Robust MM-Estimation and Inference in Mixed Linear Models

Mixed linear models are used to analyse data in many settings. These models generally rely on the normality assumption and are often fitted by means of the maximum likelihood estimator (MLE) or the restricted maximum likelihood estimator (REML). However, the sensitivity of these estimation techniques and related tests to this underlying assumption has been identified as a weakness that can even lead to wrong interpretations. Recently Copt and Victoria-Feser(2005) proposed a high breakdown estimator, namely an S-estimator, for general mixed linear models. It has the advantage of being easy to compute - even for highly structured variance matrices - and allow the computation of a robust score test. However this proposal cannot be used to define a likelihood ratio type test which is certainly the most direct route to robustify an F-test. As the latter is usually a key tool to test hypothesis in mixed linear models, we propose two new robust estimators that allow the desired extension. They also lead to resistant Wald-type tests useful for testing contrasts and covariate efects. We study their properties theoretically and by means of simulations. An analysis of a real data set illustrates the advantage of the new approach in the presence of outlying observations.