More than a random effect
A striking example lies in the Qualifications and Curriculum Authority's annual reports on GCSE and A-level enquiries and appeals. About 1 per cent of all results are queried. At GCSE, some 22 per cent get an improved grade; at A-level, about 11 per cent. But these averages disguise big differences between boards. Last summer, 23 per cent of GCSE enquiries to AQA resulted in a grade change; only 17 per cent of appeals to OCR were successful. At A-level, AQA raised one in seven grades; OCR changed just one in 13. Clearly, the odds offered by the two boards are very different.
If this were a one-off discrepancy, we could put it down to abnormality or chance. But this pattern has been constant for at least five years so it cannot be random. There must be a systematic difference in the way the two boards re-mark. The data can't tell us what that difference is, but there are only two logical conclusions: either AQA's marking is much less reliable than OCR's, or OCR's examiners are more reluctant to change grades on appeal.
Last year, The TES carried an article entitled "OCR board deters A-level re-grades" about a leaked memo encouraging examiners not to change marks without "very strong grounds". Could this be the explanation we are looking for?
George Bethell, Director of Anglia Assessment.