The Academies Illusion: What the data reveals

Henry Stewart's picture
It has become the conventional wisdom, in politics and the press, that academies are an unqualified success and the only route to transform an underperforming school. Analysis of the data does not back up this view. After a reminder of the relative increases in GCSE results in 2012, I look at how different schools use equivalents and likelihood of taking traditional subjects.

Non-academy GCSE results grew as fast as those in similar academies

The Department for Education claimed that GCSE results grew in academies in 2012 by more than five times as much as non-academies. This was based on an increase of 3.1% in academy GCSE results and just 0.6% in non-academies. However the comparison is a false one, as academies are much more likely to start from a lower % in 2011. The graph below compares similar schools, based on their 2011 GCSE results:

For those schools whose GCSE benchmark was in the 20-40% range in 2011, academies increased by 7.8% and maintained schools by 7.7%. Both are great improvements and the schools deserve to be congratulated. However it makes little difference whether the school was an academy or not. The chart shows how the likelihood to increase or decrease their results relates closely to the previous level of results.

Students in academies are more likely to take "equivalent qualifications":

"Equivalent qualifications" are qualifications like BTECs that can currently count as equivalent to GCSEs and are generally regarded as used to game the system. It was the Daily Telegraph that first spotted that "Academy schools 'inflate results with easy qualifications'", finding that 11.8% of academy results were down to equivalents, compared to 6% for non-academies. The gap actually widened in 2012:

2012 Results5 ACEMGCSE onlyDifference

Note: "5 ACEM" refers to % achieving 5 GCSEs including English and Maths.

But this is perhaps an unfair comparison. Sponsored academies are more likely to have started with lower results and schools with lower results are more likely to use equivalents. Do academies make more use of easier equivalents when compared to similar schools? To compare similar schools I have split the DfE data set into ranges according to their 2011 GCSE results, 0-20%, 20-40% and so on. There are only 3 schools in the 0-20% range so the key range for looking at schools in need of improvement is the 20-40% range:

Use of Equivalents2012 5ACEM2012 GCSE onlyDifference
Academies, 20-40% band41.2%23.9%17.3%
Non-academies, 20-40% band42.1%30.2%11.9%


For similar schools, academies are still much more likely to achieve their league table GCSE results through equivalents. The 2013 set of exams represent the last year when equivalents can be used in this way. From 2014 very few non-GCSEs will be counted as equivalent, and then only as equivalent to one GCSE. Will we see a dramatic fall in GCSE results in many schools, especially academies, in 2014?

Students in academies are less likely to take history or geography GCSE:

HumanitiesAll schools20-40% Range


One of Michael Gove's favourite topics is the lack of history teaching in British schools. Yet the schools in which students are least likely to study history or geography are the sponsored academies - even when compared to similar schools, in the same results band.

Students in academies are less likely to take a language GCSE:

LanguagesAll schools20-40% Range


The same is true of languages. Students are less likely to take a language GCSE if they are in a sponsored academy.

Academies: More of their GCSE count come from equivalents

20-40% rangeNo. of GCSEsw/out equiv
Academies               11.4                        5.2
Non-academies               11.4                        6.2


In both academies and non-academies in this range (of 2011 GCSE results) students take an average 11.4 GCSEs if equivalents are included. However this figure reduces dramatically, and falls by a greater amount in academies, once equivalents are taken out. Students take an average 5.2 actual GCSEs in these academies and 6.2 in these non-academies. This has big implcations for results in 2014. Have these schools changed their strategies for the current Year 10s, and increased the number of GCSEs they are taking?

Conclusion: Academies are not transformative

The supporters of academies like to paint a picture of a completely different sort of school, bringing new opportunities to communities whose schools would otherwise be failures. The data does not back up this view. Students in sponsored academies are less likely to take the humanities and language GCSEs that our Secretary of State is so keen to promote. Many are doing well and have seen significant growth in GCSE results. However this increase is just as large in similar non-academies and is more likely in academies not to be based on actual GCSE exams.


Data Notes

All the data here is taken from the DfE dataset on the 2012 GCSE results, released in January 2013. There were 73 academies and 175 non-academies in the 20-40% range used here. This is all schools who achieved between 20% and 40% in 2011 for the % of students achieving 5 GCSEs including English and Maths (including equivalent qualifications).

The more recent converter academies have not been included here. This is partly because there were very few (just 16) in the 20-40% range in 2011. But also because these converted principally from Good or Outstanding schools and so do not address the issue of how the schools in need of improvement are tackling their challenges.


Add new comment