Robert Erin Weiss

I teach classes in Bayesian Data Analysis (Biostat 234), Bayesian Theory (Biostat 202C), and Longitudinal Data Analysis (Biostat 236). My classes have computing and writing components, because statistical practice requires that you calculate inferences and communicate what you have learned. I particularly like working in an academic environment because of the variety of projects I get to work on and because I get to work with many great students on research projects, paper writing, and as dissertation advisor. I've supervised 30 Biostatistics doctoral dissertations to date and have helped doctoral students from numerous departments with their dissertations! 

Education


  • PhD, Statistics, University of Minnesota, Minneapolis, MN, 1989
  • MS, Statistics, University of Minnesota, Minneapolis, MN, 1987
  • B.Math, Mathematics, University of Minnesota, Minneapolis, MN, 1984

Areas of Interest


My biostatistical research interests include Bayesian methods, meta-analysis, longitudinal data analysis, and hierarchical modeling generally. Bayesian methods are useful in all data analysis offering flexibility and simplicity in statistical modeling. Problems with meta-analysis include abstracting data from papers, multiple data points from single studies, and missing data. Issues in longitudinal data include (i) having a multivariate observation or a spatially distributed observation at each time point, (or both!) (ii) nonlinear time trends and (iii) covariance modeling. I work in model development (Figuring out what is the right model for a given data set), model checking (I have a model, but is it right and how can I improve it?), and model selection (Here's a bunch of models, which one is best?).

Motivation for much of my research comes from the statistical difficulties seen in my collaborative work, particularly in the analysis of human behavior, emergency medicine, and ophthalmology. In HIV research, my colleagues and I study HIV+ and HIV-at-risk individuals, substance use disorders, and linkage to care. For Emergency Department research, we've mostly recently been studying bad cardiac outcomes after syncope (fainting). In Ophthalmology, my colleagues and I are studying declining visual acuity and loss of nerve layers in the macula in people with moderate to advanced glaucoma. Visual acuity is a functional measure; it is a spatially located measurement of how well a person can see, how well they can function. Macular nerve layer thickness is a structural measurement, a physical measurement of the health of the macular tissue. These functional measures may be measured on a grid of points called superpixels, or on a circle of points called sectors. 

Selected Courses


  • Biostat 202C: Bayes Theory
  • Biostat M234: Applied Bayesian Inference
  • Biostat M236: Analysis of Repeated Measures
  • (Biostat 251: Multivariate Biostatistics)
  • (Biostat 411: Correlated Data)

Selected Publications