After more than 20 years of existence in the US, and having failed to deliver on expectations of performance, charter schools have finally started to show marginal improvements across the board, a major new study reveals.

However, the research also shows that almost a third of charters - independent, state-funded schools - still deliver a “significantly weaker” performance in mathematics than their traditional state-maintained counterparts, while a fifth lag behind in reading.

Even the overall boost in performance is attributed to widespread closures of failing charters and an increase in the number of high-performing ones, rather than a wholesale improvement of existing schools.

The findings were published last week in the latest National Charter School Study, a widely cited report produced every four years by the Center for Research on Education Outcomes (Credo), part of Stanford University in California.

The approach to charter schools and their performance varies considerably from state to state. However, the findings will prove of interest to politicians not only in the US but also around the world, as more schools are placed in the hands of organisations outside direct government control. England, Sweden and New Zealand have all made moves to transfer management of schools to external providers.

In 2009, Credo issued the first comprehensive research into the performance of charter schools. The study, which considered 16 states, showed that just 17 per cent of charter schools outperformed mainstream schools, whereas 37 per cent performed “significantly worse”.

The latest study analysed the test scores of charter schools in 26 states as well as in New York City. It shows that, on average, the schools have made improvements, but it says that this is mainly down to the closure of poorly performing charters and a drop in the standards of local state-maintained schools.

“The 2013 Credo study finds that charters in the original 16 states have made modest progress in raising student performance in both reading and mathematics, caused in part by the closure of 8 per cent of the charters in those states in the intervening years since the 2009 report as well as declining performance in the comparison traditional public schools over the same period,” the document states.

The results prompted criticism of charters from the US’s teaching unions. The National Education Association said that the study calls into question “the wisdom of investing large amounts of taxpayers’ money in the creation of charter schools in hopes of improving schools”.

And Randi Weingarten, president of the American Federation of Teachers, said the results show that charters are no better than mainstream schools. “The message here, no matter how it is framed, is that 20 years after the start of the charter school movement, even with all the private energy and public policy cheerleading it has engendered, students in charter schools roughly perform the same as students in the rest of public education - not the leaps and bounds that were promised,” she said.

But Laura McInerney, a former teacher and now a Fulbright scholar researching the charter school movement in the US, said that the performance of charters depends mainly on the state. “A lot of the states have now been able to learn from the earliest adopters and do a very good job, such as those in Maine,” she said. “But somewhere like Arizona is notorious for its problems, as it said that the market is the way forward without any accountability. And three years later (the system) was found to be riddled with fraud and now many (charter schools) have to be closed down.”

Meanwhile, some of England’s first free schools, which were based largely on the US charter school movement, are already showing signs of underperformance.

Discovery New School in West Sussex was judged to be inadequate by schools inspectorate Ofsted last month. In March, three free schools - Batley Grammar School in Yorkshire, Sandbach School in Cheshire and Kings Science Academy in Bradford - received “requires improvement” notices from inspectors.

CHARTING THE CHARTERS

6,000 charter schools exist in the US.

2.3m students are taught in charters.

READING PERFORMANCE

25% of charters studied were significantly stronger than traditional schools.

56% were no different.

19% were significantly weaker

MATHEMATICS PERFORMANCE

29% of charters studied were significantly stronger than traditional schools.

40% were no different.

31% were significantly weaker.