Professor Stijn Vansteelandt BSc MSc PhD
- Stijn Vansteelandt's Contacts
- Keppel Street
- WC1E 7HT
I am Belgian and graduated as Master in Mathematics at Ghent University in 1998. I obtained a PhD in Mathematics (Statistics) in 2002 at the same university, and subsequently did postdoctoral research in the Department of Biostatistics of the Harvard School of Public Health. I joined the London School of Hygiene and Tropical Medicine in 2017, and also hold a position as Professor of Statistics in the Department of Applied Mathematics, Computer Science and Statistics at Ghent University, Belgium.
I develop statistical methods for inferring the causal effect of an exposure on an outcome from experimental and observational data under minimal and well-understood assumptions. My work focuses on a variety of topics in biostatistics, epidemiology and medicine, such as the analysis of longitudinal and clustered data, missing data, adjustment for baseline covariates in randomised experiments, mediation and moderation/interaction, instrumental variables, family-based genetic association studies, time-varying confounding, analysis of outcome-dependent samples and phylogenetic inference.
My recent research is primarily aimed at making causal inferences less vulnerable to the weaknesses (imprecision, finite-sample bias and susceptibility to model misspecification) of simple inverse probability weighted estimators that dominate causal inference research. I aim to realise this either by improving inverse probability weighted estimation (see my work on bias-reduced double-robust estimation) or by popularising and extending alternative estimation methods, such as g-estimation, that avoid inverse weighting. More recently, I am also developing statistical methods for post-regularisation inference (that is, how to obtain honest confidence intervals and p-values that acknowledge the uncertainty due to variable selection).
A characteristic feature of my research is the use of semi-parametric models.
- Statistical methods
- Causal Inference
- Missing Data
- data-adaptive methods
- semi-parametric methods