ASSESSING GENERALIZED LINEAR MIXED MODELS USING RESIDUAL ANALYSIS

Abstract. A nonparametric smoothing method for assessing the adequacy of generalized linear mixed models (GLMMs) is developed. The proposed method is based on smoothing the residuals over continuous covariates to avoid the partition of continuous . GLMMs are regarded as conditional models, whereas GEE models are treated as marginal models. Generalized linear models represent a class of fixed effects regression models for different types of response variables including continuous, dichotomous, and counts, while GLMMs are obtained from generalized linear models by incorporating random effects into the linear predictors. Some papers related tothe topic of random-effects models can be referred to as Stiratelli et al. [5], Schall [6] and Zeger and Karim It is critical for all types of regression models to assess model adequacy before making any further inferences on model parameters. Ritz [8] developed goodness-of-fit tests for mixed models based on classical goodness-of-fit statistics; see Stephen