Optimal tests for random effects in linear mixed models

In the past decade, mixed-effects modeling has received a great deal of attention in applied and theoretical statistical literature. They are very flexible tools analyzing repeated measures, panel data, cross-sectional hierarchical data. However, complex nature these models motivated researchers to study different aspects this problem. One which is test significance random effects used model unobserved heterogeneity population. The method likelihood ratio based on normality assumption error term been proposed. does not necessarily hold practice. paper, we propose an optimal so-called uniform local asymptotic detect possible presence linear mixed models. We show that proposed statistic is, consistent, locally asymptotically even for require traditional comparable classical L.ratio-test when standard assumptions met. Finally, simulation studies real data analysis also conducted empirically examine performance procedure.