Reviewed by Leila Zenderland
Science, May 18, 1990 v248 n4957 p884(3)
LEILA ZENDERLAND Department of American Studies, California State University, Fullerton, CA 92634
Robert Joynson The Burt Affair
On 24 October 1976, the London Sunday Times published a front-page story headed "Crucial data was faked by eminent psychologist." "The most sensational charge of scientific fraud this century," it began, "is being levelled against the late Sir Cyril Burt. . . . Leading scientists are convinced that Burt published false data and invented crucial facts to support his controversial theory that intelligence is largely inherited." The ensuing scandal forms the subject of Robert Joynson's book. Joynson's research is likely to inspire at least one more round in this controversy, for he argues that Burt has been unjustly maligned.
Even in a field as prone to public controversy as intelligence testing, the Burt affair forms an exceptionally dramatic and disturbing episode. Burt, a brilliant mathematician, author of Factors of the Mind (1940), and editor of the British Journal of Statistical Psychology, had been a pioneer in educational psychology. His studies of intelligence also contained data on the largest number of identical twins raised apart--key evidence in debates between hereditarians and environmentalists. Knighted in 1946, Burt was Britain's most honored psychologist during his lifetime. As "long as psychology remains a subject of scientific inquiry," psychologist Leslie Hearnshaw noted in his eulogy following Burt's death in 1971, "he will live in its halls of fame" (p. 27).
By 1972, however, environmentalist Leon Kamin had begun to raise serious questions about Burt's science in answering hereditarian Arthur Jensen. Burt's articles on intelligence testing, Kamin observed, were suspiciously lacking in basic information, such as place, time, and type of test administered, and were filled with highly improbable coincidences, such as correlations that remained exactly the same--to three decimal places--even when sample sizes had more than doubled. By 1974, Kamin and Jensen had reached rare agreement: Burt's "correlations are useless for hypothesis testing," wrote Jensen; his numbers "are simply not worthy of our scientific attention," charged Kamin (p. 162).
Charging fraud was another matter. Kamin's writing, however, had interested Times reporter Oliver Gillie, who began searching for Burt's research assistants, Margaret Howard and Jane Conway. Finding no evidence of their existence, past or present, and being informed that these were probably pseudonyms invented by Burt, Gillie broke the story.
Caught in the ensuing crossfire was Hearnshaw, who at the time was working on a biography commissioned by Burt's sister. Hearnshaw agreed to examine the new charges, and when published in 1979 his study, Cyril Burt, Psychologist, described a gifted scientist whose early research had probably been genuine, but whose data, largely destroyed by wartime bombings, had probably been partly fabricated in postwar writings. Burt's personal and published papers since the 1940s, Hearnshaw concluded, suggested a pattern of deliberate deceit in claims about his role in his field's history, the quantity of new data collected, and the number of assistants helping him. In fact, of 40 different "authors" who published material in the journal Burt edited, over half may have been Burt himself, writing under pseudonyms. In light of Hearnshaw's findings, the British Psychological Society declared Burt guilty of fraud.
Joynson now proposes that the Burt investigation be reopened. The Society's actions, he argues, were premature, for the charges remain to be proven. Burt, he maintains, will be exonerated.
Joynson argues his case like a wily defense lawyer. Burt must be presumed innocent, he insists, until proven guilty beyond a shadow of a doubt. And the burden of proof must lie with the prosecution. Moreover, the standards of admissible evidence must be narrowed. Hearsay must be disallowed. Memories by contemporaries should be admitted only when accompanied by written documentation from the period in question. And all written evidence should be unambiguous. The defense has an easier task: it must prove only that one or more explanations besides fraud are possible.
Joynson's explanations for the many charges against Burt range from the plausible to the incredible. Burt, he argues, may simply have used outdated methods, or exercised poor judgment. Or he may have fallen victim to political enemies, jealous colleagues, or even a conspiracy of gossip-mongerers. (At one point Joynson shows that several of those who now believe Burt guilty, including hereditarian Hans Eysenck, had once worked at Maudsley Hospital, where they may have heard unsubstantiated rumors about Burt's untrustworthiness.)
In this case, the best defense is a good offense. Joynson's main target is Hearnshaw's biography, which he blames for turning the tide against Burt and contends is full of errors and unproven assumptions. He focuses on four of Hearnshaw's main charges: that Burt changed the historical record; that his kinship studies, with their suspicious correlations and missing assistants, suggest fraud; that he lied about the sources for his final papers; and that he suffered from mental illness.
Joynson spends much time examining Burt's historical claims. Did Burt, as Hearnshaw contends, try to diminish Charles Spearman's contributions to factor analysis to promote his own? Hearnshaw's book documents Burt's pattern of distorting evidence to exaggerate his own priority; Joynson argues with him, footnote by footnote. Burt may have become "less deferential" toward Spearman, Joynson concedes, but this is hardly an indictable offense.
The kinship studies present a more central challenge, and here Joynson is most unconvincing. Critics have found Burt's published claims difficult to challenge precisely because he is so vague in supplying details. Nonetheless, of 64 correlation coefficients reportedly calculated on new samples in Burt's 1966 study, as many as 30 are overtly suspicious, for they repeat figures found in earlier articles. The most famous of these--an intelligence correlation of .771 reported in 1955 for 21 pairs of identical twins raised apart, and in 1966 for 53 pairs--Joynson believes may have been "a genuine coincidence" (p. 155). As for the others, Joynson proposes, Burt may simply have inserted old figures if he had no new data, without realizing the need to mention that these came from different samples. In any case, Joynson accuses critics of paying too much attention to Burt's suspicious figures and too little to the 34 that are new. After all, Joynson reasons, if repeated coefficients suggest fraud, then "by the same logic we must also now argue that the appearance of a new coefficient suggests that the data are genuine" (pp. 156-157). Such logic speaks for itself.
In considering Burt's "missing assistants," Joynson endorses an explanation offered by Burt's former colleague Charlotte Banks. Like Hearnshaw, Joynson believes that Howard and Conway may have been volunteer social workers whom Burt met before the war. Both also agree that Burt probably conducted no major new twin studies after 1950. The data in Burt's later papers, Joynson proposes, were prewar materials--materials misplaced in Burt's many moves, gradually rediscovered, and then published. This was possible, he argues, for Burt's secretary was, in Banks's words, "very accurate but couldn't file for toffee, and was very sensitive to criticism." "I am sure he would have promised her not to say the material had been lost," writes Banks (p. 180). To Joynson, this story suggests "possible answers" explaining postwar publications credited to Burt's prewar assistants.
Joynson's most serious charge against Hearnshaw concerns Burt's final papers. In 1969, Burt published test results reportedly gathered between 1914 and 1965 showing school performance declining. Hearnshaw quotes an interview that supposedly appeared in The Guardian in which Burt claimed that his tests had been given regularly to hundreds of schoolchildren--a claim Hearnshaw then proves to be a lie. Joynson, however, has been unable to locate any such interview. Such a charge must be answered. Nonetheless, even if found, Joynson argues, the interview will not prove decisive, since newspaper accounts are notoriously unreliable.
Joynson's final strategy is to discredit Hearnshaw's explanations for Burt's behavior--mental illness and childhood influences. There is no independent evidence, Joynson contends, that Burt suffered any mental instability. Moreover, whatever survival instincts Burt manifested had probably been learned not from the "'gamin' subculture" that Hearnshaw claims he had known as a child living near the slums but from academic life, which "revolves around backbiting, innuendo, second-hand gossip, and abuse of confidence" (p. 256).
Such a defense leaves a strange effect. Joynson may believe that he has exonerated his client; Burt, however, hardly leaves this courtroom with his reputation intact. Ironically, Joynson's "innocent" Burt emerges as an even less likable character than Hearnshaw's "guilty" Burt. Such a verdict, however, is acceptable to Joynson; Burt's methods may have been less than admirable, he argues, but they were short of criminal.
Joynson's arguments are sure to invoke detailed rejoinders from those now called "anti-Burters." Of more concern than Burt's posthumous reputation, however, is the broader question of standards of evidence, both scientific and historical, raised here. Joynson's research contains nothing to challenge the current consensus that, as scientific evidence, Burt's data are unacceptable. Moreover, even if one believed Burt innocent of conscious wrongdoing, the fact that such data were used in debates over educational policies and went unchallenged until the 1970s would still be scandalous.
Historians, however, can rarely invoke such strict standards in admitting evidence. Unfortunately, like Hearnshaw, they must often draw their conclusions from incomplete records, ambiguous writings, and the memories of contemporaries--the same materials Joynson uses to construct his alternative explanations. Burt may never have received his day in court; his place in history, however, must now be judged by the work he left behind.
