In this report, the Education Policy Institute has undertaken an in-depth analysis of the Free Schools programme in England. 

Free Schools in England is the most detailed, independent, assessment yet of this new and much debated reform of the school system. The analysis uses the latest data on Free Schools to assess the impact of the programme on several measures, including pupil performance, inspection outcomes, popularity with parents, composition of pupils from different backgrounds and the extent to which the schools are addressing shortages of school capacity and high quality places.

You can download the report here.



  • In spite of the growth of the programme since 2011, two thirds of areas in England are not within a reasonable distance of either a primary or secondary free school.
  • Free schools are helping to meet the need for new school places – and growth has been higher in areas of “basic need”.
  • However, the programme has been ineffective in targeting areas of low school quality – indeed free school places are more likely to be found in areas of high performance (such as London) than in the areas of low school performance (such as the North East). Some of this is, however, explained by the need for new places in London to address population growth.
  • Free schools are more likely to be located in areas of disadvantage, but disadvantaged pupils in these areas are less likely to be admitted than would be expected. In the most deprived areas, 24 per cent of reception aged pupils in free schools were eligible for free school meals versus 32 per cent in other schools.
  • Free school pupils are much more likely to have a first language that is other than English than pupils in other state funded schools. In primary free schools just 33 per cent of pupils are white British, compared with 67.2 per cent of pupils nationally.
  • Free schools appear to be less popular with parents than all other school types, measured by parental preference data. However, free schools appear to become more popular with parents the longer they are open.
  • Free schools have not yet established themselves as the preferred local school for parents – where the nearest local school is a free school, just 22 per cent of pupils at primary and secondary level attend that school – the lowest of all school types.
  • It is not yet possible to conclude whether free schools are more effective in improving pupil attainment than other schools.
    • Free school Ofsted judgements are better at primary, similar to other schools at secondary, and considerably worse for special and alternative-provision schools.
    • Attainment and progress at the end of primary school is so far poor, but this is based on data from a small number of schools that may not be representative of the programme as whole. At secondary, the average free school Progress 8 score was the joint highest of all school groups (matching converter academies), but this may reflect the higher progress of pupils who are over-represented in free schools (such as those with English as an additional language).
  • This analysis has not considered issues of value for money of the programme nor the impact of post-16 study or destinations

Detailed findings:

Do free schools address growing demand for school places?


  • The free schools programme has been reasonably successful in linking new schools to areas experiencing shortages – in particular at primary level.
  • In areas with the greatest need of additional secondary places, free schools have created an extra 76 places per 1000 pupils. But they have also generated an extra 20 places per 1000 pupils in areas that already have a surplus of school places. The same tendency is, however, true when we look at other new schools.
  • Overall, we find that it is the expansion of existing schools – rather than free schools – that has generated the most new school places and these are more closely linked with population pressures.
  • The free schools programme has not been successful at targeting new school places in areas where school standards are low. New primary free school places are as likely to be set up in high performing areas as low performing areas, and in secondary education, new free school places are slightly more likely to be set up in high performing areas. This is also driven by the regional bias of free schools – with many being set up in high performing London.
  • To address this lack of prioritisation of underperforming areas, the programme could place a greater emphasis on introducing schools in areas that both need new places and are lower performing.

Are free schools popular with parents?


Since their introduction in 2011, claims have been made that free schools are popular with parents:

  • However, the parental preference data shows that where parents are applying to free schools, they are unlikely to be their first preference:

– at primary free schools just over 35 per cent of applications to free schools were as a first preference for parents – the lowest of any school type

– just under 29 per cent were a first preference at secondary school – again, the lowest of any school type.

  • Around half of pupils in England still attend their nearest school – yet this figure is significantly lower for free schools. Out of all pupils who have a free school as their nearest school, just under a quarter of these pupils go to that school. These rates are the lowest of any school type in England.
  • It important to note however that these schools have only recently come into being – and also that free schools have been set up as providers may want to offer a different education from what is available in other local schools. It may not be the intention of these schools, therefore, to be the natural first choice for all local school parents.
  • Moreover, the longer a free school has been open, the greater the proportion of local pupils that will apply, and the more likely that it will be as a first preference.

Are free schools serving disadvantaged pupils?


When it comes to attracting pupils from disadvantaged backgrounds, there is a mismatch between the pupils that go to free schools compared with the local area:

  • The overall percentage of disadvantaged pupils in free schools – those on free school meals – is close to the national average:

– In primary free schools in 2017, 13.3 per cent of pupils were eligible for free school meals compared to 14.7 per cent of pupils nationally.
– In secondary free schools, 14.2 per cent of pupils were eligible for free school meals compared to 13.8 per cent of pupils nationally.

  • Free schools are more likely to be set up in areas of high disadvantage, compared to areas of low disadvantage:

– At primary, 15 per cent of primary free school places are in the most deprived areas while only 8 per cent are in the least deprived areas.
– Likewise, 13 per cent of secondary free school places are in the most deprived areas, while only 8 per cent are in the least deprived areas.
– At both primary and secondary, there is a larger share of free school places in the most deprived areas compared with all other schools.

  • The evidence therefore supports the Government’s claim that many free school places have been created in areas of high disadvantage.
  • However, taken together, this suggests that, while free schools are more likely to be set up in the most disadvantaged areas, disadvantaged pupils in these areas are less likely to be admitted to these schools.
  • This is particularly the case for primary free schools. In the most deprived areas, only 24 per cent of reception aged free school pupils are eligible for free school meals versus 32 per cent in other schools. 

Free schools admit a disproportionate number of pupils who do not have English as their first language (EAL):

  • In free schools, this figure for EAL pupils in 2016 stood at 39.4 compared to 20.6 per cent nationally in primary, and 24.9 versus 16.2 per cent in secondary.
  • Pupils in free schools are less likely to be from white British backgrounds than pupils in other state-funded schools. In primary free schools 33.0 per cent of are white British, compared to 67.2 per cent nationally. These figures are 44.9 versus 69.5 per cent in free secondary schools.
  • This is partly driven by the number of faith based free schools and the concentration of free schools in London. There are also high numbers of EAL students in free schools outside London.

Ofsted outcomes and pupil performance


There is insufficient evidence to reach definitive conclusions on the effectiveness of free schools:

Ofsted outcomes

  • Compared to other schools, there is little difference between free schools and other schools in the proportion of good and outstanding schools at both primary and secondary level. However, primary free schools are much more likely to be rated as outstanding – while for secondary schools there is little difference.
  • Special and alternative provision free schools are less likely to be rated as outstanding.
  • However, as large numbers of free schools have not yet been inspected, it is important to allow time for the programme as whole to become more established, before conclusions can be drawn.

 Pupil Performance

  • It is also too early to conclude whether free schools have a positive or negative effect on individual pupil performance.

– Outcomes in free schools at the end of primary (KS2) are relatively poor but these statistics are derived from a small number of schools, many of which are former independent schools (that have historically not had to take part in national curriculum assessments).o  In 2016, Entry to the E-Bacc was higher in free schools than other schools (by almost 10 percentage points) but achievement of the E-Bacc was only marginally higher (by just 1.3 percentage points).
– After controlling for pupil characteristics, attainment in free schools was actually slightly behind a group of ‘similar pupils’ but this difference was equivalent to one third of a grade in one subject.
– In 2017, pupil progress in secondary using the new ‘Progress 8 measure’, is higher in free schools than any other school group than Converter Academies. We have been unable to undertake a similar detailed analysis on the 2017 results because the pupil characteristic data is not yet available. Once we control for pupil characteristics, there may be little difference between free schools and other schools.