Epidemiology Methods Workshop
Using the ‘C’ word: The implications of using explicit causal frameworks in epidemiologic research and teaching
Wednesday, January 11, 2011
All slots have been filled via lottery
Interaction, Effect Modification and Synergy: A Rose by Any Other Name?
The workshop will begin with a discussion of the implications of a “potential outcomes” (aka counterfactual) approach for understanding which model best represents what we mean by “synergy” as a causal concept.
Directed Acyclic Graphs (DAGs): Useful Tools or Needless Complexity?
This workshop module will focus on describing what DAGs are, how to build them, and how to use them for making decisions about confounder control. We will examine the extent to which a DAG approach to confounding would change our decisions about what variables are confounders.
Mediation: What Is It, and How Do We Find It?
We will discuss mediation in the context of the causal concepts of the potential outcomes and sufficient component cause models. The problems with “traditional” methods of assessing mediation will be examined and some alternative methods described.
In all three parts of the workshop, the focus will be on understanding the types of problems for which these statistical analytic tools were designed and the implications of using these tools for the interpretation of results, rather than the mechanics of the specific statistical approaches themselves.
Space is limited to 25 participants; participation is open to all students and faculty able to commit to the full workshop, though priority will be given to doctoral students. If necessary, a lottery may be instituted. Sign up via Catalyst WebQ through 12/6/11.
Sponsored by the UW Department of Epidemiology
Updated on November 30, 2011