20 April 2017

Why grammar schools aren’t selective enough

By Jamie Martin

England’s grammar school debate, with its distortions of evidence and each side’s inability to empathise with the other’s motivation, epitomises the worst of partisan political argument.

The good news is that by understanding why the debate is so polarised, and looking dispassionately at data regarding the science of intelligence and the impact of schooling, we can probably satisfy at least some of the motivations on both sides. In so doing, we can produce education reform to improve the lives of individuals and the country at large.

To first understand the polarisation, it’s helpful to consider the evolutionary and moral psychology of how people form political beliefs. Jonathan Haidt articulates and synthesises this effectively in his book The Righteous Mind. Evidence shows that we use logic only to offer supporting arguments for (mainly heritable) instinctive beliefs. These beliefs are based on what Haidt calls a moral matrix made up of six strands; the grammar school debate appears to mainly turn on two of these: care/harm and fairness/cheating.

The opponents of grammar schools – generally but not exclusively in the centre or on the Left of politics – are primarily motivated by a sense of care towards an oppressed group (pupils from poorer backgrounds), and a sense of fairness (anger that wealthier pupils gain advantage from superior education prior to the grammar school selection test and at grammar school).

When evidence is produced showing that grammar schools disproportionately educate wealthier pupils, and that poorer pupils at non-selective schools perform worse than control groups, this presses a hot button in grammar opponents’ instinctive morality.

The proponents of grammar schools – usually on the Right – are primarily motivated by fairness but, crucially, in a different way (the lack of opportunity for bright pupils from poorer backgrounds). They tend towards prioritising maximal utility (equality of opportunity and economic efficiency) over minimising the costs for weaker groups. Evidence that bright pupils from poorer backgrounds do better in grammar schools than comprehensives resonates with them in a way that the worsened performance of poorer pupils in non-selective schools doesn’t.

From two honourable and understandable motivations, therefore, springs a deeply polarised argument. Finding a resolution requires understanding where data shows the two groups are wrong, and where they are right.

The primary error of the opponents of grammar schools lies in their belief that success in standardised tests, such as the 11+, is stacked in favour of those with financial and social capital. They ignore the significant utility (to all of society) of prioritising the education of the very gifted.

Among those working empirically in relevant fields, there is, in fact, a broad consensus that intelligence is largely heritable and that standardised tests are very good predictors of potential. The (limited) evidence indicates tutoring for tests, such as the 11+ or US SAT, makes a small difference relevant only to the most marginal candidates. Research also indicates very talented people have a disproportionately beneficial impact on society and that focused education helps develop their talent.

The primary error of the proponents of grammar schools is to focus on a small number of positively affected pupils – those from poorer backgrounds who pass the 11+and thrive – rather than the two far larger groups with less positive outcomes. The first of these is the group of (disproportionately poorer) pupils who suffer from the under-performance of the etiolated schools located close to grammars which have creamed off the best pupils and teachers.

The second group is of the relatively wealthy pupils who make up the majority of those benefiting from grammar school education but appear to benefit less from it than poorer pupils who attend. They, therefore, fail tests either of equality or utility – because they aren’t benefiting as much as poorer peers, and owing to this group’s size the majority are by definition not part of the cognitive elite responsible for high creative output and societal impact.

Proponents also miss the fact that by not testing personality traits (also heritable and testable and very important to future outcomes) the 11+ is too narrow a test.

So is it possible to develop a policy more in line with the scientific evidence, taking account of groups suffering negative externalities, that appeals to the moral motivations of both groups? Pleasingly, yes.

The answer lies in the aforementioned evidence that the greatest utility for all society lies not in special education for the top quartile of pupils, but in the 1-2 per cent most talented (2-3+ standard deviations above the average). These students go on to make a disproportionate contribution to society (in a range of fields). This has been evidenced and explored by papers on smart fraction theory as well as in longitudinal studies, such as Van Der Built University’s of Mathematically Precocious Youth.

The creation of a series of smart fraction schools for the top 1-2 per cent would satisfy grammar school proponents’ desire for maximal utility of the talents of the brightest. Selecting the pupils using standardised assessments that are as culturally neutral and resistant to tutoring as possible (the US SAT shows this can be achieved to a reasonable extent), and that take into account the impact of personality traits – as well as aptitude – should go some way to satisfying both sides’ ideas of fairness.

Since the policy would affect such a small number of teachers and pupils, it would produce none of the negative externalities of the mainstream grammar schools (etiolation of other schools; disproportionate benefit to a large number of only moderately talented but wealthier pupils).

Standardised aptitude and personality tests of all pupils at 11 (or even better earlier) would also offer the opportunity (and a positive externality to grammars) to target resources on the genuinely most in need (ie the lowest scorers) in a way that should appeal to both the care motivation of opponents of grammars, and the utility motivation of proponents.

If you want to change the world, first you have to understand it. Respecting the motivations and arguments of your opponents and the data and findings of those working empirically in the area is usually a good place to start.

If both sides of the grammar schools debate can do that, then we can achieve what I believe they both want: an education system that helps us improve individuals’ lives while building a stronger economy and fairer society.

Jamie Martin founded Africa’s first EdTech incubator (www.edtech.org.za) and is former adviser to Michael Gove