frequencynod32

TechnicalarticleWileyHandbookofScienceandTechnologyforHomelandSecurityHumanBehaviorandDeceptionDetectionMarkG.
Frank,UniversityatBuffalo,StateUniversityofNewYorkMaureenO'Sullivan,UniversityofSanFranciscoMelissaA.
Menasco,UniversityatBuffalo,StateUniversityofNewYorkKeywordsBehavior;Cognition:Deception;Detection;Emotion;Judgment;Lying;Malfeasance;Memory.
AbstractHumanintelligenceisthekeytostoppingterrorism,andthereforeitisessentialtoknowwhentheinformationobtainedisfalse.
Thischapterbrieflyoutlinestheresearchonbehavioralcluestodeception,aswellasresearchonpeople'sabilitiestospotdeceptiononceithashappened.
Wefindthatthereisnoclueorcluepatternthatisspecifictodeception,althoughtherearecluesspecifictoemotionandcognition.
Ingeneral,behavioralcluesareonlylimitedintheirabilitiestoidentifydeceptionandthattherearestillbehavioralmeasurementissuesthatmayplagueresearchondeception.
Moreover,acloserexaminationofthelaboratoryresearchsuggestsmanyresearchstudiesarenotrelevanttosecuritycontexts,thustheresearchliteraturemayunderestimatetheusefulnessofbehavioralinformation–particularlyfortheutilityofidentifyingemotionalandcognitivestates.
Wealsofindthatmostpeople,unaidedbytechnology,cannotdetectliesfrombehaviorinformation,butthatsomegroupsdoshowsignificantlyhigherlevelsofaccuracy–althoughmoreresearchisneededtounderstandwhy.
Weconcludethatmoredirectedinteractionwithscientistsandpractitioners–inbothlabworkandintherealworld,increatingrealworlddatabases,inidentifyingbaseratesformalfeasantbehaviorinsecuritysettings,inoptimizingtrainingandinidentifyingpreexistingexcellencewithinsecurityorganizationscanmorerapidlycapitalizeontheusefulnessofbehavioralinformationinsecuritysettings.
Terrorismatitscoreisahumanendeavor.
Humanbeingscultivatewhattheyhate,planandthenexecuteterroristattacks.
Thus,anyinformationthatcanaidtheintelligenceorsecurityofficertoweightheveracityoftheinformationheorsheobtainsfromsuspectedterroristsorthoseharboringthemwouldhelppreventattacks.
Thiswouldthennotonlyaddanotherlayertoforceprotectionbutwouldfacilitatefutureintelligencegathering.
Yetthefacetofacegatheringofinformationthroughsuspectedterrorists,informants,orwitnessesisrepletewithobstaclesthataffectitsaccuracysuchasthewelldocumentedshortcomingsofhumanmemory,honestdifferencesofopinion,aswellaswhatisthefocusofthischapter-outrightdeception[1].
Theevidencesuggeststhatindaytodaylifemostliesarebetrayedbyfactorsorcircumstancessurroundingthelie,andnotbybehavior[2].
However,thereareArticleID:SY07Page1TechnicalarticleWileyHandbookofScienceandTechnologyforHomelandSecuritytimeswhendemeanorisallaHomelandsecurityagenthasathisorherdisposaltodetectsomeonewhoislyingabouthisorhercurrentactionsorfutureintent.
Becausealieinvolvesadeliberate,consciousbehavior,wecanspeculatethatthiseffortmayleavesometrace,sign,orsignalthatmaybetraythatlie.
Whatintereststhescientist,aswellassocietyatlarge,isa)Aretherecluesperceptibletotheunaidedeyethatcanreliablydiscriminatebetweenliarsandtruthtellers;b)Dothesecluesconsistentlypredictdeceptionacrosstime,typesoflies,differentsituations,andcultures;andifa)andb)aretrue,thenc)Howwellcanourcounter-terrorismprofessionalscanmakethesejudgments,andcantheydothisinrealtime,withorwithouttechnologicalassistanceScientificoverview–behavioralsignsofdeception.
Todatenoresearcherhasdocumenteda"Pinocchioresponse;"thatis,abehaviororpatternofbehaviorsthatinallpeople,acrossallsituations,isspecifictodeception(e.
g.
,[3]).
Allthebehaviorsidentifiedandexaminedbyresearcherstodatecanoccurforreasonsunrelatedtodeception.
Generallyspeaking,theresearchondetectingliesfrombehaviorsuggeststhattwobroadfamiliesofbehavioralcluesarelikelytooccurwhensomeoneislying-cluesrelatedtoliars'memoryandthinkingaboutwhattheyaresaying(cognitiveclues),andcluesrelatedtoliar'sfeelingsandfeelingsaboutdeception(emotionalclues)[3,4,5,6,7,8].
Cognitiveclues.
Alieconceals,fabricates,ordistortsinformation;thisinvolvesadditionalmentaleffort.
Theliarmustthinkharderthanatruthtellertocoverup,createeventsthathavenothappened,ortodescribeeventsinawaytoallowmultipleinterpretations.
Additionalmentaleffortisnotsolelythedomainoftheoutrightliar,however;apersonwhomusttellanuncomfortabletruthtoanotherwillalsoengageinadditionalmentalefforttocomeupwiththeproperphrasingwhilesimultaneouslyreducingthepotentialnegativeemotionalreactionoftheother.
Thisextraefforttendstomanifestitselfwithlongerspeechlatencies,increasedspeechdisturbances,lessplausiblecontent,lessverbalandvocalinvolvement,lesstalkingtime,morerepeatedwordsandphrases,andsoforth[9].
Researchhasalsoshownthatsomenonverbalbehaviorschangeasaresultofthismentaleffort.
Forexample,illustrators–handorheadmovementsthataccompanyspeech,andareconsideredbymanytobeapartofspeech(e.
g.
,[10])–willdecreasewhenlyingcomparedtotellingthetruth[11,12].
Anotherwayinwhichcognitionisinvolvedintellingalieisthroughidentificationofnaturalisticmemorycharacteristics.
Thismeansthatexperiencedeventshavememoryqualitiesthatareapparentupondescriptionthataredifferentfromeventsthathavenotbeenexperienced(the"Undeutschhypothesis"[13]).
Eventsthatwerenotactuallyexperiencedfeaturemoreambivalence,havefewerdetails,apoorerlogicalstructure,lessplausiblity,morenegativestatementsandArticleID:SY07Page2TechnicalarticleWileyHandbookofScienceandTechnologyforHomelandSecurityarelessembeddedincontext.
Liarsarealsolesslikelytoadmitlackofmemory,andhavelessspontaneouscorrections(reviewedby[8,9]),andmayusemorenegativeemotionwords,andfewerselfandotherreferences[14].
Mentaleffortcluesseemtooccurmoreinthedeliveryofthelie,whereasmemoryrecallcluestendtooccurinthecontentofthelie.
Wenotethatnotalllieswilltaxmentaleffort;forexample,itismuchlessmentallytaxingtoansweracloseendedquestionlike"Didyoupackyourownbags"withayesornothantoansweranopenended"Whatdoyouintendtodoonyourtrip"Moreover,acleverliarcanappearmorepersuasiveifheorshesubstitutesanactualexperiencedeventastheiralibiratherthancreatinganentirelynewevent.
Thismaybewhyarecentgeneralreviewpaper[9]foundconsistentnon-homogenouseffectsizesforthesementaleffortandmemorybasedcuesacrossthestudiestheyreviewed,astheparticularparadigmsusedbyresearchersvariedgreatlyintheextenttowhichtheliesthatwerestudiedmentallytaxedtheliars.
Emotionalclues.
Liescanalsogenerateemotions,rangingfromtheexcitementandpleasureof"pullingthewooloversomeone'seyes"tofearofgettingcaughttofeelingsofguilt[4].
Darwin[15]firstsuggestedthatemotionstendtomanifestthemselvesinthefacialexpressions,aswellasinthevoicetones,andthatthesecanbereliableenoughtoaccuratelyidentifyemotionalstates.
Researchhassinceshownthatforsomeexpressions–e.
g.
,anger,contempt,disgust,fear,happy,sadness/distress,orsurprise–culturesthroughouttheplanetrecognizeandexpresstheseemotionsinboththefaceandvoicesimilarly[16].
Totheextentthataliefeatureshigherstakesforgettingcaught,wewouldexpecttoseemoreofthesesignsofemotioninliarscomparedtotruthtellers.
Ifthelieisapoliteliethatpeopletelloftenandeffortlessly,therewouldbelessemotioninvolved(e.
g.
,[17]).
Meta-analyticstudiessuggestthatliarsdoappearmorenervousthantruthtellers,withlessfacialpleasantness,highervocaltension,highervocalpitch,greaterpupildilationandfidgeting[9].
Ifthelieitselfisaboutemotions–e.
g.
,tellingsomeonethatonefeelscalm,wheninfactoneisnervous–theresearchshowsthatsignsofthetrulyfeltemotionappearinthefaceandvoicedespiteattemptstoconceal,althoughthesesignsareoftensubtleandbrief[18,19].
Measurementissues.
Oneissueinmeasuringliesignsistomakeclearwhatismeantbythetermscognitionandemotion.
Forexample,indeceptionresearchthetermarousalisusedinterchangeablywithemotion,butoftenreferstomanydifferentphenomena:anorientingresponse(e.
g.
,[20]),anexpressionoffear(e.
g.
,[21]),amoreindeterminateaffectsomewherebetweenarousalandemotion(e.
g.
,[22];seealsodiscussionbyWaid&Orne[23]),aswellasphysiologicalstatesasdifferentasstress,anxiety,embarrassmentandevenanger[24].
ArticleID:SY07Page3TechnicalarticleWileyHandbookofScienceandTechnologyforHomelandSecurityAsecondissueinmeasuringliesignsistoclarifythelevelofdetailofmeasurementaswellastospecifywhythatlevelofdetailmayormaynotcorrelatewithlying[25].
Manymeta-analysesofbehavioraldeceptioncluesreportinsignificanteffectsizes,butthevarianceamongeffectisnothomogenous(e.
g.
,[3;9,26;27;28]).
Forexample,somestudiesinvestigatedbehavioratthemostelementalphysicalunitsofmeasurementsuchascountingthemovementsinthehands,feet,arms,legs,torso,eyemovements,eyeblinks,pupildilation,lippressing,browloweringorraising,lipcornerpuller(smiling),fundamentalfrequency,amplitude,pauses,filledpauses,responselatency,speechrate,lengthofresponse,connectorwords,uniquewords,self-references,andsoforth.
Otherstudiesinvestigatedbehavioratthemostelementalpsychologicalmeaningunitsofmeasurement.
Someoftheseincludedmanipulators-whichinvolvetouching,rubbing,etc.
,ofvariousbodyparts–whichcouldbecomposedofanumberofhand,finger,armmovements,butwhichwerescoredfortheoreticalratherthanmerelydescriptivereasons.
Otherpsychologicallymeaningfulunitsofmeasurementincludeillustrators,whichaccompanyspeechtohelpkeeptherhythmofthespeech,emphasizeaword,showdirectionofthought,etc.
,oremblems,whicharegesturesthathaveaspeechequivalent,suchasaheadnodmeaning"yes",orashrugmeaning"I'mnotsure",orfacialemblemssuchaswinking.
Thepsychologicalmeaningunitsmightalsoincludevocaltension,speechdisturbances,negativestatements,contextualembedding,unusualdetails,logicalstructure,unexpectedcomplications,superfluousdetails,selfdoubt,andsoforth.
Finally,otherstudiesinvestigatedbehavioratthemostinterpretative/impressionisticunitlevel,whicharefurtherunarticulatedcompositesofthephysicalandthepsychologicalmeaningunitsdescribedearlier.
Someofthesebehaviorsimpressionisticvariablesincludefidgeting,involvement,bodyanimation,posture,facialpleasantness,expressiveness,vocalimmediacyandinvolvement;andspokenuncertainty;plausibility,cognitivecomplexity(again,seereviewby[9]).
Theproblemofcourseisthatasonemovesfromphysicaltoimpressionisticmeasures,itwouldseembecomehardertomakethosejudgmentsreliably.
Thisisnotalwaysthecasethough;e.
g.
,theterm"smile"hasrarelybeendefinedinresearchreports,yetindependentcodersaretypicallyabove.
90reliabilitywhencodingsmiles(see[29]forareview).
Althoughresearchsuggeststhatpeoplecanbemoreaccuratewhentheyemployindirectinferencestodeception(e.
g.
,Doesthepersonhavetothinkhard[30]),'gut'impressionstendtobeuncorrelatedwithaccuracy[26].
Thissuggeststhatwemustbecautiousaboutcluesattheimpressionisticlevel,andthatitmaybemoreproductivetostudythemattheirpsychologicallevelwheretheymightbemoremeaningfultounderstandingdeception.
Prognosisongeneralizabilityofdeceptionfindingsacrosstime,lies,situations,andcultures.
Itissafetoconcludethatalthoughtherearesomecluesthatbetrayalieatratesgreaterthanchance,noneofthemareexclusivetodeception.
ThisconclusionArticleID:SY07Page4TechnicalarticleWileyHandbookofScienceandTechnologyforHomelandSecurityappliestomachinebasedphysiologicalapproachesaswell.
However,theoriginsofthesesigns–mentaleffort,memory,andemotion-areuniversal.
Thissuggeststhatifthecontextinwhichtheinformationisgatherediscontrolled,anddesignedtodifferentiallyaffectliarsandtruthtellers,itwouldincreasegreatlythechancesofbeingabletodistinguishpeoplewithdeceptiveintentfromthosewithtruthfulintent.
Polygraphexaminationhasdonethisbycontrollingtheirquestionstyletoimprovehitrates,buttodatethishasnotbeendonesystematicallyinbehavioralstudies.
Thusitseffectsareunknown,butwecanspeculatebaseduponwhatweknowaboutnormal,truthfulhumanbehavior.
Ifthelieisofnosignificancetotheperson,withnocostsforgettingcaught,andinvolvesasimpleyesornoanswer,oddsaretherewillnotbemanycluestodistinguishtheliarandthetruthteller.
Ifthesituationhassignificancetotheperson,thereareconsequencesforgettingcaught,andthepersonisrequiredtorecountaneventinanopenendedquestion,thenwewouldexpectmorecluestosurfacethatwoulddistinguishtheliarfromthetruthteller.
Thismaybeacurvilinearrelationship;asituationofextraordinaryhighmentaleffortandemotion-e.
g.
,oneinwhichapersonisbeingbeaten,screamedat,andthreatenedwithexecution-willgenerateallthe"lieclues"describedearlier,butequallyinliarandtruthteller.
Nonetheless,informationaboutmentaleffort,experiencedmemory,andemotioncanbeveryusefulcluestoHomelandSecuritypersonneltoidentifybehavioral"hotspots"[4]thatcanprovideinformationaboutissuesofimportancetothesubject.
Acounter-terrorismIntelligenceofficerwhoknowswhenasubjectisfeelinganemotionorthinkinghardcanknowwhattopicstopursueoravoidinaninterview,whetherthesubjectisfabricating,concealinginformation,ormerelyfeelinguncomfortablewiththetopic,althoughtruthful.
Scientificoverview–abilitiestospotliars.
Researchoverthepast30yearssuggeststhattheaveragepersonisslightlystatisticallybetterthanchanceatidentifyingdeception,butnotpracticallybetter.
Themostrecentreviewofover100studieshasshownthatwhenchanceaccuracyis50%,theaveragepersonisapproximately54%accurate[31].
Thereareanumberofreasonsforthispoorability;amongthempoorfeedbackindailylife(i.
e,apersononlyknowsabouttheliestheyhavecaught);thegeneraltendencyforpeopletobelieveothersuntilprovenotherwise(ie,a'truthbias';[32]),andespeciallyafaultyunderstandingofwhatliarsactuallylooklike(ie,thedifferencebetweenpeople'sperceivedcluestolyingcomparedtotheactualclues;[26]).
Generalabilitiesofspecializedgroups.
Mostofthestudiesreviewedwerelaboratorybasedandinvolvedobserversjudgingstrangers.
Butsimilarresultsarefoundevenwhentheliarsandtruthtellersareknowntotheobservers(alsoreviewedby[31].
Iftheliesbeingtoldarelowstakes,sothatlittleemotionisarousedandtheliecanbetoldwithoutmuchextracognitiveeffort,theremaybefewcluesavailableonwhichtobaseaArticleID:SY07Page5TechnicalarticleWileyHandbookofScienceandTechnologyforHomelandSecurityjudgment.
Butevenstudiesofhighstakeslies,inwhichbothliarsandtruthtellersarehighlymotivatedtobesuccessful,suggestanaccuracylevelthatisnotmuchdifferentthanchance.
Researchthatexaminedunselectedprofessionalsinvolvedinsecuritysettings–police,federalagents,andsoforth–havetypicallyfoundthattheytooarenotanymoreaccurateintheirabilitiestospotdeceptionthanlaypeople(e.
g.
,[27,33,34,35,36]).
However,withinthesestudiestherehavebeenahandfulofgroupsthathaveperformedbetterthan60%accurateonbothliesandtruths,andwhatthesegroupsaredoingmightbeinformativeforHomelandSecurityapplications.
ThefirstgroupidentifiedwasagroupofSecretServiceagentswhonotonlyweresuperior,asagroup,indetectingliesaboutone'semotions,butthosewhoweremoreaccurateweremorelikelytoreportusingnonverbalcluesthanthosewhowerelessaccurate.
Theauthors[33]speculatedthattheSecretServiceagentsweremoreaccuratethantheothergroupsbecausetheyweretrainedinscanningcrowdsfornonverbalbehaviorsthatdidnotfit,andtheyalsodealtwithassassinationthreats,manyofwhichweremadebymentallyillindividuals.
Unlikemostpoliceofficerswhoseassumptionofguiltinsuspectsishigh[37],reflectingtheexperienceoftheirdailywork,SecretServiceagentsinterviewedsuspectswheretheyknewthebaserateoftruedeaththreatswaslow.
Thesecondsetofgroupsidentifiedincludedforensicpsychologists,federaljudges,selectedfederallawenforcementofficersandagroupofsheriffs[34].
Acommonalityamongthesegroupsseemedtobetheirveryhighmotivationtoimprovetheirliedetectingskills.
Athirdsetofgroupsidentifiedwerepoliceofficersexaminingreallifelies,whoshowed65%overallaccuracyindetectingliesandtruths[38].
Individualdifferences.
Aswithanyability,researchsuggeststhatsomepeoplearebetterabletodetectdeceptionthatothersinhighstakelies(e.
g.
,[39]);thisskilldoesnotseemtotranslatetolowerstakelies[32].
Oneelementofbetterskillinhigherstakesettingsistheabilitytojudgemicro-momentarydisplaysofemotion[33,39].
Othergroupswhoshowedbetterthan60%accuracyincludedpeoplewithlefthemispherebrainlesionsthatpreventedthemfromcomprehendingspeech[40],andthosesubjectswhoscoredhigheronatestofknowledgeofcluestodeceitwerealsomoreaccuratethanthosewhodidnot[41].
Adifferentapproachhasbeentoidentifyindividualswhoobtainhighscoresonliedetectiontestsandstudyingthemindetail[42].
Aftertestingmorethan12,000peopleusingasequentialtestingprotocolinvolvingthreedifferentliedetectionaccuracymeasures,O'SullivanandEkmanidentified29highlyaccurateindividuals.
Theseindividualshadakindofgeniuswithrespecttotheobservationofverbalandnonverbalclues,butsincegeniusoftenconnotesacademicintelligence,theexpertliedetectorswerelabeled"truthwizards"tosuggesttheirspecialtalent.
Althoughthistermisunfortunateinmistakenlysuggestingthattheirabilitiesareduetomagicratherthantalentandpractice,thetermdoesreflecttherarityoftheirabilities.
OneofthefirstfindingsoftheWizardProjectwasaprofessionArticleID:SY07Page6TechnicalarticleWileyHandbookofScienceandTechnologyforHomelandSecurityspecificsensitivitytocertainkindsoflies.
About1/3ofthewizardswerehighlyaccurateonallthreeofthetestsused.
Anotherthirddidverywellontwoofthetests,butnotonthethird,inwhichpeopleliedortoldthetruthaboutwhethertheyhadstolenmoney.
Nearlyallofthesewizardsweretherapistswhohadlittle,ifany,experiencewithliesaboutcrime.
Ontheotherhand,theremainingthirdofthewizardswerelawenforcementpersonnel–policeandlawyers–whodidverywellonthecrimeliedetectiontest,butnotonatestinwhichpeopleliedortoldthetruthabouttheirfeelings.
Comparedwithamatchedcontrolgroup,expertliedetectorsaremorelikelythancontrolstoattendtoawidearrayofnonverbalbehaviorsandtobemoreconsciouslyawareofinconsistenciesbetweenverbalandnon-behaviors.
Althoughexpertliedetectorsmakealmostinstantaneousjudgmentsaboutthekindofpersontheyareobserving,theyarealsomorecautiousthancontrolsaboutreachingafinaldecisionabouttruthfulness.
Criticalneedsanalysis.
ResearchonhumanbehavioranddeceptiondetectioncanmakeausefulcontributiontoHomelandSecurityneedsaslongasscientistsandpractitionersunderstandwhatitistheyareobserving–signsofthinking,signsoffeeling.
Thisruleappliestoautomatedapproachesthatmeasurephysiologyaswell.
Evenwiththislimitation,traininginbehavioralhotspotrecognitionmaymakesecuritypersonnelbetteratspottingthosewithmalfeasantintent.
Othercriticalneedsinclude:Morerelevantlaboratoryparadigmsandsubjects.
Wemustrecognizethatgeneralmeta-analysesoftheresearchliterature,althoughuseful,arelimitedintheirapplicabilitytosecuritycontexts,sincesuchanalysestendtocombinestudiesthatfeatureliestoldwithfewstakesandcognitivedemandswiththosewithhigherstakesandstrongercognitivedemands.
Thus,weshouldbemoreselectiveaboutwhichstudiestoexamineforcluesthatmaybeusefulorrelevanttosecuritycontexts.
Thisalsomeansitisimportantforscientiststodevelopresearchparadigmsthatmorecloselymirrorthereallifecontextsinwhichsecuritypersonnelwork.
Althoughlaboratorysettingsarenotaspowerfulasrealworldsettings,highstakelaboratorydeceptionsituationscanprovideinsightswiththebestchanceofapplicability.
Consistentwiththisapproach,twocurrentairportsecuritytechniquescapitalizeonbehaviorsidentifiedbyresearchstudiesonstress,withanecdotalsuccess(ie,TSA'sScreeningPassengersbyObservationTechniquesandtheMAStatePoliceBehavioralAssessmentSystem).
OnewaytofacilitatethistypeofprogressistohaveHomelandSecuritypersonneladviselaboratoryresearch,aswellasallowresearcherstospendon-the-jobtimewithSecuritypersonnel.
Webelievepairingtheresearchersandpractitionerswouldeventuallyresultincallsforlaboratorystudiesfeaturinghigherstakestotheliars,differentsubjectpopulationsbeyondUS/Europeans(asresearchsuggeststhatpeoplecandetectArticleID:SY07Page7TechnicalarticleWileyHandbookofScienceandTechnologyforHomelandSecuritydeceptioninotherculturesatratesgreaterthanchance;[43,44]),anddifferinginterviewlengthssuchasexaminingshorterinterviews(i.
e,a30-90secondsecurityscreening)andlongerinterviews(i.
e.
,a1-4hourintelligenceinterview).
Examinationandcreationofrealworlddatabases.
Therehavebeenveryfewstudiesofrealworlddeception(e.
g.
,[38]),yetthetechnologicalcapabilityexiststocreatemanymore.
Thebiggestproblemwithrealworlddataisdeterminingthegroundtruth(wasthepersonreallylying,ordidtheytrulybelievewhattheyjuststated).
Estimatinggroundtruth–ascomparedtoknowinggroundtruth–willslowdowntheidentificationofanypatternsorsystems.
Clearcriteriamustbeestablishedaprioritodeterminethisgroundtruth.
Forexample,confessionsofmalfeasanceareagoodcriterion,butfalseconfessionsdohappen.
Catchingsomeonewithcontraband(i.
e.
,a'hit')isalsoagoodcriterion,butoccasionallythepersonmaybetruthfulwhenheorshestatesthatsomeonemusthavesnuckitintotheirluggage.
Moreover,academicsshouldadviseonthecaptureandrecordingofthesedatabases,toinsurethatthematerialsareabletobeexaminedbythewidestnumberofresearchersandresearchapproaches.
Forexample,mostofthepoliceinterviewvideowehaveseenisofsuchpoorqualitythatwecannotanalyzefacialexpressionsinanydetail.
Itisonlywhenthesedatabasesarecombinedwiththelaboratoryworkthatwecanmoresharplyidentifybehaviorsorbehavioralpatternsthatwillincreasethechancesofcatchingthosewithmalfeasantintent.
Tooptimallyusethisinformationthough,wemustalsoexamineindetailknowncasesoffalsenegativesandfalsepositivesaswellascorrecthitstodeterminewhymistakesweremadeinthesejudgments.
Groundtruthbaserates.
Securitypersonneldonotknowisthebaseratesformalfeasanceintheirsettings.
Althoughitmaybelogisticallyimpossibletohandsearcheverypieceofhandluggageinabusyairport,orfolloweveryinvestigativelead,itwouldbeessentialtoknowthisbaserateinordertoascertaintheeffectivenessofanynewbehavioralobservationaltechnique.
Thiswouldalsopermitmoreusefulcost-benefitanalysesofvariouslevelsofsecurityandtraining.
Alesssatisfyingbutstillusefulwaytoascertaineffectivenessistocomparehitratesforcontrabandforthoseusingvariousbehavioralobservationtechniqueswiththosewhoarestoppedrandomly(aslongasthedayoftheweekandtimeoftheday/yeararescientificallycontrolled).
OptimizingTraining.
Themostrecentmeta-analysisoftheresearchliteratureontrainingpeopletoimprovedeceptiondetectionfrombehaviorhasshownthatacrossover2,000subjects,therewasamodesteffectfortraining,despitetheuseofsubstandardtrainingtechniques[45].
Thisobviouslysuggeststhatbettertrainingtechniqueswillyieldlargerimprovementsinpeople'sabilitiestosortouttruthfromlie.
OneArticleID:SY07Page8TechnicalarticleWileyHandbookofScienceandTechnologyforHomelandSecuritytrainingchangewouldbetotrainonbehavioralcluesthatarederivedfromsimilarsituationsandsupportedbyresearch.
Forexample,onestudytrainedresearchsubjectstorecognizeasetofbehavioralcluesthatarebelievedtobeindicativeofdeception,andareoftentaughttolawenforcementpersonnelassignsofdeception,althoughmanyofthesesignsarenotsupportedbythescientificliterature[46].
Thisstudyreporteda10%decreaseinaccuracyforthegroupsreceivingsuchtraining.
Therefore,thefirststepinadequatetrainingistoidentifywhatinformationisusefulfortraining(seeabove).
Thesecondstepistodeterminethemosteffectivewaytodeliverthatinformation.
Forexample,whatisthetrainingdurationthatmaximizescomprehension–onefullday,3fulldays,ormoreShoulditbedoneinagrouporselfstudyDoesitneedsimplerepetition,ormorecreativeapproaches,andhowmanytrainingitemsareneededDoesitneedtobereinforcedatparticularintervalsHowmanycluesshouldbetaught–i.
e.
,atwhatpointdoyouoverwhelmtraineesHowdoyoutraininsuchaswayastoimproveaccuracywithoutoverinflatingconfidenceThesearejustafewofthequestionswithunknownanswers.
IdentifyingExcellence.
Anothercriticalneedistoidentifywhowithinrelevantorganizationsshowssignsofexcellence,throughtheirhigherhitratesorwhateverotherclearcriteriacanbeapplied.
Thisstrategyissimilartothestrategyofthe"wizards"study[42].
Onecautionisthattodate,mosttestingmaterialwillbelaboratoryexperimentbased,andthegeneralizabilityofthatinformationtorealworldcontextsisnotperfect.
Anexaminationoftheconvergentvalidityoflaboratorytestsofdeceptiondetectionandothermorenaturalisticapproachmeasures(peerratings,fieldobservationsinairportsorotherpointsofentrywithaccuracydeterminedbytherateofcontraband"hits"byindividualscomparedtorandomselection)wouldbeagreatstart.
Futureresearchdirections.
Theaforementionedcriticalneedssuggestseveralresearchquestions,butbynomeansisthatsectioncomprehensive.
Aswepeerintothefuture,thereismuchworktodo.
Apartiallistoffuturedirectionssuggestsweshould:Examinetheroleoftechnologyinfacilitatingbehavioralobservation.
Anumberofcomputervisionalgorithmsarenowavailablethatcanaidobservation,suchasrecognizingemotionalexpressionsintheface(e.
g.
,[47]).
Whatisunknownishowrobustthesealgorithmsareinrealworldcontexts.
Whatisalsounknownishowbesttocombinetechnologicalobservationofbehaviorwithhumanjudgment.
WouldtherebeatendencyforhumanstooverrelyuponthetechnologyovertimeIdentifytheoptimalenvironmentalsetupforsurveillance,whetherwithtechnologyortheunaidedeye.
ThisincludesproxemicplacementofArticleID:SY07Page9TechnicalarticleWileyHandbookofScienceandTechnologyforHomelandSecuritytables,lines,stanchions,otherindividuals,andsoforth.
Onegoalwouldbetocreateanenvironmentthatwouldreducethetypicalstressfeltbythenormaltraveler,whichwouldhopefullyincreasethesalienceofanysignofstressexhibitedbythemalfeasanttoincreasethechancesitcanbeobserved.
Identifyoptimalinteractionstylebetweensecurityagentsandthepublic.
Onecanaggressivelyquestionandthreatentravelers,butthatmightrenderbehavioralobservationuselessduetotheoverallstressengendered.
Arapportbuildingapproach(e.
g.
,[48])mightbebetter,butthisneedsmoreresearch.
Identifytheoptimalinterviewstyle.
Phrasingofquestionsisimportantinobtaininginformation,butthishasnotbeenresearchedintheopenliterature.
Smallchangesinphrasing–e.
g.
,openversuscloseended-mightaddtotheadditionalcognitiveburdenoftheliarandthuscouldbeuseful.
Theorderofquestionswillalsobeimportant,aswellwhetheroneshouldmakeadirectaccusation.
Butonlyadditionalresearchwilltell.
Identifytheoptimalwaytocombinebehavioralclues.
Researchtendstoexamineindividualbehavioralcluestoascertaintheireffectiveness,yetmoremodernneuralnetworkandmachinelearningapproachesmaybesuccessfulinidentifyingpatternsandcombinationsofbehaviorsthatbetterpredictdeceptioninparticularcontexts.
Identifythepresenceofcountermeasures.
Aninevitablesideeffectofthereleaseofanyinformationaboutwhatbehaviorsarebeingexaminedbysecurityofficerstoidentifyriskierindividualsinsecuritysettingsisthatthisinformationwillfinditswayontotheinternetorotherpublicforum.
Thismeansapotentialterroristcanlearnwhattodoandwhatnottodoinordertoescapefurtherscrutiny.
Theproblemisthatwedon'tknowyetwhetheronecanconcealalltheirbehaviorsinthesereallifecontexts.
Moreover,someofthesebehaviors,likeemotionalbehavior,ismoreinvoluntary[16]andshouldbehardertoconcealthatmorevoluntarybehaviorlikewordchoice.
Thusitremainsanopenquestionastowhetherapotentialterroristcancountermeasureallofthecriticalbehaviors.
Spacelimitationsprecludeanexhaustivelistofneeds,futuredirections,andresearch.
Ingeneral,theresearchsuggeststhattherearelimitedcluesthatareusefultosortingoutliarsandtruthtellers,butmostpeoplecannotspotthem.
However,acloserexaminationofthisliteraturesuggeststhatsomebehavioralcluescanbeusefultosecuritypersonnel,andsomepeoplecanspottheseclueswell.
Wefeelthatitmaybeultimatelymostproductivetoexpandourthinkingaboutbehavioralcluestodeceittoincludethinkingaboutbehavioralcluestoaperson'sreality–cluesthatsomeoneisrecountingatruememory,isthinkinghard,orishavinganemotiontheywishtohide.
Thiswouldenableasecurityofficertomakethemostaccurateinferenceabouttheinnerstateofthepersontheyareobserving,which,whencombinedwithbetterinteractionandinterviewingtechniques,wouldenablethemtobetterinfertherealreasonsforthisinnerstate,beitintendingusharm,tellingalie,ortellingthetruth.
ArticleID:SY07Page10TechnicalarticleWileyHandbookofScienceandTechnologyforHomelandSecurityReferences1.
Haugaard,J.
J.
&Repucci,N.
D.
(1992).
Childrenandthetruth.
InS.
J.
Ceci,M.
DeSimone-Leichtman,andM.
E.
Putnick(Eds)Cognitiveandsocialfactorsinearlydeception.
Hillsdale,NJ:Erlbaum.
2.
Park,H.
S.
,Levine,T.
R.
,McCornack,S.
A.
,Morrison,K.
,&Ferrar,M.
(2002).
Howpeoplereallydetectlies.
CommunicationMonographs,69,144-157.
3.
Zuckerman,M.
,DePaulo,B.
M.
,&Rosenthal,R.
(1981).
Verbalandnonverbalcommunicationofdeception.
InL.
Berkowitz(Ed.
)Advancesinexperimentalsocialpsychology.
Vol14,(pp.
1-59).
SanDiego,CA:AcademicPress.
4.
Ekman,P.
(1985/2001).
Tellinglies.
NewYork:Norton.
5.
Ekman,P.
,&Frank,M.
G.
(1993).
Liesthatfail.
InM.
Lewis&C.
Saarni(Eds.
)Lyinganddeceptionineverydaylife.
(pp184-200).
NewYork:GuilfordPress.
6.
Hocking,J.
E.
,&Leathers,D.
G.
(1980).
Nonverbalindicatorsofdeception:Anewtheoreticalperspective.
Communicationmonographs,47,119-131.
7.
Knapp,M.
L.
,&Comadena,M.
E.
(1979).
Tellingitlikeitisn't:Areviewoftheoryandresearchondeceptivecommunication.
HumanCommunicationResearch,5,270-285.
8.
Yuille,J.
C.
(Ed.
)(1989).
CredibilityAssessment.
Dordrecht,Netherlands:Kluwer9.
DePaulo,B.
M.
,Lindsay,J.
J.
,Malone,B.
E.
,Muhlenbruck,L.
,Charlton,K.
,&Cooper,H.
(2003).
Cuestodeception.
PsychologicalBulletin,129,74-112.
10.
McNeill,D.
(1992).
HandandMind.
WhatGesturesRevealaboutThought.
Chicago:ChicagoUniversity.
Press.
11.
Ekman,P.
,&Friesen,W.
V.
(1972).
Handmovements.
JournalofCommunication,22,353-374.
12.
Vrij,A.
(1995).
Behavioralcorrelatesofdeceptioninasimulatedpoliceinterview.
TheJournalofPsychology,129,15-28.
13.
Undeutsch,U.
(1967).
BeurteilungderGlaubhaftigkeitvonAussagen.
InU.
Undeutsch(Ed.
),HandbuchderPsychologie.
Bd.
II:ForensischePsychologie(pp.
26-181).
Goettingen,VerlagfurPsychologie.
14.
Newman,M.
L,Pennebaker,J.
W.
,Berry,D.
S.
,&Richards,J.
M.
(2003).
LyingWords:PredictingDeceptionFromLinguisticStyles.
PersonalityandSocialPsychologyBulletin,29,665-675.
15.
Darwin,C.
(1872/1998).
Theexpressionoftheemotionsinmanandanimals.
NewYork:Oxford.
(3rdEdition,w/commentariesbyPaulEkman).
16.
Ekman,P.
(2003).
Emotionsrevealed.
NewYork:HenryHolt.
17.
DePaulo,B.
M.
,Kashy,D.
A.
,Kirkendol,S.
E.
,Wyer,M.
M.
,&Epstein,J.
A.
(1996).
Lyingineverydaylife.
JournalofPersonalityandSocialPsychology,70,979-995.
18.
Ekman,P.
,Friesen,W.
V.
,&O'Sullivan,M.
(1988).
Smileswhenlying.
JournalofPersonalityandSocialPsychology,54,414-420.
ArticleID:SY07Page11TechnicalarticleWileyHandbookofScienceandTechnologyforHomelandSecurity19.
Ekman,P.
,O'Sullivan,M.
,Friesen,W.
V.
,&Scherer,K.
(1991).
Invitedarticle:Face,voice,andbodyindetectingdeceit.
JournalofNonverbalBehavior,15,125-135.
20.
deTurck,M.
A.
,&Miller,G.
R.
(1985).
Deceptionandarousal:Isolatingthebehavioralcorrelatesofdeception.
HumanCommunicationResearch,12,181-201.
21.
Frank,M.
G.
(1989).
Humanliedetectionabilityasafunctionoftheliar'smotivation.
Unpublisheddoctoraldissertation,CornellUniversity,Ithaca,NY.
22.
Burgoon,J.
E.
,&Buller,D.
B.
(1994).
Interpersonaldeception:III.
Effectsofdeceitonperceivedcommunicationandnonverbalbehaviordynamics.
JournalofNonverbalBehavior,18,155-18423.
Waid,W.
M.
,&Orne,M.
T.
(1982).
Thephysiologicaldetectionofdeception.
AmericanScientist,70,402-409.
24.
Steinbrook,R.
(1992).
Thepolygraphtest:Aflaweddiagnosticmethod.
NewEnglandJournalofMedicine,327,122-123.
25.
Frank,M.
G.
(2005).
Researchmethodsindetectingdeceptionresearch.
InJ.
Harrigan,K.
Scherer,&R.
Rosenthal(Eds.
).
HandbookofNonverbalBehaviorResearch.
London:Oxford.
(pp341-368).
26.
DePaulo,B.
M.
,Stone,J.
,&Lassiter,D.
(1985).
Deceivinganddetectingdeceit.
InB.
R.
Schlenker(Ed.
)Theselfandsociallife.
(pp.
323-355).
27.
Vrij,A.
(2000).
Detectingliesanddeceit:Thepsychologyoflyingandtheimplicationsforprofessionalpractice.
Chichester:JohnWiley.
28.
Zuckerman,M.
&Driver,R.
E.
(1985).
Tellinglies:Verbalandnonverbalcorrelatesofdeception.
InW.
A.
Siegman&S.
Feldstein(Eds.
)Multichannelintegrationofnonverbalbehavior.
(pp.
129-147).
Hillsdale,NJ:Erlbaum.
29.
Frank,M.
G.
(2003).
Smiles,lies,andemotion.
InM.
Abel(Ed).
Thesmile:Forms,functions,andconsequences.
NewYork:TheEdwinMellenPress.
(pp15-43).
30.
VrijA.
;EdwardK.
;&BullR.
(2001).
Policeofficers'abilitytodetectdeceit:Thebenefitofindirectdeceptiondetectionmeasures.
LegalandCriminologicalPsychology,6,185-19631.
Bond,C.
F,Jr.
,&DePaulo,B.
M.
(2006).
Accuracyofdeceptionjudgments.
PersonalityandSocialPsychologyReview,10,214-234.
32.
DePaulo,B.
M.
,&Rosenthal,R.
(1979).
Tellinglies.
JournalofPersonalityandSocialPsychology,37,1713-1722.
33.
Ekman,P.
,&O'Sullivan,M.
(1991).
WhocancatchaliarAmericanPsychologist,46,913-920.
34.
Ekman,P.
,O'Sullivan,M.
,&Frank,M.
G.
(1999).
Afewcancatchaliar.
PsychologicalScience,10,263-26635.
DePaulo,B.
M.
,&Pfeifer,R.
L.
(1986).
On-the-jobexperienceandskillatdetectingdeception.
JournalofAppliedSocialPsychology,16,249-267.
36.
Kraut,R.
E.
,&Poe,D.
(1980).
Behavioralrootsofpersonperception:Thedeceptionjudgmentsofcustomsinspectorsandlaymen.
JournalofPersonalityandSocialPsychology,39,784-798.
ArticleID:SY07Page12TechnicalarticleWileyHandbookofScienceandTechnologyforHomelandSecurity37.
Meissner,C.
A.
&Kassin,S.
M.
(2002).
:"He'sguilty!
":Investigatorbiasinjudgmentsoftruthanddeception.
LawandHumanBehavior,26,469-480.
38.
Mann,S.
,Vrij,A.
,&Bull,R.
(2004).
Detectingtruelies:Policeofficers'abilitiestodetectsuspects'lies.
JournalofAppliedPsychology,89,137-149.
39.
Frank,M.
G.
,&Ekman,P.
(1997).
Theabilitytodetectdeceitgeneralizesacrossdifferenttypesofhighstakelies.
JournalofPersonalityandSocialPsychology,72,1429-1439.
40.
Etcoff,N.
L.
,Ekman,P.
,Magee,J.
J.
,&Frank,M.
G.
(2000).
Superiorliedetectionassociatedwithlanguageloss.
Nature;405(11May),139-139.
41.
Forrest,J.
A.
,Feldman,R.
S.
,&Tyler,J.
M.
(2004).
Whenaccuratebeliefsleadtobetterliedetection.
JournalofAppliedSocialPsychology,34,764-780.
42.
O'Sullivan,M.
,&Ekman,P.
(2004).
Thewizardsofdeceptiondetection.
InGranhag,P.
A.
,&Stromwell,L.
(Eds.
),TheDetectionofDeceptioninForensicContexts(pp.
269–286).
Cambridge,UK:CambridgeUniversityPress.
43.
Bond,C.
F.
Jr;Atoum,A.
O.
(2000).
Internationaldeception.
Personality&SocialPsychologyBulletin,26,385-39544.
Bond,C.
F;Omar,A;Mahmoud,A;&Bonser,R.
N.
(1990).
Liedetectionacrosscultures.
JournalofNonverbalBehavior,14,189-204.
45.
Frank,M.
G.
,&Feeley,T.
H.
(2003).
Tocatchaliar:Challengesforresearchinliedetectiontraining.
JournalofAppliedCommunicationResearch,31,58-75.
46.
Kassin,S.
M;&Fong,C.
T.
(1999).
"I'minnocent!
":Effectsoftrainingonjudgmentsoftruthanddeceptionintheinterrogationroom.
Law&HumanBehavior,23,499-516.
47.
Bartlett,M.
S.
,Littlewort,G.
,Frank,M.
G.
,Lainscsek,C.
,Fasel,I.
,&Movellan,J.
(2006).
Fullyautomaticfacialactionrecognitioninspontaneousbehavior.
JournalofMultimedia,6,22-35.
48.
Collins,R.
,Lincoln,R.
,&Frank,M.
G.
(2002).
Theeffectofrapportinforensicinterviewing.
Psychiatry,Psychology,&Law,9,69-78.
FurtherreadinglistEnteryour"FurtherReading"listhere.
Thissectionwillcontainadditionalrecommendedreadings(books,articles,websites,etc.
)thatarenotincludedinthereferencesection.
Forsomearticles,thissectionmayincludereferencesforspecifictechniquesorprotocols;forothers,alistingofbookchaptersthatprovidegeneralinformationonthetopicmaybesufficient.
Referencesinthereadinglistarenotnumbered,butotherwiseshouldfollowthestyleoutlinedfortheReferencessection.
Cross-referencesSee:Socialandpsychologicalaspects;Terrorismrisk:characteristicsandfeatures;ArticleID:SY07Page13TechnicalarticleWileyHandbookofScienceandTechnologyforHomelandSecurityDeterrence;Physicalsecurity;modelsandcountermeasuresGlossarytermsMalfeasanceHotSpotsMicroExpressionContentBasedCriteriaAnalysisDemeanorPinocchioresponseMeta-AnalysisArticleID:SY07Page14