Skip Navigation

Course Directory

Causal Mediation Analysis

Summer Inst. term
Mental Health
1 credits
Academic Year:
2021 - 2022
Instruction Method:
Asynchronous Online with Some Synchronous Online
Mon 06/14/2021 - Wed 06/16/2021
Class Times:
  • M W,  10:00am - 12:00pm
Auditors Allowed:
Yes, with instructor consent
Grading Restriction:
Letter Grade or Pass/Fail
Course Instructors:
Trang Nguyen

Understanding of fundamental statistical concepts such as linear and logistic regression


Are you looking into mechanisms of a causal effect, studying an intervention with components targeting different causal pathways? Or are you writing a mediation aim for an NIH grant, or writing the mediation paper in your dissertation? If you have questions about what mediation analyses even mean, nevermind how to estimate mediation effects, come join us!

Provides guidance on a thoughtful mediation analysis aiming to study the mechanisms through which exposures have their effects on outcomes or to study the effects of potential interventions on variables on the causal pathway. Connects definitions of causal effects in mediation analysis, including (in)direct and others, to real-world research questions. Explains the assumptions required to identify those effects in experimental and observational studies. Illustrates how some of these effects are estimated.

Learning Objectives:

Upon successfully completing this course, students will be able to:

  1. List several types of causal effects in mediation analysis including (in)direct effects (controlled, natural and interventional) and a broad class of interventional effects
  2. Connect a real-world research question to an effect definition from among these types
  3. Explain the three main types of assumptions required for identification of causal effects in mediation analysis (consistency, unconfoundedness, and positivity) and relate them to the effect(s) of interest
  4. Implement estimators of mediation effects
Methods of Assessment:

This course is evaluated as follows:

  • 20% Participation
  • 20% Reflection
  • 60% Problem sets

Instructor Consent:

No consent required