David Cox was one of a kind: a brilliant scientist with encyclopedic knowledge of statistical theory and methods; and an unassuming man who would always find time to help others with their research. Soon after his death last year, I wrote a very brief account of David’s work (Firth et al., 2022), including mention of the immense support and encouragement that he gave to younger statisticians over several decades. Here I make some related notes of a more personal kind, about my late friend and academic ‘father.’
My first encounter with David was when he gave the first-day welcome lecture in the MSc Statistics course at Imperial College. While that memory is hazy, I recall what I think was his main message: to do well in the MSc, and specifically to go beyond simply mathematics (that is to say, to learn how to think statistically), we students would all need to work very hard. David not only had a brilliant mind, he also had a visibly strong work ethic and was a superb role model in that regard. Some years later when he had agreed to supervise my PhD, I quickly learned that the best time of day to meet with David Cox was before 9 a.m.—not just because his secretary would arrive at 9 with a diary full of other demands, but also because David started early and he liked to see students doing the same.
Further into the MSc course came my first experience of David’s legendary handwriting. David gave lectures on components of variance, in such contexts as split-plot, split-split-plot and yet more exotic study designs. Those lectures were sparkling, with lots of ideas I had not met before; but they were hard to follow, because his writing on the blackboard was awful. Every word he wrote would start out small and then tail off into invisibility. Thankfully, David always seemed pleased when asked for clarification, and so that became a habit for me; and then he would write more legibly for a few seconds at least.
During my PhD years, David Cox spent a period working in Australia. His arrangement with me, as with other PhD students at the time, was that I would send a short report on progress and any difficulties I faced, and he would write back. That was all done via hand-written airmail letters; a there-and-back cycle would take 2 weeks if we each replied promptly. My ‘short’ report was invariably at least two pages long, probably because I felt the need to reassure David that I was working hard! His reply was typically just a few lines, but that was plenty. He had read my report, and he would usually either encourage me to continue along the same path or else gently suggest “You might like to look at xxx...” to tell me that I should be doing things differently. The “xxx” was sometimes old work from the 1930s, but equally often a paper that had just appeared in a recent issue of one of the journals. David seemed to know about absolutely everything. (His letters from Canberra also brought together his PhD students as a mutual support group in London: we needed one another’s help to decipher his writing!)
That was all in the 1980s. Fast forward now to 2007, with David in his 80s. A dozen or so statistics research groups in the United Kingdom and Ireland had come together to form a combined PhD first-year teaching program, the Academy for PhD Training in Statistics (APTS). We had the idea of involving David Cox somehow, if he was willing. It turned out he was very willing. David gave the inaugural course of APTS lectures, on Statistical Inference (based on Cox, 2006) to a group of around 80 new PhD students. The students loved it, and so did David. It was exhausting for him, but his enthusiasm for the subject and for educating young researchers carried him through. His lectures were wonderful (and remarkably, David’s blackboard writing had somehow become legible!). David’s contribution in those early years of APTS was huge, and the fact that APTS thrives (now with around 180 students each year) is due in large part to his inspirational work at the start.
Probably I am not alone in still being guided by David, even now. When unsure about any idea, I often just ask myself ‘What would David Cox do?’
David Firth has no financial or non-financial disclosure to share for this article.
Cox, D. R. (2006). Principles of Statistical Inference. Cambridge University Press. https://doi.org/10.1017/CBO9780511813559
Firth, D., Reid, N., Mayo, D., & Battey, H. (2022). Remembering Sir David Cox, 1924–2022. Significance, 19(2), 30–37. https://doi.org/10.1111/1740-9713.01632
©2023 David Firth. This article is licensed under a Creative Commons Attribution (CC BY 4.0) International license, except where otherwise indicated with respect to particular material included in the article.