Friday, 7 March 2014

NIck Wright: Genetics and schools policy


In this article, Nick Wright looks at the work of Michael Gove's former special adviser Dominic Cummings and places it in the context of the history of genetic determinist theories of education. 

Ever since the intellectual case for school selection based on intelligence quotient testing was found to be grounded in faked statistics there has been an unceasing quest to find a new theoretical basis for a school system based on differentiating children.

The talmudic text deposited at the DfE by Michael Gove's departing policy advisor suggests that genetics may be the new great white hope for those who need the appearance of intellectual rigour to cloak their preference for selection. Only a political innocent would think that an argument that dismissed the value of much education, trashed teachers and claimed for heritability a privileged place in determining children's ability would go unchallenged.

Michael Gove's special advisor or Spad, Dominic Cummings – the author of a 250 page reflection on government, politics and education* – is reportedly returning to his career in the ministerially-favoured 'free schools' sector.

Already the practical case for such schools – freed from curriculum requirements, the necessity to employ qualified teachers, pay agreed pay rates or appoint head teachers with teaching experience – is unravelling more speedily than the faked science that underpinned the introduction of the 11 plus.

To recap – for any ministerial aide whose Oxford degree in Ancient and Modern History missed out the 20th century – Cyril Burt, knighted in 1946 for services to psychological testing – based his conclusion that the superior performance in his tests of upper class children in private preparatory schools was due to their greater inherited intelligence. He reached this conclusion in 1909 and by the time he died (in 1971) the idea that tests at the end of primary education were an accurate predictor of general intelligence, and by extension, future academic performance, vocational aptitude and positioning in the social structure, was under assault.

Burt based much of his research on the performance in intelligence tests of monozygotic twins.

Within a few years both the identity (even existence) of his supposed collaborators in field work and the statistical reliability of his key evidence was questioned. By the end of the decade his intellectual reputation was irredeemably tarnished and his friend and biographer Leslie Hearnshaw had concluded that Burt's data – critically important in the political case for selection and the post-war tri-partite division of secondary education – was fraudulent or could not be relied on.**

Remarkably, the evidence of this fraud has been challenged and a lively polemic has ensued in which the strongest suggestions that Burt's findings remain valid have come from people who still cleave to the proposition that the genetical heritability of intelligence as measured by IQ scores has decisive utility in education policy.

While this puts the question firmly into the contested terrain of ideology and politics Cummings claim that: “There is strong resistance across the political spectrum to accepting scientific evidence on genetics. Most of those that now dominate discussions on issues such as social mobility entirely ignore genetics and therefore their arguments are at best misleading and often worthless" is a transparent bid to claim the evidential high ground.

Relying on the work of US behavioural geneticist Robert Plomin he claims for genetic factors a decisive role in determining intelligence and educational attainment.

Plomin enters the back story at this point. In a speculative Spectator article this July he takes a swipe at Leon Kamin, the author of a 1970's book The Science and Politics of IQ which analysed Burt's statistics and found the correlation coefficients of monozygotic and dizygotic twins' IQ scores were the same to three decimal places, across articles – even when Burt more than doubled his sample.

Plomin seems uncomfortable with the mechanistic application of his research to contemporary political controversies. In an illuminating BBC Radio 4 exchange with the geneticist Steve Jones he was at pains to make the point that no necessary policy implications follow from his work and  emphatically stated: “I am not advising on policy.”

By way of contrast Dominic Cummings makes larger claims for this evidence.

"Work by one of the pioneers of behavioural genetics, Robert Plomin, has shown that most of the variation in performance of children in English schools is accounted for by within school factors (not between school factors), of which the largest factor is genes. Scores in the phonics test show ~70% heritability; scores in National Curriculum reading and maths tests at 7, 9, and 12 show ~60-70% heritability; and scores in English, Maths and Science GCSEs show ~60% heritability in a just completed twin study (the GCSE data will be published later in 2013).


Cummings is keen to discount environmental factors in shaping pupil performance and emphasise the primacy of genetic factors.

“In contrast, the overall effects of shared environment (including all family and school influences shared by the twins growing up in the same family and attending the same school) accounts for only about a third of the variance of GCSE scores. Educational achievement in school is more heritable than IQ in English school children: i.e the heritability of what is directly taught is higher than what is not directly taught. Perhaps differential performance in educational achievement is heritable because it is taught: that is, roughly similar schools teaching the same material reduces a major source of environmental variation, therefore the variation that remains is even more due to genetic variation.”

And he is is equally keen to discount family wealth, social position and class as factors:

“Similarly, this paper (Science, 23/4/2010) shows how good teachers improve reading standards for all but this means that the variance that remains is more due to genetic differences. This leads to a conclusion almost completely at odds with prevailing conventional wisdom in political and academic debates over education: differences in educational achievement are not mainly because of ‘richer parents buying greater opportunity’ and the successful pursuit of educational opportunity and ‘social mobility’ will increase heritability of educational achievement.”

To collapse the entire corpus of  'progressive' educational thought into the simple concept that educational achievement arises from privileged access is a transparently rhetorical flourish that abandons any claim to rational debate.

Behind it lies a an unsavoury strain of thinking with a compelling claim to historical continuity. It was, after all, Cyril Burt, who in his 1909 study argued that:

“Wherever a process is correlated with intelligence, these children of superior parents resemble their parents in themselves being superior. Proficiency at such tests does not depend on opportunity or training, but on some innate quality. The resemblance in degree of intelligence between the boys and their parents must, therefore, be due to inheritance. We thus have an experimental demonstration that intelligence is hereditary.” 1909, p 181.

Burt's 1909 comparative measure of parental intelligence had the virtue of simplicity. He simply assumed it from their profession and social position taking intellectual and upper class parents of his sample cohort to be more intelligent than the tradesmen.

Sixty years later, when quizzed about his measure of parental intelligence he reported that: “The intelligence of the parents was assessed primarily on the basis of their actual jobs, checked by personal interviews.”

Today of course, rich parents – even those seized by the insight that the intelligence of their offspring is mostly inherited – are still inclined to trust the education of their children not to the cash strapped and highly invigilated state sector but to a more relaxed environment where generously-endowed private schools with smaller class sizes and better paid teachers enjoy more intimate and  productive relations with the most prestigious universities. Lower down the pecking order ambitious parents will move might and main (and address) in order to get their children into a school perceived to be better.

An inevitable consequence of the current fashion for behavioural genetics is the hunt for the 'intelligence gene'.  Much hope is invested in the work of a young Chinese researcher Zhao Bowe at the biotech research centre BGI Shenzen where a large array of DNA sequencing machines is available.


The research programme – which involves an international group of collaborators – is aimed at discovering the genetic code for intelligence by distilling information from the genomes of thousands of prodigies. Large claims are made for the project; that it will identify the genetic basis of IQ and allow for large scale embryo screening including interventions to raise the IQ potential of unborn children.

Sceptics naturally predict failure for this utopian project. They point to the complexity involved in the relatively simpler scheme to isolate the genetic determinants of height in which nothing much emerged until the DNA sample exceeded ten thousand. After herculean efforts scientists have now tracked down something close to a thousand genetic variations that can be connected to variations in height with some claim to universal applicability across different population groups.


Zhao's is a bold initiative and one with no easily determined time scale. If the interest in this project objective arises from a desire for scientific and rational policies which might identify and educate a larger cohort of exceptionally gifted children, perhaps one drawn from a wider social spectrum than conventional school systems and testing regimes deliver, it might have wide appeal.

If, however, it was to be deployed to exclude some children from access to the best locally available education its appeal would be greatest to those who dismiss the potential of the great majority of children to develop beyond expectations. Cummings is frank in his advocacy of special educational privileges for children who are identified as especially gifted.

From the standpoint of school improvement professionals, concerned with the education of that great majority of children who do not enjoy the class room conditions of the elite, the priority is to devise strategies that realise the educational potential of the many.

If it would be perverse to ignore science that showed a powerful role for heritability  in IQ test results it would be equally perverse to discount the significance of the uncontroversial finding that Cummings references: 'that good teachers improve reading standards for all'.

Education has a wider role than simply identifying and developing exceptionally gifted pupils. Such children have to live in a real world where the fullest development of each child's potential is the condition for the fullest development of society as a whole.

Behavioural genetics has a serious public relations problem in education circles that is not going to be easily dispelled by partisan polemics of the type deployed by Michael Gove's attack dogs.

Neither will it be rehabilitated if it is deployed to perpetuate an education system that allows privileged elites, whether those elites are selected by parental wealth or genetic endowment, to prosper while the needs of society as a whole, and of the individual child, are neglected.

A deeper understanding of the role of heritability factors in the development of the child is an undoubted common good. But a fixed gaze on the genome must not be at the expense of a 20/20 vision of education for all.

*      http://static.guim.co.uk/ni/1381763590219/-Some-thoughts-on-education.pdf

**    Hearnshaw, L.S. (1979). Cyril Burt: Psychologist. Ithaca, New York: Cornell University Press.

No comments:

Post a Comment