作者:标准资料网 时间:2024-05-15 15:03:51   浏览:9299   来源:标准资料网

【英文标准名称】:StandardPracticeforEstimatingThurstonianDiscriminalDistances
【原文标准名称】:评定瑟斯顿识别距离的标准实施规程
【标准号】:ASTME2262-2003(2009)
【标准状态】:现行
【国别】:美国
【发布日期】:2003
【实施或试行日期】:
【发布单位】:美国材料与试验协会(US-ASTM)
【起草单位】:E18.03
【标准类型】:(Practice)
【标准水平】:()
【中文主题词】:
【英文主题词】:Forced-choicedatasets;Thurstoniandiscriminaldistances;Unidimensionalequal-variancemodel
【摘要】:Undertheassumptionsofthemodel,theThurstonianmodelapproachtomeasuringtheperceiveddifferencebetweentwosamples(whetheroverallorforaspecificattribute)isindependentofthesensorymethodusedtocollectthedata.Convertingresultsobtainedfromdifferenttestmethodstod''valuespermitstheassessmentofrelativedifferencesamongsampleswithoutrequiringthatthesamplesbecomparedtoeachotherdirectlyorthatthesametestmethodsbeusedforallpairsofsamples.Thurstonianscalinghasbeenappliedto:Creatingahistoricaldatabasetotrackdifferencesbetweenproductionandreferencesamplesoverperiodsinwhichdifferenttestmethodswereusedtomeasurethedifference,Comparingtherelativesensitivitiesofdifferentusergroupsandconsumersegments,Comparingtrainedpanelsthatusedifferentmeasuringtechniques,Comparingtherelativesensitivitiesofconsumersversustrainedpanels,Comparingdifferentmethodsofconsumertesting(forexample,CLTversusHUT,preferenceversushedonicscales,etc.),andComparingdifferentdiscriminationtestmethods.1.1ThispracticedescribesprocedurestoestimateThurstoniandiscriminaldistances(thatis,d''values)fromdataobtainedontwosamples.Proceduresarepresentedforfourforced-choicemethods(thatis,thetriangle,theduo-trio,the3-alternative-forced-choice(or3-AFC)andthe2-AFC(alsocalledthedirectionaldifferencetest)),theA/Not-Amethod,theSame-Differentmethodandfordataobtainedfromorderedcategoryscales.Proceduresforestimatingthevarianceofd''arealsopresented.Thus,confidenceintervalsandstatisticaltestscanbecalculatedford''.1.2Theproceduresinthisdocumentpertainonlytotheunidimensional,equal-variancemodel.Other,morecomplicatedThurstonianmodels,involvingmultipledimensionsandunequalvariancesexistbutarenotaddressedinthisstandard.Theprocedureforforced-choicemethodsislimitedtodichotomousresponses.TheprocedurefortheA/Not-Amethodassumesequalsamplesizesforthetwosamples.TheprocedurefortheSame-Differentmethodassumesequalsamplesizesforthematchedandunmatchedpairsofsamples.Forallmethods,onlyunreplicatedtestsareconsidered.(Testsinwhicheachassessorperformsmultiple(thatis,replicated)evaluationsrequiredifferentanalyses.)1.3Thurstonianscalingisamethodformeasuringtheperceptualdifferencebetweentwosamplesbasedonaprobabilisticmodelforcategoricalchoicedecisionmaking.Themagnitudeoftheperceiveddifference,x03B4;,canbeestimatedfromtheassessors''categoricalchoicesusingthemethodsdescribedinthispractice(SeeAppendixX3foramoredetaileddescriptionofThurstonianscaling).1.4Intheory,theThurstonianx03B4;doesnotdependonthemethodusedtomeasurethedifferencebetweentwosamples.Assuch,x03B4;providesacommonscaleofmeasureforcomparingsamplesmeasuredunderavarietyoftestconditions.Forexample,Thurstonianscalingcanbeusedtocompareproductsmeasuredunderdifferenttestconditions,tocomparepanels(trained,consumerorboth)thathaveevaluatedthesamesamples(usingthesameordifferenttestmethods)andtocomparetestmethodsontheirabilitytodiscriminatesamplesthatexhibitafixedsensorydifference.1.5Thisstandardmayinvolvehazardousmaterials,operationsandequipment.Thisstandarddoesnotpurporttoaddressallofthesafetyconcerns,ifany,associatedwithitsuse.Itistheresponsibilityoftheuserofthisstandardtoestabl......
【中国标准分类号】:A52
【国际标准分类号】:67_240
【页数】:47P.;A4
【正文语种】:英语