ScholarWorks@숙명여자대학교

초록

Respondents to panel surveys are commonly divided into continuous and noncontinuous groups based on their response patterns. In this study, we propose an estimator based on composite weights as an effective nonresponse adjustment method to reduce the bias of noncontinuous response groups. We derive the properties of the proposed estimator, such as its bias and mean squared error, and then compare its efficiency with that of alternative estimators. We present the results of simulations demonstrating that the proposed estimator exhibits less variance than the conventional method of directly using the response rate of a noncontinuous response group. It also exhibited a lower bias than that obtained using the response rate of the continuous response group. The composite weighting method used in the proposed estimator showed stable results in terms of minimizing extreme weights, indicating that it may be considered highly effective for noncontinuous response groups in panel surveys. © 2023, Korean Statistical Society.

키워드