Steve Fienberg: “Statistics is what I do!”
Photo: Carnegie Mellon University

Stephen (Steve) E. Fienberg was the “superman of statistics” (L. Wasserman), “a hero of the statistics profession” (M. Straf), “the ultimate public statistician” (A. Carriquiry, E. Lander), and “the best kind of Bayesian” (E. George). His life and illustrious career ceased at the age of 74, on December 14, 2016, after a four-year battle with cancer.

Steve was Maurice Falk University Professor of Statistics and Social Science at Carnegie Mellon University (CMU) and co-director of the Living Analytics Research Centre, with additional appointments in the Machine Learning Department, Heinz College, and Cylab. For his outstanding contributions to the profession, he received the COPSS and Wilks awards, the NISS Distinguished Service Award, the Lise Manchester Award from the Statistical Society of Canada, the Founders’ Award from the American Statistical Association, and the Zellner Medal from the International Society for Bayesian Analysis. He was an elected member and lifetime National Associate of the National Academies (National Academy of Sciences and National Research Council), an honor bestowed for his extraordinary contributions to the Academies. He was an elected fellow of IMS, ASA, the Royal Society of Canada, American Academy of Arts and Sciences, American Academy of Political and Social Science, American Association for the Advancement of Science and an elected member of the International Statistical Institute. Elyse Gustafson, who worked with Steve on IMS matters for nearly 20 years, describes his extraordinary impact: “Steve’s involvement was endless: he was President of the IMS, Editor of the Annals of Applied Statistics, and chair of nearly every IMS committee at some point. But what he really did for the IMS can’t be easily judged by where he served, it was how he did it. He led with his heart and his love for the organization. The IMS is stronger because of his leadership and vision.”

Steve’s impact on research, education and the practice of statistics is astonishing in its breadth. He published seven books and over 340 papers in leading journals in statistics, sociology, and machine learning, and wrote hundreds of editorials and reviews. Among the many areas where Steve made foundational contributions is categorical data analysis. His first technical paper appeared in the Annals of Mathematical Statistics in 1968 on the geometry of an r × c contingency table. CMU students will fondly remember him bringing a physical 3D tetrahedron to class to demonstrate the geometry of log-linear models. Throughout his career, he continued to work on the geometry of exponential families and shone new light on the potential of algebraic statistics. Discrete Multivariate Analysis: Theory and Practice (co-authored with Y. Bishop and P. Holland, 1975), and his 1980 book, The Analysis of Cross-Classified Categorical Data, are classics in the field. Going far beyond this seminal work, Steve made fundamental contributions to causal inference, mixed membership models, forensic science, statistics and the law, surveys, census and other problems related to the federal statistics on accuracy of prediction and population size estimation. He was consistently at the forefront of many developments in statistical sciences, including the foundations of inference and Bayesian analysis.

Steve’s outstanding leadership and vision is reflected in new journals that he spearheaded. He created Chance magazine with Bill Eddy in 1988. He co-founded the Journal of Privacy and Confidentiality in 2009, and served as its Editor-in-Chief until his last days. When Brad Efron was asked by the IMS to start a new journal, Annals of Applied Statistics (AOAS), the first person he approached was “the top social science statistician in the US,” Steve Fienberg. AOAS was created in 2007, and Steve held multiple editorial positions, including Editor-in-Chief (2013–16). The Annual Reviews of Statistics and Its Applications was launched under Steve’s leadership in 2014. Nancy Reid and Stephen Stigler recall: “His energy and enthusiasm for it were contagious, and his encyclopedic knowledge and vast network of colleagues were invaluable.”

He was a tireless champion of the importance of statistics, promoting it to the broader scientific community, government, media, policymakers, and the general public. Steve always insisted on establishing the rightful presence of statistics in many other disciplines by publishing across disciplines, creating and supporting multi-disciplinary centers, advocating for inclusion of statisticians on committees of the Academies, and serving as a role model for many around the world. Of his own role models, in a 2013 Statistical Science interview, Steve said: “Both Fred [Mosteller] and Bill [Kruskal] were Renaissance men and I didn’t know how I would do things in the same way they did, but it became very clear to me that just doing papers in the Annals and JASA wasn’t enough.” Steve’s actions matched his words. The list of real-life problems where he left his mark, in his own words, includes: the National Halothane Study, Sun-Times Straw Poll, 1970 Draft Lottery, social indicators, criminal justice, crime victimization, election night forecasting, social networks, legal applications, cognitive aspects of survey design, census undercount, privacy and confidentiality, disability and the National Long-Term Care Survey, administrative record linkage, human rights estimation, and the bank, telecom and theme park collaborations at the Living Analytics Research Center. “He was one of the very few scientists whose contributions were truly multidisciplinary in being relevant both in Statistics and the Social Sciences, and thus he set a standard for all of us working in these fields. Most importantly, studying his work helped [me] to understand when a mathematical argument is needed and when it is not,” said Tamás Rudas of the Hungarian Academy of Sciences.

Stephen Fienberg was a giant in Statistics, but it was his human side that has left a deeper impact on many of us. He was an exceptional and tireless mentor. He advised 46 PhD students and many postdocs and junior faculty. Together with his wife Joyce, he included his mentees into their extended family and built an extensive network of friends across the globe. Amidst his many professional activities, he always found time and energy for mentoring. He would help you unravel statistical incantations, and with his encouragement, many have become leaders in their fields. Invariably, his advice also included a few extra tips on good wines, great books, and special travel spots. Steve’s mentorship style defied boundaries of geography or formal affiliations—many others, who were never his students or worked in the same university, also consider him as a mentor, and all emphasize his generosity, the unwavering support and dedication, deep respect to all individuals no matter their academic age or stature, and a remarkable ability to have insights into other people’s work.

Steve was born on November 27, 1942, in Toronto, Canada. He received his bachelor’s degree in mathematics and statistics in 1964 from University of Toronto, where in his first statistics course Don Fraser introduced him to thinking geometrically about statistics, a skill that stayed with him throughout his career. Steve earned his PhD in Statistics from Harvard University in 1968 under the supervision of Fred Mosteller. During his first job at the University of Chicago, he met Bill Kruskal who influenced his work on surveys and involvement with the Committee on National Statistics (CNSTAT). The next eight years, he was at the University of Minnesota where he chaired the Department of Applied Statistics. In 1980, he became a Professor of Statistics and Social Science in the Department of Statistics at CMU, his academic home for the next 36 years. During that time, he served as a department head and dean, had numerous visiting positions, and made a brief return to Canada as a vice provost at York University.

Steve enjoyed life to its fullest. He adored his family, and was especially fond of his six grandchildren. He was an avid ice hockey player, half-joking that information on three variables—Pittsburgh location, age over 60, and being a hockey player—would disclose his identity with high certainty. He enjoyed traveling and attending meetings around the world. It was often easy to pick him out of the crowd because of his prominence and his affinity for wearing bright-colored shirts. He was not only a lively dinner companion but also a conscientious guest. At a 1990 conference in Taipei, where Steve was one of four statisticians to receive a key to the city of Taipei, he so outdid everyone in the numerous rounds of traditional banquet toasts that he was subsequently introduced as “Professor Fienberg, who has ocean capacity.” He truly had enormous capacity for everything he did, especially for his work and our profession. Many of us have marveled at his multitasking ability, his exceptional high energy and endless enthusiasm. In Ed George’s words: “We have lost a dear friend, one of the greatest champions of statistics, whose wonderful influence on all of us will be missed but never forgotten.” This echoes what many of us feel.

In mid-October 2016, CMU celebrated Steve’s transfer to emeritus status. At the end of a two-day program, Steve gave a talk to reflect on his career. His opening slide, quoting Groucho Marx (“Those are my principles, and if you don’t like them… well, I have others”), while capturing his personality and sense of humor, also contained truth: Steve was a man of strong principles and opinions, and had an unwavering belief in the importance of statistical science and its role in advancing other fields. These qualities made him a hero of the statistics profession. Though Steve would simply say, “Statistics is what I do!”

—

Written by Elena Erosheva and Aleksandra Slavkovic

We are especially grateful to J. Fienberg, J. Abowd, A. Carriquiry, M. Cuellar, B. Efron, E. George, T. Gneiting, A. Goldenberg, E. Gustafson, R. Mejia, B. Murphy, N. Reid, A. Rinaldo, T. Rudas, M. Sadinle, S. Stigler, M.Straf, J. Tanur, L. Wasserman, and many others, who shared with us their recollections, insightful and personal comments about Steve and his lasting legacy on the IMS, statistics and the sciences.