The IMS Bulletin published in Volume 51, Issue 3, an extensive and thorough account of the scientific work of Sir David Roxbee Cox (David, for short). Written by Heather Battey and Nancy Reid, it sketches David’s contributions to stochastic processes and statistical inference; these are indeed the fields where most of his scientific work is situated.
I had close contact with David for a period of about 50 years. In this complement to his obituary, I would like to add a mention of an influential book, written in a somewhat different spirit, Multivariate Dependencies: Models, Analysis and Interpretation, by David Cox and Nanny Wermuth, published in 1996.
Let me now turn to an aspect which I will call David’s personality. By this, I mean that in his judgments and statements he was always absolutely independent, open, rigorous and honest. He would never follow a trend. I am going to illustrate this by three examples.
Firstly, the doubly stochastic Poisson process, which we both worked on, although from different angles. It was me who called it the “Cox process”. David protested; he did not want the credit. Indeed, a relatively unknown author had already used such a construction in a very particular problem of applied statistics.
The second example concerns the First World Congress of the Bernoulli Society in Tashkent in 1982. I was Chairman of the Program Committee and I asked David to be Deputy Chairman. His decisions on the choice of speakers were, of course, excellent. Above all, his judgments about possible speakers were never influenced by non-scientific factors, for example by “This person belongs to such-and-such school, or to such-and-such country,” or by “He or she is a friend of an influential person.” Not surprisingly, the Program Committee was indeed a few times under pressure to take such factors into account.
Finally, David was able to change his mind when confronted with new facts; he was not afraid of being regarded as fickle. For example, in 1981, Argentina had invited the International Statistical Institute (ISI) to hold the 43rd World Statistics Congress in Buenos Aires. This would have been during the military dictatorship in Argentina which lasted from 1976 to 1983. David and I were in a committee that had to advise the ISI whether to accept this invitation. Everybody, including David, was against it, except me. I had worked in Argentina and I argued that the young statisticians there had very much looked forward to this congress because it would allow them to break up their isolation. David changed its mind and voted for holding the Congress. It ended up taking place and was useful; I do not think that the military government profited much from it.