The Government’s exam watchdog has ordered an immediate inquiry into teachers’ claims that results have been “fixed” to balance the books in a bid to quash accusations of falling standards. Schools have complained that good students were severely marked down, particularly in psychology, history and English coursework. The criticism has focused mainly on exam board OCR.

Graham Able, head of Dulwich College, London, said: “Dozens of schools have been affected. It seems the decision was made to be meaner with coursework to overcompensate for high AS grades last year.”

An English teacher at an independent school in Shropshire said: “Three students received straight As throughout their two years, only to then receive a U grade on the last module. The boundary for a U was set at 57 per cent of the total mark, which is ludicrous.”

In one state school, 17 out of 26 students received U for A-level psychology coursework. There were also very low grades on one particular paper.

A head of sixth form at a state school in Shropshire also believes grades awarded by AQA in A2 history coursework have been deflated. “It is the first time in my 17-year teaching career when none of my students gained an A grade,” she said.

Awarding committees set grades in each subject by looking at the difficulty of the paper, the spread of marks and the proportion of students gaining them. They also take into account the distribution of grades in previous exams. Grade boundaries are often adjusted year-on-year.

However, the process was particularly complicated this year because it was the first year of the new A-level so there were no direct comparisons to be made. Instead, all the boards took into account the grade spread which could be expected from the cohort’s previous GCSE results. Results from the final year of the old-style A-level and last year’s AS results were also taken into account.

When the initial A2 results were combined with the AS results from last year, the distribution of grades was strikingly higher than those achieved in the traditional A-level. As a result, the original A2 grade boundaries were shifted, making the standard “more demanding” so that overall grades were in line with 2001 A-level results.

A senior figure from one awarding body said: “In previous years you could predict with confidence because you had a benchmark. This year the whole thing came out higher. Once you add all the six modules up, they take on a life of their own.”

An OCR spokeswoman said that teachers might not have fully appreciated the harder demands of the A2, but insisted the exam board had followed the rules in the normal way.

The exam board has received just over 4,000 requests for re-marks, compared to 1,618 at this time last year. A spokesman said the rise could be explained by the introduction of the new modular A2. The percentage of A-level entries gaining grades A to C this year was 65 compared to 59 last year. A fifth of entries were graded A, up 2 percentage points on 2001.

* Exam board OCR awarded more top grades in AS and A-level qualifications than the other two English boards this year.

An analysis of A-level results in the main subjects shows the Cambridge board awarded more A to C grades in nine out of 14 subjects.

At AS level, OCR gave out more top grades in eight of the subjects. Edexcel awarded the highest proportion of grade As.

Edexcel was more likely to award A to C grades and A grades in French, German and Spanish A-level than the other boards. In science subjects, AQA awarded the highest grades.

There were substantial differences between the boards in some syllabuses. In ICT A-level for instance, 34 per cent of those taking the OCR paper gained the top three grades, with 4 per cent getting As. By contrast, Edexcel gave 51 per cent of entries A to C and 13 percent gained A grades.

Gaps between the boards of more than 10 percentage points occurred in A-level general studies, geography and RE and AS maths.

Many top private schools favour OCR syllabuses which explains the higher proportion of top grades.