Statistics

Statistics at Princeton was the creation of Luther P. Eisenhart and Samuel S. Wilks. Its history falls naturally into four periods: Wilks alone, World War II, postwar under Wilks, the early years as a department.

Dean Eisenhart's vision brought Wilks to Princeton's Department of Mathematics in 1933. For the next eight or ten years, Wilks was the lone statistician in that department, building a sequence of courses, stimulating undergraduates, and starting to turn out Ph.D.'s. Much of his important research was carried out and published during this era. His influence on a variety of areas of application grew rapidly, ~~and his editorship of the Annals of Mathematical Statistics carried that journal through its crucial and formative decades. His pattern of ``write mathematics, and do applications'' became well established and was thoroughly conveyed to his students.

During World War II Wilks became deeply involved in the activities of the National Defense Research Committee, both in the Applied Mathematics Panel and at Princeton, where SRG-P (Statistical Research Group-Princeton) and its branch at Columbia (``SRG-P Jr'') involved graduate students T. W. Anderson, P. J. McCarthy, F. Mosteller and D. F. Votaw, and brought in R. L. Anderson, W. G. Cochran, A. M. Mood, L. S. Savage, J. W. Tukey, J. D. Williams, and C. P. Winsor for varying periods. (Many of these names were to become familiar to statisticians in the decades that followed.)

Merrill Flood became responsible for the Fire Control Research Group, involving in addition to mathematicians, physiologists, and engineers, C. P. Winsor, an engineer-turned-physiologist-turned-statistician, J. W. Tukey, a chemist-turned-topologist-turned-statistician, and G. W. Brown, a mathematician-turned-statistician.

During these years the flavor was, of course: ``do crucial applications, and any mathematics that can help.''

This increase in the number of statisticians made statistics much more visible in Princeton (and vice versa) and must have played a part in the administration's approval of a Section of Mathematical Statistics in the Department of Mathematics. This section, which initially consisted of Wilks, Tukey (part-time Princeton, part-time Bell Telephone Labs), and usually a visitor, grew in 1956 by the addition of Francis J. Anscombe (a former visitor). The old flavor of ``write mathematics and do applications'' continued, although some were now much more willing not to look to mathematicians as the source of approval. Undergraduate courses and interest grew somewhat, and there were steady streams of Ph.D.'s and faculty research.

The Korean War stimulated the establishment of the Analytical Research Group, of which Forman S. Acton was director from 1952 to 1956, and about half of whose work was statistical. George E. P. Box was then brought to Princeton as director of the Statistical Techniques Research Group (1956-1959), whose participants included E.M.L. Beale, J. Stuart Hunter, Colin Mallows, Mervin E. Muller, and Henry Scheff‚. Again there was a larger critical mass and mutual stimulation, with a further broadening of interests and a further recognition of mathematics as important but not all embracing.

After Wilks's death in 1964, with encouragement from some mathematicians and the administration, an independent Department of Statistics was formed in 1966. Tukey was its (part-time) chairman from 1966 to 1969 and after an administrative review and approval of expansion, he was succeeded by Geoffrey S. Watson in 1970. A new stream -- undergraduate majors -- joined the streams of Ph.D.'s, faculty research, and involvement in statistical applications; emphasis on undergraduate study became comparable with the emphasis on graduate study.

Frederick F. Stephan, who had come to Princeton in 1947 in Sociology, transferred to the new department, broadening its interests significantly. J. Stuart Hunter, who had returned to teach in the School of Engineering in 1968, also took an active role for a time. John Hartigan returned in 1964 to help hold the fort and, after Watson's arrival, the department was strengthened by Donald McNeil (1971-1976) and Peter Bloomfield (1971, acting chairman 1976-1977) among others.

Tukey's interest in data analysis (regarded as encompassing statistics), now that the pressure of needing to teach ``mathematics'' was removed, led first to a freshman course in Exploratory Data Analysis and then to a book (preliminary edition 1970-71, first edition 1977) emphasizing simple arithmetic and careful thinking. In December 1974 a PDP 11-40 computer was installed. Both these steps illustrate the continued broadening in the Department's attitude, where very mathematical questions remain important but other Ph.D. theses focus on specific applications or even on the analysis of important sets of data.

Robustness (in the technical sense of that word) and time series continue to be strong interests in the department, as they have been for decades. More recently, collaboration with the Office of Population Research and the Center for Environmental Studies has been prominent. The department is proud that the first two chairmen of the Department of Statistics, both at Harvard and at Yale, came from Princeton and that, in the early seventies, those of its undergraduate majors who wished to do their graduate study in statistics were able to do so at Harvard, Stanford, or Yale.

John W. Tukey


From Alexander Leitch, A Princeton Companion, copyright Princeton University Press (1978).

Go to Search A Princeton Companion