Multivariate generalized linear models using R

Abstract

Multivariate Generalized Linear Mixed Models Using Rpresents robust and methodologically sound models foranalyzing large and complex data sets, enablingreaders to answer increasingly complex researchquestions. The book applies the principles of modelingto longitudinal data from panel and related studiesvia the Sabre software package in R. The authors firstdiscuss members of the family of generalized linearmodels, gradually adding complexity to the modelingframework by incorporating random effects. Afterreviewing the generalized linear model notation, theyillustrate a range of random effects models, includingthree-level, multivariate, endpoint, event history,and state dependence models. They estimate themultivariate generalized linear mixed models (MGLMMs)using either standard or adaptive Gaussianquadrature. The authors also compare two-level fixedand random effects linear models. The appendicescontain additional information on quadrature, modelestimation, and endogenous variables, along withSabreR commands and examples. In medical and socialscience research, MGLMMs help disentangle statedependence from incidental parameters. Focusing onthese sophisticated data analysis techniques, thisbook explains the statistical theory and modelinginvolved in longitudinal studies. Many examplesthroughout the text illustrate the analysis ofreal-world data sets. Exercises, solutions, and othermaterial are available on a supporting website.