Diagnostic Measures in Restricted Linear Mixed Models

In this paper, we have presented the diagnostic measures to identify outlier and influential observations in restricted linear mixed models. Several methods of diagnosis have been used as a means of identifying influential observations, such as the Cook’s distance, the likelihood distance, and covariance ratio methods. Through a parametric bootstrap procedure, we have obtained empirical distributions of the test statistics, and additionally as a way to evaluate the performance of the test statistics one specified simulation study has been used. Finally, a numerical example is given to illustrate the theoretical results.

Keywords:

Case deletion , Diagnostic measures , Influential observations , Linear mixed models , Mean shift outlier model