The Sutton Trust's sums on state school students at university do not add up, says Sam Freedman
The Sutton Trust regularly argues that there are 3,000 state school students "missing" from leading universities. This magic number has been adopted as fact by media commentators and continually crops up in policy debate.
The figure represents the difference between the number of state school students at 13 "top" universities and the "target" figure represented by their Higher Education Funding Council for England participation benchmark.
Yet it is underpinned by assumptions.
Are the Hefce benchmarks accurate? Are the Sutton Trust's 13 universities representative? Is it right to assume, as the Sutton Trust implies, that all 3,000 are from disadvantaged backgrounds?
Until last year, the Universities and Colleges Admissions Service scores, on which Hefce benchmarks are based, were capped, leading to inaccurate benchmarks for universities that had many applicants gaining four or more As at A level.
Additionally, the Sutton Trust uses the 1997-98 benchmarks as a base in its reports. Yet in 1998-99, Hefce changed its definition of an independent school. This means those original benchmarks "are not comparable" with those of subsequent years.
Furthermore, the Sutton Trust ignores Hefce's location-adjusted benchmarks, which take into account the fact that some universities are in areas with a higher proportion of independent schools. For example, in London 14 per cent of pupils go to independent schools compared with 6 per cent nationally.
The trust's choice of "leading" universities is also suspect. It does not contain any university from the North West, Wales or Northern Ireland.
It omits nine Russell Group universities, including Manchester (which finished above eight of the Sutton Trust institutions in a recent Times Higher list).
It also excludes 12 non-Russell Group universities that The Times Higher considers among the best 200 in the world. Students at these institutions are among the 3,000.
If the "missing" figure is calculated with all the other leading universities included and draws on the location-adjusted benchmarks, the magic number drops to 1,900. Given the inaccuracies of the Hefce benchmarks, it is possible that there are, in fact, no missing students.
Even if one took the 3,000 at face value, the Sutton Trust and the media make an unwarranted connection between state school attendance and economic disadvantage.
But numerous state schools have a high middle-class intake and many independent schools have bursary programmes that allow working-class pupils to attend. Research from Cambridge University has shown that the class balance between applicants from independent and state schools is almost identical.
In terms of preventing discrimination, it is more useful to look directly at social class. Hefce provides a benchmark that does just that. Using this, the figure tumbles to 857 for the Sutton Trust's 13 universities - which might not make quite as good a headline as the "missing 3,000" but is more accurate.
The trust's claim to support genuinely fair access based on ability is laudable. But achieving this goal cannot be helped by manipulating data that were never supposed to be used to set "targets".
False figures create an unsubstantiated pressure on universities that can only harm genuinely fair access in the future.
Sam Freedman is research officer at the Independent Schools Council.