In 1933, Bartlett was recruited by Egon Pearson to the new statistics department at University College, London. Pearson was already working with Jerzy Neyman. Also in the college were Ronald A. Fisher and J. B. S. Haldane. Bartlett was stimulated by all of them, most of all by the work of Fisher, criticising some of it (for example, fiducial inference) while developing other parts (for example conditional inference). Relations between the two men fluctuated; sometimes Bartlett was in Fisher's good books, but often not. In 1934, Bartlett became statistician at the Imperial Chemical Industries (ICI) agricultural research station at Jealott's Hill. Not only did he deal with practical problems but he worked on statistical theory, as well as on problems in genetics but he became interested in the characterisation of intelligence. He remembered Jealott's Hill as the best working environment of his career. Bartlett left ICI for the University of Cambridge in 1938 but at the outset of World War II was mobilised into the Ministry of Supply, conducting rocket research alongside Frank Anscombe, David Kendall and Pat Moran.
After the war Bartlett's renewed Cambridge work focused on time-series analysis and stochastic process. With Jo Moyal he planned a large book on probability, but the collaboration did not work out and Bartlett went ahead and published his own book on stochastic processes. He made a number of visits to the United States. In 1947 he became professor of mathematical statistics at the School of Mathematics at the University of Manchester where he not only developed his interests in epidemiology but also served as an able and active administrator. In 1960, he took up the chair of statistics at University College, London before serving the last eight years of his academic life as professor of biomathematics at the University of Oxford. He retired in 1975.
After his retirement Bartlett remained active in statistics, visiting the Institute of Advanced Studies at the Australian National University several times. He had married Sheila, daughter of C. E. Chapman, in 1957, the couple parenting a daughter. Bartlett died in Exmouth, Devon.
(1934) The vector representation of a sample. Proc. Camb. Philos. Soc., 30, 327–340.
(1936) Statistical information and properties of sufficiency. Proc. Royal Soc. Lond. A 154, 124–137.
(1937) Properties of sufficiency and statistical tests. Proc. Royal Soc. Lond. A, 160, 268–282. (reprinted with an introduction by D. A. S. Fraser S. Kotz & N. L. Johnson (eds) Breakthroughs in Statistics, volume 1. Springer, New York. 1992.)
(1938) Methods of estimating mental factors. Nature, 141, 609–610.
(1939) A note on tests of significance in multivariate analysis, in Proceedings of the Cambridge Philosophical Society
(1941) The statistical significance of canonical correlation. Biometrika.
