Fixed Effects in General Linear Mixed Effect Modelling in Longitudinal Studies: Bootstrap and Modified Classical Procedures

Equivalent asymptotic procedures can yield conflicting inference when applied to the same dataset (Aubin & Cordeiro, 2000). Among the class of likelihood ratio tests, modified likelihood ratio tests (e.g., Bartlett, 1937) have been proposed and have documented improved performances over classical procedures. Modified F tests have also been proposed (e.g., Kenward & Roger, 1997), and have been recently been shown to compare favorably with modified likelihood ratio tests (Fouladi & Shieh, 2002). Bootstrapping is generally proposed as another means by which to obtain improved procedures under varied sampling distributions and conditions. In the current paper, the performance characteristics of bootstrap procedures based on residual resampling are contrasted with the performance of modified asymptotic test of fixed parameters in mixed linear models of data from a small sample repeated measures study. In particular, the rejection probabilities of these procedures under varied conditions are documented.