Research into the pattern of variability in the outcomes of school and college GCSE results.
Ofqual is today (22 April 2016) publishing research into the pattern of variability in the outcomes of school and college GCSE results as one way of understanding the extent of volatility in the qualifications system.
Overall, in years when specifications and overall cohorts are stable, one might expect the majority of centres with entries in successive years to have very similar outcomes.
Most centres display little year-on-year variation. However, some centres display large year-on-year variation and some commentators have expressed concerns that the comparable outcomes approach to awarding, in managing ‘grade inflation’, might be having a differential effect on some centres.
Examples often given are of those operating in a more challenging context, for example, those with a significant percentage of students from low socio-economic status backgrounds or with speakers of English as an additional language.
Factors which appear to be associated with centre variability are:
centre type: selective and independent centres have less variability in grades A* - C than other centre types. This is most likely because of the distribution of grades within these centres tends to be comfortably within grades A* - C rather than straddling or centering on the C/D grade boundary;
entry stability: centres with stable year-on-year entry numbers are likely to show less variability than those with changes in entry size between years.
Factors which do not seem to have any relationship with centre variability are:
- the proportion of students entitled to free school meals within the centre;
- the deprivation index of the centre;
- the proportion of students with English as an additional language in the centre.
We intend to conduct further, more sensitive, analysis when student level data for prior attainment, free school meal eligibility and the deprivation index become available to better understand whether these systematic factors are associated with year-on-year variability.
Published: 22 April 2016