https://bayesian.org/wp-content/uploads/2024/03/2403.pdf
From the March 2024 issue of the ISBA bulletin reproduced below:-
OBITUARY FOR PROFESSOR THOMAS LEONARD (1948-2023)
Diego Andres Perez Ruiz
diego.perezruiz@manchester.ac.uk
With profound sadness, we announce the passing of Thomas Hoskyns Leonard, a distinguished
and pioneering statistician whose contributions to the field have left an indelible mark. A retired
professor from the Universities of Wisconsin-Madison and Edinburgh, Thomas Leonard’s career was
marked by ground-breaking work in Bayesian statistics and a deep commitment to interdisciplinary
research.
Thomas Leonard’s academic journey began at Imperial College London, where he earned a BSc in
Mathematics with First Class Honours in 1970. Immediately thereafter, he completed an MSc in
Statistics from University College London with Distinction and a PhD in Statistics from the same
institution. His early career saw him as a Lecturer at the University of Warwick, where he co-founded
the Department of Statistics and the innovative MORSE (Maths, Operations Research, Statistics, and
Economics) degree programme with Robin Reed and P.J. Harrison.
In 1979, Thomas Leonard moved to the United States, joining the University of Wisconsin-Madison
as an Associate Professor, later becoming a full Professor. His tenure at UW Madison was marked
by significant contributions, including his work at the Mathematics Research Centre. He worked
on improving the Bayesian components of both the teaching and research programmes, alongside
Kam Wah Tsui and Michael Newton. In 1995, he returned to the U.K. to lead the Department of
Mathematics and Statistics at the University of Edinburgh, a position he held until 2001.
A true visionary, Thomas Leonard co-founded the International Society for Bayesian Analysis (ISBA)
in 1992, alongside Arnold Zellner and Gordon Kaufman. In 2013, he was elected as a Founding
Fellow of ISBA. He also served as a committee member for multiple statistical societies, including
the Edinburgh Section of the Royal Statistical Society, contributing to numerous interdisciplinary
debates and initiatives.
Thomas Leonard’s academic work
Thomas Leonard’s academic work was both prolific and influential. His research in Bayesian Categorical Data Analysis is considered seminal, and his contributions to other areas have had farreaching impacts. His work on multinomial shrinkage estimators drew parallels with the cryptanalysis techniques of Alan Turing and his assistant I.J. Good, showcasing a profound connection
between historical methodologies and modern statistical analysis. Leonard’s approach to density
smoothing was similarly inspired by Good’s work, underscoring the depth and innovation of his
research and beautiful ideas.
His seminal contributions to Bayesian Categorical Data Analysis have been highly praised by leading
statisticians such as Alan Agresti and David Hitchcock and further expanded upon by distinguished
scholars Nan Laird and Terry Speed. His publications, including his co-authored book ”Bayesian
Methods: An Analysis for Statisticians and Interdisciplinary Researchers” and his work on the history of Bayesian statistics, have been widely recognised. Together with John Hsu, Leonard developed the conditional Laplacian approximations, offering a computationally accessible method forBayesian inference that has significantly advanced the field and is widely used today.
His constant search for understanding, and his tenacity as a researcher, led him to develop some of
the most difficult and influential theories in Bayesian Statistics. With Hsu, Chiu, and Tsui, Leonard
co-invented the matrix linear covariance model, a development recognised as foundational by econometricians James LeSage, Kelley Pace, and Manabu Asai. In his 1999 publication with John Hsu,
Leonard presented innovative alternatives to the expected utility hypothesis in economics, challenging traditional theories of risk aversion and offering new perspectives on economic behaviour.
Leonard’s interdisciplinary work extended into family medicine, where, in collaboration with Richard
L. Brown, Orestis Papasouliotis, and others, he played a crucial role in establishing the Wisconsin
Substance Abuse Programme. His influential work in obstetrics, particularly his 1981 invited discussion paper with Jim Low and colleagues, transformed the prediction of fetal metabolic acidosis,
marking a significant advancement in the field. His collaborations in geophysics with Ian Main and
others have been recognised as ground-breaking, further illustrating the wide-reaching impact of his
research.
Beyond Academia
Beyond academia, Thomas Leonard was a passionate advocate for mental health, providing expert
testimony and evidence to the Scottish Government. He submitted a large amount of anecdotal and
statistical evidence to the Scottish Parliament regarding the widespread devastating side effects of
psychiatric medications and electroconvulsive therapy.
His commitment to societal issues was evident in his extensive work on the history of eugenics,
including a significant submission to the Commission of Inquiry into the History of Eugenics at
University College London (UCL).
Thomas Leonard’s legacy is immortalised not only in his contributions but also in the lives of the
students and peers he inspired and mentored. His passing is a great loss to the statistical community
and to all who had the privilege of knowing and working with him. Thomas Leonard was an extraordinary person, kind, gentle, and generous in ways that few ever have the privilege to encounter
in their lifetime. We extend our deepest condolences to his family, friends, and colleagues.
His memory and contributions will continue to inspire and guide future generations of statisticians.
Tom, you will be profoundly missed.
