Economists are accustomed to drawing conclusions from statistical studies and aggregations of data. It's hard to reduce economic behavior to controlled, double-blind studies, and laboratory rats aren't necessarily a good model for corporate rats. So it came as no surprise to me that some Cornell University economists thought they might get a handle on the elusive cause of childhood autism by studying rainfall and the availablity of cable television. Working from the assumption that children spend more hours watching television in households that have cable TV, and in locations where high rainfall keeps them indoors, and observing significantly higher rates of autism in communities with a confluence of those conditions, the researchers suggested early television viewing as a possible trigger for autism spectrum disorders.
When I first read about the study, I was reminded of a story Peter Drucker tells, in his marvelous, autobiographical, historical commentary, Adventures of a Bystander, about an outstanding statistics teacher at the University of Minnesota.A little bald-headed bearded man, almost a dwarf, he ran a doctoral seminar in which he projected tables and graphs onto a screen without any legend or explanation. "Look at the figures," he'd say, "and tell me what they tell you." The students would point out this irregularity in a distribution or that periodicity, this pattern or that internal contradiction, and the little man would nod, smile, argue, and get across a great deal about numbers as a grammar without even belaboring the point. Then he'd flash on the screen two series of figures that obviously belonged together; they showed a close correlation, almost one to one, over long time periods. "Clearly," all the students said, "these two series are causally related." "That's what every statistician would say," the little man responded, "but perhaps you can tell me what the relationship could be. This series," pointing to the table on the left, "is the annual herring catch off Newfoundland; and that one," pointing to the right, "is the number of illegitimate children born the same year in North Dakota."
It may be that the Cornell professors' study has produced only a "correlation" of this type. However, the really interesting story is not the report, but the reaction to it. Television promoters and viewers are digusted; parents of autistic children are offended; and the study has been mocked, ridiculed, and satirized. It is not true because it cannot be true. Let someone report a study that indicates one race or sex might be inherently predisposed to excel another in some area, and he will find he has called the Furies down upon his head. Dare to suggest that day care might be harmful to children and your research will be attacked, because such a conclusion hits us in our most vulnerable place. The famed science fiction writer, Isaac Asimov, showed no patience for those who base their beliefs on faith rather than on objective science; nonetheless, I heard him say that he believed the universe to be cyclical, ever expanding and contracting, rather than a single expansion after one Big Bang, solely because he did not want to believe that the universe would ever come to an end. Modern American society does not want to believe that television might be harmful; therefore any such suggestion must be laughed out of court.
With my own anti-television bias, I must be careful to resist the temptation not to recognize the limitations of such statistical research. I will, however, state that it should not be laughed at. No one else has come up with a convincing explanation for the alarming rise in autism cases, and the researchers' ideas deserve serious consideration—especially in light of their observation that the Amish, whose distinctives include not watching TV, have unusually low rates of childhood autism. This study does not prove that allowing young children to watch television for long hours increases their risk of developing autism. But it is a possibility worth further investigation.
