A0383
Title: Data integration with nonprobability sample: Semiparametric model-assisted approach
Authors: Danhyang Lee - Baylor University (United States) [presenting]
Sixia Chen - University of Oklahoma (United States)
Abstract: A novel semiparametric model-assisted estimation method is introduced that integrates data from both probability and nonprobability samples, thereby facilitating robust and efficient inferences regarding finite population parameters. To mitigate selection bias, whether ignorable or nonignorable, associated with the nonprobability sample, a flexible semiparametric propensity score model that extends beyond the missing at-random assumption is proposed. The approach employs a pseudo-profile-likelihood method to estimate the propensity score model. Subsequently, a difference estimator is constructed utilizing the probability sample as a foundation, where the proxy values of the study variable for the finite population are derived from the nonprobability sample using the estimated propensity score model. The asymptotic properties of the proposed estimators are presented, and formulae for variance estimation are provided. Through a series of simulations and a real data application, the proposed estimation procedure is validated, and its superiority over some existing estimators is demonstrated.
