摘要
由于常规生存模型假定删截和事件发生相互独立,导致研究者常常忽略删截案例中可能还包含经历事件的风险为零的"长期存活者"。忽略这一总体异质性会导致有偏的参数估计。文章介绍并应用分割总体生存模型来探究初婚初育史对中国育龄妇女的二孩生育间隔的影响。结果发现,在控制个体特征和所处文化背景特性的情况下,对于有可能生育二孩的妇女,初婚年龄的推迟、初育为女孩的情况都将导致二孩生育间隔的缩短;所属初育队列越晚、初育孩子健康、初育后哺乳时间越长、间隔期内人工流产次数越多则都会显著延长二孩生育间隔。
Standard Survival Models always assume that censoring is independent of the event under study. This makes researchers sometimes omit such a phenomenon that some long-term survivors,who will never experience the event,are treated as the censored observations who will eventuallyexperience the event. Omittingthis population heterogeneitywill bias the estimated parameters. Usingthe data from 'National Demographic and Reproductive Health Survey in 1997' ,the paper introduces and applies the split population survival model to the study on the effects of first marriage and first birth and their results on second birth interval for Chinese women at childbearing. The results indicate that,holding the individual characteristics of women and the nature of cultural context in which the women live constant,the birth space of second birth is affected significantly statistically by year of first marriage,age of first marriage,sex of first birth,health status of first birth,length of breastfeeding for first birth,and amount of induced abortion women experienced between first and second births.
出处
《中国人口科学》
CSSCI
北大核心
2010年第1期36-45,共10页
Chinese Journal of Population Science
基金
国家社科基金重点项目"中国人口发展与经济社会可持续发展战略研究"(项目批准号:08ARK001)阶段性研究成果
作者简介
巫锡炜 北京大学社会学系，博士研究生。
