Good neighbours: do free schools help others improve?

"Free Schools don't just raise the performance of their own pupils—they raise standards in surrounding schools in the area too"

David Cameron, 9 March 2015

Free schools are once again in the headlines, with the Prime Minister promising more under a future Conservative government.

The announcement earlier this week coincided with publication of a report on the impact of free schools by Policy Exchange, an independent educational charity, which concludes:

"Free Schools are raising standards for other pupils across their local communities, especially in some of the poorest performing schools."

This is a strong claim given the limitations of the evidence, especially the small numbers of schools involved in the research. The report notes that performance improves in weaker schools, but "higher performing schools make less progress and the very highest drop back." These important points seem to have got lost in the discussion that followed, and are not reflected in Policy Exchange's headline conclusion and the Prime Minister's claim.

Do free schools raise standards in other schools?

To answer this question, Policy Exchange looks at a new free school and finds the three nearest similar schools. They do this for all free schools approved in a particular year, and track the progress of those nearby schools from the year the free school is approved up to 2014.

We're talking about small groups here: typically between 20 and 30 free schools of each type open each year; that means comparing between 60 and 90 neighbouring schools. We'd expect figures to be quite variable given these small sample sizes and Policy Exchange advise treating the results with caution.

For secondary schools that were performing worse than average, those with a free school nearby improved more than similarly-attaining schools nationally. But by the same measure, successful schools with a free school nearby improved less than other such schools nationally. These effects cancel each other out so overall neighbouring schools perform in line with the national level.

The picture is similar for primary schools near free schools: the lowest-performing made more progress while higher performing schools made less progress than would be expected nationally.

Policy Exchange told us:

"We make it very clear — in the report and in press interviews — that there seems to be a pattern suggesting free schools raise standards for neighbouring schools, especially in the most underperforming schools. We have not said that free schools lift results for every school in a local authority."

There is no proof that free schools themselves were the cause of any change in performance in other schools nearby.

Policy Exchange suggests that free schools might be raising standards through competition with local schools. But there's nothing in this argument specific to free schools: local competition could also be achieved by opening a new maintained school or academy in the area, or by the arrival of a new headteacher in an existing school. The report doesn't compare the effects of new free schools with other new schools, so there's nothing to show whether free schools are necessarily better.

Do free schools raise standards for themselves?

As we've said previously, it's too early to say whether free schools fare better than other schools under Ofsted inspections. Just 76 schools, less than a third of the total, have been inspected and given judgements so far. That figure excludes two free schools which were judged inadequate before later closing.

Of those that are open, not enough free schools have exam results yet to begin comparing them to other schools.

Keeping free schools in perspective

Free schools are only a small part of the picture, with just over 250 currently open out of more than 24,000 schools in England. As the Education Select Committee pointed out, even opening a further 500 free schools would only bring the proportion to 3% of all schools in England, and that assumes that no other schools open in the same period, which seems unlikely given the projected increases in primary and secondary student populations.

After digesting ten sessions of oral evidence and around 140 written submissions, the Education Select Committee reported in January 2015 that:

"it is too early to draw conclusions on the quality of education provided by free schools or their broader system impact" Education Select Committee, 21 January 2015

That's the fairest summary of the evidence we've seen so far.