B0175
Title: Assumption-lean quantile regression
Authors: Georgi Baklicharov - Ghent University (Belgium) [presenting]
Stijn Vansteelandt - Ghent University and London School of Hygiene and Tropical Medicine (Belgium)
Christophe Ley - University of Luxembourg (Luxembourg)
Abstract: Quantile regression is a powerful tool for detecting varying associations across different parts of the dependent variable's distribution. However, when using quantile regression to parameterize the conditional association between an exposure and an outcome, given covariates, two potential issues are often ignored. Firstly, the exposure coefficient estimator may not converge to a meaningful quantity when the model is misspecified, and secondly, variable selection methods may induce excess uncertainty, rendering inferences overly optimistic. These issues are addressed by introducing a nonparametric main effect estimand that still captures the (conditional) association of interest, even when the quantile model is misspecified. This estimand is estimated using the efficient influence function under the nonparametric model, allowing for the incorporation of data-adaptive procedures such as variable selection and machine learning. The approach provides a flexible and reliable method for detecting associations that is robust to model misspecification and excess uncertainty induced by variable selection methods.
