“Who is the statistician or scientist you admire the most?”

As for the previous ISBA Bulletin, Luke Bornn sent me this (impossible!) question, to be answered in a few hundred words. First, let me exclude all living statisticians and scientists to avoid making a choice among all those people I admire and hurting anyone’s feeling (and also because this is somehow unfair to younger researchers). So let us stick to dead individuals! Second, I am quite hesitant to choose between a scientist (broad category!) and a statistician (restrictive category!). Again, let me [first?] stick to statisticians, avoiding the impossible choice between Albert Einstein, Marie Curie, Srinivasa Ramanujan, Henri Poincaré, Evariste Galois, Ada Byron,  and others…

Even in this smaller statistical world, I remain hesitant between many names, the two primary contenders being Pierre-Simon de Laplace and Harold Jeffreys. They both share an image of universal scientists (or “honnêtes hommes”) in that they dealt with many topics on a wide-ranging scale and developed new methodologies that impacted the field for years and even centuries. They both worked in statistic as a side occupation, their main field being mathematics and astronomy, and geophysics, respectively. Nonetheless, maybe thanks to their outsider quality, they imagined fundamental changes to our field. Not only did Laplace re-derive Bayes’ theorem but he also saw to its implementation and reflected on the choice of priors. Similarly, Jeffreys set the ground for the construction of reference priors in an essential piece of work. While I could claim some low-level kinship with Laplace (since we both spent our childhoods in Pays d’Auge, Normandy, went to study in Caen,  and end up living in neighbouring cities south of Paris [where an RER stop is named after him]), and. more seriously, while I tremendously enjoyed Laplace’s Traité Philosophique sur les Probabilités, I nonetheless end up voting for Harold Jeffreys as I have a great fondness for Theory of Probability. Which remains in my opinion the most important book written for Bayesian analysis. (This post just appeared in the ISBA Bulletin of March 2011, along more inspired pieces by fellow Bayesians.)