Blundell R, Windmeijer F
Department of Economics, University College London, UK.
Health Econ. 1997 Jul-Aug;6(4):439-43. doi: 10.1002/(sici)1099-1050(199707)6:4<439::aid-hec276>3.0.co;2-b.
For small group sizes, the GLS estimator in multilevel models is biased and inconsistent when the random cluster effects are correlated with the regressors. A fixed effects approach, conditioning on the cluster effects, provides consistent estimates for the slope parameters. The two estimators are equivalent when group sizes are large. The same results obtain for two-stage estimation procedures that allow for some of the regressors to be simultaneously determined with the dependent variable. The GLS and fixed effects estimators are applied to data on acute care hospital utilization in the UK, allowing for health authority district effects.
对于小组规模，当随机聚类效应与回归变量相关时，多层次模型中的广义最小二乘（GLS）估计量存在偏差且不一致。一种固定效应方法，以聚类效应为条件，可为斜率参数提供一致的估计。当小组规模较大时，这两种估计量是等效的。对于允许一些回归变量与因变量同时确定的两阶段估计程序，也会得到相同的结果。GLS估计量和固定效应估计量被应用于英国急性护理医院利用情况的数据，并考虑了卫生当局地区效应。
