BIOGRAPHY 12.1 William S. Gosset (1876 -1937)
William Sealy Gosset was born in Canterbury, England, the latest descendant of an old Huguenot family that had left France after the revocation of the Edict of Nantes. He studied at Winchester, then at Oxford, where he focused on mathematics and natural sciences. Upon graduation, he joined Arthur Guinness and Son, a Dublin brewery, and he remained employed there throughout his life, ultimately becoming chief brewer at a new brewery in London.
Early on, Gosset saw a need for careful scientific analyses of a variety of processes, from barley production to yeast fermentation, all of which profoundly affected the quality of the brewery's final product, beer. His firm sent him to study under Karl Pearson (Biography 14.1) at the University College, London. At the time, the theory of estimation based on large samples had been fully worked out, but Gosset noticed a void with respect to small-sample estimation theory. Small samples, however, were typical of Gosset's work at the brewery. So Gosset developed the theory himself. In a now famous 1908 paper, The Probable Error of a Mean, he noted that s is an erratic estimator of s when n is small; hence, customary measures of the precision of estimates were invalid for small samples and unknown s. His paper presented the sampling distribution of a statistic now known as Student's t and introduced small-sample estimation by means of the t-distribution family. It is difficult to overestimate the importance of this achievement. It has proven fundamental to statistical inference as it exists today, not only in the realm of estimation, but also in hypothesis testing and the analysis of variance. Sir Ronald Fisher (Biography 13.1), who held great admiration for Gosset and shared with him a keen interest in agricultural experimentation, quite aptly called Gosset "the Faraday of statistics," for Gosset had a similar ability to grasp general principles and develop them further by applying them to practical ends.
All but one of Gosset's numerous papers were published under the pseudonym "Student" to protect the advances in his firm's quality control from nosy competitors. For many years, an air of romanticism surrounded the appearance of "Student's" papers, and only a few individuals knew his real identity, even for some time after his death.
Sources: Dictionary of Scientific Biography (New York: Scribner's, 1972), pp. 476-477; International Encyclopedia of Statistics, vol. I (New York: Free Press, 1978), pp. 409-413.
Copyright © 2003 South-Western. All Rights Reserved. Disclaimer