Walter Dempsey

Walter Dempsey is an Associate Professor of Biostatistics at the University of Michigan School of Public Health and Assistant Research Professor at the Institute for Social Research. His research develops statistical methods for digital health, focusing on experimental design for multi-stage decision making, modeling of complex longitudinal data, and analysis of relational network structures.

Education:

• Ph.D. in Statistics, University of Chicago (2015)
• B.Sc. in Mathematics, Statistics and Economics, University of Chicago (2009)

Research Focus: Dr. Dempsey's work integrates statistical theory with health applications, particularly in mobile health (mHealth) technologies. His methodological research spans three interconnected areas: (1) Designing adaptive trials for health decision-making; (2) Developing hierarchical latent variable models for intensive longitudinal data from wearables and sensors; (3) Creating statistical frameworks for analyzing interaction networks that satisfy invariance principles while capturing empirical behavior patterns.

Publication Trends: His recent work demonstrates strong emphasis on network modeling, mobile health interventions, and causal inference methods. Publications frequently appear in top statistics and machine learning venues including JASA, Biometrika, ICML, and NeurIPS, with consistent focus on developing interpretable models for health applications.

Student Advising:

• Hera Shi (PhD, 2023) - Time-varying treatment effects in micro-randomized trials
• Yuhua Zhang (PhD, 2023) - Statistical methods for network data
• Madeline Abbott (Current PhD) - Latent variable models for intensive longitudinal data
• Easton Huch (Current PhD) - Robust Bayesian methods for causal inference

Laboratory: Leads the Dempsey Lab developing statistical methodologies for digital health, with projects spanning network analysis, survival modeling, and adaptive intervention design.