A-Level results: Government scores a ‘U’ for public relations

The signs leading up to A-Level results day weren’t looking good.

Crisis comms 101 is to take control of the narrative, writes Tom Try
Crisis comms 101 is to take control of the narrative, writes Tom Try

On 7 August, an analysis of the algorithm and data used by England's exam regulator, Ofqual, to distribute grades after the cancellation of exams due to the pandemic found that 39 per cent of the grades given by teachers were likely to be adjusted down.

Fast-forward 10 days and the front pages exclaimed: “Government forced into humiliating exams U-turn”.

What happened is an expert lesson in how not to do crisis comms.

It all began with an algorithm.

In the absence of actual exams, Ofqual devised an algorithm to distribute grades based on students' predicted grades to ensure that national results were "broadly similar” to previous years.

Sounds great, but the problem is that students consistently underperform.

Their dog/cat died, they split up with their girlfriend/boyfriend, they didn’t read the question – in fact, 40 per cent of students don’t make their predicted grades.

Deciding who would underperform is where Ofqual’s method unravelled.

It decided that a school’s performances over the previous three years should take precedence over a teacher’s prediction of a student's grade.

Simply put, as a result, a bright student from an underperforming school was far more likely to have their results downgraded.

The facts were clear: the algorithm was biased, and 730,000 A-level results were just about to be announced based on that algorithm.

It was an impending disaster.

Crisis comms 101 tells us to take control of the narrative – so that’s what Number 10 decided to do.

With barely 36 hours to the publication of the results, they announced a “triple lock” to allow students to retake or appeal.

But rather than fix the underlying bias, it put the onus on the students to complain.

The fundamental issue remained, and the narrative was very much out of control.

Perception is reality and in the media, doubly so – you need to listen to what’s being said.

Most years we’re treated to uplifting stories of students overcoming hardship to secure a place at Oxbridge – and I’m sure there are plenty of those stories untold – but this was ‘the year the algorithm ruined my life’. Dreams dashed, and all because the computer said no.

In an interview two days after the results were published, Gavin Williamson said: “This is it. No U-turn, no change.”

But the very next day young protesters marched on Number 10 shouting “f**k the algorithm”.



Students, parents, teachers – even some Tory MPs – were up in arms.

Backed into a corner, only 10 days after the methodology was published, Number 10 did exactly what it said it wouldn’t do.

It about-turned and decided the teacher-assessed grades would be accepted after all.

Gavin Williamson said he’s “incredibly sorry for the distress” caused – a reputation in tatters and his job on the line.

What does this teach PR professionals?

The mistake made by the comms team at Number 10 is they didn’t listen. The writing was on the wall from day one.

Sticking to your own agenda is a fundamental error in crisis comms – always.

By the time they realised, it was too late.

There’s a wider matter at play too. If a team of human moderators had diligently graded each entry and come up with a similar outcome, would the story have been different?

Perhaps the true issue is we’re just not ready to let the algorithm take control – not yet, at least.

Tom Fry is head of data analytics at Resonance


Have you registered with us yet?

Register now to enjoy more articles and free email bulletins

Register
Already registered?
Sign in