Why do private schools achieve better exam results?

The perceived advantages that private schooling confers are once again in the news. Last week, the Rt Hon Matthew Hancock MP, the Paymaster General, suggested that companies should ask job applicants whether they went to private school. The announcement provoked a number of articles condemning the apparent inequality that private schooling produces - both in terms of academic attainment and future labour market outcomes.

This blog focuses on why private school students achieve better exam results. On the face of it, the advantages from attending private school are stark. In 2014-15, 19% of A-level entries at private schools were graded A*, compared with a national average for all schools of 8%. Similarly, 32% of private school GCSE papers achieved A* compared to a national average of 8%. While these figures show a strong association between private schooling and exam attainment, they do not necessarily show that private schools cause better attainment.

Private schools may admit students with certain prior characteristics that cause them to do better in exams. In other words, private schools may not add value, but simply benefit from the type of students that attend them. To disentangle this, you need to compare the exam results of students with similar characteristics who attend both private and state schools. A number of studies have attempted to do this, controlling for prior characteristics such as academic selection upon entry or socio-economic background. Then you will be able to see if private schools do really add value to academic attainment.

Academic selection

Most private schools are academically selective. That is, they employ some form of entrance exam, usually the Common Entrance Examination. The more elite the private school, the more academically selective they are likely to be. For instance, for entry in 2015, the Prime Minister’s former school Eton rejected around 750 applicants who failed to achieve a high enough entrance exam score.

Controlling for academic selection has proved reasonably easy for researchers. This is because there are a number of state schools in the United Kingdom which still select students by academic ability: grammar schools.

Studies which have found that the difference in exam results between state and private schools substantially reduces once they are both selective. Sullivan and Heath, for example, compared the GCSE results of students at academically selective state schools to those of private schools. The found that the exam results of selective state schools were similar, but not quite equal, to the exam results of private schools. In contrast, the difference between non-selective state schools and selective state schools or private schools was much greater. Similarly, a further study estimated that approximately 80% of the difference in GCSE performance between private schools and state schools was due to academic selection. These studies suggest that a significant proportion of the gap between exam results at state and private schools can be explained by the prior academic ability of private schools’ students. This implies that private schools are not necessarily adding much value.

Socio-economic background

Private schools admit students from more affluent backgrounds. Studies have shown that cognitive ability is significantly associated with parental socio-economic background. Fee-charging private schools may be simply benefitting from admitting more wealthy students than from causing any better outcomes.

Controlling for socio-economic background, in addition to academic selection, has proved much more difficult for researchers. A study published earlier this year found when prior academic ability and IDACI (an index of deprivation which measures the proportion of children under the age of 16 that live in low income households in a particular neighbourhood) were controlled for, the average GCSE difference between students studying at private schools and those studying at state schools fell from two GCSE grades to 0.64 of a GCSE grade.

The study was limited because IDACI is measured at a postcode level. The study therefore cannot be used to identify the exact socio-economic grade of students within that postcode. If this was possible, the differences between private and state schools students may reduce the GCSE grade difference further. The National Pupil Database (NPD), which does include the socio-economic background of specific students - for example, by using the Free School Meals (FSM) proxy, cannot be used since submission to the NPD is voluntary for private schools.

Regardless, the available evidence suggests that the gap in exam scores between private and state schools is greatly reduced once you compare students from similar neighbourhoods.

Other confounding factors

While academic selection and socio-economic background explain a significant proportion of private schools perceived advantaged, they do not appear to explain it all. The aforementioned study which controlled for IDACI and prior ability still found that a 0.64 grade difference persisted. This may be partly explained by weak indicators of deprivation and prior ability. However, it may also be explained by other characteristics that private school students possess. For instance, parents from higher socio-economic backgrounds may be able to provide greater support to their children throughout their education through, for example, access to extracurricular activities. There also may be unobserved causal effects, such as parental academic attainment, which contribute to private schools’ advantage.

The only method of controlling for such variables would be to randomly assign a large group of children to private and state schools. Unfortunately, no research has found a way to carry out such a study in the United Kingdom. There are, however, some interesting international studies. The US state of Louisiana runs a scheme called the Louisiana Scholarship Program (LSP). The LSP provides vouchers for disadvantaged students at low-performing Louisiana public schools to attend private schools of their choice. The vouchers are allocated randomly through a lottery system. A study from December last year compared the educational outcomes of lottery winners and lottery losers. The study found that lottery winners did significantly worse in examinations than lottery losers. This suggests that private schooling does not add value to randomly selected pupils in Louisiana. We should, of course, be wary of applying American research to British schools. However, in the absence of a randomised study from the UK, the research does offer an interesting insight into the effectiveness of private schooling.

Conclusion

Private schools achieve significantly better exam results than state schools. However this attainment gap does not necessarily mean that private schools add value. They may simply benefit from the type of students that attend them. Evidence suggests that private schools benefit from being academically selective, this allows them to admit only academically excellent students. In contrast, most state schools are not selective and thus admit students with a range of academic abilities. Moreover, private school students are much less likely to come from deprived backgrounds than their state school counterparts. When these two variables are controlled for, much of the difference between private and state schools in exam performance is explained. Designing randomised trials to test all the possible variables associated with public and private schools has proved difficult in the UK. However, evidence from the Louisiana Scholarship Program  suggests that private schools do not add value in the United States, at least.

James Dobson is a researcher at Bright Blue