这个问题涉及处理嵌套数据技术答案question.pdfVIP

Thisquestionconcernsthetechniquestodealwithnesteddata.Thefirstpoint

askswhyOLSregressionisproblematicinanalyzingmultiple-leveldata.Thesecond

pointasksustolisttwowaysinhandlingmultiple-leveldata.Thethirdpointrefersto

theapplicationofoneoftheprocedure(e.g.HLM)inrealworldresearch.Thelast

pointconcernshowtoassessmodelfitandnumbersofvariables.

Inansweringthisquestion,IwillfirstoutlinetheassumptionsofOLSregression

andexplainthereasonswhytheseassumptionsareviolatedbymultiple-leveldata,

takingdisaggregationasanexample.Second,Iwilldemonstratetwostatistical

techniquestohandlenesteddata:HLMandWABA,alongwiththeirappropriate

conditionsandlimitations.Third,IwillfocusonHLM,andusethepaperofBommer

andDierdorff(2007)regardingtheOCB-performancerelationstoillustratethestep-

by-stepanalyticstrategyofHLM.In,Iwilldemonstratehowtoassess

goodnessofmodelfitandnumbersofvariables.Morespecifically,inOLSregression,

22

R,ΔRandFtestarenecessary.Formulti-levelsituations(e.g.HLM),pseudoRare

required.AmongAIC,BICandloglikelihoodratiotest,alikelihoodratiotest

requiresnestedmodels.

OLSRegressionAssumptionsandViolations

Organizationsarehierarchicalinnature,consistedofmultiplelevels.According

toHofmann(1997),organizations“haveindividualsnestedwithinworkgroups,work

groupsnestedindepartments,departmentsnestedinorganizations,andorganizations

nestedinenvironments.”Thus,itisnecessarytofigureoutthewaysinwhichwecan

betterinvestigateandtestthesehierarchicalrelationshipsacrossdifferentlevels.More

specifically,Hofmann(1997)definedhierarchicalrelationshipsasthe