These days, I focus on overparameterized prediction models, second-generation p-values, and the role of scientific principles in data-driven methods. This website has historically highlighted foundational work in statistical evidence metrics but is now gradually shifting toward prediction modeling.

​​

I am a data scientist and biostatistician whose work spans statistical theory, biomedical research, and the development of interdisciplinary data science programs. I'm big on second-generation p-values, an improved p-value that resolves the mismatch between how classical p-values are defined and how they are routinely misinterpreted in practice. They perform well and more faithfully capture the scientific principles of deductive reasoning, requiring scientists to articulate a meaningful null hypothesis rather than defaulting to a functionless straw man. My research spans theory and practice, with foundational contributions to likelihood-based inference, mediation modeling, missing data in prediction models, clinical trials analysis, false discovery rates, and ROC curve methodology.

​

My collaborative research spans a wide range of domains, including cancer diagnosis and screening, clinical trials, radiology, nephrology, functional neuroimaging, structural biology, and women’s health. I try to consistently emphasize the foundations, interpretability, and reliability of predictive models, with a focus on improving decision-making in high-stakes biomedical settings.

 

As Director of the PRISM Lab (“Prediction, Inference, and Scale in Data Science Models”), I lead and mentor PhD students developing methods that bridge statistical innovation and biomedical application. I am a Fellow of the American Statistical Association and the American Association for the Advancement of Science. My work has been recognized with the Chancellor Award for Research at Vanderbilt University and the Spinoza Chair in Medicine at the University of Amsterdam.

​​