top of page

CAGs, TAGs and mutant algorithms




In 2020 and 2021, formal 'sit-down' school exams in England were cancelled because of the Covid-19 pandemic. Students, though, still needed to be recognised for their learning over many years of schooling, and to be able to progress to their next stage of education, or employment.


How could students be awarded fair grades?


That's a question that, with hindsight, the government botched.


Many of the problems, though, were evident at the time, and far better answers were quite possible. From March 2020 until August 2021, I was a close observer of these events as they happened, and I published a series of blogs commenting 'in real time', as shown here. And afterwards, I wrote a book - Missing the Mark: Why so many school exam grades are wrong, and how to get results we can trust - published by Canbury Press in August 2022, Chapters 10, 11, 12 and 13 of which tell the full, sorry, story.


Let me note that a key feature of many of these blogs is not the blog itself, but the richness of the comments - my thanks to all those who made such important, and truly excellent, contributions.


 

 

Written evidence to the House of Commons Education Committee (submitted 4 April 2020, made public 4 May 2020)

 

 

 

 

Written evidence to the House of Commons Education Committee (submitted 18 June 2020, made public 1 July 2020)

 

Written evidence to the House of Commons Education Committee (submitted 26 June 2020, made public 7 July 2020)

 

 

Written evidence to the House of Commons Education Committee (submitted 3 July 2020, made public 15 July 2020)

 

 

Halfon is right: Ofqual has more to do (co-authored with Rob Cuthbert, HEPI, 14 July 2020)

 

 

 

Written evidence to the House of Commons Education Committee (submitted 28 July 2020, made public 8 September 2020)

 

 

Written evidence to the House of Commons Education Committee (submitted 12 August 2020, made public 8 September 2020)

 

CAGs rule OK (HEPI, 18 August 2020)

 

 

 

Written evidence to the House of Commons Education Committee (submitted 25 September 2020, made public 6 October 2020)

 

Why the rank order contingency plan is flawed (Schools Week, 4 November 2020)

 

 

 

 

 

 

An idea for managing this year’s unreasonable appeals process  (Sixth Form Colleges Association, 28 June 2021)

 

There were, of course, any number of blogs and journal articles authored by others, notably Rob Cuthbert's HEPI blog of 10 August 2020. Media coverage too was extensive; let me mention here just two particularly influential newspaper stories.


This piece by Liz Lightfoot, published by the Guardian on 20 June 2020, was one of the earliest reports of a particularly vicious injustice of the algorithm, the 'history trap', whereby a talented student would be denied a high grade simply because no student at that school had been awarded a high grade over the previous three years.


And on 8 August 2020 - that's the Saturday prior to the announcement of the A level grades on the following Thursday - the Guardian's front page main headline was "Nearly 40% of A level predictions to be downgraded in England". The accompanying piece told of Huy Duong, whose son was awaiting his A level results. Combining the skills of a mathematician and a masterful detective, Huy had put together some fragments of information provided by the exam regulator, Ofqual. As a result, he made an astonishing forecast: that the next Thursday's results would show that 39% of the grades that had been suggested by teachers (known as 'CAGs') would be over-ruled, and downgraded, by the algorithm.


This article created uproar across the country, and over the next few days, triggered a series of events, including some last minute changes to the appeals process.


On results day, Thursday 13 August, Huy's forecast was confirmed as being absolutely right.


The national uproar then became even louder, and a few days later, on Monday 17 August, the results of the algorithm were discarded, and the CAGs reinstated.


Comments


Recent Posts
Archive
bottom of page