Bayesian Methods
Professor Emeritus Wesley Johnson’s research focuses on Bayesian parametric, nonparametric and semiparametric methods as well as on inference for survival analysis, for longitudinal and spatial data, and for diagnostic outcome data and protocols. He also develops asymptotics for Bayesian inference and develops methods for constructing informative and partially informative prior distributions in complex models. His primary area of interdisciplinary work is veterinary and human epidemiology.

Diagnostic Testing Protocols
“One of my long-term application areas involves the development of diagnostic testing protocols and methods,” says Professor Emeritus Johnson. Related to this is the development and implementation of biomarker models for disease diagnosis. A second area of long-term applications is the analysis of longitudinal data including the Study of Women Across the Nation (SWAN) data. This is a multicenter project that is continuing after 20-plus years of studying health outcomes related to women during the menopausal transition. “I have no doubt that this work has real world impact in terms of practice of medicine in women’s health.”

From Theory to Practice
“My main goal has been to develop methods, and the software to implement those methods, that people would actually use,” says Professor Emeritus Johnson. “Up to now, this has been accomplished to a large degree through my work on diagnostic testing.” He has given more than a dozen workshops over the last 20 years on this topic, and the software that he and his colleagues have developed is widely used. He continues his efforts to improve the art of incorporating the science behind the data into Bayesian modeling. “I continue to pursue a number of more theoretical projects with the expectation that they will eventually be found to be useful in practice.”