A1102
Title: A two-stage method for integrating probability and non-probability samples with misclassified covariates
Authors: Hua Shen - University of Calgary (Canada) [presenting]
Abstract: Probability samples in clinical research often lack critical, accurately measured variables, while non-probability samples provide detailed data but are not representative and include measurement errors. The aim is to present a novel two-stage estimation technique to integrate these sample types, addressing misclassification in categorical covariates using a latent-variable framework without validation data. Through simulations, the method demonstrates superior accuracy over traditional approaches. Its effectiveness is validated on a real dataset to improve inference quality in studies involving misclassified covariates.
