Randomized controlled trial is the gold standard for causal inference yet it is not always feasible. Multi-wave longitudinal data is increasingly used for causal inference. Standard methods such as regression yield biased results in the presence of time-varying confounders. In this workshop, I will demonstrate using Inverse probability treatment weight (IPTW) and marginal structural model (MSM) to adjust for time-varying confounders.
Examples in this workshop will be based on research in psychology, public health and epidemiology, and the analyses will be conducted in R and StatsNotebook.
Link for R: https://www.r-project.org/
Dr. Gary Chan is a statistician and epidemiologist at the National Centre for Youth Substance Use Research, University of Queensland. He is the chief developer of StatsNotebook, an open source R based statistics app. His research focuses on statistical software development, causal inferences, and the epidemiology of substance use among young people. He has served as a consultant at the United Nations Office on Drugs and Crime to evaluate the data collection methodology on global substance use data, and as a biostatistical consultant for the West Moreton Health Service.