It was the first indication given by the Government that grammar schools might not offer the elite education their supporters claim, and prompted a triumphant article by Lord Hattersley in one newspaper last week.

But now ministers have muddied the waters again with a new set of data suggesting that secondary moderns do best of all.

The data was published in Hansard, in a series of written answers by education minister Baroness Blackstone to Conservative peer Baroness Blatch.

The first answer revealed that the top-performing quarter of comprehensive pupils all got five good GCSEs last year - compared to only 96 per cent of grammar-school pupils. The comparison reflects badly on grammars because in selective areas they cream off roughly a quarter of pupils.

But a follow-up answer reveals that secondary modern pupils outperform the bottom 75 per cent of comprehensive students. Around 33 per cent of secondary modern upils get five GCSEs at grades A*-C, compared to 27 per cent achieved by the bottom three-quarters in comprehensives.

Comparing point scores in both cases produced the same result. In other words, comprehensives might be better for the brightest, but secondary moderns are better for the less gifted.

A spokesman for the Conservatives said the data confirmed that the Government was wrong to try to abolish schools that were performing perfectly well.

But Professor David Jesson of York University, who researched in detail results from selective and non-selective schools, said the comparison was misleading.

“There are relatively few secondary modern schools and the name gives no indication of the school. Most are not in inner-city areas, they are in areas such as Kent and Lincolnshire and are likely to have an ability profile which is higher than that of many comprehensive areas,” he said.

The Department for Education and Employment, whose statistical analysis unit produced the figures, declined to elaborate on them. The debate is likely to continue.