StatIstical Methods for Observational Studies - STA542
Course Objectives
- Determine the correct statistical analysis for a given set of data.
- Utilize statistical software to analyze linear models and correctly interpret the output.
- Utilize statistical software to perform logistic regression for PROPENSITY SCORE models and correctly interpret the output.
- Utilize statistical software to analyze OBSERVATIONAL OUTCOME using PROPENSITY SCORES as covariates, weights, or matched samples as well as correctly interpret the output.
- Utilize statistical software to analyze OBSERVATIONAL OUTCOMES models, though INSTRUMENTAL VARIABLES, and correctly interpret the output.
- Utilize statistical software to perform PROPENSITY SCORE MODELS for OBSERVATIONAL models with more than TWO GROUPS through covariate adjustments.
- Discuss goodness-of-fit techniques for PROPENSITY SCORE MODEL AND INSTRUMENTAL VARIABLES.
- INTRODUCTION TO RUBIN CAUSAL MODEL, STRUCTURAL NESTED MEAN MODEL, and POTENTIAL OUTCOMES FRAMEWORK.
- Communicate the results of these statistical analyses in a concise, simple way that would be understandable to a non-statistician.
Course Topics
- Analysis of randomized trials per the content presented in STA512.
- Regression modeling, building and diagnostics.
- Experimental designs versus Observational designs.
- Observational modeling techniques such as propensity score techniques, instrumental variables, and potential outcomes.
- An introduction to the Rubin Causal Model.
- Using SAS Procedures PROC GLM, PROC REG, PROC QLIM, PROC SYSLIN, and PROC IML.
Example Syllabus