Advanced Methods in Observational Studies: Inference
June 22-26, 2015
8:30 a.m. – 12:00 p.m.
Course number: 340.709.11
Expands beyond introductory level epidemiologic concepts and methods material. Using examples from the published literature, this course will emphasizes inference, evaluation interpretation and the ability to critically evaluate issues related to comparison of populations and synthesizing inferences from observational studies including: descriptive and analytical goals for observational study inference; the counterfactual model for defining exchangeability, cause, and confounding; graphical and analytical tools for assessing causal pathways and investigating their underlying assumptions, including comparing classical and modern approaches for assessing and addressing confounding and mediation.
Student Evaluation: 4 brief daily quizzes, each of which will have both in-class and take-home components; each daily quiz will be weighted equally
Learning Objectives: Upon successfully completing this course, students will be able to:
- Utilize and illustrate a framework for distinguishing different inferential goals of an epidemiological study
- Define concepts and terminology in causal inference for epidemiology and develop graphical approaches (e.g., DAGs) for models that integrate confounding and mediation effects
- Illustrate, interpret, and contrast ‘classical’ (e.g., regression) approaches for addressing confounding with modern techniques (e.g., propensity-score and inverse-weighting methods)
- Identify and evaluate approaches for implementing a basic sensitivity analysis
Prerequisite: Students must have 1 term of epidemiology (340.601, or equivalent with consent of instructor) and 1 term of biostatistics including knowledge of multivariable regression; taking 340.708 is preferred but not a prerequisite.
Grading Restriction: Letter Grade or Pass/Fail
Course Materials: Provided in class and through CoursePlus