Exam results 2023: how to make sense of 'grade deflation' as A grades fall

Students across England are receiving lower grades than they might have done in 2022. The percentage of A or A* grades given for A-levels has fallen from 35.9% to 26.5% in 2023. Pupils’ grades are closer to those from before the pandemic, when 25.2% of results were at A or A*.

In 2021, 88.2% of all grades were a C or above; in 2022, this figure was 82.1%. In 2023, it’s down to 75.4% – slightly lower than 2019, when 75.5% of grades were at C or above.

These lower outcomes are the result of the government’s action to reverse the grade inflation of 2020 and 2021 – a consequence of assessment changes during the pandemic.

Pupils who would have sat their exams in 2020 and 2021 were instead awarded grades based on teacher assessments. These were very much higher than in previous years, and appear be inflated above what would have been achieved had they sat regular examinations.

Put simply, grade inflation means awarding students higher grades than they deserve – where their just deserts might be determined on the basis of their achievements prior to the final examinations and their recent efforts on the course.

But it’s tricky to pin down exactly what each student deserves, especially when their learning is disrupted by something like a global pandemic. On the other hand, then, deflation appears to be giving pupils less than they deserve – which is surely undesirable.

Bringing results down

As school assessment returned to normal after the pandemic, the government, through its exams agency Ofqual, resolved to bring results in England back to where they were before. In 2022 grades were adjusted to a 2019-2021 midpoint, and this year back to 2019 standards. This means that the higher grades will have been more difficult to achieve.

Devolved governments in Wales and Northern Ireland are bringing grades back to pre-pandemic levels more gradually.

To understand how this is being done, we need to look at how grades are awarded. A-level and GCSE grading practice falls somewhere between two standard ways of awarding grades: criterion-referenced and norm-referenced.

A criterion-referenced system means that pupils are given credit for what they can demonstrate they have learned or understood on the syllabus. So, for example, getting 18 our of 20 on a test would automatically be an A grade, no matter what anyone else sitting the test scores. Driving tests are a good example of criterion-referenced examinations.

In contrast, norm-referenced grading (at its simplest) assigns grades according to where a candidate’s results sit when compared to the rest of their cohort. The top 10% may get an A, the next 20% a B and so on. It would mean that every year, the same proportion of pupils would get a top grade, and the same proportion would fail.

The reality of the current A-level and GCSE system is that it is neither of these, but a mix of the two.

Setting grades

Grade boundaries are finally set only after pupils have sat their examinations and most of the work has been marked. The grade boundaries are informed by the national curriculum, the examination boards’ specifications – and the lead examiners’ detailed subject knowledge.

They will know that a typical grade B geography student, for example, will be able to explain particular concepts with a certain degree of competence. This is the criterion-referenced part of the system. Across the whole geography syllabus – or any syllabus – there are many criteria, so a kind of best fit averaging approach has to be applied by the examiners.

Then, all of the results nationally are considered and possibly adjusted, so they match the proportion of grades awarded at each level for a predetermined standard – in this case, the 2019 results. This is the norm-referenced part, and how Ofqual could be sure, ahead of the publication of results, that they would be roughly in line with 2019 figures.

Ofqual also stated that they have put “protection” in place, so that students will get the grades they would have had before the pandemic even if their performance is a little weaker. In other words, the grade boundary adjustment makes up for any knowledge missing because of the pandemic.

But it remains to be seen which young people will be most affected by the reduction in grades. For a long time there have been significant achievement gaps between disadvantaged children and their more affluent peers. The likelihood is that these gaps will widen in 2023, as disadvantaged children have suffered more absences from school since the pandemic. Such differences won’t show until the full release of data in the autumn.

What is clear is that the 2023 results – in line with 2019’s grade distributions – will certainly leave more pupils disappointed than in the previous two years when results were inflated above normal expectations.

Chris Rolph does not work for, consult, own shares in or receive funding from any company or organisation that would benefit from this article, and has disclosed no relevant affiliations beyond their academic appointment.