【英文标准名称】:PracticeforStatisticalAssessmentandImprovementoftheExpectedAgreementBetweenTwoTestMethodsThatPurporttoMeasuretheSamePropertyofaMaterial
【原文标准名称】:测量某种材料相同性质的两种试验方法间预期一致性的统计学评估和改进措施的实施规程
【标准号】:ANSI/ASTMD6708-2007
【标准状态】:作废
【国别】:美国
【发布日期】:2007
【实施或试行日期】:2007-11-27
【发布单位】:美国国家标准学会(US-ANSI)
【起草单位】:ANSI
【标准类型】:()
【标准水平】:()
【中文主题词】:协议;评估;改进;材料性能;材料试验;石油产品;特性;统计;统计学;试验方法;测试
【英文主题词】:Agreements;Assessment;Improvement;Materialbehaviour;Materialtests;Petroleumproducts;Property;Statistical;Statistics;Testmethods;Testing
【摘要】:Thispracticecoversstatisticalmethodologyforassessingtheexpectedagreementbetweentwostandardtestmethodsthatpurporttomeasurethesamepropertyofamaterial,anddecidingifasimplelinearbiascorrectioncanfurtherimprovetheexpectedagreement.ItisintendedforusewithresultscollectedfromaninterlaboratorystudymeetingtherequirementofPracticeD6300orequivalent(forexample,ISO4259).Theinterlaboratorystudymustbeconductedonatleasttenmaterialsthatspantheintersectingscopesofthetestmethods,andresultsmustbeobtainedfromatleastsixlaboratoriesusingeachmethod.Note1-ExamplesofstandardtestmethodsarethosedevelopedbyvoluntaryconsensusstandardsbodiessuchasASTM,IP/BSI,DIN,AFNOR,CGSB.Thestatisticalmethodologyisbasedonthepremisethatabiascorrectionwillnotbeneeded.Intheabsenceofstrongstatisticalevidencethatabiascorrectionwouldresultinbetteragreementbetweenthetwomethods,abiascorrectionisnotmade.Ifabiascorrectionisrequired,thentheparsimonyprincipleisfollowedwherebyasimplecorrectionistobefavoredoveramorecomplexone.Note2-Failuretoadheretotheparsimonyprinciplegenerallyresultsinmodelsthatareover-fittedanddonotperformwellinpractice.Thebiascorrectionsofthispracticearelimitedtoaconstantcorrection,proportionalcorrectionoralinear(proportional+constant)correction.Thebias-correctionmethodsofthispracticearemethodsymmetric,inthesensethatequivalentcorrectionsareobtainedregardlessofwhichmethodisbias-correctedtomatchtheother.Amethodologyispresentedforestablishingthe95%confidencelimit(designatedbythispracticeasthebetweenmethodsreproducibility)forthedifferencebetweentworesultswhereeachresultisobtainedbyadifferentoperatorusingdifferentapparatusandeachapplyingoneofthetwomethodsXandYonidenticalmaterial,whereoneofthemethodshasbeenappropriatelybias-correctedinaccordancewiththispractice.Note3-Inearlierversionsofthisstandardpractice,theterm"cross-methodreproducibility"wasusedinplaceoftheterm"betweenmethodsreproducibility."Thechangewasmadebecausethe"betweenmethodsreproducibility"termismoreintuitiveandlessconfusing.Itisimportanttonotethatthesetwotermsaresynonymousandinterchangeablewithoneanother,especiallyincaseswherethe"cross-methodreproducibility"termwassubsequentlyreferencedbynameinmethodswhereaD6708assessmentwasperformed,beforethechangeinterminologyinthisstandardpracticewasadopted.Note4-Usersarecautionedagainstapplyingthebetweenmethodsreproducibilityascalculatedfromthispracticetomaterialsthataresignificantlydifferentincompositionfromthoseactuallystudied,astheabilityofthispracticetodetectandaddresssample-specificbiases(see6.8)isdependentonthematerialsselectedfortheinterlaboratorystudy.Whensample-specificbiasesarepresent,thetypesandrangesofsamplesmayneedtobeexpandedsignificantlyfromtheminimumoftenasspecifiedinthispracticeinordertoobtainamorecomprehensiveandreliable95%confidencelimitsforbetweenmethodsreproducibilitythatadequatelycovertherangeofsamplespecificbiasesfordifferenttypesofmaterials.
【中国标准分类号】:E30
【国际标准分类号】:75_080
【页数】:
【正文语种】:英语