Grammars are key to success for the deprived

Grammar schools are a sure-fire route to success for the poorest children - if they can break the middle-class stranglehold on places, research shows today.

But while selection did not stack the odds against poor children in pure academic terms, the picture was different when it came to the social class mix in grammar schools.

Professor Paul Gregg and Adele Atkinson, of the Leverhulme Centre for Market and Public Organisation, looked at the results of tests taken at age 11 with the same children's GCSE results from 2002.

They compared pupils who lived in parts of England with grammar schools and secondary moderns with others who lived in regions that were similar except for the fact they had comprehensives instead.

For the first time, the researchers were able to use government data showing whether every pupil was poor enough to qualify for free school meals or not.

Grammar school pupils did better than those with similar primary test scores who went on to comprehensives, the study found. There was little "substantial" difference in performance between selective and nonselective-local education authorities.

Those who failed the 11-plus scored "slightly worse" at GCSE in selective LEAs compared with those in the non-selective authorities.

Mrs Atkinson said: "The results suggest there are much stronger positives from grammar schools than there are negatives from secondary moderns."

But when they came to look at the parental wealth effect, the findings highlighted the social gulf in grammar schools. In the 19 English LEAs where selection was the norm, an average of 12 per cent of children were poor enough to qualify for free meals. But in grammar schools, the figure was two per cent.

The study confirmed earlier research which has consistently shown that a high concentration of poor children in a school depressed results. The difference was even greater in secondary moderns with high proportions of children from cash-strapped families. "This polarisation drives all the under-attainment in the nongrammar schools and about a third of the benefits to grammar school pupils," the researchers said.

If grammar school selection did not "segregate" children socially, their pupils would still be high achievers. But comprehensives would lose their disadvantage and their results would be closer.

The minority of children from poor families who made it to grammar school did "exceptionally well", bumping up their GCSE scores by eight points, the equivalent of rising from a B to an A in the same number of subjects - or getting one whole extra A*.

Professor Gregg and Mrs Atkinson speculated that the financial burden of preparing children for one or more versions of the 11-plus could be to blame.

Middle-class parents with the desire and the cash to ensure their children did well were taking more than their fair share of places, according to purely academic criteria.