Well, you read it here first. It’s a couple of weeks now since retired headteacher and Voice of the North author Bernard Trafford first wrote about predictions emerging from statistical experts (among which he doesn’t count himself) that this year’s public exam results would be dodgy. Now that fear appears to be gaining credence, as this year’s exam candidates face Edu-Chaos and injustice.
It started with the recent results (always out earlier than the rest) from the International Baccalaureate (IB), the increasingly popular A Level substitute that requires 18-year-olds to study a broad range of subjects at differing levels. As many feared, those results are down.
The exam body, the IBO, used teachers’ gradings, candidates’ individual and schools’ collective past performance, and its own algorithms based on all of the above, to come up with the grades it awarded. They universally disappointed.
As challenges, complaints and outrage spread, politicians, teachers, parents and candidates are starting to wonder anxiously what might happen to August’s domestic GCSE and A Level results – because the UK’s exam watchdog, Ofqual, has adopted an almost identical statistical method for awarding results in this summer without written exams.
The problem lies in the grades put forward for each candidate by teachers (whose fault it definitely isn’t!), misleadingly dubbed “predicted grades” by media.
Yes, in normal years (remember those?) the government’s burdensome accountability system, including Ofsted, makes teachers estimate pupils’ target grades: exam boards demand predicted grades as part of quality assurance; and schools send universities predicted grades to aid their selection process.
Teachers making those predictions assume a candidate will put in the necessary work, keep healthy and have a following wind behind them on the day of the exam. In other words, they’re “all being well” predictions.
…vs. “estimated” grades
This year, by contrast, teachers have done something quite different. In April they were required to assess in more detail, indeed mathematically, each pupil’s level of attainment in April, with evidence. Next, each school produced a rank order for all pupils studying that subject: not an easy task, since the teaching groups in that subject will have been taught by different teachers. Moreover, Ofqual provided minimal guidance.
Nonetheless, these estimated grades still tend to be optimistic: no teacher can randomly suppose that a candidate will get ill, have an accident or lose a relative in May/June.
Thus, as my statistician friend Dennis Sherwood wrote recently, an element of optimism is inevitable even in these meticulously estimated rates. Since Ofqual cannot countenance the “grade inflation” that must ensue, it will mark everyone down.
The Commons Education Select Committee, chaired by Tory MP Robert Halfon, is concerned that disadvantaged pupils – particularly those from BAME or deprived backgrounds or with special needs – will be particularly hard hit. It’s probably right.
As it happens, though, even those in the most privileged schools risk being marked down unfairly: if their year-group is significantly stronger than last year’s in the same school (and it happens), Ofqual won’t allow it to outperform last year’s cohort.
Tough, you might say. We couldn’t hold written exams under lockdown: so what can we fall back on, if not statistics?
Me, I prefer human beings and the knowledge one human being (the teacher) has of another (the pupil). But, as we all know, nowadays algorithms rule: and algorithms will triumph this year, to the detriment of countless exam candidates.
No right of appeal
The Select Committee was also exercised about the fact that bias and discrimination will prevent wronged candidates from mounting any successful appeal. My understanding is that no one can complain or appeal. It’s possible that individual schools will be able to challenge procedures: my long experience suggests that even the most forceful or influential school will get nowhere.
Ofqual brightly reminds dissatisfied candidates that they can (re-)sit a full suite of exams in October or November. Quite who will prepare them, when they haven’t had a lesson since March, or even how schools maintaining post-Covid safety levels will have space or manpower to accommodate and supervise exams remains a mystery.
We’re on the brink of seeing great wrongs done to exam candidates this year: moreover, notwithstanding the legitimate concerns of Robert Halfon’s Select Committee, it won’t be only the disadvantaged who are, well, done down.
Too many ambitious candidates risk seeing their ambition of becoming a barrister turn to training as a barista. If that’s the cost of maintaining the integrity of an already creaking exam system, it’s not an acceptable one.