It has been a deeply strange year, with the calendar half-empty for most of us and many of the events we associate with summer cancelled. Few have had an emptier summer than school-leavers – the unlucky cohort of this year’s 17- and 18-year-olds. Not for them the revision, the sleepless nights, the cursed rote learning, the scrutiny of past papers and mark schemes. In what seemed like every schoolchild’s dream, exams were cancelled in late March.
They were replaced by a series of detailed teacher assessments of what grades their students would have obtained had A-levels gone ahead: “Centre Assessment Grades” (CAGs). They were meticulous and contained more than 40 data points, including measurement of past academic performance, mock exams and GCSE results. These were submitted by schools and colleges in mid-to-late June. In the interim, the government asked the English exams quango Ofqual (and its equivalents across the UK) to find a way to do two things. First, to compensate for potential teacher over- estimation (and occasional underestimation) of student attainment, and thereby prevent “grade inflation”. Second, to standardise outcomes with previous years, to retain, as the Prime Minister said, “the credibility” of the system and give confidence to universities and employers.
This was a difficult task: how to assess for exams that had not taken place? After much discussion, Ofqual devised a series of calculations (“algorithms”) that took into account a candidate’s CAGs, as well as data about their previous academic record and how their school had performed in the recent past. Even though these safeguards seemed – and probably are – worthy, they suffered from a fatal set of assumptions, both about the reliability of the data and about the predictive power of the statistics involved.
Rumours about what might happen with the CAGs circulated for weeks. In recent days, first in Scotland, then across the rest of the UK, the horror became clear. Hundreds of thousands of students had been downgraded from their teachers’ assessments. Many lost university places on the basis of “results” for exams that were never sat. Worse, much of the downgrading was based on the historically poor performance of particular schools, lessening the potential for individual excellence to be reflected. In contrast to state schools, many private schools, with smaller classes, came out well.
On the day the A-level results were published in England, I met Mithushan. He is the teenager you hear about in every politician’s speech about the value of education and hard-working families. The son of Sri Lankan immigrants, he came to the UK as a child. He scored a perfect set of GCSE grades at an improving comprehensive school in Kingston, south-west London. Nationally, he was one of only 250 students to do so. He had strong A-level predictions and won a place at Gonville & Caius College, Cambridge, to read medicine. His father, a former civil war refugee, works in a petrol station. His parents are deeply proud of him.
But the algorithm could not compute. The school’s past performance, on which it based its standardisation, did not tally with what Mithushan “should” achieve. It therefore downgraded his CAGs to one A* and three As. For Caius College, that wasn’t good enough to study medicine. “No one had ever gone to Cambridge from my school,” Mithushan told me. “My school hoped I’d be the first. I just wanted to make everyone who believed in me proud. Most of all, I wanted to make myself proud.” Before long, as I reported on individual stories such as Mithushan’s for BBC Newsnight, my inbox was filling up with messages from devastated students and parents. I had little of worth to say to them. We shared a common incredulity: that it should be these outstanding students, who had overcome so much, that were punished in the name of consistency, resulting in an outcome that went against the thrust of education policy for the past decade.
There are many detractors of both the exams system and its incarnation after Michael Gove’s reforms in the 2010s. But surely its one unalloyed virtue was that it could work for the bright but disadvantaged, who could use it as a springboard to escape their circumstances. It was a system built for outliers. And yet here we had a temporary replacement that was designed to suppress those outliers where they occurred.
Initially, politicians seemed impervious to what was happening. The Education Secretary Gavin Williamson, even as he announced the government’s U-turn on 17 August by stating that teacher estimates would be used for exam results in England, said: “We had the constant reassurance that the system, the algorithm was the right algorithm, to ensure a fair distribution of grades, right across England.”
This was a sentiment echoed by other politicians from other parties. But they and Williamson missed the point entirely. There was no right algorithm; no algorithm could have done the job that was being asked of it. Other European countries had no such problem. Some, such as Germany, even managed to hold their exams. Others moved to some form of CAG and indeed saw grades rise (as in Norway and France). In the year of the pandemic, many polities seemed neither concerned about grade inflation nor motivated to design a system that pretended that those exams had gone ahead.
Yet without exception that is precisely what all four nations of the UK did, with a blind faith in the power of data and technology. All parties who govern in the UK – the Conservatives, Labour, the SNP, the DUP, Sinn Féin and even the Liberal Democrats (who front the education ministry in Wales) – are culpable. It seemed to occur to no one, anywhere, what the obvious deterministic frailties of this system might be; that even if a set of algorithms could ever predict with certainty how an individual might perform then, for reasons of politics and, yes, morality, they probably shouldn’t.
It is hard not to dwell on the fact that at the heart of this government is a man who mandates the reading of a book entitled Superforecasting for political advisers, who lauds the scientific method and its predictive capacities, and who has forged a remarkable electoral record through the employment of data to understand how voters think. We cannot know the extent of Dominic Cummings’ involvement in this sorry episode, and it may be that he was not part of it at all. But his approach encapsulates a method of governing that was on full display throughout.
For teachers, far from being an aberration, it was the apotheosis of how the Department for Education (DfE) and politicians had understood schooling for a decade – a decade of English education policy in which Cummings had been a central player as adviser to Michael Gove. One headteacher described the events of recent days as the “ideological end point of the strategic culture war within English state education”.
He added: “The DfE, lacking in direct classroom experience, believes that there is a truth in data, which can provide direction to the messy business of educating young people. Teachers, on the other hand, tend to be more values-driven. Thus data and values are locked into a long-term battle for the soul of education. It is possible to have both and it is the job of the Secretary of State to achieve this educational alchemy on a system-wide basis.” The exams fiasco demonstrates the weaknesses of this form of technocracy – of “dataocracy” – and of how politics and its practitioners are beholden to it.
There is an assumption that algorithms are inherently neutral, an apolitical mechanism for politicians to make difficult decisions and ensure fair outcomes. But that is always untrue. Today, too many of our politicians seem unable to argue from first principles, and to identify and arbitrate political problems in the terms in which they should be grounded: justice, desert, equality and freedom. The philosopher Michael Freeden argued that an ideology was simply a lens through which a person sees and orders the political world around them. A politician, viewing this through a Conservative lens, would have identified these problems immediately: that the principle of desert and respect for the individual would demand an alternative. It is perhaps a consequence of a polity where trust in politics has sunk so low that, as the former Supreme Court judge Jonathan Sumption has argued, the natural result is to reduce everything to matters of legalism and, I would contend, to a political mathematics, too.
Perhaps, for all of the charges against this government of populism (of which there is some) and deep ideology (of which there is less), the most surprising thing about the present cabinet is how deeply unpolitical many of its members are – Williamson included. We now have a higher education system in disarray. It was obvious that the U-turn would come, as it did earlier in Scotland, because the alternative, which was to withstand mounting political pressure and devise an appeals system to solve potentially hundreds of thousands of cases within a fortnight, was harder. In reverting to the CAGs, the government has thus converted a profound political problem into a perhaps unanswerable logistical one for the higher education sector.
Many universities do not have capacity for all the students who have now met their offers. They are faced with incorporating this bumper crop of students at a time when they are legally obliged to enact significant social-distancing measures across campus. These universities can console themselves that they will not face a funding shortfall. But there are potentially existential problems for those institutions at the bottom of the pecking order. Suddenly, they may find themselves significantly under capacity. One university vice-chancellor told me he could think of at least half a dozen institutions that could face bankruptcy without significant financial support from the government. Some are even chartering flights to India and China to ensure international students can take up their places.
But this is the stuff of relatively normal politics – of budgets and balance sheets. These are problems that can be solved, should the political will exist. I am troubled far more by what this crisis has revealed about how our politicians now seem to think; about how little, in the name of systems and faceless bureaucracies, the fates of individuals can matter; and about how hollow grand rhetoric about social mobility can be.
And I cannot forget Mithushan’s face, or his story. I was filming at his school when the U-turn was announced. He texted the news that Cambridge would honour his place. His head of sixth form cried. But they have more travails to come. Those in Year 12 have an even less enviable educational inheritance. Despite students having had six months off school, the government insists they will take exams in ten months’ time, when universities will be full to bursting. Exporting disadvantage from one year to the next is hardly satisfactory.
I cannot escape how Mithushan looked and acted. He wasn’t angry, or furious, as I would have been at his age. (I, too, was the first in my family to go to university, and attended a comprehensive school. Had time and happenstance been different, our places could have easily been reversed.) He was completely defeated. Everything he said and did told me that this was exactly what he was expecting: that things were never going to work out; that of course they wouldn’t, for someone like him. That it was what he expected of the system: to mark him down and shut him out. It was chilling.
I am left with a reminder of how monstrous the state can be: how impersonal, how alienating. I reflect on how often, in this time of Covid-19, that has been forgotten: how the state can be responsible for systems that steal the dignity of those it purports to serve or protect, all in the name of the greater good. In reporting on this story, I have come to understand far better what happened with the Windrush scandal, for in spirit, if not in substance, this episode had many of its hallmarks: the same need to “standardise”, the same impersonal regard for circumstance, the same rigidity, the same fingerprints of alienation. I worry that those fingerprints won’t leave Mithushan and those caught up in these events. I worry, too, how often the state acts like this, and that, in an era in which it exerts more power than ever, often no one will even notice.
Lewis Goodall is policy editor at BBC Newsnight
This article appears in the 19 Aug 2020 issue of the New Statesman, Failed