Sir Ronald Aylmer Fisher

 

Sir Ronald Aylmer Fisher was born February 17, 1890 in East Finchley, London to a fine arts auctioneer named George Fisher and Katie (Heath). None of his ancestors showed any strong mathematical inclinations, with the small exception of his uncle who graduated as a Wrangler from Cambridge. As a child, Fisher’s eyesight was already poor and this inhibited his private reading. In 1904, when he was 14 years old, his mother died, and that same year he received a scholarship in mathematics to Harrow. Because of his eyesight, he received tutoring in math under G. H. P. Mayo. In 1909 he won a scholarship in mathematics to Cambridge University and graduated in 1912 with a B. A. in Astronomy. He was awarded a studentship for one year and studied theory of errors under F. J. M. Stratton and also more astronomy and quantum physics under J. Jeans.

In the spring of 1911, the formation of the Cambridge University Eugenics Society was born at Fisher’s urging. He served on the Council and also spoke there on the theories of the day; Darwin’s theory of evolution and natural selection. Fisher had a great interest in evolutionary theory, especially genetics. This was where much of his statistical interest came into play, both at Cambridge and throughout the rest of his life.

In April of 1912, Fisher published his first paper introducing what would later become known as the method of maximum likelihood. The remainder of that summer, Fisher took up correspondence with W. S. Gosset ("Student") in relation to a part of his equation for standard deviation. During this correspondence, he ended up making the distinction between sample and population means. In the theoretical or pure statistics, large samples, sometimes even infinite (using ideal situations), were taken from the population. Almost the entire population itself was taken as the ideal sample, making little error, but making applying those techniques impractical. R. A. Fisher was interested in the practical applications of statistics, dealing with small quantities in a sample, not infinitely large ones.

He also recognized degrees of freedom and proved a geometrical formulation that Gosset had been working on. Some general admiration for Fisher’s work by Gosset and others was tempered with the fact that it was apparent that Fisher did not like to make mistakes, and had a hard time admitting that he had. This insistence, however, drove Fisher to redraw and explain in many different terms the points he was trying to make, and in the process he thought of theories that he and others would expand upon later on.

Upon graduating, Fisher took a statistical job with the Mercantile and General Invesment Company in the City of London while training for the Territorial Army in 1913. In 1914, Fisher began teaching high school mathematics and physics in order to serve his country as he was rejected from the army because of his eyesight. At this time, his statistical work drew the attention of Karl Pearson, a famous statistician at the time.

Pearson published Fisher’s article on general sampling distribution and directed studies in his own labs to create a cooperative study. This study was published in 1917, including a criticism of Fisher’s paper that he was not informed of. Fisher was little known then, and the fact that Pearson had gone over his head and criticized the method of maximum likelihood that Fisher had used, and labeled it inverse inference, something Fisher had deliberately tried to avoid. Fisher had gone out of his way to avoid using inverse inference and the fact that Pearson had not understood him injured his pride. This eventually led to a violent confrontation between these two. Pearson then ignored Fisher’s re-explanation and also some other very important and brilliant theories. Fisher was drawing attention to a mistake Pearson was making in degrees of freedom in specific tables, and Fisher was actually right. However, because of Pearson’s fame, the papers regarding these theories were withdrawn from the Royal Society’s consideration in 1916 because of the tension between Fisher and Pearson. Fisher had even tried a new method of proof, knowing that Pearson disagreed with his original explanation, but these were also ineffective. In 1919, Pearson offered Fisher a job in his department, but Fisher declined because of these reasons.

Instead, Fisher took a as a statistician at Rothamsted Experimental Station which was mostly involved in agricultural research. He started with a study on wheat and continued to work with all sorts of agricultural data, becoming well known in his field. Here he developed the analysis of variance, which brought into light the problems of not just the data but the actual experiments. This led to the science of experimental design and ultimately to his famous Statistical Methods for Research Workers.

In 1917, Fisher had married Ruth Eileen and in 1919 they had their first child, George, which Fisher became quite attached to. They had several more children, and they all lived the subsistence farm life, reading aloud and having deep conversations in the evenings. He self taught his children as long as possible pushing them to question and learn -- the ideals that he thought were so important. He tried often as possible to bring his children into his work, even allowing George to help him with his genetic mouse stocks.

The 1920’s were a time when Fisher tackled the problem of natural selection and population genetics. He wrote the famous Genetical Theory of Natural Selection. He truly believed in natural selection -- and that it should be studied by itself, not in correlation with evolution theory. His wife wrote the entire book down at home, and this took the place of the usual family discussions for quite awhile.

While at Rothamsted he had provided methods in statistics that researchers needed to deal with real life variation that was always present in agricultural and biological experimentation. He wrote two more very important papers, one in 1922 on maximum likelihood, and one in 1925 concerning small observational samples obtained from scientific experiments. He clarified the distinction between sample statistics and population values. These ideas gave researchers many tools to deal with variants, small sample sizes and more precise estimations.

In 1929, Fisher became a Fellow of the Royal Society as a mathematician and he then continued researching amidst summer teaching at Iowa State University in 1931 and 1936. In 1933, Pearson retired and Fisher accepted the position as Galton Professor of Eugenics at Cambridge University. There was some friction between his laboratory and the Statistics department, which was now run by E. S. Pearson and J. Neyman (who had also disagreed with Fisher earlier). The relation of the two departments continued to deteriorate.

The Rockefeller Foundation eventually set up a small unit for human serological research in 1935. With this, Fisher evaluated the Rh blood groups and in the process predicted two new antibodies. In 1939, Fisher’s department moved to Rothamsted after being forcibly evicted from London at the beginning of the war and as there was no work, they dispersed. With no unit or funding left, he accepted an election to the Balfour Chair of Genetics at Cambridge in 1943 and was offered a professorial residence. He was forced to move all of his genetic stock and equipment to the residence offered to him at Rothamsted and to leave his family in Harpenden. This year also brought tragedy as his son George died while servng his country in the Royal Air Force.

After the war, funding for research projects in general was limited, if not obliterated for some sections. He continued to support statistics and its practical uses in experimentation and also continued to explore his own varied interests. He took the opportunity to travel a bit and after he retired in 1957, he joined in research with E. A. Cornish in 1959 in Adelaide, Australia. He died there on July 29, 1962.

Among his numerous honors, memberships, and awards Fisher received throughout his lifetime, he received three medals of the Royal Statistical Society, the Darwin Medal (1948) and the Copley Medal (1956). He received honorary degrees from numerous Universities, and was a member in over 20 academies, societies, and institutes. He was even created Knight Bachelor by Queen Elizabeth in 1952.

With all of his contributions that shaped the world of statistics, he made it as we know it today. G. A. Barnard wrote that "to attempt, in a short article, to assess the contributions to the subject by one largely responsible for its creation would be futile.…He ascribed to the Student idea that, while there must always be uncertainty involved in statistics, this need not imply any lack of precision—the uncertainty may be capable of precise quantitative assessment…Fisher did much to give this idea form and reality."