William Stewart examines the reports of a shocking performance gap between able pupils at private and state schools.
Comprehensive failure" and "Brightest are failed by state schools" screamed last week's headlines.
The research that inspired them was equally sensational, on first glance at least.
Professor David Jesson, of York university, uncovered a seemingly shocking gap in the outcomes between the state and independent sectors for the most gifted pupils in England. But other academics have suggested the picture may be less clear-cut. And after their concerns were put to the author by The TES, he conceded that other factors could indeed be at work.
Professor Jesson used pupils' points scores in key stage 2 national tests taken in 2000 to identify the top 5 per cent "very able" pupils in the country.
There were 37,000 of them, the vast majority, 30,000, in the state sector with another 7,000 in independent schools. He then tracked the cohort's progress through to the age of 16, and looked at the proportion gaining five A or A* grades in this year's GCSEs.
This time the numbers in independent schools were higher still, at 7,500, but in state schools they had plummeted by a third to 20,000.
A-level results showed that trend accelerating, with 7,600 independent pupils gaining three A grades, compared to 13,000 from state schools. By the time of Oxbridge entry the sectors had virtually levelled out, with 3,500 state pupils and 2,600 from independent schools.
It was, Professor Jesson said, evidence not only of a state-independent divide but also of such a divide involving pupils who were of similar ability, and suggested a "severe talent drain".
The Sutton Trust leapt on the research as proof of what it had been saying all along, that an educational apartheid exists between state and private schools.
Sir Peter Lampl, the trust's chairman, said: "If anything the findings understate the extent of the apartheid, because they are based on identifying the most able on achievement at 11, when bright state-school students are considerably behind their independently- educated peers."
The right-wing press portrayed the research as an indictment of the Government's record in education. What was slightly surprising was that it did not further emphasise another aspect of the research which, it could quite easily be argued, makes the case for the return of selection.
The national test data supplied by the Government allowed Professor Jesson to pinpoint which school each of the 30,000 very able state pupils attended, as identified by key stage 2 test results.
A breakdown demonstrated a strong variation in outcomes for those children depending on how many pupils of similar ability were taught with them.
In English secondary schools with between one and five very able pupils sitting their GCSEs in 2005, they scored an average of four A or A* grades.
In schools with from six to 10 very able pupils, they scored an average of five AA* grades, rising to six for schools with very able cohorts of between 11-20, seven for schools with 21-40 very able pupils and an average of eight AA* grades for the schools that had more than 40 very able pupils in a cohort.
So a straightforward case, then - state schools with high concentrations of able pupils (that is, grammar schools) do best by them. Well, up to a point. As Professor Jesson notes, schools with low numbers of very able pupils also have high numbers of free school meals, and vice versa. So it may be high levels of disadvantage rather than a school's performance that make the difference.
The suggestion that independent schools do better with very able pupils than state schools is also more complicated than it first appears. To begin with, the Oxbridge entry figures Professor Jesson used to build up his overall picture are unlikely to reflect state-school performance accurately .
As he himself points out, a Sutton Trust study in 2003 showed that consistently higher proportions of identically-qualified candidates from independent schools were admitted to the 13 most prestigious universities than were their state-school counterparts.
So differences between the sectors at this stage may have more to do with university admission procedures than how well state schools are doing, and indeed there is no guarantee that very able pupils from state schools will want to attend Oxbridge.
Moreover, Professor Jesson argues that a closer analysis of the figures reveals that the disparity within the state school system is not always so great.
He looked at what happened to the upper quartile of the very able pupils in each school. For example in a school with four pupils in the national top 5 per cent, how did the very best pupil perform?
Looking at this top 25 per cent of the best, schools with between one and five very able pupils scored an average of seven AA* GCSE grades. That only rose slightly to an average of nine AA* GCSE passes for all schools with more than 20 very able pupils in a cohort.
For Professor Jesson, an associate director of the Specialist Schools and Academies Trust, this is a cause for optimism. It demonstrates to state schools that "you can do the best for your pupils" and gives them something to aim for.
Other academics differ, and suggest that the upper-quartile figures merely show that the most able pupils will do well whatever system they happen to be educated in.
But, they add, apparent differences between state and independent school outcomes for the most able 5 per cent need not be bad news for state schools, because Professor Jesson is not necessarily comparing like with like.
Professor Peter Tymms, director of Durham university's curriculum, evaluation and management centre, said: "I would expect independent schools to have a slight edge over state schools, blow for blow. But the size of the effect (shown in Dr Jesson's research) is more than I would expect."
He points out that the work looks only at pupils who achieve above a certain points score and fails to take account of any variation in ability within the top 5 per cent. The group will contain the extraordinarily able alongside those with more prosaic talents.
It is, he says, perfectly possible that independent schools have a disproportionately large number of these brightest of the brightest. This is particularly likely as the key stage 2 tests were taken in the primary phase, when no state schools select but many independent schools do so.
Professor Tymms also argues that the same test scores at key stage 2 will not necessarily reflect the same level of ability for children in the two sectors.
The pressure on state schools to perform well in the tests and maximise their league table positions means they will do everything to squeeze the best possible performance out of pupils. But he says that while many independent schools participate in the national tests because parents are keen for them to do so, they do not have to take them quite so seriously.
They have other concerns, such as public-school entrance exams, and are unlikely to spend as much time training pupils to perform well in them.
So pupils in independent schools are likely to have a higher natural ability than state-school counterparts with the same test scores.
Lastly, as does the National Union of Teachers, he points out that Dr Jesson's research does not take account of the impact that pupils' home backgrounds can have on their performance in school.
So for a variety of reasons the differences in outcomes for the most able may be very little to do with the actual performance of English schools.
The criticism has been taken on the chin by Professor Jesson. He accepts that background may be a factor and that the possibility of independent schools having a disproportionately high number of the best of his very able top 5 per cent is "quite likely to be true".
"It is very clear indeed that where pupils in those (independent) schools do have key stage 2 measures they tend to be very high," he said.
It was unfair to identify state schools as failing and that was not what his research had set out to do. He said he was disappointed that that was how it had been portrayed in the media. The aim was positive, to show schools what could be done if policies were introduced to make the most of the abilities of very able pupils.
"It is a glass half-full, half-empty issue," he said. "The fact that state schools with large numbers of these pupils achieve such high levels should encourage schools with smaller numbers to do the same."