Home » Departments and Centres » Academic catalogue » Course detail

Open sidebar menu

Causal Inference with Longitudinal Data using G-Methods (SPS-WS-RI-CAU-24)

SPS-WS-RI-CAU-24

Purpose

G-methods, originally developed in epidemiology, are gaining traction in the social sciences as powerful tools for addressing causal inference challenges in longitudinal settings. These methods are specifically designed to handle repeated exposures over time, time-varying confounders, and treatment-confounder feedback loops. By accounting for dynamic confounding, G-methods such as inverse probability of treatment weighting (IPTW) and the G-formula can, under certain assumptions, estimate causal effects of multiple exposures.
In this seminar, we will explore scenarios where G-methods are essential and focus on two key approaches: IPTW and the G-formula. To deepen our understanding, we will simulate data-generating processes that include various forms of dynamic confounding and demonstrate how results differ when using other longitudinal analysis methods (e.g., fixed effects models) compared to G-methods. Finally, participants will have the opportunity to apply these techniques to a dataset of their choice or simulate data that mirrors empirical settings, allowing hands-on practice with the methods discussed.

Register for this course