报告题目：Using Auxiliary Information in Probability Survey Data to Improve Pseudo-Weighting in Non-Probability Samples: A Copula Model Approach

报告时间：2023年07月07日（周五）

线下报告地点：数学楼2-2会议室 9：30-11：00

线上腾讯会议：278 939 678

报告摘要：Although probability sampling has long been regarded as the gold standard for survey methods, nonprobability sampling, such as online opt-in surveys, have gained popularity due to their convenience and cost-effectiveness. However, nonprobability samples can introduce estimation bias due to the unknown nature of the underlying selection mechanism. In this talk, we present a parametric approach to integrate probability and nonprobability samples using shared ancillary variables. It assumes that the joint distribution of ancillary variables follows a latent Gaussian copula model, and logistic regression is used to model the mechanism by which population units enter the nonprobability sample. The unknown parameters in the copula model are estimated through the pseudo maximum likelihood approach, and the logistic regression model is estimated by maximizing the sample likelihood constructed from the nonprobability sample. Our simulation results demonstrate that the proposed method effectively corrects selection bias in nonprobability sample by consistently estimating the underlying inclusion mechanism. By leveraging additional information from the nonprobability sample, the combined method provides a more efficient estimation of the population mean than using the probability sample alone. A real data application is provided to illustrate the practical use of the proposed method.

报告人简介：薛兰，美国俄勒冈州立大学统计系教授，于2005年在密西根州立大学获得博士学位。主要研究方向包括非/半参数回归，函数型数据分析，网络数据分析，时空数据分析等，研究内容涉及统计学，数学以及图像分析和大数据等相关问题。论文发表于 Annals of Statistics, Journal of the American Statistical Association, Biometrics, Statistica Sinica 等重要期刊，现任Statistica Sinica, Journal of the American Statistical Association, Biostatistics and Epidemiology, 曾任 Sankhya series B 期刊的编委。