Multilevel Modelling of Country Effects: A Cautionary Tale
Multilevel Modelling of Country Effects: A Cautionary Tale
复制标题
DOI：
10.1093/esr/jcv059
复制
发表时间：
2016-02-01
影响因子：
3.2
通讯作者：
Jenkins, Stephen P.
中科院分区：
文献类型：
--
作者：
Bryan, Mark L.;Jenkins, Stephen P.
国家对个人结果的影响通常使用适用于统一的多国数据集的多层次（分层）模型进行分析，如ESS，EU-SILC，EVS，ISSP和SHARE。我们指出的问题，国家的影响，似乎没有得到广泛重视的评估，并发展我们的论点，使用Monte Carlo模拟分析多层次的线性和logit模型。由于每个国家内的个人样本量很大，但只有少数国家，分析人员可以可靠地估计个人层面的影响，但对概括国家影响的参数的估计可能不可靠。多层次建模方法不是万能药。
Country effects on outcomes for individuals are often analysed using multilevel (hierarchical) models applied to harmonized multi-country data sets such as ESS, EU-SILC, EVS, ISSP, and SHARE. We point out problems with the assessment of country effects that appear not to be widely appreciated, and develop our arguments using Monte Carlo simulation analysis of multilevel linear and logit models. With large sample sizes of individuals within each country but only a small number of countries, analysts can reliably estimate individual-level effects but estimates of parameters summarizing country effects are likely to be unreliable. Multilevel modelling methods are no panacea.