Issue 196 of SOCIALIST REVIEW Published April 1996 Copyright © Socialist Review
The research of Cyril Burt which underpinned intelligence testing in education, was shown to be false over 20 years ago. John Parrington looks at a new attempt to rehabilitate its author, just at a time when intelligence testing and selection are back on the agenda
Before the introduction of comprehensive education in the 1960s and 1970s all children destined for state schools had to take the 11-plus. This creamed off a small minority into the grammar schools while the vast majority were labelled as failures. The 11-plus was based on the infamous intelligence quotient (IQ) test. Its supporters believed that IQ tests measured intelligence and that IQ is hereditary. They relied on the false assumption that a person's intelligence is a fixed thing, like the amount of water in a bucket, which can be measured and graded accordingly. Such crude reasoning is also at the heart of the new Tory tests.
In Britain IQ tests were first popularised by Cyril Burt, an educational psychologist and one of those responsible for devising the 11-plus. Burt claimed his 40 years of research proved a child's intelligence was mainly inherited from its parents and that social circumstances played only a minor role. His research formed the basis of education policy for half a century-from the 1920s until the 1970s. Yet only a year after his death in 1971, evidence began to emerge that Burt was a fraudster who had simply invented results to fit his theories about the hereditability of intelligence.
Burt carried out studies of IQ in identical twins separated at birth and raised apart. He chose such individuals because they provide the only really adequate natural experiment for separating genetic from environmental effects on humans. However, such cases are extremely rare. There have been only three other published studies of this kind besides Burt's. No one had previously been able to assemble more than 20 pairs of twins.
Burt's study stood out because he claimed to have found 53 pairs, more than twice the total of any previous attempt. Burt also showed an extremely high hereditability of 80 percent for IQ performance. It was no wonder that his work was eagerly seized upon by all those who shared his rigid hereditarian views.
Things started to unravel soon after Burt's death, when it was shown by respected US psychologist Leon Kamin that Burt's figures constituted a statistical impossibility. 'A liar and a fraud,' was Kamin's verdict. This charge was borne out when it was found that Burt's two female 'collaborators', who supposedly collected and processed his data, had never worked with him and probably never existed! Eventually even Burt's friend and official biographer, Leslie Hearnshaw, was forced to accept that the charges of fraud were justified.
The exposure of Burt as a fraudster was of enormous significance in the movement to get rid of the 11-plus and selective education, and replace them with the comprehensive system. The result of the shift was a vast improvement in education for all children. Now, as the Tories try to roll the clock back in our schools and Labour turns its back on nonselective education, there has been an attempt to revive the discredited IQ tests and the reactionary ideas behind them. Central to this, a campaign is under way to rehabilitate Cyril Burt by his supporters.
The campaign began a few years ago with the appearance of books by two British authors, psychologist Robert Joynson and sociologist Ronald Fletcher. Both concluded that the charges against Burt had not been proved and that, at the very least, he deserved the benefit of the doubt. The newspapers seized on these claims. Yet Joynson and Fletcher's arguments turned out to be insubstantial. The best they could come up with was a claim that one of Burt's 'women assistants' might have been identified.
Now a new book has appeared, a collection of essays entitled Cyrll Burt--Fraud or Framed. It claims to be much more objective than previous books about Burt. Yet all of its contributors clearly share Burt's view that intelligence is largely hereditary. Two of them, Hans Eysenck of London University and US psychologist Arthur Jenson, have been Burt's most dogged defenders. This is hardly surprising, given that both of them originally used his studies to build their careers and back up their reactionary theories about race, class and IQ.
In this instance, their contributions merely provide a certain amusement value. Eysenck, who was Burt's star pupil and well known right winger, here provides a character assessment of Burt. Full of hysterical outbursts against the 'politically correct left wing militants' who Eysenck is convinced were responsible for Burt's downfall, this is obviously meant to be a sympathetic portrayal. Yet it also includes such revelations as the fact that Burt wrote articles attacking his own students' work, but under a pseudonym, in the journal he himself edited!
Serious character defects aside, the main case against Burt remains the charge that his work contains evidence of systematic fraud. Jenson, who at the height of the US civil rights movement argued that blacks were genetically inferior to whites, attempts in his article to challenge this evidence. In fact he merely recycles the extremely weak arguments first put forward by Joynson and Fletcher.
This is unfortunate, because one of the positive features of this book which emerges in later articles and which in the end does make it interesting reading, is that there is a rebuttal here for practically every single one of Joynson and Fletcher's claims, including definitive evidence that Burt's 'assistants' existed only in his imagination. Moreover, in these later articles evidence emerges that Burt's fraud may be much more extensive than was ever suspected when the original allegations were made.
It now seems certain that the fabrication of data and invention of 'assistants' first shown in his twin studies also appear to be a feature of much of his other work, such as in the influential study which claimed declining educational standards were due to progressive teaching methods.
At some points in the book it is implied that Burt's fraud was that of a failing, sick old man. In fact a cavalier disregard for the facts seems to have been characteristic of Burt's work long before he descended into outright fraud. In his very first study in 1909 which looked at intelligence and social class, he concluded that it made very little difference whether children were born in a slum or a leafy suburb--intelligence was largely inherited. In support of this, he claimed that the IQs of fathers and sons were very similar, but provided no parental test scores. When questioned 50 years later about how he had measure parental intelligence, it emerged that Burt hadn't--he had merely assumed it from their social standing! It is hard to imagine a more circular reasoning. Yet soon after this Burt was made the official psychologist of the London County Council, responsible for the administration and interpretation of mental tests in London's schools.
The fact that further damning evidence against Burt can emerge from a book whose contributors all seem fairly sympathetic to his general views deals a significant blow to current attempts to rehabilitate him. However, it also limits the usefulness of this book for those seeking to challenge the whole notion that intelligence is largely genetic. A claim made repeatedly in this book, even by those contributors who are most critical of Burt, is that the conclusions he reached about the hereditability of intelligence retain their validity despite his exposure as a fraudster.
In support of such a claim it is said that there are many studies other than Burt's which support the idea that intelligence is inherited. Yet the credibility of such studies has already been destroyed by critics like Kamin, not on the grounds of fraud, but because they do not stand up to serious scientific scrutiny.
One of the reasons that Burt's study was so often quoted was because it claimed no similarity between the social circumstances of the families that raised separated twins. In contrast, in the two other early studies of this kind, it has become apparent that twins supposedly separated at birth were not raised apart at all. They were raised by members of the same extended family, in the same village. They played together and went to school together.
Other adoption studies of IQ which are said to demonstrate the effect of genes have their own experimental difficulties, including the failure to match children by age, extremely small sample size, and biased selection for study. It is claimed in this book that the most recent study meets many of the criticisms levelled at previous work. Yet although its investigators claimed the twins they studied had grown up entirely separately, they had previously reported that twins apparently separated at birth wore the same number of rings on their fingers, called their dogs by the same names and married women with the same names, which would imply some social interaction between the twins at some point in their lives--unless the authors are seriously suggesting there is a hereditary gene for naming your dog.
If IQ studies are flawed in their scientific method, a much more fundamental flaw is the basic assumption that inspires them. This is the notion that intelligence is a natural attribute, a fixed quantity, marginally affected by the social environment but essentially unchangeable. In fact such a view is rejected nowadays by most serious psychologists. 'As quaint as the archaic belief in the four humours that control human temperament,' is how Michael Howe, psychology don at Exeter University, describes it.
Most scientists studying the way human minds work now agree that intelligence and human thought processes are multi-faceted and complex phenomena. It has been demonstrated that there is no automatic connection between ability in one area and another. And undermining the IQ testers' claims to measure general intelligence, recent research shows that children can easily be trained in the skills needed to score high marks.
Of course it would be madness to say there is no such thing as natural ability. But the truth is that such abilities are dramatically shaped by the social environment. An ordinary primary school student today can correctly add a column of figures vastly more quickly than the most intelligent Ancient Roman mathematician, who had to struggle with cumbersome Xs, Vs and Is. The real problem with selective education is that it limits most children's learning environment and stifles their development before it's even begun.
Cyril Burt--Fraud or Framed?
by N Mackintosh (Oxford University Press £19.99)