版權(quán)說明:本文檔由用戶提供并上傳,收益歸屬內(nèi)容提供方,若內(nèi)容存在侵權(quán),請進(jìn)行舉報(bào)或認(rèn)領(lǐng)
文檔簡介
CognitiveBiases
inthe
RiskMatrixWilliamSiefert,M.S.
EricD.SmithBoeingSystemsEngineeringGraduateProgramMissouriUniversityofScienceandTechnology?2019SmithWilliamSiefert,M.S.“Fearofharmoughttobeproportionalnotmerelytothegravityoftheharm,butalsototheprobabilityoftheevent.”Logic,ortheArtofThinkingAntoineArnould,1662ConsequencexLikelihood=RiskRiskgraphingHyperboliccurvesHyperboliccurvesinlog-loggraphiso-risklines
5x5Risk“Cube”O(jiān)riginalCurrentObjectivevs.Subjectivedata"Campfireconversation"piecePresentSituationRiskmatricesarerecognizedbyindustryasthebestwayto:consistentlyquantifyrisks,aspartofarepeatableandquantifiableriskmanagementprocessRiskmatricesinvolvehuman:NumericaljudgmentCalibration–location,gradationRounding,CensoringDataupdatingoftenapproachedwithunderconfidenceoftendistrustedbydecisionmakersGoalRiskManagementimprovementandbetteruseoftheriskmatrixConfidenceincorrectassessmentofprobabilityandvalueAvoidanceofspecificmistakesRecommendedactionsHeuristicsandBiasesDanielKahnemanwontheNobelPrizeinEconomicsin2019"forhavingintegratedinsightsfrompsychologicalresearchintoeconomicscience,especiallyconcerninghumanjudgmentanddecision-makingunderuncertainty.“Similaritiesbetweencognitivebiasexperimentsandtheriskmatrixaxesshowthatriskmatricesaresusceptibletohumanbiases.AnchoringFirstimpressiondominatesallfurtherthought1-100wheeloffortunespunNumberofAfricannationsintheUnitedNations?Smallnumber,like12,thesubjectsunderestimatedLargenumber,like92,thesubjectsoverestimatedObviatingexpertopinionTheanalystholdsacircularbeliefthatexpertopinionorreviewisnotnecessarybecausenoevidencefortheneedofexpertopinionispresent.HeuristicsandBiasesPresenceofcognitivebiases–eveninextensiveandvettedanalyses–canneverberuledout.Innatehumanbiases,andexteriorcircumstances,suchastheframingorcontextofaquestion,cancompromiseestimates,judgmentsanddecisions.Itisimportanttonotethatsubjectsoftenmaintainastrongsensethattheyareactingrationallywhileexhibitingbiases.TerminologySubjectiveParametersLikelihood(L)Consequence(C)SubjectiveProbability,π(p)Utility(negative),U-(v)Shownon:Ordinate,YaxisAbscissa,XaxisObjectiveParametersObjectiveProbability,pObjectiveValue,v5x5Risk“Cube”O(jiān)riginalObjectivevs.Subjectivedata"Campfireconversation"pieceLikelihoodFrequency
ofoccurrenceisobjective,discreteProbabilityiscontinuous,fiction"Humansjudgeprobabilitiespoorly"[CosmidesandTooby,2019]Likelihoodisasubjectivejudgment (unlessmathematical)'Exposure'byprojectmanagertimelessConsequence,CObjectiveConsequencedeterminationiscostlyRangeofconsequenceTotallife-cyclecostMil-Std882d$damageHumanimpactEnvironmentLawCatastrophic>$1MDeath,DisabilityirreversibledamageViolateCritical:$1M-$200KHospitalizationto>=3personnelReversibledamageViolateMarginal:$200K-$10KLossofworkdays;injuryMitigationdamageNegligible:$10K-$2KNolostworkday;injuryMinimaldamageCaseStudyIndustryriskmatrixdata1412originalandcurrentriskpoints(665)TimeoffirstentryknownTimeoflastupdateknownCost,ScheduleandTechnicalknownSubjectmatternotknownBiasesrevealedLikelihoodandconsequencejudgmentMagnitudevs.Reliability[GriffinandTversky,1992]MagnitudeperceivedmorevalidDatawithoutstandingmagnitudesbutpoorreliabilityarelikelytobechosenandusedSuggestion:DatawithuniformsourcereliabilitySpeciousnessofdataObservation:riskmatricesaremagnitudedriven,withoutregardtoreliabilityExpectedDistributionfororiginalriskpointsinRiskMatrix?BivariateNormalUniform:1.EstimationinaPre-DefineScaleBias
Responsescaleeffectsjudgment[Schwarz,1990]Twoquestions,random50%ofsubjects:Pleaseestimatetheaveragenumberofhoursyouwatchtelevisionperweek:__________X_____________1-45-89-1213-1617-20MorePleaseestimatetheaveragenumberofhoursyouwatchtelevisionperweek:__________X_____________1-2 3-45-67-89-10MoreLikelihoodMarginalDistributionofOriginalPoints123455827275428840Normaldistributionwithμ=3.0,σ=0.783833067633038?=actual–normal20-5878-422(Χ2=22,LogisticΧ2>~10rejectH0H0=NormalEffectofEstimationinaPre-DefinedScale
‘Peopleestimateprobabilitiespoorly’[CosmidesandTooby,2019]Consequence/SeverityamplifiersEffectofEstimationinaPre-DefinedScale
‘Peopleestimateprobabilitiespoorly’[CosmidesandTooby,2019]Consequence/SeverityamplifiersSeverityAmplifiersLackofcontrolLackofchoiceLackoftrustLackofwarningLackofunderstandingManmadeNewnessDreadfulnessPersonalizationRecallabilityImminency5x5RiskMatrixSituationassessment5x5RiskMatricesseektoincreaseriskestimationconsistencyHypothesis:CognitiveBiasinformationcanhelpimprovethevalidityandsensitivityofriskmatrixanalysisProspectTheoryDecision-makingdescribedwithsubjectiveassessmentof:ProbabilitiesValuesandcombinationsingamblesProspectTheorybreakssubjectivedecisionmakinginto:preliminary‘screening’stage,probabilitiesandvaluesaresubjectivelyassessedsecondary‘evaluation’stagecombinesthesubjectiveprobabilitiesandutilitiesHumansjudgeprobabilitiespoorly*SubjectiveProbability,π(p)
smallprobabilitiesoverestimatedlargeprobabilitiesunderestimatedπ(p)=
(pδ)/[pδ+(1-p)δ](1/δ) p=objectiveprob. 0<δ≤1Whenδ=1,π(p)=p=objectiveprobabilityusualvalueforδ:
δ=0.69forlosses
δ=0.61forgainsGainsandlossesarenotequal*SubjectiveUtility
Valuesconsideredfromreferencepointestablishedbythesubject’swealthandperspectiveFramingGainsandlossesare subjectivelyvalued1-to-2ratio.Forgains:U+(v)=Ln(1+v)Forlosses:U-(v)=-(μ)Ln(1–cv) μ=2.5
c=constant
v=objectivevalue
ImplicationofProspectTheoryfortheRiskMatrixANALYSESANDOBSERVATIONS
OFINITIALDATA
Impedimentsfortheappearanceofcognitivebiasesintheindustrydata:IndustrydataaregranularwhilethepredictionsofProspectTheoryareforcontinuousdataQualitativedescriptionsof5rangesoflikelihoodandconsequencenon-linearinfluenceintheplacementofriskdatumpoints Nevertheless,theevidenceofcognitivebiasesemergesfromthedata2.DiagonalBiasAnticipationoflatermovingofriskpointstowardtheoriginRiskpointswithdrawnfromtheoriginupwardandrightwardalongthediagonalRegressionon1412OriginalPointsInterceptSlopeR2.20.220.223.ProbabilityCenteringBias
LikelihoodsarepushedtowardL=3SymmetrictoafirstorderImplicationofProspectTheoryfortheRiskMatrix3a.AsymmetricalProbabilityBiasSubjectiveprobabilitytransformationπ(p)predictsthatlikelihooddatawillbepushedtowardL=3LargeprobabilitiestranslateddownmorethansmallprobabilitiesaretranslatedupReducedamountoflargesubjectiveprobabilities,comparatively1234558272754288404.ConsequenceBias
ConsequenceispushedhigherEngineeridentifieswithincreasedrisktoentirecorporation'Personal'corporateriskStatisticalEvidenceforConsequenceBias
MaxatC=4C=1significantlylessthanC=5counts C=2significantlylessthanC=4ConsequenceOriginalDataPoints1234520145538599110Normaldistributioncomparison:χ2=600,df=40.0probabilityConsequencesmoothedConsequenceincreased,→,byAmplifiersH0=NormalConsequencetranslationLikelihoodmitigationrecommendationsEngineersandManagementTechnicalriskhighestpriorityScheduleriskcommunicatedwellbymanagementCostrisklikelihoodlessfrequentlycommunicatedbymanagement.
HighercognizanceofcostriskwillbevaluableattheengineeringlevelLikelihoodmitigation1.Technical2.Schedule3.CostConsequenceMitigationEngineers:ScheduleconsequenceseffectcareersTechnicalconsequenceseffectjobperformancereviewsCostconsequencesareremoteandassociatedwithmanagementHighercognizanceofcostriskwillbevaluableattheengineeringlevelConsequencemitigation1.Schedule2.Technical3.CostCONCLUSIONFirsttimethattheeffectsofcognitivebiaseshavebeendocumentedwithintheriskmatrixClearevidencethatprobabilityandvaluetranslations,aslikelihoodandconsequencejudgments,arepresentinindustryriskmatrixdataSteps1)thetranslationswerepredictedbyprospecttheory,2)historicaldataconfirmedpredictionsRiskmatricesarenotobjectivenumbergridsSubjective,albeituseful,meanstoverifythatriskitemshavereceivedrisk-mitigatingattention.DataCollectionImprovementsContinuumofdatafromRiskmanagementto(Issuemanagement)OpportunitymanagementDifferentdatabasesyearsofdataineachTimeWaterfallRiskchartsSuggestionsforriskmanagementimprovementObjectivebasisofrisk:FrequencydataforProbability$forConsequenceLong-term,institutionalrationalityTeamapproachIterationsPublicreviewExpertreviewBiasesanderrorsawarenessFutureworkConfirmationofthepresenceofprobabilitybiases,andvaluebiasesinriskdatafromotherindustrie
溫馨提示
- 1. 本站所有資源如無特殊說明,都需要本地電腦安裝OFFICE2007和PDF閱讀器。圖紙軟件為CAD,CAXA,PROE,UG,SolidWorks等.壓縮文件請下載最新的WinRAR軟件解壓。
- 2. 本站的文檔不包含任何第三方提供的附件圖紙等,如果需要附件,請聯(lián)系上傳者。文件的所有權(quán)益歸上傳用戶所有。
- 3. 本站RAR壓縮包中若帶圖紙,網(wǎng)頁內(nèi)容里面會有圖紙預(yù)覽,若沒有圖紙預(yù)覽就沒有圖紙。
- 4. 未經(jīng)權(quán)益所有人同意不得將文件中的內(nèi)容挪作商業(yè)或盈利用途。
- 5. 人人文庫網(wǎng)僅提供信息存儲空間,僅對用戶上傳內(nèi)容的表現(xiàn)方式做保護(hù)處理,對用戶上傳分享的文檔內(nèi)容本身不做任何修改或編輯,并不能對任何下載內(nèi)容負(fù)責(zé)。
- 6. 下載文件中如有侵權(quán)或不適當(dāng)內(nèi)容,請與我們聯(lián)系,我們立即糾正。
- 7. 本站不保證下載資源的準(zhǔn)確性、安全性和完整性, 同時(shí)也不承擔(dān)用戶因使用這些下載資源對自己和他人造成任何形式的傷害或損失。
最新文檔
- 2025年公共設(shè)施窗簾清洗消毒服務(wù)合同范本3篇
- 2024版汽車檢測臺租賃合同
- 2024石材外墻干掛勞務(wù)服務(wù)合同標(biāo)準(zhǔn)版6篇
- 2025年度特色飲品店門面房租賃及新品研發(fā)合同3篇
- 2025年度圓形冷卻塔能源管理服務(wù)合同4篇
- 2024版基礎(chǔ)建設(shè)融資借款協(xié)議模板版
- 2025年度水電工程質(zhì)保期服務(wù)合同4篇
- 2025年度學(xué)校圖書館窗簾升級改造合同4篇
- 2025年度生態(tài)修復(fù)工程承包樹木合同協(xié)議書4篇
- 2024石材行業(yè)品牌推廣與營銷合同3篇
- 領(lǐng)導(dǎo)溝通的藝術(shù)
- 發(fā)生用藥錯(cuò)誤應(yīng)急預(yù)案
- 南潯至臨安公路(南潯至練市段)公路工程環(huán)境影響報(bào)告
- 綠色貸款培訓(xùn)課件
- 大學(xué)生預(yù)征對象登記表(樣表)
- 主管部門審核意見三篇
- 初中數(shù)學(xué)校本教材(完整版)
- 父母教育方式對幼兒社會性發(fā)展影響的研究
- 新課標(biāo)人教版數(shù)學(xué)三年級上冊第八單元《分?jǐn)?shù)的初步認(rèn)識》教材解讀
- (人教版2019)數(shù)學(xué)必修第一冊 第三章 函數(shù)的概念與性質(zhì) 復(fù)習(xí)課件
- 重慶市銅梁區(qū)2024屆數(shù)學(xué)八上期末檢測試題含解析
評論
0/150
提交評論