From School Grades to Student Grades: The Case for Reforming Exam Grading

I write this having just heard the news that Ofsted’s single-word grades for school inspections are to be scrapped. No longer will we see those banners proudly proclaiming “Outstanding”; no longer will teachers fear the opprobrium of “inadequate”.

Ofsted’s “landmark reform”

The opening paragraphs of the DfE press release announcing this “landmark school reform” read:

“Single headline grades for schools will be scrapped with immediate effect to boost school standards and increase transparency for parents, the government has announced today.

Reductive single headline grades fail to provide a fair and accurate assessment of overall school performance across a range of areas and are supported by a minority of parents and teachers.”

This clearly states, in relation to school inspections, that “headline grades fail to provide a fair and accurate assessment”.

Might these words also be true in relation to another key aspect of education, the grades awarded to students in exams?

How “fair and accurate” are school exam grades?

Let’s look at the evidence – evidence being the operative word, for here is a statement, given in evidence at a hearing of the Commons Education Select Committee on 2 September 2020, by Dr Michelle Meadows, at that time Ofqual’s Executive Director for Strategy, Risk and Research:

“There is a benchmark that is used in assessment evidence that any assessment should be accurate for 90% of students plus or minus one grade. That is a standard benchmark. On average, the subjects were doing much better than that. For A-level we were looking at 98%; for GCSE we were looking at 96%, so we did take some solace from that.” (Q997)

I understand that to mean that 98% of A level grades are “accurate plus or minus one grade”, implying that 2% are two or more grades adrift; for GCSE, the corresponding numbers are 96% and 4%.

2% and 4% seem like small numbers. 

But a few weeks ago, 816,948 A level, and 5,677,941 GCSE, grades were awarded in England. So about 16,000 A level grades, and 225,000 GCSE grades, as awarded, were wrong by at least two grades – which certainly was important to those 240,000-odd students. Especially since these wrong grades result from legitimate differences in academic opinion, with all marking done in full compliance with the mark schemes. They are therefore not associated with any “marking errors”, and so a “review of marking” will confirm the originally-awarded grades, even though they are wrong by at least two grades.

A somewhat similar, but much simpler, statement was made at the same meeting by Dame Glenys Stacey, Ofqual’s then Chief Regulator, who referred back to the words of Dr Meadows:

“It is interesting how much faith we put in examination and the grade that comes out of that. We know from research, as I think Michelle mentioned, that we have faith in them, but they are reliable to one grade either way. We have great expectations of assessment in this country.” (Q1059)

Great expectations indeed. But, to my mind, not unreasonably so, for students’ destinies depend on the grades awarded. So how can grades which are, at best, “reliable to one grade either way” be reliable enough?

Unfairness at the 3:4 grade boundary

This unreliability has some very real consequences, most notably at the 3/4 grade boundaries for GCSE English and Maths.

If, at best, a certificate showing “GCSE English, Grade 3” is only “reliable to one grade either way”, then this can only mean “Perhaps the grade the candidate truly deserves is grade 3. Or maybe grade 2. Or even grade 4. No one knows.”

Add to that the implications of the statement made by Dr Meadows, which tells us that about 4% of those ‘awarded’ grade 3 actually merited a grade at least two grades different – perhaps a 5, or a 1. 

To make that real. This year, 22.4% of the cohort of 782,022 students in England were awarded grade 3 – that’s rather more than 175,000 students. 4% of those merited a grade at least two grades higher, or two grades lower, so a reasonable assumption is 2% in each direction. That implies that about 3,500 students merited at least grade 5. Yet their certificates show grade 3, and it is on that grade 3 that they are judged. And forced to re-sit. Remember too that these grade errors happen even if there are no “marking errors”, so even if all those 3,500 students were to challenge the grade on their certificate, a “review of marking” will confirm that two-grades-wrong grade 3.

If that isn’t totally unfair, please tell me what is.

It seems to me that, in England (for which there is explicit evidence, as summarised here), GCSE, AS and A level grades are both inaccurate and unfair.

Here’s to a future press release!

So I look forward to a (not too far distant!) future DfE press release which will read:

“Single symbol grades for students will be scrapped with immediate effect to boost confidence and increase transparency, the government has announced today.

Reductive single symbol grades fail to provide a fair and accurate assessment of overall student performance, despite their being supported by a number of vested interests.”

Written by Dennis Sherwood is an independent management consultant, and author of ‘Missing the Mark – Why so many school exam grades are wrong, and how to get results we can trust’ (Canbury Press, 2022)