Statistics prof questions 'The Bell Curve'

Three years after The Bell Curve commanded national attention, Stephen Fienberg, a statistics and social science professor at Carnegie Mellon University, offered a scientific rebuttal to the controversial book Tuesday in College Hall. His talk, "Intelligence and Success: Is It All in the Genes?" followed the recent publication of Intelligence, Genes & Success, a collection of responses to The Bell Curve which Fienberg contributed to and co-edited. According to Fienberg, the central question about The Bell Curve is, "Was there science there, and if so, what was it and can we rely on it?" After summarizing the argument presented by Murray and Hernnstein in The Bell Curve, Fienberg said that the book rests on a "syllogism"--if IQ is largely genetic and correlates with socio-economic status, then socio-economic status is genetically determined. Such an argument contains major public policy implications, Fienberg said. If The Bell Curve is correct, educational resources should be expended exclusively on the intelligent and government social services are a waste of money. But the most radical implication, according to Fienberg, is that the intellectual elite should withdraw into sheltered enclaves to protect themselves from the masses. In response, Fienberg talked about both the history of the book and the specific statistical measures it uses. Specifically, while Murray and Hernnstein take the existence of a single measure of intelligence, IQ, as a given, Fienberg said the subject is the source of much controversy in social science circles. He noted that both IQ and the statistical tests used in The Bell Curve have their roots in eugenics, the controversial science of race. According to many scientists, the problem with a single measure of intelligence goes beyond the IQ test itself. The prevalent theory today is that intelligence cannot be measured by any one number and that instead tests should measure a variety of types of intelligence. Fienberg questioned some of the statistical tests used in the book as well. The most problematic statistic in The Bell Curve, according to Fienberg, is Murray and Hernnstein's contention that genes determine 60 to 80 percent of a person's IQ. But this number includes the unvarying genetic influence common to everyone, producing a deceptively high result. If a person had a different set of genes, for example, the potential change in IQ would be no more than about 30 percent. This shifts the terms of debate from saying, "'There is a genetic component to intelligence,' to saying, 'Yes, but it's not so large'," Fienberg said. Referring to the argument that educating those with poor IQ scores is futile, Fienberg pointed to the so-called Flynn effect, a statistic which shows IQ scores to be rising worldwide -- an inexplicable observation if IQ is truly genetically based. In ultimately rejecting the policies advocated by Murray and Hernnstein, Fienberg stressed that "while we do take the arguments in The Bell Curve seriously, they're badly flawed and have little use for policy decisions in America?. Charles Murray is just not a statistician." After the talk, Mike Dresnack, a Wharton doctoral student, said he had come to hear a statistician's view of The Bell Curve and was convinced by the refutation. Another audience member said Fienberg's statistical arguments were beyond question but his opinions on public policy are purely subjective. Fienberg agreed, noting, "Real policy conclusions are not? something you can extract from data that will tell you 'yea' or 'nay'."

Please note All comments are eligible for publication in The Daily Pennsylvanian.