Ask most people what sunk the RMS Titanic and they will probably tell you it was an iceberg. It’s true that on 14 April 1912 the luxury liner hit an ice field in the north Atlantic. But the cause of the disaster was the result of a catalogue of errors.

As the public inquiries found at the time, the Titanic had a flawed design in her flooding compartments (rendering the famous “unsinkable ship” very sinkable). There were not enough lifeboats and, above all, there was human error at nearly every stage in the process of getting the vessel to her maiden voyage.

Ultimately, it was the experienced captain Smith who took the fateful decision to maintain full speed ahead to New York, knowing very well the dangers of floating icebergs in the cold spring ocean. The combined impact of a flawed design and bad decision-making brought about a catastrophic maritime disaster that people still talk about over a century later. 

In the context of the summer exams fiasco, the Titanic makes a great allegorical tale.

Tes FE podcast: The exams fiasco, with UCU's Jo Grady

Opinion: Be warned, Mr Williamson – students won't forget this

GCSE resits: how can we prepare students for September?

The GCSE and A-level results fiasco

Education secretary Gavin Williamson is the new head of the White Star Line. Ofqual chairman Roger Taylor is captain Smith. Indeed, it was in April this year when the secretary of state took the fateful decision to issue England’s independent regulator with the design plans for this year’s grading model. I wrote in April that Williamson had done something quite wrong-headed and unprecedented. I warned then that what he was proposing could sound the “death knell” for public confidence in Ofqual.

While the crisis leaves big question marks hanging over the ministerial futures of Williamson and Nick Gibb, the regulator is hardly immune. Taylor was on the bridge of the ship when the algorithm struck. It is still yet to be seen whether he decides to do what captain Smith did – perhaps the honourable thing – and go down with his metaphorical ship.

In June, the Commons Education Select Committee, chaired by Robert Halfon, held a number of hearings where it became obvious that the type of statistical modelling put forward by Ofqual was going to result in far from plain sailing conditions on results days. The chief regulator, Sally Collier, was repeatedly pressed to provide assurances about the impact of tutor calculation and standardisation on disadvantaged groups. The appeals system was placed under detailed scrutiny.

I appeared in front of the same select committee by video link. Collier, who appeared semi-detached with some of her answers, deferred several questions to Ofqual staff subordinates. It was obvious to participants in the session that this tactic wasn’t going down well with Halfon. Following the committee’s deliberations, MPs made a series of practical recommendations. Several of them have still not been implemented, including the call on ministers to properly resource awarding organisations to cope with the net additional costs of implementing this summer’s exceptional arrangements.

Above all, there was absolutely no proper consultation by ministers at the start of the crisis. Williamson and Number 10 had options: they could have made a virtue of bringing education leaders, the regulator and awarding bodies together to look at what alternative arrangements were possible. They simply didn’t.

Alternative arrangements

Some European countries, like Germany, went ahead with exams, albeit socially distanced. In the four nations of the UK, the art of the possible was rendered impossible, by a set of snap political decisions made without the involvement of the wider education sector, learners or parents.

The UK went into full lockdown on 23 March, with restrictions beginning to ease by 10 May. Based on what had happened in Spain and Italy, the government knew the likely trajectory of the pandemic and how quickly the spread would be eased and the curve flattened. 

Epidemiologists have found that young people are largely unaffected by the disease. Similarly, educational settings were discovered not to be places where a super spread of coronavirus could take hold. A major scientific study of schools in France, for example, concluded in late April: “Children did not spread the infection to other students or to teachers or other staff at the schools.”

Public Health England has recently confirmed these findings in relation to a study of educational settings in this country. Two under-15s in Britain have died of Covid 19-related infections. It’s no wonder the chief medical officer, Professor Chris Witty, is confident of the need for all schools to reopen in September.

All this means that Williamson could have ordered a delay to the examinations by about eight weeks and young people would still have been able to pick up, by late summer, a set of results achieved under the usual examination conditions. After all, secondary schools and colleges were technically open right up until the end of July. With an appropriate Covid 19-secure plan, the resources of the education sector could have been marshalled to achieve such a goal. In retrospect, perhaps, it will go down as an irony that massive city convention centres were requisitioned to provide Nightingale NHS hospitals at vast taxpayer expense. In the end, even at the height of the coronavirus curve, they were hardly ever used. These cavernous facilities, on the outskirts of most major cities, could have been used to host appropriately socially distanced GCSE and A-level exams.

Other options, including secure online exams, were not even placed on the table because ministers seemed determined to pursue a more macho course of action. They decided to order the cancellation of the examinations and leave Ofqual to worry about the technical details. Even then, ministers were very specific about the parameters that the independent regulator had to work within.

Two critical government edicts, set out in formal written directions from Williamson, were cast in stone: Ofqual was ordered to accept a system of teacher-predicted grades, but grade inflation was to be avoided by the application of a standardisation model in line with school and college performance in previous years.

A fundamentally flawed approach

The entire education sector, including ministers, knew that combining both approaches was fundamentally flawed. As far back as 2016, Dr Gill Wyness, of University College London, had carried out an empirical study of A-level predicted grades. She found 75 per cent of grades on Ucas applications had been overpredicted. Bright working-class kids had particularly suffered with grades marked down by tutors compared with their more affluent peers.

So, it was obvious at the very beginning of the pandemic that relying on centre-assessed grades (CAGs) was going to result in significant grade inflation and inequality across the board. Of course, Ofqual had the answer: it would apply an algorithm based on the mean performance in each subject, by centre, over the previous three years. The problem with the Ofqual algorithm, like all forms of artificial intelligence, is that is dreadful at dealing with outliers. You only have to think about the daily struggle with predictive text on a smartphone or computer to understand the limitation of algorithms. They’re great when text is adjusted correctly, utterly frustrating when it is not.

The infuriated victims of the Ofqual algorithm were faultless young people, some of whom had been given a U (unclassified) in their A levels for an exam they were never allowed to sit. Students in large centres, like FE colleges, were awarded grades way out of kilter with what centres had put forward to awarding organisations. The larger the cohort, the more egregiously the algorithm appeared to behave. Meanwhile, candidates entered from independent fee-paying schools seemed to be getting an easy ride. It led to mass outrage on social media and demonstrations outside the Department for Education in Westminster. 

No doubt a massive row will rumble on about who is to blame and who should be held to account. The debate still has some way to go. One thing, though, is certain: for all the imperfections of academic exams, they are, on the whole, the fairest way of assessing a national cohort of students. For sure, we need to emerge from the crisis with a proper inquiry about the resilience of terminal assessments and about no coursework being allowed in academic studies.

We need to recognise that tutors have a role to play in ongoing structured assessment, even if unconscious bias and grade inflation is the inevitable consequence of pure CAGs. Above all, we should never forget that the great thing about the English exams system, including one of the reasons why it is copied and envied around the world, is that it only really works when direct government intervention is kept out of the way.

Like the decision to cancel this summer’s exams has shown, politicians going back decades have a terrible record at passing the litmus test of public confidence in qualifications. After all, Titanic was an avoidable disaster, and, so it seems, was this summer’s exams fiasco. 

Tom Bewick is the chief executive of the Federation of Awarding Bodies. This blog was written in a personal capacity.