For foundational contributions to statistical science that have shaped both the theoretical underpinnings and applied practice of the discipline across more than four decades, Peter McCullagh is the third recipient of the David R. Cox Foundations of Statistics Award, presented by the American Statistical Association. McCullagh will receive the award and deliver a lecture titled “What Is a Regression Model?” at the Joint Statistical Meetings in Boston Massachusetts at 10:30 a.m. on August 5.
McCullagh is the John D. MacArthur Distinguished Service Professor of Statistics at The University of Chicago. He is best known for co-authoring the textbook Generalized Linear Models, winner of the inaugural Karl Pearson Prize (now the Founders Prize) in 2013. His 1980 paper in the Journal of the Royal Statistical Society Series B on regression models for ordinal data introduced cumulative logit models that have found broad application across the biological and social sciences. The paper was selected as the JRSSB retrospective paper for presentation at the 2022 RSS International Conference, celebrating 50 years of the proportional hazards model. McCullagh has contributed four read papers to the Royal Statistical Society, underscoring the consistent originality and conceptual reach of his research.
Beyond generalized linear models, McCullagh has pursued a deeper foundational program to place statistical modeling on a rigorous logical and algebraic footing. His 2002 discussion paper “What Is a Statistical Model?” in the Annals of Statistics proposed an axiomatic framework in which treatment effects are defined as group elements perturbing outcome distributions from baseline, a formulation that exposes logical inconsistencies in much of the mainstream methodology literature. His monograph, Tensor Methods in Statistics, introduced algebraic tools for high-dimensional inference, while his 2022 book, Ten Projects in Applied Statistics, applies decades of statistical wisdom to challenging real-world problems, offering critiques of causal inference and statistical modeling. Throughout his career, McCullagh has exemplified the combination of theoretical depth and practical insight that characterized David Cox’s own approach to the discipline.
McCullagh’s influence extends across probability theory, biostatistics, and computational statistics. His contributions to quasi-likelihood, local sufficiency, approximate conditional inference, Cauchy-family parametrizations, and the structure of exchangeable arrays have shaped the field for decades. He is a Fellow of the Royal Society and a recipient of the Guy Medal in Silver from the Royal Statistical Society. He was also recognized by COPSS in 1990 as an outstanding statistician under the age of 40.
The David R. Cox Foundations of Statistics Award was created in 2022 through an endowment by Deborah G. Mayo, professor emerita of philosophy at Virginia Tech.
Leave a Reply