What impact do grammar schools have on educational performance? And do they offer a real opportunity for bright children from disadvantaged backgrounds? New research from the CMPO, based at the University of Bristol, by Adele Atkinson and Professor Paul Gregg finds that:

• Both proponents and critics of selection are overstating their case: on average, there is little difference in achievement between pupils in local education authorities (LEAs) that still have selection and similar pupils in comparable non-selective LEAs.
• Grammar school pupils outperform comparable children in non-selective LEAs – but non-grammar school pupils underperform compared to their counterparts in non-selective LEAs.
• Selection does work in favour of bright pupils from poor backgrounds – but only a small minority actually make it into grammar schools.

The research uses data from the national pupil database – the Pupil Level Annual School Census – to track children’s test and exam results from the ages of 11 to 16. It looks at whether LEAs that still have grammar schools achieve better GCSE results for their pupils, as well as which children benefit from academic selection.

The data reveal a huge gulf between the population mix of pupils as a whole and the mix in LEAs with grammar schools. In the 19 LEAs in England that retain substantive selection (each typically with 25% of pupils at grammar schools), just 6% of all pupils eligible for free school meals (an indicator of children living in low-income households) attend grammar schools while 26% of all other children within the LEA gain a grammar school place. 12% of pupils in non-grammar schools in these areas are entitled to free school meals; the figure is only 2% in the grammar schools.

The disadvantage of poor children applies even to those of the highest ability. Using Key Stage 2 (KS2) test scores at age 11 as an independent measure of children’s ability, the researchers find that of those in the top three KS2 groups, just 32% of those eligible for free school meals attend grammar schools compared with 60% of better-off children. Poorer children in selective LEAs are only half as likely to attend a grammar school as other children with the same underlying ability. This pattern is also true of children with special needs and those for whom English is a second language.

So does getting into or missing out on a grammar school place make a difference to attainment within selective LEAs? And what happens to the poorer children who disproportionately miss out on grammar school places?

The research finds that children attending grammar schools are doing very well compared with peers with the same KS2 results in comparable non-selective areas, while those not getting into grammar schools within selective LEAs are performing slightly worse. So despite the overall result that there is no substantial difference on average, there are significant benefits to the elite 25% who attend grammar schools and a small degree of detriment to the other 75% of pupils in selective LEAs.

The research then investigates whether this result is driven by the fact that children with special needs, learning deficits or from poorer backgrounds are concentrated in the non-grammar schools. What emerges is that large concentrations of poor children in a school reduces overall attainment and that selection results in far fewer poor children in grammar schools and far more in non-grammar schools than in comprehensive areas.

This polarisation drives all the under-attainment in the non-grammar schools and about a third of the benefits to grammar school pupils. In other words, if the selection process did not segregate the affluent into grammar schools and the poor into the remainder, then selection would be leading to gains among those at grammar schools with no disadvantage for the rest compared with comprehensive education.

Finally, the research explores the effects of selection on children on free school meals. In turns out that the small minority of poor pupils who make it into grammar schools do exceptionally well, getting nearly eight grade points more – equivalent to eight GCSEs being raised from a C to a B. Those not attending grammar schools do no worse than their peers in non-selective LEAs. So selection does work in favour of bright pupils from poor backgrounds if they can get into the grammar schools in the first place. But even among the very able poorer children, only a small minority make it.

It is distinctly possible that the under-representation of poorer children in grammar schools stems from each grammar school operating separate admissions policies and sometimes exams. This places more onus on parents to apply to the grammar school and prepare children for the tests, a process that fell to primary schools when admissions were more standard. It is possible that this more pro-active parent choice approach is leading to a gulf in access between affluent and poor children.

Selective Education: Who Benefits from Grammar Schools?’ by Adele Atkinson and Paul Gregg is published in the Autumn 2004 issue of the bulletin of the Centre for Market and Public Organisation