Sir Ronald Aylmer Fisher

Sir Ronald Aylmer Fisher

The English statistician Sir Ronald Aylmer Fisher (1890-1962) introduced fresh ideas into the planning and interpretation of quantitative biological experiments. He was a pioneer in the mathematical theory of genetics.

Ronald Fisher was born in London on Feb. 17, 1890, and was educated at Cambridge University, where he specialized in mathematics and physics. In 1919 he was appointed to the new post of research statistician at the Rothamsted Experimental Station in Hertfordshire. His primary task there was the analysis and reinterpretation of a 66-year backlog of records on continuous agricultural experiments and associated meteorological data. In discharging this duty he revolutionized existing statistical techniques, and he expressed the new outlook in Statistical Methods for Research Workers (1925), which was to become, and to a large extent remains, the bible of applied statistics.

A central theme of Fisher's work in statistics was hypothesis testing. Many experiments, especially biological ones, are essentially devised to check whether or not some agent has a determinate effect on a test organism. This necessitates a direct and fair comparison between the organism tested with the agent (test organism) and the organism without it (control organism). Because of normal biological variation, it is unlikely that the two organisms will be the same even if the agent is wholly ineffective. Therefore the question arises whether an observed difference, in favor of the test organism, is entirely due to chance variation or is mainly due to the agent. Fisher termed the basic "no effect" situation the null hypothesis, which the experiment is designed to check. Having satisfied himself that the null hypothesis is untenable, the experimentalist can go on to make quantitative estimates of the effect of the agent—especially if he has been wise enough to test it at different levels of application.

Another and cognate area that Fisher transformed was the planning of experiments in which considerable variation of results is to be expected and in which ancillary factors have to be taken into account. For example, in crop fertilization trials highly localized soil variations may exist, and Fisher developed schemes for the random allocation of subplots in ways that such nuisance factors are minimized without harm being done to the probabilistic model on which the hypothesis testing depends.

Darwinism versus Mendelism

Aside from his practical work on the methodology of experimentation, Fisher wrote perceptively on the rationale of statistical inference. He also established the exact distribution of several important statistical functions. At the same time he was probing a quite different matter: the quantitative side of the theory of natural selection. He became a leader in the reconciliation of Darwinism and Mendelism, and in 1930 he published his second great book, The Genetical Theory of Natural Selection. Ever since the rediscovery of Gregor Mendel's experimental evidence for particulate inheritance and the subsequent fashioning of the theory of genes, there had been an uneasy feeling that these ideas did not tie in with Charles Darwin's theory of natural selection. Fisher was one of the first to tackle the tough mathematical problems and conceptual difficulties in this area. He discovered just how the frequency of particular genes in a given population will fluctuate under the influence of natural selection. In the book he also put forward his views on eugenics; it is a classic of population genetics.

An important by-product of Fisher's work on genetics was his practical and theoretical interest in human blood grouping. In 1935 he set up a blood-grouping unit in London, and one of the outcomes was the unraveling of the mode of inheritance of Rhesus groups.

Fisher left Rothamsted in 1933 and moved on to professorships in London and, later, Cambridge University. He had been made a fellow of the Royal Society in 1929 and was knighted in 1952. After his retirement he emigrated to Australia, where he died on July 29, 1962.

Further Reading

A detailed biography of Fisher appeared in the Royal Society, London, Biographical Memoirs of Fellows of the Royal Society, vol. 9 (1963). Fisher's work in statistics is discussed in James R. Newman, ed., The World of Mathematics (4 vols., 1956), and Lancelot Hogben, Statistical Theory: The Relationship of Probability, Credibility and Error (1957).

Additional Sources

Box, Joan Fisher, R. A. Fisher, the life of a scientist, New York: Wiley, 1978. □