Parametric and semi-parametric bootstrap-based confidence intervals for robust linear mixed models

The linear mixed model (LMM) is a popular statistical for the analysis of longitudinal data. However, robust estimation and inferential conclusions LMM in presence outliers (i.e., observations with very low probability occurrence under Normality) not part mainstream data analysis. In this work, we compared coverage rates confidence intervals (CIs) based on two bootstrap methods, applied to three methods. We carried out simulation experiment compare CIs different conditions: 1) without contamination, 2) contaminated by within-, or 3) between-participant outliers. Results showed that semi-parametric associated composite tau-estimator leads valid decisions both uncontaminated This being most comprehensive study estimators LMM, provide fully commented R code all methods example.