Selection bias and sensitivity analysis
17页1、1 SelectionBiasandSensitivityAnalysis1 OverviewofselectionbiasRosenbaum ssensitivityanalysisStatalab runningsensitivityanalysiswithrbounds 2 ReadingsforThisSession Guo S Fraser M 2009 PropensityScoreAnalysis ThousandOaks CA SagePublications Inc Sections8 1 8 3 8 6ofChapter8 Rosenbaum P R 2002 Observationalstudies 2nded NewYork Springer Chapter4 Rosenbaum P R 2005 Sensitivityanalysisinobservationalstudies InB S EverittandD C Howell Eds EncyclopediaofStatisticsinBehavioralScience pp 1809 1814 NewY
2、ork JohnWiley Sons Ltd 3 Overviewofselectionbias 4 ImportancetoStudySelectionBias Althoughthenewmethodsofpropensityscoreanalysis PSA addresssomeofthecorechallengesfoundincausalinferencesofobservationalstudies theycannotcompetewithrandomizedcontrolledtrial RCT onakeyissue PSAonlycontrolsforselectionbiasgeneratedbymeasuredcovariates whileRCTbalancedataandcontrolsforbiasgeneratedbybothmeasuredandunmeasuredcovariates Itisforthisreason researchersusingPSAshouldbealertedbyselectionbiases particularlyb
3、ythosegeneratedbytheunmeasuredcovariates 5 SourcesofSelectionBias Self selectionResearcher sselectionTheclassicexample thefailuretopredictHarryS Truman s1948presidentialvictoryThemodernexample internetsurveyNonprobabilitysamplingAdministrator sselectionGeographicselectionorgeographicmismatchMeasurementerror nonrandomerrorRater seffectinlongitudinalstudySelectionduetoattrition 6 OvertBiasversusHiddenBias Rosenbaum 2002 p 71 Anobservationalstudyisbiasedifthetreatedandcontrolgroupsdifferpriortotrea
4、tmentinwaysthatmatterfortheoutcomesunderstudy Anovertbiasisonethatcanbeseeninthedataathand forinstance priortotreatment treatedsubjectsareobservedtohavelowerincomesthancontrols Ahiddenbiasissimilarbutcannotbeseenbecausetherequiredinformationwasnotobservedorrecorded Overtandhiddenbiasesthataffectcausalinferencearecontrolledinwell implementedrandomizedexperiments Balancingthedataonhiddenbiasisespeciallyaprobleminobservationalstudies 7 Rosenbaum ssensitivityanalysis 8 TheBasicIdea 1 Correctingthehi
5、dden selectionproblemessentiallyinvolvestakingstepssuchasrespecifyingananalyticmodelbyusingadditionalvariables collectingadditionaldata redesigningastudyasarandomizedexperiment ormakingotherendeavors Evidently thesestrategiescanbeexpensiveandtime consuming Itisoftenpreferable then tostartbyconductingsensitivityanalysestoestimatethelevelofbias Thistypeofanalysisseekstoanswerthequestion Howsensitivearefindingstohiddenbias Althoughsensitivityanalysisisexploratory itisanimportantstepinanalysesusingt
《Selection bias and sensitivity analysis》由会员共***分享,可在线阅读,更多相关《Selection bias and sensitivity analysis》请在金锄头文库上搜索。
2024-03-27 17页
2024-03-27 31页
2024-03-27 26页
2024-03-27 34页
2024-03-27 18页
2024-03-27 25页
2024-03-27 24页
2024-03-27 25页
2024-03-27 26页
2024-03-27 27页