Wednesday, July 16, 2025

8:00 AM - 12:00 PM

Grand Room 2

---

Details

Description: Missing data are a common problem in epidemiological studies that can lead to bias if there is a violation in the assumptions about the causes of missing data that underlie (implicitly or explicitly) the analysis. Yet, the traditional forms of missing data assumptions (e.g. “missing at random”) are difficult to assess in practice where it is common for multiple variables to have missing data. This workshop will take participants through the steps of a new roadmap for handling missing data that is grounded in the use of causal diagrams to facilitate the specification and assessment of missingness assumptions, and thereby decide on the most appropriate approach to handle missing data. Prerequisites and intended audience: The target audience is epidemiologists with some statistical background, including some knowledge of regression methods. Presenters: Margarita Moreno-Betancur, Clinical Epidemiology & Biostatistics Unit, Murdoch Children’s Research Institute & University of Melbourne Ghazaleh Dashti, Clinical Epidemiology & Biostatistics Unit, Murdoch Children’s Research Institute & University of Melbourne