The New Challenge : From a Century of Statistics to an Age ofCausationJ

Some of the main users of statistical methods { economists, social scientists, and epidemiologists { are discovering that their elds rest not on statistical but on causal foundations. The blurring of these foundations over the years follows from the lack of mathematical notation capable of distinguishing causal from equational relationships. By providing formal and natural explica-tion of such relations, graphical methods have the potential to revolutionize how statistics is used in knowledge-rich applications. Statisticians, in response, are beginning to realize that causality is not a metaphysical dead-end but a meaningful concept with clear mathematical underpinning. The paper surveys these developments and outlines future challenges. 1 A Century of Denial Francis Galton's discovery of correlation, at the end of the nineteenth century Galton, 1888], dazzled one of his students, Karl Pearson, generally considered the founder of modern statistics. The year 1911 saw publication of the third edition of Pearson's The Grammar of Science , which contained a new chapter titled \Contingency and correlation-the insuuciency of causation." This is how Pearson introduces the new topic: \Beyond such discarded fundamentals as`matter' and`force' lies still another fetish amidst the inscrutable arcana of modern science, namely, the category of cause and eeect" Pearson, 1911, p. iv]. And what does Pearson substitute for the archaic notion of causation? Correlations and contingency tables!! He states (ibid, p. 159), Such a table is termed a contingency table, and the ultimate scientiic statement of description of the relation between two things can always be thrown back upon such a contingency table. : : : Once the reader realizes the nature of such a table, he will have grasped the essence of the conception of association between cause and eeect. Thus, Pearson categorically denies the need for a concept of causal relation independent of or beyond correlation. He held this view throughout his life and, accordingly, did not mention causation in any of his technical papers. His objection to animistic concepts such as \will" and \force" was so erce and his rejection of determinism so absolute that he exterminated causation from statistics before it had a chance to take root. Pearson's crusade innuenced markedly the direction of statistical research and education in the twentieth century, also known as \The Statistical Century." The Encyclopedia of Statistical Science Kotz and Johnson, 1982], for example, devotes 12 pages to correlation but only 2 pages to causation, and one of those pages is spent …