Sixth-formers who were disappointed with their English literature GCSE result may believe that they would have scored higher marks if their school had chosen different set texts - but statistical analysis suggests this would have made virtually no difference.

John F Bell of the University of Cambridge Local Examinations Syndicate tested the fairness of the GCSE English literature examination by analysing the grades that 764 candidates from 394 exam centres achieved in the higher-tier paper offered by the Midland Examining Group in 1995.

The MEG syllabus included five prose texts (After the First Death, The Mayor of Casterbridge, Animal Farm, Debbie Go Home and Of Mice and Men), four works of drama (The Crucible, An Inspector Calls, Macbeth and Journey’s End), as well as two selections of poetry. But candidates only had to produce evidence of having studied four texts (prose, poetry and drama, and one work of Shakespeare) - and were examined in three.

It was subsequently claimed that The Mayor of Casterbridge and Macbeth had proved to be harder than the other texts. But John Bell demonstrates - in a paper published in the latest issue of the Oxford Review of Education (Volume 23 Number 4) - that the marking scheme allowed for such disadvantages.

Although question choice did have some statistically significant effects, they were small. “For the majority of candidates the difference was less than 1 per cent, plus or minus,” he says. “There was no evidence that candidates were seriously disadvantaged by their choice of texts or questions.”

Correspondence: John F Bell, Research and Evaluation Division, University of Cambridge Local Examinations Syndicate, 1 Hills Road, Cambridge.