The Garden of England has historically been a peaceable sort of place. Even the Battle of Britain was several hundred feet above ground. But these are turbulent times for secondary schools in Kent and the neighbouring Medway towns which once belonged to the county. With 39 grammar schools between them, this corner of south-east England is the largest expanse of selective education remaining in Britain and it is under attack.

Charles Clarke, the Education Secretary, has said that he is concerned about the effects of academic selection at 11. An influential group of Labour MPs, including Frank Dobson, the former health secretary, have started a campaign to get the abolition of grammar schools into the next manifesto. Parents in Northern Ireland are getting written notices explaining that their selective system is to be abolished. In England the education select committee is set to publish a report on admissions that may well make uncomfortable reading for a government that once promised “no more selection” but has ensured that the 164 remaining grammars are in good health. In fact, the number of pupils grammar schools educate has increased by more than a third in the past decade, the equivalent of 11 new schools.

The arguments about selective education, and Kent in particular, acquired fresh vigour with the appointment of Charles Clarke in 2002. No sooner had he arrived at the Department for Education and Skills, than he was questioning the academic performance of a county which is, by most standards, rather affluent. He also agreed to meet pro-comprehensive campaigners, a symbolic step for a government which has done its best to bury the issue.

At first glance, Kent’s performance seems perfectly respectable. Although the county has some of the least successful schools in the country, it also has some of the best, and can by no means be dismissed as a dud authority.

Overall, its pupils seem to show reasonable progress. At key stage 2 the percentage achieving level 4 is below the national average (although this could be affected by the fact that so many primaries concentrate hard on passing the 11-plus instead of the national tests). Yet by GCSE, 54 per cent of Kent pupils achieved five or more A* to C GCSEs in 2002, compared with 53 per cent for statistical neighbours and just under 50 per cent nationally.

According to Graham Badman, Kent’s director of education, its performance this year places it “in the top quartile of all local authorities”. He adds: “Since the 1990s, Kent has consistently outperformed the majority of authorities in terms of the number of young people leaving our schools with no qualifications against a background of more looked-after children (8 per cent of the national total) than any other authority. Yes, we have a tail of underachievement but, sadly, so does every other authority.”

Mr Badman, however, who has previously expressed reservations about academic selection at 11, does not pretend that he is satisfied. He is attempting to improve the authority’s performance by grouping schools together into 22 “clusters”, each one funded with pound;30,000.

A report from the Office for Standards in Education, commissioned by Mr Clarke, suggests that all is not well. The inspectors’ findings, published in January 2003, gave important ammunition to opponents of grammars. It found that fewer Kent schools were judged “very good” in inspections than the national average (13.1 per cent against 16.5 per cent).

Its schools were substantially more likely to require special measures or to have serious weaknesses. The county was ranked fifth out of seven comparable authorities for the proportion of 18-year-olds going on to higher education. The proportion of schools getting poor results was far higher than nationally.

At the time, there were 14 schools in Kent requiring special measures: 10 primaries and four secondaries. This represents 2.1 per cent of Kent primaries and 3.8 per cent of secondaries. The comparative national figures were 1.1 per cent and 1.4 per cent.

Groups such as Case, the Campaign for State Education, were delighted, and placed the report alongside another helpful study, this time from Professor David Jesson at York university. Comparing the educational outcomes of Kent and Medway with what would be expected if they were comprehensive, he concluded that they “show substantial evidence of low performance in GCSE outcomes” (“A review of structure and performance of secondary education in Kent and Medway”, 2002).

Nationally, however, where grammars tend to be spread more thinly, the selection evidence is more confusing. In results released at the start of this year comprehensives showed better value-added ratings between key stage 3 and 4 than their grammar schools rivals. Yet in the KS3 ratings, published a few months earlier, grammars far outstripped comprehensives in the improvements between the ages of 11 and 14.

Research findings by Ian and Sandie Shagen (“The impact of specialist and faith schools on Performance”, 2002 NFER) and the DfES (“Pupil progress in secondary education by school type in England”, 2001 June 17, 2002) also suggest that selective areas outperform their comprehensive neighbours in the progress made from KS2 to 3 but, importantly, that the reverse is true at GCSE.

In 2003, a report by the National Audit Office found that, again, selective schools made a significant improvement to KS3 results, but made no appreciable difference at GCSE. It suggested that grammars had little effect on the performance of other, non-selective schools in the same locality - to the consternation of Case. Yet the Green Paper, Schools Building on Success (2001), found that some of the country’s lowest-attaining schools are secondary moderns in selective areas.

Much-quoted figures from the then Department for Education and Employment, have suggested that the top 25 per cent of the ability range does better in comprehensives than in grammars. But these figures have also been much criticised as statistically flawed.

There is some evidence that grammars do particularly well for the lower-ability range, the “borderline” students who just made it through the 11-plus. But the same studies also suggest that the very brightest pupils fare better at comprehensives.

So the arguments based on academic results are confused and confusing. But they are not the only statistics around. One fact that remains clear is this: nationally, only 2.7 per cent of grammar school pupils are entitled to free school meals. These tend to be the children of the unemployed. The average for secondary schools in England and Wales is 17.1 per cent. The contrast is even greater in Kent, where entitlement to free meals is lower still, running at closer to 2 per cent. Grammars, then, remain substantially more middle-class than their neighbours, something that does not tally with the government drive to boost the performance of children and students from disadvantaged backgrounds. Nor does it seem that the selection mechanism, the 11-plus exam, is producing a fair or scientific distribution of grammar-school places.

This social skewing also suggests there could be damaging effects for secondary modern schools if, as a consequence, they were to end up with a disproportionately low number of middle-class families repeatedly identified as an important element in school success. The NAO conclusion that there is no observable sign that this has happened deserves further investigation.

But with ministers campaigning for more working-class undergraduates, it is the social statistics rather than the GCSE results that are likely to form the ammunition of the future.