1999 Hartley Lectures

Department of Statistics
Texas A&M University

The Texas A&M University Department of Statistics

presents the

H.O. HARTLEY MEMORIAL LECTURES

April 6, 7, and 8, 1999

Lecturer: Adrian Raftery
Professor of Statistics and of Sociology, University of Washington

The H. O. Hartley Memorial Lectures were established by the Texas A&M University Department of Statistics in 1988 to honor the memory of Herman Otto Hartley. Hartley accepted an appointment as Distinguished Professor at Texas A&M in 1963, founded Texas A&M's Institute of Statistics, and served as its director until his retirement in 1977. Hartley built his initial faculty of four into a group of 16, directed more than 30 doctoral students, and published over 75 papers during this period. He served as president of the American Statistical Association in 1979. Professor Hartley died on December 30, 1980.

Hartley was well known for his work on the foundations of sampling theory, and also made important contributions to mathematical optimization, estimation with incomplete data, estimation of variance components, and establishment of safe doses in carcinogenic experiments. Hartley collaborated with Egon Pearson to produce the classic two-volume Biometrika Tables for Statisticians.

Hartley earned a Ph.D. degree in mathematics at Berlin University in 1934, and a Ph.D. degree in statistics under John Wishart at Cambridge University in 1940. He taught at University College, London and Iowa State College before coming to Texas A&M. Hartley was deeply committed to all phases of his profession, including education, research, and delivery of knowledge and advice to users of statistics. H. O. Hartley was not only brilliant academician, but also a warm and caring human being. His legacy continues to have a profound influence on Texas A&M students and the larger statistical community.

The H. O. Hartley Memorial Lectures are held every other year, and were first given in 1988 by Peter J. Diggle. Subsequent lecturers have been Bradley Efron, E. J. Hannan, Sir David R. Cox and Wayne A. Fuller. This year the Department of Statistics is honored to have Professor Adrian Raftery present the Hartley Lectures.

Dr. Raftery is a Professor of Statistics and of Sociology at the University of Washington. He received his doctorate from the Université de Paris VI in 1980. He was on the faculty at Trinity College in Dublin from 1980-1986. He joined the faculty at the University of Washington in 1986. Professor Raftery has been a visiting faculty member at Université de Paris VI and INRIA, in France.

His many honors include being elected a Fellow of the American Statistical Association. He won the 1996 American Statistical Association Award for Outstanding Statistical Application. He was the 1997 Institute of Mathematical Statistics Special Invited Lecturer. He also won the Population Association of America 1998 Clifford C. Clogg Award for contributions to population statistics. He is the Coordinating Editor and Applications and Case Studies Editor of the Journal of the American Statistical Association, as well as the Editor of Sociological Methodology.

He has done exciting work in both applied and theoretical statistics. He has approximately 90 refereed publications in various journals. Some of the statistical topics of his research are time series, change points, model based clustering, spatial point pattern research, and Bayesian model selection. These topics have been applied to many fields, some of which are sociology, demography, whale population estimation and dynamics, and various environmental problems. His broad scope of work is demonstrated by the diverse agencies for which he has received support. These include the National Science Foundation, Office of Naval Research, North Slope Borough Research Contract, Environmental Protection Agency, and U.S. West.

STATISTICAL INFERENCE FOR DETERMINISTIC SIMULATION MODELS
4:00 p.m. Tuesday, April 6, 1999
206 Memorial Student Center

Deterministic simulation models are used in many areas, including the making of environmental and other policy decisions, atmospheric science, engineering, pharmaceutical research and demography. They tend to be complex, and to require the specification of many inputs. This if often done in an ad-hoc manner, and little attention has been given to taking proper account of uncertainty and evidence about the inputs and outputs to the model. I will describe the Bayesian melding method, which brings together ideas from modeling, measure theory and the pooling of expert opinions.

Reception to follow...

MODEL-BASED CLUSTERING AND IN SPATIAL POINT PROCESSES
4:00 p.m. Wednesday, April 7, 1999
255 Chemistry Building

In recent years, it has been realized that cluster analysis, long a collection of ad hoc techniques, can be profitably recast in terms of a statistical model: a finite mixture of probability distributions.
I will review some of this work, and include successful applications from imaging for automated textile manufacturing, medical image analysis, medical diagnosis, minefield detection and the detection of seismic faults from earthquake catalogs.

VARIABLE SELECTION AND BAYESIAN MODEL AVERAGING IN EPIDEMIOLOGICAL CASE-CONTROL STUDIES
4:00 p.m. Thursday, April 8, 1999
Room 150, Blocker

We propose Bayesian model averaging as a formal way of taking account of model uncertainty in case-control studies. This yields an easily interpreted summary, the posterior effect probability, and out simulation study indicates this to be reasonably well calibrated in the situations simulated. The methods are applied and compared in the context of a previously published case-control study of cervical cancer.

Copyright © Unless otherwise stated, all contents of these web pages are copyright by their respective authors or by the Department of Statistics, Texas A&M University, 1995-1999. Comments and questions about this site should be sent to web@stat.tamu.edu.