A "tutor-proof" 11-plus test has failed to improve the likelihood of poor and ethnic minority pupils being accepted to grammar schools, a report claims.
Campaigners in Buckinghamshire also say that the reformed entrance test, which is based more closely on the primary curriculum, has increased so-called "11-plus tourism", whereby pupils from other counties travel to take the exam as practice for similar tests in their own areas.
A report from campaign group Local, Equal, Excellent - based on official statistics obtained through Freedom of Information requests - highlights figures suggesting that the tutor-proof test may not be boosting the chances of certain groups.
The data suggests that children of black Caribbean and mixed whiteblack Caribbean heritage in the town of High Wycombe did particularly badly in the first year the exam was sat, in September 2013. The combined success rate for these groups fell from 15 per cent to 5 per cent.
Children of Pakistani heritage also appeared to do worse in the town: they made up 13 per cent of children gaining grammar school places in the year prior to the new test, but just 11 per cent under the first year of the new test.
The exam, versions of which are now used by a third of grammars in England, has also had little success in improving access for children eligible for free school meals, campaigners said.
The figures were released as arguments surrounding grammar schools resurfaced, with Kent County Council reviving plans for a new school in Sevenoaks and home secretary Theresa May backing a new grammar in her constituency of Maidenhead in Berkshire.
The Buckinghamshire figures also show a 30 per cent rise, from 2,443 to 3,183, in the number of pupils from outside the county who took the test in the first year it was sat - a trend that has been maintained this year. The rise in local children taking the test was just 8 per cent.
Open University lecturer Katy Simmons, who helped to compile the report, told TES that the exam should be "withdrawn" unless it could be made fairer. There was still a great disparity in children's ability to prepare for the test, she said, and pupils from prep schools were more than three times as likely to pass than pupils from state primaries. "The evidence so far suggests that the new exam has been little more than an expensive exercise in spin," she added.
Philip Wayne, headmaster of Chesham Grammar School and chair of the Buckinghamshire Grammar Schools group of headteachers, which commissioned the new 11-plus, said they had worked hard "to ensure that the test was deemed appropriate for Buckinghamshire cohorts and was fair for children whatever their gender, ethnicity or social background".
But he admitted it was a complex issue and other measures, such as giving priority to free school meals children who passed the exam, had a role to play in widening intake.
He added that there was no "immediate obvious solution" to pupils from outside the area taking the test. "We are aware that some indeed will be doing so to give their child a practice run for a test elsewhere in the country, and we cannot prevent children from sitting the test because of where they live," he said.
The test was developed by the Centre for Evaluation and Monitoring at the University of Durham. Dr Sue Stothard, head of assessment development, said that until detailed analysis was completed, it would be impossible to conclude how effective the exam had been in opening up access to grammars. She added that 11-plus tourism pre-dated the new test.