RETROSPECTIVE RETROSPECTIVE

Stephen Fienberg: Superman of statistics Larry Wassermana,1

Stephen E. Fienberg died on December 14, 2016 after career, bringing ever greater depth and breadth to a 4-year battle with cancer. He was husband to Joyce, the area. Ultimately, he led the effort to import father to Anthony and Howard, grandfather to six, tools from algebraic geometry into statistics to re- brother to Lorne, mentor, teacher, and a prolific veal subtle and exotic properties of statistical researcher. Most of all, he was a tireless promoter of models. the idea that the field of statistics could be a force of It is hard to summarize Steve’s work because of its good for science and, more broadly, for society. astonishing breadth. A list of topics he worked on is a Steve was born in in 1942. After receiving tour of the field: networks, graphical models, data pri- his bachelor’s degree in mathematics at the University vacy, forensic science, Bayesian inference, text analysis, of Toronto in 1964 he went on to earn a PhD in statistics statistics and the law, the census, surveys, cybersecur- at Harvard, which he completed in 1968. After stints ity, causal inference, the geometry of exponential fam- at the University of Chicago and the University of ilies, history of statistics, mixed membership, record Minnesota, he landed in the Department of Statistics linkage, the foundations of inference, and much more. at Carnegie Mellon in 1980, where he stayed for the Steve’s impact goes far beyond his published work rest of his career, except for two years when he in statistics. He was inducted into the National Acad- served as vice president at York University. He wore emy of Sciences in 1999 but was involved in working many hats over the years: department head, dean, with the Academy for over 40 years. He served on 35 cofounder of the Center for Automated Learning and committees and panels and for many years was the Discovery (which eventually became the first ma- cochair of the Report Review Committee. Steve over- chine learning department in the world), codirector saw the reviews of hundreds of reports each year. He of the Living Analytics Re- chaired the 1989 report “Statistical Assessments as search Centre at Singapore Evidence in the Courts” and the Committee to Review Management University, mem- the Scientific Evidence on the Polygraph, which ex- berofCyLab(Computerand posed the flaws of polygraphs. He chaired the Com- Communications Security Center), mittee for National Statistics for six years. By my count, and codirector of the Center Steve testified before Congress at least a dozen times. for Statistics and Applications Through his interdisciplinary work, government work, in Forensic Evidence, to name and service at the National Academy Steve made the a few. case that statistics is a critical component in all aspects Steve published hundreds of science and policy analysis. of papers and seven books. Despite his busy schedule, Steve found plenty of time He is best known for his work for mentoring. He had 46 PhD students who have gone on log-linear models. These on to successful careers in academia, government, and models are the foundation industry. He supervised many postdocs and mentored for virtually all data analysis junior faculty. His office door was quite literally always that involves categorical var- open. He worked tirelessly behind the scenes to make iables. His book Discrete Mul- sure his students and colleagues received accolades. He tivariate Analysis, written with took a quiet pleasure in the success of others. Yvonne Bishop and Paul Holland, Steve’s life was not just about work. He enjoyed fine is the bible on this topic. Walk dining, good wine, and the company of friends. If you into any statistician’s office needed a restaurant recommendation, you went to and you will find it on the Steve. I remember a dinner with Steve a few years Stephen E. Fienberg. Image courtesy of shelf. Steve continued to work ago. After a long evening of wine, food, and good Carnegie Mellon University. on this subject throughout his conversation, I had to go home and sleep. Steve had

aCarnegie Mellon University, , PA 15213 Author contributions: L.W. wrote the paper. The author declares no conflict of interest. 1Email: [email protected].

3002–3003 | PNAS | March 21, 2017 | vol. 114 | no. 12 www.pnas.org/cgi/doi/10.1073/pnas.1700160114 Downloaded by guest on September 25, 2021 a different plan. He would go home to his wife, Joyce, My brief description of Steve’s life may smack of sneak in another hour of work, and then head off to his exaggeration. However, if anything, my account is in- late-night hockey game. This was typical: Steve could complete. It has been suggested by some that per- find time for work, friends, family, food, wine, and haps there was more than one Steve Fienberg. This sports in a single day. His exceptional energy was his would explain a lot. The fact is, he was unique, and we trademark and a source of wonder to his colleagues will never know how one person could accomplish so and friends. much in one lifetime.

Wasserman PNAS | March 21, 2017 | vol. 114 | no. 12 | 3003 Downloaded by guest on September 25, 2021