摘要
对时变性因素的处理方法将直接影响到人口迁移选择性分析结果的可靠性。文章利用第五次全国人口普查数据,解析了处理时变性因素的三种方法(调查时点法、局部倒推法和全面倒推法),剖析了不同方法下数据结构、统计模拟及实证结果的变化与差异。研究表明,不同数据处理方法将导致不同的数据结构,且导致时变性变量的均值和方差出现变化;时变性变量和非时变性变量对因变量作用的力度和统计显著性产生了明显的影响;当对全面倒推法下分年的迁移人口数据进行分析时,发现教育的选择性与传统方法出现差异,这与经验观察相符。
Time-varying factors have been attached highly importance in the methodology of migration research.Using the 2000 Chinese population census micro data,this paper presents three methods(survey time method,partial retroversion method and holistic retroversion),which are different from each other in terms of theoretical implication but closely related in the way of data structure.The variable choice in methods demonstrates a sharp contrast effect in data structure in the means and variance of the variables.Besides,it also alters the strength and significance of independent variables on the dependent variable.The holistic retroversion benefits the research in finding a diverse pattern of educational selectivity for migration other than a uniform one of positive selection,which is consistent with the empirical observation.
出处
《中国人口科学》
CSSCI
北大核心
2008年第2期65-71,共7页
Chinese Journal of Population Science
