Every year, at about this time, Ofqual publishes a wealth of statistics relating to reviews of marking and grade changes following the summer GCSE, AS- and A-level exams in England.

Many of this year’s headline numbers are similar to those of recent years, but there are some welcome new features, too.

Firstly, there is now just a single data source, rather than the two that had been used hitherto. The numbers presented are now more consistent, but comparisons with data published in the past need to be treated carefully.

Secondly, information relating to short-course GCSEs is no longer included.

Thirdly, and perhaps most notably, two new tables have been added, providing further valuable information:

• Table 1, showing the numbers and percentages of component marks that were subject to a review of marking, and the numbers and percentages of marks changed.
• Table 24, showing the numbers and percentages of students who challenged a grade, and the numbers and percentages for whom one or more grades were changed.

Reviews and grade changes in 2025: key insights

The headline numbers for this year, and the years since the return to pre-Covid standards, are shown below.

Overall, then, the percentage of grades challenged has fallen somewhat from 5.1 per cent last year to 4.6 per cent, but those fewer challenges seem to have been more “successful”: the percentage of challenges resulting in a grade change has risen from 22.5 per cent to 24.1 per cent.

That said, the overall percentage of grades changed has remained as 1.1 per cent, which might beguile the unwary into inferring that the remaining 98.9 per cent were, therefore, right.

Perhaps so; but only if every one of the 95.4 per cent of the awarded grades that were not challenged did not contain an error, lurking, undiscovered. You may judge for yourself the likelihood of that.

New information on marking errors

Under the current rules for challenges, a grade can be changed only if a review of marking discovers a marking error, which, on correction, gives a new total mark within a grade different from that originally awarded.

The number of marking errors is, therefore, important, for surely the hallmark of a trustworthy exam system is that this number is small.

Ofqual is to be congratulated on improving the visibility of this number in its newly introduced table, the key features of which, for the summer 2025 exams, are shown as follows:

As can be seen, for this summer’s GCSE exams, there were almost 15 million component marks, of which some 404,845 - 2.7 per cent of the total - were subject to a review of marking.

This led to the discovery and correction of over 150,000 marking errors - this being about 1 per cent of the total number of marks given, and 37.3 per cent of the marks reviewed.

This last number - 37.3 per cent - is not shown in Ofqual’s Table 1, nor are the corresponding numbers for AS and A level (44.3 per cent) and the combined total (38.8 per cent).

I find these last numbers the most important. If nearly 40 per cent of the reviews of marking discover and correct a marking error, how many marking errors are there in the 97.1 per cent of marks that were not reviewed?

Let me repeat my congratulations to Ofqual for its improved disclosure this year.

That said, there is yet more that can, and in my opinion should, be disclosed - so perhaps that might happen next year.

Dennis Sherwood is the author of Missing the Mark: why so many school exam grades are wrong, and how to get results we can trust

You can now get the UK’s most-trusted source of education news in a mobile app. Get Tes magazine on iOS and on Android