歡迎來(lái)到天天文庫(kù)
瀏覽記錄
ID:7288931
大?。?65.50 KB
頁(yè)數(shù):14頁(yè)
時(shí)間:2018-02-10
《group differences in generalized linear models》由會(huì)員上傳分享,免費(fèi)在線閱讀,更多相關(guān)內(nèi)容在工程資料-天天文庫(kù)。
1、Chapter9GroupDifferencesinGeneralizedLinearModelsTimF.LiaoAbstractThischapterdealswithmakingcomparisonsbetweenTxedgroupsintheframeworkofgeneralizedlinearmodels.First,webrie?yintroducegeneralizedlinearmodels,themostcommontypeofregressionmodels.Next,wediscus
2、sasimplesystemforanalyzinggroupdifferencesinregression.Weprimarilyfocusontwotypesofcomparisons?analyzingdifferencesintheparametervectorsofthelinearpredictoranddifferencesintheunderlyingdistributionsforthegroupsinthemodel.Toillustratesuchcomparativemethodsf
3、orgroupdifferences,weperformanalysesusingreal-worlddata.Theoretically,groupdifferencesinregressionestimatescanbeviewedasanexampleofconditionalcausality.Practically,testinggroupdifferencesinregressionmayseemanyusefulapplicationsinsocialscienceresearch.Intro
4、ductionThischapterdealswiththetopicofmakinggroupcomparisonsinregressionanalysis.Byògroupcom-parisons,ówemeanexaminingthedifferencesincertainstatisticalpropertiesbetweenTxedgroups.òRegressionanalysisóreferstoanalysisapplyinganymemberofthefamilyofstatistical
5、modelsknownasgeneralizedlinearmodelsortheirextensions.Forcomparingunderlyingdistributions,theframeworkofgeneralizedlinearmodelsisconvenientbecausecertainmodelsmayassumedifferentstatisticaldistributions.Makingpropercomparisonsisfundamentallyimportantincondu
6、ctingempiricalsociologicalresearch,becauseòvirtuallyalltheorieswithempiricalramiTcationsimplysomeformoranotherofcomparisonó(Lieberson1985:45).ThischapterpresentsasystematicmethodformakingcomparisonsbetweenTxedgroupsintheregressionsetting,followingtheprinci
7、plesforsuchpurposesinLiao(2002).Wedo,however,gobeyondtheearliertreatmentbyexaminingmorecloselydistributionaldifferencesbetweengroupsandbetweenlinkfunctionsaswellaseffectsdecompositioninthesystematiccomponent.Mostoften,dataanalystsseektocompareparametersacr
8、ossTxedsocialgroupssuchasgender,race,andnationalitybytestingtheirhypothesizedequality.This,however,representsonlyone,albeitarguablythemostcommon,typeofcomparison.Inaddition,onemaybeinterestedincomparingacross
此文檔下載收益歸作者所有