階層隨機效應模型常用來分析具有相關性的群集資料，而階層隨機效應模型組成分配的選擇常常是為了計算的方便。 然而，本文的重點是在階層隨機效應模型的組合分配是否影響其最終的邊際分配具有可強韌化性質。我們將更換組成模型的分配，並探討邊際分配的迴歸參數之最大概似估計量和母體平均數之最大概似估計量是否具有一致性，推斷邊際分配是否具有可強韌化性質。 ;Hierarchical random effects models are convenient for modeling clustered data. More generally, model pecifications involved in the hierarchy are based on the ease of computation. We, on the other hand, focus on the issue of robustness. More specifically, we explore how the intermediate distributions in the hierarchy affect the robustness of the ultimate marginal distribution in providing consistent estimates for the population means/regression coefficients.
