Six Sigma ttests(PPT 57)英文

6-1•Z-test(n30)•t-Test(n=30)•2-test•F-Test6-3(SampleSize30)Z-testt-testSingleVarianceDoubleVariance2-testF-testVarianceUnknown(SampleSize=30)6-4中国最大的资料库下载SixSigmaIntroducethe2-samplet-testanditsimportanceincomparingmeansDiscusspracticalvs.statisticalsignificanceDiscussthecomparisonofvariancesObjectives6-6•HypothesisTestingComparesgroupsIstheresufficientstatisticalevidencetorejectthenullhypothesis?•Aftercollectingdata,howdoweTESTit?SeveraldifferenttestsAppropriatetestdependson:ObjectiveTypeofdata•HerewewilllookattestsforvariablesOutput(continuousY)datawithattributesInput(discreteX)data6-7:m=Target1-Samplet2levelsofDiscreteX,ContinuousYH0:m1=m22-Samplet3+levelsofDiscreteX,ContinuousYH0:m1=…=mkANOVAContinuousX;ContinuousYH0:Slope=0Regression6-8’stheY?______________TypeofData?_______________What’stheX?______________TypeofData?_______________Whattypeoftoolwouldyouuse?_________________________Example:TwoLevelsofXASupervisorwantstoknowiftwocustomerservicerepsspendsignificantlydifferentamountsoftimeontheirphonecallsTimeonphoneContinuousCSRDiscrete2-Samplet-test6-9•Uset-testwhencomparing:Samplemeantoanothersamplemean–2Samplet-testSamplemeantoaspecificvalue(e.g.,target,goal,orbelief)–1Samplet-testCoveredintheappendix•中国最大的资料库下载•TestisaSignal-to-NoiseratiothatcomparessampleaveragestowhatwouldbeexpectedifH0weretrue6-10(ifapplicable)StudyShapeStudySpreadStudyCenteringORGraphicalAnalysis6-11–ComparingmeansoftwoSamples•Thisisanexampleofahypothesistestthatcomparestwosamplemeans.H0:mBob=mJaneHa:mBob≠mJane•20samplesfromeachoperator6-12•It’seasiertogetusefulgraphsandanalyseswhenthedataareallinthesamecolumn,i.e.,Asinglecolumnforthemeasurementdata(Y)Onecolumnforeachofthe‘identifiers’(X)•Inourexample,usethefollowingcolumns:Person(withBobandJaneasthevalues)Time(withallofBob’sandJane’svalues)•Readingacrossarowgivesthetimeandoperatorwhoansweredthecall6-13•t-testistheSignal-to-Noiseratiousedtoevaluatethishypothesis•t-testmakestwoassumptionsStableprocessNormaldistributionTeststillperformswellwithmoderatedeparturesfromnormality6-16–2SamplesMinitabWhattolookfororaskComparisonsInvolvingXwith2LevelsStudyStability(ifapplicable)SPCChartI-MRIsthedataineachgroupstableovertime?Lookforanytrendsorpatterns.6-17•UseanI-MRcharttoevaluatethestabilityofthedataSincetherearetwogroupsofdata,usethe“EstimateParametersBYGroupsin…”optioninMinitab6-18=4.495UCL=6.469LCL=2.521BobJane210MovingRangeR=0.7421UCL=2.425LCL=0BobJaneIandMRChartforTimebyPersonWhatdecisiondowemake?ArethereanysignificantpatternstoindicatethateitherBob’sorJane’sdataisnotfromjustonepopulation/process?6-19(ifapplicable)StudyShapeSPCChartI-MRIsthedataineachgroupstableovertime?Lookforanytrendsorpatterns.DescriptiveStats,NormalProbabilityPlotPointsshouldfallonastraightlineonthenormalprobabilityplot.t-testswillstillworkaslongasthedataisnotbadlyskewed.6-20*GoodnessofFitNormalProbabilityPlotforTimeByPersonMLEstimatesNormalityCheckBothsetsofpointsfallreasonablyclosetoastraightline6-22(ifapplicable)StudyShapeSPCChartI-MRIsthedataineachgroupstableovertime?Lookforanytrendsorpatterns.Desc