Secondary school league tables, Ofsted inspections and government improvement targets all use statistics that are based only on pupils who are registered as attending the school towards the end of their time there. School leaders therefore have an incentive to remove children from their rolls before the January of GCSE year, when ‘census’ data are collected, if they think the pupils will not do well. The government insists that regulation prevents this happening, but past investigations (see here and here) have indicated that it does, even if the practice isn’t widespread.

The research organisation Education Datalab recently published research suggesting that ‘pupils leaving can have a very flattering impact on the league table results of a school’. Among the year group who did their GCSEs in 2015, there were 87,000 moves during their five years at secondary school. In 125 of England’s 3400 secondaries, the published results would have fallen by at least five percentage points if the grades of the pupils who left early had been included in the calculations, at a weight proportional to the time they spent in each institution. In two cases the difference rose to a reputation-transforming 16 points.

The research found that children who went to other state schools tended to have poor results. A previously unidentified group of 20,000 children left state education completely; only 6 per cent of them went on to achieve five good GCSEs. Individual school year groups can shrink by up to a quarter between year 7 (the first year of secondary school) and GCSEs in year 11.

Pupils can leave a school early for a variety of reasons. They may be expelled, or parents may decide to take a children out of struggling institutions. But ‘in some cases’, Datalab concluded, ‘pupils are being “managed out” of mainstream schools … with the effect of boosting the league table performance of the school which the pupil leaves.’

In 2014, the exams regulator Ofqual asked teachers to take part in a survey about schools gaming the grade-driven accountability system. A quarter of the 545 teachers who came forward said they had experience of ‘students being removed from the school roll so as to avoid their results contributing to [school performance] measures’. Asked to rate the ‘acceptability’ of this from a moral perspective, on a scale of one to ten, the respondents on average gave it a ‘one’: least acceptable.

I’ve heard stories of headteachers (a minority, certainly) having conversations with parents along the lines of: ‘Why don’t we move your child so I don’t have to put an official exclusion on his record?’ One headteacher came to me several years ago with details of how a school which had won national praise for improving its results had been expecting his and other institutions to take on struggling pupils whom it had asked to leave. ‘Just had a parent in with their son,’ he said.

Son has had a handful of behaviour issues and levels of attainment are very low. Boy told he was taken in by [management at his former school] and told to sign a contract [to leave the school]. Dad said it was so unpleasant he just wanted to walk out and was very upset. I’m admitting the boy. He is unlikely to get five good GCSEs, so that’s how his former school’s results are better than ours, and have improved so rapidly.

The parents of affected pupils tend not to go to the media, so individual cases typically go unreported.

It seems likely that malpractice is going on in only a small minority of institutions. But the tactic is the logical end point of a system, in operation since the advent of league tables in 1992, founded on the assumption that the way to improve schools is to publish statistical indicators of institutional performance and remorselessly reward or penalise them for their results.