Intermediate Biostatistics: Analysis of Discrete Data

This course provides a working knowledge of statistical methods suitable for data with discrete response values. Among the topics students will explore are statistics for contingency tables, Poisson and negative binomial regression, propensity scores, instrumental variables, principal components analysis, bootstrapping, cross-validation, and model building, all with an emphasis on epidemiologic applications. Students may use either R or SAS statistical software.

Topics Include

• Univariate analyses of discrete data
• Confounding and interaction
• Mantel-Haenzel techniques
• Logistic regression
• Modeling predictors in logistic regression
• Building hypothesis-driven models
• Propensity scores
• Building predictive models
• Multinomial and ordinal logistic models
• Regression for matched data: generalized estimating equation and conditional logistic

This course was formerly numbered HRP261