EVER SINCE Margaret Thatcher allowed parents to choose schools in the late 1980s, academics and policy-makers have argued about the impact that competition has had on the social mix in classrooms.
Many have argued that market forces cause social segregation, because middle-class parents monopolise places at "good" schools. As a result other schools are forced to take a disproportionate number of poorer pupils.
But others have suggested that the previous system - where schools took pupils from their local area - was even more unjust, because only the affluent could afford to move to be near a good school.
But this month the findings of an influential study are in line with a growing body of research that argues that England's schools are indeed more socially divided than ever.
Education officials last year took heart from statistics produced by Stephen Gorard and John Fitz, of Cardiff University, which suggested that schools had in fact become more socially mixed.
The result contradicted previous work and was considered particularly important because it was the first large statistical study, coming after many smaller projects.
But the latest data from Dr Gorard will disappoint ministers. Although he still insists that school populations became more mixed between 1989 and 1997, he now says that the situation has deteriorated since Labour came to power.
Gorard and Fitz have found that the poorest pupils (defined as those who qualify for free school meals) are now more concentrated in certain schools.
They have devised an "index of segregation", which estimates the number of disadvantaged children who would need to change schools for there to be an even spread of disadvantage.
In 1989, 35.4 per cent of pupils receiving free school meals would have had to change schools to achieve this even spread. By 1997 this had fallen to 30.1 per cent. But 1999's figures show a rise to 32.1 per cent - the same as 1993's figure.
However, pupils with special needs, English as an additional language and those from an ethnic minority have continued to be more evenly spread.
Dr Gorard believes that the state of the economy could partly explain the increase: during the recession of the early 1990s schools may have appeared more mixed, as more children qualified for free school meals. But local admissions policies have also contributed to the increase.
Dr Gorard says: "Our tentative conclusion is that the increase in polarisation is partly linked to the economic cycle. But there are many local facors which also have an impact. There is still a lot of work to be done to find a full explanation for this segregation."
Gorard and Fitz's latest findings bring them roughly in line with their rivals Alex Gibson of Exeter University and Sheena Asthana, of Plymouth University, who have long argued that market forces cause social polarisation in schools.
The two pairs of academics have vehemently attacked each other's methods. Gibson has called Gorard's measure of segregation unreliable; Gorard disagrees with Gibson's conclusion that the gap between the worst and best schools is widening. Their spat will come to a head later this year when each publishes a paper in an educational journal alongside a commentary attacking the other's approach.
Gibson and Asthana's research concludes that high-performing schools are attracting a growing proportion of affluent pupils (TES, April 30, 1999).
They analysed the GCSE results of 1,584 comprehensives and found that the results of the top 10 per cent of schools were improving at nearly three times the rate of the bottom 10 per cent, in terms of raw scores (ie the top schools improved by around six percentage points and the bottom by around two). They also analysed pupils' home postcodes and concluded that a school's position in its local pecking order was the key determinant of how quickly its social composition and exam results improved.
Dr Gibson believes that Gorard and Fitz's analysis is not only flawed but dangerous, as it has reassured policy-makers that the market is doing no harm.
He says: "It is not just that Gorard and Fitz are wrong, but that they are dangerously wrong.
"Their dismissal of concerns about polarisation is imbued with the authority of large-scale quantitative analysis. There is a danger that the wealth of evidence cataloguing the detrimental consequences of marketisation is being eclipsed.
"Policy-makers have been resistant enough to this evidence without the succour offered by Gorard and Fitz's deeply flawed index of segregation. This is not just an arid debate about numbers - if we are to going to have policy based on the evidence then it is vital that we get a true picture of what is happening in schools."
Dr Gorard himself argues that more research is needed to determine why segregation is worsening in some local authorities but not others. But he believes that markets may have had much less impact on education than is popularly believed. He said: "The advent of choice may be truly both less beneficial than some advocates suggest, and less harmful than some critics fear."
Dr Gorard's study will be published in due course. For details e-mail: