It’s probably one of the things that secondary school leaders most fear: a bolt-from-the-blue collapse in their GCSE results.

While Ofqual’s “comparable outcomes” system pegs GCSE results at a national level to previous years’ performance, big swings - both negative and positive - can and do happen at school level.

“Centre-level variability” is the jargon used by Ofqual to refer to year-on-year changes, and the exams regulator has today published research looking at what causes it.

Here’s how stability and volatility affects different types of school:

1. Schools with high numbers of grade C/D candidates

Ofqual said that “how centre ability profile affects centre variability is not entirely clear”. But it added that “centres with very high or low ability profiles are more likely to experience lower level of variability in outcomes than centres with high proportions of grade C or grade D candidates”.

2. Schools with changes in pupil numbers

The research found that “centres with a change in the number of students between years are more likely to experience variability in outcomes”.

3. Schools where there was a swing last year

Whether a centre had stable or volatile results in one year is a predictor of their stability or variability in the following year. Ofqual said that schools that experience positive volatility in one year are likely to experience negative volatility in the next: “This is probably because such centres have a high proportion of candidates who are clustered around the grade C/D borderline (rather than, for example, clustering around grade B) and so more year-on-year variability is inevitable even if a small change in the entry size and/or ability profile.”

4. Disadvantaged schools

According to Ofqual, “some commentators have suggested that the comparable outcomes approach to awarding, in successfully managing unwarranted grade inflation, might be having a differential effect on those centres operating in a more challenging context”.

However, Ofqual found “that measures of socio-economic status have little or no bearing on centre variability”.

“This indicates that the comparable outcomes approach to awarding does not have a systematic negative impact on centres with higher proportions of low socio-economic status candidates,” the report adds.

5. Schools with changes in pupil ability

Perhaps unsurprisingly, Ofqual found that “the large majority of observed variability” in school-level outcomes “can be attributed to changes in the ability profiles and other characteristics of successive years of candidates within centres”.

“Only a small proportion of the observed centre variability may be associated with the indeterminacy of the examinations system,” it concludes.