Treść książki
Przejdź do opcji czytnikaPrzejdź do nawigacjiPrzejdź do informacjiPrzejdź do stopki
AgnieszkaZięba,JanKordos1
Comparingthreemethodsofstandard
errorestimationforpovertymeasures
ABSTRACT
Indicatorsofpovertyandsocialexclusionareanessentialtoolformonitoring
progressinthereductionoftheseproblems.Forthisreason,asetofpovertymeasures
commonlynamedasLeakenindicatorsattheEuropeanCouncilSummitinDecember
2001wasestablished.MostoftheseindicatorsarecalculatedaccordingtotheEurostat
recommendations,usingdatafromEuropeanStatisticsonIncomeandLiving
Conditions(EU-SILC).Complexsampledesignofthissurveyrequiresapproximate
methodsofstandarderrorestimation,generallylinearizationorreplicationtechniques.
InourstudythreesuchmethodsforchosenLaekenindicatorsarepresented.We
compareresultsofbootstrap,jackknifeandlinearizationmethods.
Introduction
Assessmentofvariabilityinestimatorsofunknownparametersisone
ofthemostimportantissuesinstatisticalinference.Itisacommonpractice
toreporttheestimatesofunknownparameterwithestimatesofthevariance
orthestandarderror[Shao,2003].Derivationofsuitableestimatorofthe
varianceofthegivenstatisticisdifficultincaseofnon-linearstructureof
estimatedmeasureandcomplexsurveydesign[Sitter,1992].Incomepoverty
measuresaresuchkindofmeasuresandarecalculatedaccordingtothe
Eurostatrecommendations,usingdatafromEuropeanStatisticsonIncomeand
LivingConditionsEU-SILC[Eurostat,2007].Usingthissurveyitisnecessary
totakeintoconsiderationthattheselectionofthesurveysampleisdoneby
two-stagestratifiedsamplingwithdifferentselectionprobabilitiesatthefirst
stage.Thiscaserequiresapproximatestandarderrorestimation.Itwouldbe
desirabletoprovideonemethodologyofassessingaccuracyofestimation.In
ourstudythreefrommostcommonmethodsforfivepovertyindicatorsare
presentedandcompared:bootstrap,jackknifeandlinearizationmethods.
1
WarsawSchoolofEconomics.
