades of the 39% of those who saw their grades down-graded – the vast majority of whom, don’t forget, would have likely underperformed their predicted grades anyway – has naturally vastly increased the number getting A*/As and/or going to their first choice university. In Scotland, the five-year average for passes is between 75% to 79%. This year it was 89%. This has produced chaos in their university system – a chaos that Williamson’s u-turn, forced onto him by outraged parents and media hysteria, has brought south. Its consequences are numerous, and in the long-term will prove detrimental to many disadvantaged students’ life chances.
University admissions departments suddenly must find new places for students who have qualified following the u-turn, as well as dealing with those who both wish to swap from their second choices to their first and those who got in before the u-turn. Which, I shall never tire of mentioning, was more than last year. Though Worcester College was praised for letting in all offer-holders regardless of their results, Oxford colleges are now announcing that Williamson’s u-turn means some students will be deferred until next year. This is no surprise, as detailed below. Now 38% of students will have passed with an A or A* – far higher than usual. If we assume many would have been over-optimistic assessments from helpful teachers, this makes a mockery of the grading system: those are not necessarily results students would have got in an exam. How fair is that on previous years’ students, who were predicted As and A*s but lost out on their first choices because their results weren’t up to scratch? Rather than be let down by the system, the u-turn has largely benefitted this year’s cohort in spite of the pandemic.
Over the longer-term, those who lose most from the u-turn are those disadvantaged students argued to have lost most to Williamson’s original policy. Though the government has assuaged the upset of many this year who feel hard done by, they have, in the process, disadvantaged next year’s contingent of A-Level students. As Toby Young outlines in this week’s issue of The Spectator, usually good universities over-offer places in the knowledge that many students won’t make their predicted grades – this year, for example, Oxford made 3,900 offers for 3,287 places. Before the u-turn, roughly the same number would have missed their grades as in previous years. But now Oxford and other universities must find hundreds more places for successful applicants. This means accepting many more students than in an average year, and as they lack the space or staff to accommodate them all, that means many students deferred until next year. That’s fewer places in 2021 for next year’s A-Level students. That means a tougher competition for places without the free pass of being taken on predicted results.
If we accept private school applicants have an advantage at getting to good universities their state-educated counterparts lack, that means fewer A-Level students from more disadvantaged backgrounds attending good universities from next year’s cohort. They will have tougher competition for fewer places, so fewer children from disadvantaged backgrounds will be going to university overall. Since the numbers involved are so large, that will mean a long-term backlog for universities scrabbling to accommodate more students. As fewer applicants will now be going to universities lower down league tables as a consequence of being able to go to their first-choices, more universities will face potential financial crises this year. Their closure means fewer university places for future years. A u-turn demanded out of a mistaken idea of “fairness” will therefore have the longer-term effect of entrenching educational inequality. Williamson will not be to blame for that. It will be the fault of those who did not look at this year’s results in the round and forced on the Education Secretary a u-turn he should not have made.
Williamson deserves criticism. His biggest mistake was closing schools and avoiding exams. His next mistake was failing to persuade the public Ofqual’s original policy was the right one. An awful lot of rancour could have been prevented if the government simply got out the message that this year’s original results were better than last year’s. Yes, some pupils didn’t do as well as they thought they would. But that happens every year. Finally, he deserves criticism for u-turning in such a way as to cause chaos both now and in the future. But he is not to blame for a media that spun an outcome that improved on last year’s into a national crisis. Whilst he should have stuck to his guns, he deserves to stay in place to clear up the mess from a u-turn forced upon him, to sort out the crisis it has caused in universities and stare down the anti-teaching teaching unions to get our schools reopened. He is not a heartless moron, but a victim of this pandemic and the difficult choices it has thrown up, like the rest of us.