It is with great sadness that the statistical community learned of the death of Sir David Cox in January 2022. Throughout the world, tributes have been paid to David’s exceptional achievements in using statistics to advance science and human welfare. To honor his memory and capture David’s unique recognition in the statistical world, Harvard Data Science Review commissioned this special theme.
David Cox received his education at St. John’s College, University of Cambridge and a PhD in 1949 from the University of Leeds. After several years of research work in the wool industry, he became a lecturer in mathematics at the University of Cambridge as well as a visiting researcher at Princeton University and University of Berkeley. After 10 years at Birkbeck College, University of London, he then worked as a professor of statistics for more than 20 years at the Imperial College of Science and Technology in London and as Head of the Department of Mathematics before he moved in 1988 to become Warden of Nuffield College in Oxford, England. He continued there with intensive research until his sudden death in 2022. His wife, Joyce, and their four children, Andrew, Joan, Stephen, and John, survive him.
The outstanding contributions by David Cox to the theory and applications of statistics have a pervasive influence and have opened up entire new research areas. His written work has a profound influence on science at large. For instance, the introduction of what now is called ‘Cox regression’ in survival analysis has started a research area with now many books and thousands of papers on statistical theory and on statistical practice. It has changed the way survival studies in medicine and technology are performed and evaluated.
His many (over 20) monographs have been translated into French, German, Japanese, Russian, and Spanish. He has supervised more than 60 doctoral students and has helped and encouraged many more; many of whom have become leading researchers themselves. He has also collaborated with and inspired numerous colleagues. In addition, he has served the profession extensively, for instance as President of the Bernoulli Society, President of the Royal Statistical Society, and President of the International Statistical Institute—and most significantly as editor of Biometrika for 25 years.
David was awarded the inaugural International Prize in Statistics in 2016. Commenting on this award, Susan Ellenberg, the chair of the International Prize in Statistics said: “Professor Cox changed how we analyze and understand the effect of natural or human-induced risk factors on survival outcomes, paving the way for powerful scientific inquiry and discoveries that have (had an impact on) human health worldwide” (“International Prize,” 2016).
In this special theme, former students and close colleagues share their personal recollection of interactions with David Cox. In addition, we asked young colleagues to contribute a short note, those who had received the Royal Statistical Society David Cox Research Prize: a prize that was bestowed by David Cox, using the money he received in 2016 with the International Prize in Statistics. David also requested that it be named the David Cox Research Prize only after his death.
The kaleidoscope of personal memories and reflections that have been assembled paint an impressionist portrait of Cox’s unique personality and ways of working. His modesty, his generosity with his time and encouragement of young researchers, his enthusiasm for new ideas, his incisive mind and sharp intellect, his quirky sense of humor, and his immense energy shine repeatedly through the recollected anecdotes.
In his contribution, David Firth (2023) recounts the “immense support and encouragement that [David] gave to younger statisticians over several decades,” an observation echoed by Vernon Farewell (2023), Dominik Rothenhäusler (2023), Thomas Berrett (2023), and Christiana Kartsonaki (2023). As commented by Peter McCullagh and Vernon Farewell, Cox encouraged his students to read wide, explore “all sorts of useful but not directly relevant topics” (McCullagh, 2023) and “foster[ed] exposure to the broad spectrum of statistical research” Farewell (2023).
Cox’s “synoptic view of statistics” beautifully qualified by Bradley Efron (2023) as “Mozartian” by reference to Mozart’s reputation for writing all the 17 staffs of a symphony at once, is recounted by many. Andrea Rotnitzky (2023) vividly describes how after a seminar where she posed a question for which she has no answer, Cox “offered a solution to a complicated technical problem that made it seem easy and intuitive.” Christl Donnelly (2023) comments that “[David] couldn’t really know everything statistical, but to me it seemed like it,” a thought shared by Christiana Kartsonaki (2023). Both Anthony Davison and Nancy Reid highlight how Cox “set the tone of the statistical community from the 1960s onward” (Davison, 2023) and “was instrumental in increasing the civility of our academic exchanges” (Reid, 2023), a legacy that benefited all of us. Through Neil Shephard (2023), we get a glimpse of the wisdom of Cox’s leadership style: “provide as much intellectual space and protection as possible for young scholars, add encouragement, urge people to finish things, inspire by personal example, … and provide a stream of interesting visitors.”
Cox’s own approach to applied problems is discussed by several of the contributors. Valerie Isham and Vernon Farewell concur that Cox proceeded naturally by “first [constructing] a ‘mechanistic’ model by using a clever and insightful stochastic formulation” (Isham, 2023) or “trying to simplify the essentials of a problem to a tractable analytic framework” Farewell (2023), while Jane Hutton (2023) comments on the importance Cox places on “[thinking] about assumptions.” Ruth Keogh (2023) describes Cox’s preference for a pragmatic, or as he might have said “sensible” approach, and his “great ability . . . to break things down to the essential components.” Many comment on his ability to ask searching questions even if he appeared to be asleep, his writing style, “elegant and instantly recognizable” (Isham, 2023), and his love for concision (Herzberg, 2023). From Chengchun Shi (2023) and Tengyao Wang (2023), we see that Cox’s view on how fruitful serious contact with applications can be, even for a theoretical statistician, has been transmitted to the new generation.
We also learn many personal traits, such as his legendary handwriting where “every word he wrote would start out small and then tail off into invisibility” (Firth, 2023), his pride in optimizing journeys on the London tube (Donnelly, 2023), his “understated sense of humor; a kind of mild-mannered satire that teetered on the boundary between seriousness and absurdity” (Battey, 2023). His kind-hearted nature shone through an anecdote recounted by Andrea Rotnitzky (2023): “he found greater joy in attending a ceremony in honor of a friend than his own.” The calm and wisdom which he exercised on many occasions when faced with potentially controversial policy issues is also clearly brought out in Christl Donnelly's (2023) quote, “Don’t worry about politics. The science will endure.”
Nanny and Sylvia’s Memories
Both authors of this editorial had the privilege to interact with David over a number of years. It was a pleasure for Nanny to cooperate and publish with David for more than 30 years. For the first time in 1988, Nanny met David for hours of discussions at Imperial College when David had invited her. This was shortly after her youngest son had started school. To the later meetings in Oxford, she flew Pan Am, which left from Frankfurt at about 5 a.m. She returned the same night, shortly before midnight, so that she would not be away from her four young sons for too long. After the terrible crash at Lockerbie in December 1988, she was often almost alone on these planes. At the end of 1991, Pan Am had to stop flying completely, but this did not stop the meetings of David and Nanny.
When David came to Mainz, Germany, often together with his wife Joyce, more time was available for events aside from work. It was excellent for Nanny to see them both enjoying their stays. Already in 1992, David and Nanny received one of the prestigious Max Planck Research Awards, which kept them going to finish their 1996 book. With the additional support of the German and, later on, the Swedish Research Society, a number of research workshops with 10 to 20 colleagues were organized to meet and discuss for several days: a marvellous opportunity. For more details on their cooperation, see Nanny’s forthcoming 2023 paper in the Scandinavian Journal of Statistics, “Being inspired by David Cox.”
Sylvia met David for the first time in 1989 while they were both visiting Cornell University. Faithful to tradition, David promptly fell asleep during her seminar, which was somewhat unsettling! In the mid-1990s, Sylvia and her husband spent a memorable evening with David while he was giving a series of lectures in Paris. They took him to a chamber music concert, which David’s enjoyed though mildly remarked that it was on the long side. . . , then to a restaurant where he very much enjoyed discussing French ways of life. David was fond of the Medical Research Council Biostatistics Unit and kindly showed his support at key times during Sylvia’s directorship by sending her little notes as well as attending the Unit’s Armitage events, where he particularly enjoyed talking to the PhD students (see picture captioned “Sir David Cox featured with PhD students from the Medical Research Council Biostatistics Unit's 14th Annual Armitage Lecture & Workshop” in slideshow below). David’s passionate interest in the future of statistics and biostatistics never waned.
We thank all the contributors for giving all of us a deeper insight into David Cox’s unique legacy.
Disclosure Statement
Sylvia Richardson and Nanny Wermuth have no financial or non-financial disclosures to share for this editorial.
Cox, D. R. & Wermuth, N. (1996). Multivariate dependencies: Models, analysis and interpretation. Chapman and Hall. https://doi.org/10.1201/9781498710398
Keogh, R. H. (2023). D. R. Cox: Approach to applied statistics and the importance of being sensible. Harvard Data Science Review, 5(2). https://doi.org/10.1162/99608f92.8e56c54f
Creative Commons Attribution 4.0 International License (CC-BY 4.0)
