News, Opinion

Free Schools in England: What EPI’s report is actually telling us…


The Education Policy Institute has published a new report, Free Schools in England, but what does it tell us that we don’t already know? Editor Laura McInerney takes a look at the details.


It is now seven and a half years since the Coalition swept to power with Michael Gove at the helm as education secretary. One of his promises was a set of flashy new “free schools” which would allow parents and teachers to open innovative schools in areas with poorly performing schools with the aim of driving up standards.

So, has the promise come to fruition?

Today’s report from EPI is a serious attempt to answer this question. But what are the main findings of its 62 pages?

First, the 6 key things the report does tell us:


1. There are hardly any free schools!

Although they take up a huge amount of political oxygen, free schools make up just 2% of all state schools. And most areas of the country (two thirds) don’t have one within a reasonable distance.


2. There is insufficient data to give ROBUST conclusions about the quality of free schools

This point needs repeating over and over and over: IT IS BASICALLY IMPOSSIBLE TO SAY HOW GOOD, OR BAD, FREE SCHOOLS ARE ON THE WHOLE. Why? Because there’s only a small number of them. Plus changes in Ofsted and the testing regime complicate the whole picture. This doesn’t stop the EPI researchers from trying to come up with some conclusions. At one point the report spends eight pages trying to control for various factors just to find some kind of point. But, in the end, it’s a mixed picture, the data has serious limitations, and it is massively volatile. Anyone who goes on television and wholeheartedly says there are conclusions to draw about the quality of free schools from this report is spinning the research for more than it’s worth.


3. The very tentative conclusions are that primary schools are more likely to be outstanding, special schools and alternative provision are less likely, and secondary schools are about the same

As mentioned, this is a heavily caveated conclusion. A lot of free schools still haven’t been inspected and those that have were mostly inspected before any pupils sat national tests. Hence, their data is a little rough and ready. The EPI researchers note the judgements are more likely to be volatile.


4. A genuinely interesting point: free school pupils are more likely to speak English as an additional language but less likely to have special educational needs

This demographic – more EAL pupils and fewer special needs ones – works to free school’s advantage. EAL pupils tend to have higher progress throughout their schooling, whereas special needs pupils tend to have lower progress. This difference, however, is not recognised in performance measures. Which means when comparing straightforward progress measures, free schools have an advantage because of their intake. (Note: Don’t jump to the conclusion this is by design. The report doesn’t suggest that, and I don’t think there’s reason to believe it’s true. The more likely reason is the high number of free schools in London).


5. The really really interesting point: only 24 per cent of pupils attending free schools in the most disadvantaged areas are on free meals. The number should be around 32 per cent. 

A lot of criticisms about free schools being in the wrong areas don’t stack up. The majority are in areas with place needs, and many are in very disadvantaged communities. But this statistic shows the schools are not always attracting local pupils.

Hence, while free schools are being set up in disadvantaged areas, the report says “they are not, yet, attracting a representative number of children from disadvantaged backgrounds”.

The report analyses a whole raft of numbers on this but just about any way you cut it; there are fewer kids from poorer families in free schools than you’d expect given their location.


6. Finally, while 50% of kids across the country attend their nearest schools, this drops to just 22% when the nearest is a free school

While this might not sound great, it’s not that surprising. As the report says, many free schools were built to be “different”. For example, the first Sikh schools opened under the free school programme. Families near the new school were not Sikh sometimes did not want to send their child to the school – hence the reduced proportion of local families in free schools.

Though not surprising it does show how free schools have fragmented the local offer in places where they exist. In rural areas, where transport is only given to the nearest school, this has caused some furore. (Local authorities have dealt with it differently — some continued paying for travel, others did not). 


3 Other Findings To Use Carefully 

A problem of this report is that it comes to some conclusions which are fair, in so far as the data stacks them up, but repeating the blunt finding doesn’t do justice to the issue.


1. Lots of free schools have opened in areas where schools were already doing well

Given the original plan for free schools was driving up standards in poorly-performing areas, it may seem annoying that free schools have so often opened in places where provision is already pretty good. BUT, this is largely because far more free schools have opened in London than elsewhere.

In part, this is because the capital had the highest place need. One of the main criticisms of the project from the get-go was that it would place schools in areas that didn’t need them because schools there already had extra spaces. Hence, many were built in London because the capital did need the places. But London also has a lot of highly-performing schools. It’s therefore a slightly unfair jibe to criticise so many free schools for opening in areas where place provision was good, if the alternative was to open them in poorly-performing towns which already had half-empty schools.

2. Free schools are the least popular with parents 

It’s easy to think that because free schools have less applications from parents compared to other school type that this shows something profound about how well liked they are. Actually, all new schools, of all types, struggle to win over parents who have children at another school already. Hence, first cohorts tend to be stuffed with eldest and only children. Where a free school got lucky in an area with a massive place need, they can be over-subscribed. But it’s not alarming to see free schools initially struggling to get applicants in their initial years. Logic would predict it.


 3. ‘Things are getting better’ as the years wear on

An over-riding theme of the report is that the more recent cohorts of free schools are getting better Ofsted judgements and seem to be in better-targeted places. BUT the report never mentions that the definition of free schools changed in recent years. New academies opened by local authorities also now count as ‘free schools’. It’s not clear if the reason why things are getting better is that councils are now involved and they make better choices.


In sum, this is a great report and you should read it. But just be careful how you use those facts. Remember: THERE IS NOT ENOUGH ROBUST DATA YET!


Your thoughts

Leave a Reply

Your email address will not be published. Required fields are marked *


  1. Let’s hope the New Schools Network takes note of the warning that it’s not possible to draw firm conclusions about free schools as a group when there are so few of them. This is something I pointed out in my blog with the deliberately misleading heading ‘20% of free schools are below Progress 8 floor standard…’. (The article explained why the headline was misleading – I do not wish to be accused of deliberately spreading false news.)

    • John Connor

      Toby Young ignoring warnings? Not using data judiciously? Who’d have thought? But never mind, he’s off to wreak havoc on our universities now. Classic example of non-accountability in Tory La La Land. When has Gove ever been held to account for the devastation he has wrought?