Teachers – grade inflation – algorithms: cats know better!


Live by statistical modelling – die by statistical modelling! That’s my interpretation of the A-Level ‘scandal’. I wasn’t going to touch this with a bargepole, but the whole mess, shorn of the thicket of wailings and complaints in ‘Our MSM’, is yet another example for their anti-government crusade – a crusade on which they embark every day without doing due diligence, without doing background research.

They don’t even check out what their colleagues wrote, never mind what can be found in other papers or indeed blogs. Instead, readers are inundated with accusations that pupils from ‘poorer’ backgrounds were disproportionately downgraded by this exam exercise:

“The exam watchdog has been accused of creating an A-level grading system that punished poorer students at state schools while their wealthier, privately educated peers had their results upgraded. The proportion of grades awarded to pupils at independent schools increased more than twice as much as those handed to students at those at state schools, official figures showed.” (paywalled link)

Ooh – posh pupils did better than poor ones, how perfectly unfair! Here are some numbers, reported in The Times:

“The overall A-level results, across England, Northern Ireland and Wales, show record highs for A* and A grades. Grades for private school pupils were up by 4.7 per cent compared with last year. Sixth-form college grades were up only 0.3 per cent. Schools in the poorest areas also had more teacher assessments downgraded than those in the wealthiest. Sixth-form colleges have formally complained to ministers.” (link, paywalled

Oh dear! We are given a rebuttal though – see this snippet, published in the DT:

“The exam regulator has accused teachers of submitting “implausibly high” predicted grades amid a growing fall-out over A-level results day. On Thursday, official data revealed that close to 40 per cent of results were downgraded from teachers’ predictions (see graphic below), prompting students to complain that they had been let down by the system.” (paywalled link)

No, mustn’t criticise those teachers – must criticise ‘posh schools’! But now look at this amazing announcement, published in The Times:

“Record numbers of disadvantaged teenagers have won places at university this year despite disappointing A-level results for some. More than 210,000 have been accepted — a record number for results day, even though the number of 18-year-olds has fallen and grades have been volatile.” (link, paywalled)

So why the wailings about ‘poorer pupils being disadvantaged’? Why the wailings about unfairness and disappointments and ‘appeals’ when we read that actually, A* and A grades have risen? And – they’re going to Uni anyway, aren’t they! What is missing in these articles and reports? Questions which ought to be asked about teachers and how they grade their students!

Let’s not ask how come 40% of teachers, especially in those ‘poorer’ areas and ‘poorer’ schools, have had their assessments downgraded by this Ofqual ‘algorithm‘ because they had been ‘implausibly high’. Let’s not even ask if this implausible upgrading happened in previous years. After all, Universities and colleges need students to keep their business model running and who cares if a huge proportion of those young people end up in debt, with worthless qualifications! Who’d notice! 

Let’s also not ask how diligent all those teachers were when instructing their pupils during the CV-19 school closures. How good were their online video teaching material and methods? Moreover, shouldn’t there have been complaints from those poorer state schools about not having been given enough ‘resources’, like laptops and broadband connections for themselves and all their pupils?

There surely must have been some, or did the Teachers’ unions not care? Were they too intent on keeping schools closed until next year because that created headlines about bad BoJo trying to kill those precious teachers by exposing them to CV-19? I seem to remember those headlines … Of course, Labour had to wade in as well, in a Starmerish way:

“Sir Keir Starmer, the Labour leader, called for an urgent review of the system before next week’s GCSE results, calling it fatally flawed.” (link, paywalled

And this is what it’s all about: “The System”, created by government! It’s not about teachers and their utopian high assessments, it’s about appealing against that algorithm. Thanks to the negligence of ‘Our MSM’, not even Starmer seems to know how this algorithm came about. More on that below!

Obviously, one mustn’t voice one’s suspicion that grade inflation might have been due to teachers’ ‘generosity’, to engineer an outcome sending more students from poorer parts of the country to Uni, regardless of actual academic qualifications. As far as ‘Our MSM’ ae concerned, the whole mess is solely the fault of the Education Secretary:

“Gavin Williamson […] came under pressure to follow Scotland, where grades created by an algorithm were withdrawn and teachers’ predicted results will take their place.” (paywalled link)

It’s as if using an ‘algorithm’ is now a heinous crime. It’s as if Williamson and his henchpeople had spent their time creating something to impose on all those wonderful teachers – surely our next ‘heroes’ even though they’ve done eff all for their pupils during the Lockdown.

However, it turns out that they themselves were involved in creating this algorithm, as were their Trade Unions, as were headmasters and even pupils. I was pointed to a blog entry (here) where this was mentioned. They provided a link to the Ofqual Consultation paper where this is documented. You can find that paper here. This is a government publication. It’s not a state secret. You don’t need to use the Freedom of Information Act to read it – and it makes interesting reading!

For example, Ofqual conducted a consultation “that ran between 15 April and 29 April 2020 and which received 12,623 completed responses.”. They set out the areas to be addressed:

“We explained the aims of our proposed approach to standardising grades and the principles which underpin it, and made proposals to support these. For example, we set out our proposals to address key questions such as: how we should use statistical evidence to identify and adjust overly generous or harsh centre assessment grades; whether evidence of changes in a school or college’s past performance should inform our interpretation of this evidence; whether any issues of bias can be addressed through statistical adjustment.” (link)

Recall – this consultation took place in April, so teachers knew what was being proposed. And still they inflated their grades! Above all, Ofqual received over 12,000 replies on these points, from local authorities, schools, teachers, Unis … Interestingly, there were 78 from Teacher’s ‘representative group and union’. This wasn’t sprung suddenly on poor unsuspecting teachers. They were involved, they were consulted, and they all knew that an algorithm based on statistical models was going to be used. The data for the model was actually provided by the schools and teachers themselves.

I suggest humbly that this algorithm was more realistic that that other algorithm, the statistical model on which Lockdown was based. I ask in a most humbly fashion why it is ok to demand this Ofqual model be scrapped while that other one is still being used, false and corrupted PHE data notwithstanding, questions about the validity of CV-19 tests notwithstanding. I ask this especially since Johnson is giving with one hand – allowing further easing of restrictions – while taking away with the other, imposing fines up to £3,200 for ‘repeatedly not muzzling up’ in public (paywalled link).  

I leave you with a quite remarkable attempt at hauling another Tory minister into this A-Level mess. He’s been out of the headlines for a long time, but behind very closed doors and totally en coulisse he is apparently being touted as Johnson replacement. Dedicating a whole article today, on his attempts to tighten up the A-level grade inflation when he was education minister (from 2010 to 2014) is quite a feat.

The Times’ headline gives the game away: “A-level results: Gove’s reforms fail to prevent rise in top grades” (link, paywalled). Never mind the war he had to fight against the massed power of civil serpents and teachers’ unions. Never mind that some time has obviously passed between his time as education secretary and today. Never mind that the changes he instituted – scrapping course work for sat exams – had a result which provides one reason for the rather successful resistance of that ‘education blob’:

“One consequence of scrapping course work was that boys out-performed girls at the highest level. In contrast, under this year’s assessed grades, girls have done particularly well.” (link, paywalled)

One other point is missing here. One of his SpAds at that time was a certain Mr D. Cummings. You might have heard of him. I am amazed that The Times, the foremost belligerent in the Anti-Cummings-War, hasn’t seen fit to mention him! But at least they’ve now thrown the first little dart at Gove, so that’s ok then.

I wonder if it’ll dawn on ‘Our MSM’, on Labour, on teachers’ Trade Unions and the rest of the wailers and moaners who want to scupper this year’s A-Level result because it was ‘done by algorithm’ that we can make the same demand for scuppering Lockdown in its entirety, muzzle wearing included. That was, after all, also based on algorithms and statistical models.

‘Live by statistical modelling – die by statistical modelling’. No exception, else the cry about ‘following the science’ becomes even more tarnished.





Print Friendly, PDF & Email