Tag Archives: a level results

Embed from Getty Images

In this post: Exam Results and Gradings Ianto Stevens explained how and why examiners moderate exam results to try to smooth out inconsistencies from year to year. Some of this was to compensate for variations in the difficulty of exam papers. The overall aim was to give the same balance of grades as in previous years.

For example, suppose on a particular A Level physics paper, students across the country get markedly higher marks than in previous years. Examiners will assume that the students are of much the same ability as previous cohorts and that the variations are due to the questions asked and the marking guidelines. It is almost impossible to write a series of exam papers that produce the same range of results each time.

The problem this year was that this legitimate moderation process was applied not to actual written exam papers but to teacher’s predictions.

I spent many years teaching and organising A level and BTEC courses in a large FE college. Each year the whole of my Easter break was devoted to assessing and ranking coursework projects for some 60 A level Computing students. And, of course, I had to provide predicted grades which I based on their AS levels, mock exams and the state of their coursework at the time when the predictions had to be submitted.

I do understand why there was a reluctance to simply award the students with the grades predicted by their teachers. If that had happened then the overall grades would have been significantly higher than in previous years. The consequence would have been that a higher number of students would have met the conditions set by universities, so there was a danger that courses might have been oversubscribed. Universities always offer more places than they can fill, on the basis that not all will qualify.

So, to avoid a glut of qualified students, the raw predicted grades were treated as though they were actual marks and were moderated to bring them in line with other years. What is more, the moderation was applied at the level of an individual school or college rather than across the whole exam entry cohort – a granular application of a holistic method.

I will argue that it was not necessary to moderate the grades at all, but first let’s take a look at the procedure that was adopted. The actual predicted grades were ignored and the rankings used instead. These were mapped on to the spread of grades achieved in that subject, in that school or college, over several previous years. So if in previous years, on average, 5% had been awarded a U (ie fail), then the bottom 5% in the ranked list this year would be given a U, and so on across all the grades.