内容简介
Apreciseandaccessiblepresentationoflinearmodeltheory,illustratedwithdataexamplesStatisticiansoftenuselinearmodelsfordataanalysisandfordevelopingnewstatisticalmethods.Mostbooksonthesubjecthavehistoricallydiscussedunivariate,multivariate,andmixedlinearmodelsseparately,whereasLinearModelTheory:Univariate,Multivariate,andMixedModelspresentsaunifiedtreatmentinordertomakeclearthedistinctionsamongthethreeclassesofmodels.LinearModelTheory:Univariate,Multivariate,andMixedModelsbeginswithsixchaptersdevotedtoprovidingbriefandclearmathematicalstatementsofmodels,procedures,andnotation.Dataexamplesmotivateandillustratethemodels.Chapters7-10addressdistributiontheoryofmultivariateGaussianvariablesandquadraticforms.Chapters11-19detailmethodsforestimation,hypothesistesting,andconfidenceintervals.Thefinalchapters,20-23,concentrateonchoosingasamplesize.Substantialsetsofexcercisesofvaryingdifficultyserveinstructorsfortheirclasses,aswellashelpstudentstotesttheirownknowledge.Thereaderneedsabasicknowledgeofstatistics,probability,andinference,aswellasasolidbackgroundinmatrixtheoryandappliedunivariatelinearmodelsfromamatrixperspective.Topicscoveredinclude:AreviewofmatrixalgebraforlinearmodelsThegenerallinearunivariatemodelThegenerallinearmultivariatemodelGeneralizationsofthemultivariatelinearmodelThelinearmixedmodelMultivariatedistributiontheoryEstimationinlinearmodelsTestsinGaussianlinearmodelsChoosingasamplesizeinGaussianlinearmodelsFillingtheneedforatextthatprovidesthenecessarytheoreticalfoundationsforapplyingawiderangeofmethodsinrealsituations,LinearModelTheory:Univariate,Multivariate,andMixedModelscentersonlinearmodelsofintervalscaleresponseswithfinitesecondmoments.Modelswithcomplexpredictors,complexresponses,orboth,motivatethepresentation.