摘要
本文以2014年中国家庭追踪调查首轮并行数据为基础,采用多水平泊松回归模型,对无回答频次的影响因素进行了研究。研究结果表明,调查员的年龄、调查点发达程度以及居住环境对无回答次数有着重要影响,宏观和中观地区因素对无回答次数存在显著影响。泊松多水平回归在无回答研究的应用效果较好。最后,从技术和机制两个层面,探讨了有关追踪的建议和对策。
A multi-level Poisson regression is constructed and employed based on the parallel data,the first wave,of the2014 China Family Panel Survey,to study the impact effect of non-response frequency. The results show that,under the background of macro and meso,the non-response frequency is impacted highly by the age of interviewer,as well as the developed degree and living environment of the interviewing place. Through application,the multi-level Poisson regression is good at the study of non-response. Finally,in the technology and mechanism views,there are discussions of the tracking strategy and suggestion.
出处
《西北人口》
CSSCI
2016年第6期1-9,共9页
Northwest Population Journal
基金
国家科技部"十二五"国家科技支撑计划项目"人口与发展数学模型与综合决策支持系统"(2012BAI40B01)
国家卫生计生委<中国家庭发展追踪>项目的资助
关键词
无回答频次
多水平泊松回归
追踪措施
Non-Response Frequency
Multi-level Poisson Regression
Tracking Strategy
