Research suggests results are being skewed by over-generous examiners
EXAM standards may be threatened every time a syllabus is changed, according to research done by a leading exam board.
Examiners could over-compensate pupils for the changes - resulting in more pupils apparently doing better, according to Alastair Pollitt, the research director at University of Cambridge Local Exams Syndicate.
For example, changes in next year's English tests for 14-year-olds could lead to a fall in the standard needed to pass them, says Mr Pollitt.
However, he is confident that the UK exam boards and the Qualifications and Curriculum Authority have taken all steps possible to tackle the problem in respect of A-levels and key stage tests.
Mr Pollitt's arguments are particularly controversial because last month he denied Sunday newspaper claims, attributed to him, that A-levels and GCSEs are getting easier.
In a terse letter to the Sunday Telegraph he insisted that the study of maths quoted by the newspaper had no implications for examination standards in other subjects.
However, in another paper - presented to the International Association for Educational Assessment's conference in Barbados earlier this year - he cites the same study and then extends the argument to other subjects and exams in general.
According to the maths study examiners had quite properly made allowances for the unfamiliarity of the new A-level syllabus in its first examining year, 1986.
But they had then "forgotten" this one-off dispensation and set the 1987 standard in relation to the lowered 1986 one. Standards in subsequent years remained consistent, but at the lower level.
"If it happened in the case of A-level maths we could expect to see it wherever syllabuses change," he told the conference.
"There is a message here for governments: if you change the national syllabus frequently, you can ensure that your students' standard attainment will be seen to rise steadily, year after year!" writes Mr Pollitt.
"In fact, of course, it would just be that the exam standard was going steadily down." He adds: "If 'stepwise standards' are general, then any system of monitoring standards based on short-term comparisons is bound to lead to a decline of exam standards."
Ultimately, the best way of measuring standards over time may be to measure the difficulty of questions used in different exam papers, he suggests.
Mr Pollitt said the Barbados paper was written for an international audience as a warning of the potential pitfalls surrounding standards issues.
Boards are now required to compare current year exam standards to several previous years, and not just the last year - which should counter the "step-down" effect resulting from change