摘要
通过对美国、加拿大和中国的劳动力调查方法在抽样设计、数据收集、数据处理、数据产品和开发应用四方面进行比较分析发现,中国劳动力调查在抽样设计上虽然采用较大的样本量,但在样本代表性、样本轮换模式上存在若干问题,不利于后期的估计;在数据收集上采用传统入户询问的方式既不便于数据的录入,又增加了基层调查员的负担;在数据处理上,缺乏数据处理过程中对无回答的调整、覆盖率的校准、季节调整,尤其忽视估计量的方差估计,导致调查估计结果不够科学;缺乏对劳动力调查数据产品和开发应用。然后,从连续性调查的角度对中国劳动力调查方法进行了深入思考,提出六点建议:一是加强人口普查和劳动力调查的衔接,二是合理分配样本量,三是构建和维护劳动力调查多重抽样框体系,四是改进轮换模式,五是建立基于连续性抽样的估计体系,六是形成以劳动力调查为核心的人口住户类调查体系。
This paper makes a comparative analysis of the labor force survey in China,U. S. and Canada from the four aspects of sampling design,data collection,data processing and data products. It finds,although China labor force survey employs larger sample size,the problems on representativeness of sample and rotation scheme made it difficult for estimation; the traditional face to face interview is not convenient for data entry,which implies increasing workload of primary investigator; ignoring of non- response adjustment,coverage calibration and seasonal adjustment,especially without regard to variance estimation of estimator,which leads to unscientific survey results; products based on labor force survey are rare and their development and application is also insufficient while plenty of materials are shelved. Then,it puts forward a series of reform measures from the angle of successive sampling survey,including effective linkage of census to labor force survey,sample allocation,the multiple sampling frame creation and maintenance,sample rotation,building a system of weighing and estimation based on successive sampling; and establishing a household surveys system with the labor force survey at its core.
引文
[1]U S DEPARTMENT OF LABOR,U S DEPARTMENT OFCOMMERCE.The current population survey:design and methodology,technical paper 66[EB/OL].[2015-09-15].http://www.census.gov/prod/2006pubs/tp-66.pdf.
[2]STATISTICS CANADA.Methodology of the canadian labour force survey[M].Ottawa:Statistics Canada,2008:1-67.
[3]国家统计局.劳动力调查制度[EB/OL].[2015-09-15].http://www.stats.gov.cn/tjsj/tjzd/gjtjzd/201501/t20150105_663428.html.
[4]孙玉环.改进住户调查数据质量的多维视角[J].中国统计,2011(4):12-13.
[5]陈光慧,刘建平.基于平衡单水平轮换的连续性抽样估计方法研究[J].统计研究,2008,25(10):81-85.
[6]杨盛菁.计算机辅助电话调查技术及其应用分析[J].统计科学与实践,2012(2):49-51.
[7]LEMAITRE G,DUFOURUFOUR J.An integrated method for weighting persons and families[J].Survey methodology,1987(13):199-207.
[8]国家统计局.关于“在‘十三五’期间启用新的就业统计指标”建议的答复(摘要)[EB/OL].[2015-11-18].http://www.stats.gov.cn/tjfw/jytadf/rddbjy/201511/t20151118_1277657.html.
(1)1954年,美国普查局的Shiskin首先开发了可以在计算机上运行的季节调整程序,称为模型I,此后该程序的每一次改进都以X加上序号表示。直到1965年美国普查局推出了比较完整的季节调整程序X-11,后来世界流行的X-11-ARIMA也是以X-11程序为核心的,而X-12-ARIMA作为X-11-ARIMA的提高版,是目前较成熟的季节调整方法之一。