Section outline

  • Lecture: Confounding and Interaction III: Stratified Analysis

    Concept of weighted averages; use of stratification to form adjusted measures; managing multiple confounding paths; software implementation to detect minimally sufficient set for adjustment; limitations of stratification including time-dependent confounding/mediation

    Faculty:  Jeffrey Martin

    Location:  Mission Hall 1400

    • Session Slides:

    • Confounding and Interaction III 2016 Final Post File

    • Session Audio/Video Recording (Access restricted to registered students):

    • Recommended Reading:

      S & N: Ch. 7: sections 7.1 - 7.2.1 (p. 229 - 234) and sections 7.3.3 - 7.3.4 (p. 242 - 248)

      dagitty.net, an open source browser-based environment for creating, editing, and analyzing directed acyclic graphs

    • Optional Reading:

    • Shrier I and Platt W. Reducing bias through directed acyclic graphs. BMC Medical Research Methodology 8:70, 2008. File
      Not available unless: Your ID number contains 02
    • Suarez D et al. Differences between marginal structural models and conventional models in their exposure effect estimates. Epidemiology 22: 586-588, 2011. File
      Not available unless: Your ID number contains 02
    • Greenland S and Pearce N Statistical Foundations for Model Based Adjustments Ann Rev Pub Health 2015 File
      Not available unless: Your ID number contains 02

    • Assignment: Problem Set 11

    • esophageal cancer File
    • Problem Set 11 Confounding and Interaction III 2016 File

    • Assignment Due Date:  December 6, 2016

    • Assignment Answer Key (Access restricted to registered students):