Effects of Respondent Household and Its Head’s SES on Panel Refusal: A case study of CFPS

在梳理拒访影响因素理论模型的基础上,本文运用中国家庭动态跟踪调查(China Family Panel Study,CFPS)测试初访和测试第一次追访的并行数据和问卷数据,运用Logit模型对CFPS测试第一次追访中出现的受访家户拒访行为进行分析。分析发现,CFPS追访调查样本存在选择性拒访现象。追访中,被拒访的特点和模式与发达国家的基本一致,即受到受访家户和主事者社会经济特征、参访经历等因素的影响。分析还发现,在追访中,受访对象的拒访行为是可预测的。根据前次收集的并行数据和问卷数据,可以甄别未来可能拒访的重点对象。如能依此采取特别的样本维护策略,应可降低跟踪调查样本的损失率。

This article examines the panel refusal in a large panel survey in China， China Family Panel Study (CFPS). Logit model is used for attempting to identify the panel refusal occurred in the follow-up CFPS pilot survey. Paradata and survey data from both baseline and follow up of the CFPS pilot surveys are used. The panel refusal on unconditional distribution of socioeconomic variables and on estimates of regression coefficients are explored，and the result shows that panel refusal is highly selective in conditions of respondent household and its head’s SES and their experiences of surveys. The analysis illustrates that refusal in CFPS is quite similar in many ways to other panel studies in developed countries and is predictable. Paradata and survey data collected in the prior wave can be used to predict the follow-up refusal. 

孙妍 邹艳辉 丁华 严洁 顾佳峰 邱泽奇

欢迎阅读《社会学研究》！您是该文第3007位读者！