Abstract: Recent studeis on the fertility level in China are of controversy as various mehods and data sources have led to inconsistent results. This paper applies Preston and Coale’s (1982) variable-r method to assess the fertility level in China between 2000 and 2010. This method has been applied in various kinds of research and proved efficacious. By adopting relative age distribution data from China’s 2000 and 2010 censuses as well as fertility pattern from annual sampling surveys published by National Bureau of Statistics of China, the variable-r method under different scenarios of completeness of coverage suggests that Chinese fertility during 2000 and 2010 is around 1.60. This paper also emphasizes the imporantce of examing the underlying assumptions of this method before applying it in emprical research.

Keywords: Variable-r Methods, Data Quality, Completeness of Coverage in Census, Net Reproductive Rate, Fertility Rate

摘要： 近年来，有关中国妇女近期生育水平的研究很多，各类研究由于采用的数据和方法不同，结果不尽相同。在这些讨论中，出生人口数的数据质量和调整是各类研究结果差异较大的关键。文章使用2000年、2010年两次中国人口普查中提供的妇女人口数和国家统计局公布的10年间的生育分布模式，在考察两次人口普查女性人口漏报的基础上，应用Preston和Coale提出的变量r方法，对中国2000～2010年的净人口再生产率进行了估计。同时，文章计算出这10年间的总和生育率约为1.60左右。变量r方法对于具体的出生人口数量和育龄妇女数量的数据质量要求较低，且在多项研究中经过实践的检验。但是，在实际研究中无论应用何种人口学方法，都应注意其前提假设，仔细检查数据，减少可能存在的偏差。

关键词: 变量r方法, 数据质量, 普查漏报率, 净人口再生产率, 生育水平