Student seminar series
Xisi Zhang
PhD Candidate
Room: M3 3127
Estimating causal effects in longitudinal studies subject to missing exposures and confounders
Estimating causal effects becomes more challenging in the presence of missing data in longitudinal studies. Without carefully specified assumptions about the missingness mechanism, causal effect estimates can be biased and misleading. Existing methods primarily focus on estimating causal effects when either the exposure or the covariates are partially observed, but few address scenarios where both are missing. In this project, we specify assumptions for the missing-data mechanisms of both the exposure and covariates and develop estimation approaches under these assumptions. We then apply the proposed methods to estimate the causal effect of smoking on the development of functional limitations using data from the Health and Retirement Study (HRS).