Abstract:

Province-level mortality rates provide better indicators to describe the mortality pattern and its geographical differences than national level data. However, the limited data availability has made it hard to conduct research on provincelevel mortality and life expectancy. This paper employs ageand genderspecific mortality data, categorized by residential status (urban/rural), from the 6th Census. In its analysis, the paper uses China National Maternal and Child Health Surveillance System (MCHS) data to modify the Under 5 Mortality Rate (U5MR); applies Kannisto model to revise the mortality data of the oldest old people; and revised the mortality data for the elderly based on the comparison with Japans historical data from 1960 to 2010. In doing so, we are able to calculate life expectancy categorized by province, gender, and residential status, providing five mortality patterns according to longevity and urbanrural divide. The results show that in 2010, life expectancies at birth, 60 and 80 were 74.85, 19.44 and 6.90 respectively and that the variance of life expectancy in rural areas was bigger than that in urban areas.

Keywords: Population Census, Age-specific Mortality, Life Expectancy, The Kannisto Model

摘要：

分省死亡率研究能更准确地描述我国人口的死亡模式，度量区域之间的差异。由于基础数据的缺乏，目前对分省死亡率和预期寿命的研究还不够充分。文章系统整理了我国大陆31个省(市、区)的“六普”资料，获取了分省、分性别、分城乡的年龄别死亡数据。利用妇幼卫生监测系统分县数据，修正了5岁以下儿童死亡率，利用Kannisto模型修正了高龄人口死亡率，并通过与日本1960~2010年的单岁组死亡率比较，对各省的老年死亡率进行再修正。最后，文章测算出分省、分性别、分城乡的0岁、60岁和80岁组预期寿命，并按照长寿与城乡差异水平将死亡模式分为5类。结果显示，“六普”时期我国0岁组预期寿命为74.85岁，60岁组为19.44岁，80岁组为6.90岁；各省农村人口预期寿命的差异明显。

关键词: 人口普查, 年龄别死亡率, 预期寿命, Kannisto模型