From_Big_Data_to_Behavior_and_Back

FromBigDatatoBehaviorandBack:PsychologicalResearchinaBigDataWorldWijnandIJsselsteijn,dept.IE&ISHuman-TechnologyInteractionGroupHuman-TechnologyInteractionGroupChrisAnderson,Editor-in-Chief,Wired.com•“Thisisaworldwheremassiveamountsofdataandappliedmathematicsreplaceeveryothertoolthatmightbebroughttobear.Outwitheverytheoryofhumanbehavior,fromlinguisticstosociology.Forgettaxonomy,ontology,andpsychology.Whoknowswhypeopledowhattheydo?Thepointistheydoit,andwecantrackandmeasureitwithunprecedentedfidelity.Withenoughdata,thenumbersspeakforthemselves.”VeryQuickPopQuiz•RaiseyourhandifyourecognisethepsychologistPsychologist#1Psychologist#2Psychologist#3PsychologyandDataScience•Psychology•Scienceofunderstandinghumanthoughtandbehavior•Formulatingandtestingtheoriesoncause-effectrelations•Empiricalscience,data-driven•DataScience(fromapsychologist’perspective)•Tracking(productsortracesof)humanbehaviourin(near)real-timeandatalargescale,suchthatitimprovesourunderstandingoftheinternalandexternalforcesthatshapehumanbehaviour,andenablesmeaningfulaction.HumanBehaviourandBigDataSocialmediaandnetworksMobiledevicesCurrentdatabasesSmartobjects/IoTDatascienceisnotsomuchaboutthesheerquantityofdatathatcanbecaptured,processed,andaccessed,butratheraboutthecapacitytosearch,aggregate,andcross-referencedata,suchthatrelationalinferencescanbedrawn.SocialmediaandnetworksSmartobjects/IoTPsychologyandDataScience•Psychologycanhelptocreateinsightintothekindsofbehavioursthataremostfruitfultotrackandanalyse.•Psychologycanprovidetheoreticalandmethodologicalguidanceonbehaviouralinferencesthatcanbedrawn,distinguishingspuriouscorrelationsthatfrequentlyemergeinlargedatasetsfrommeaningfulbehaviouralpatterns.largedatasetsfrommeaningfulbehaviouralpatterns.•Psychologycanhelptoprovideuserswithpsychologicallymeaningfulandactionablefeedback,andtoelicitsustainablebehaviourchange–i.e.,takingactiononthebasisofdata.•Psychologycaninformdesignonissuesofprivacy,transparency,useracceptance,etc.BigDataandPsychology2.0•Inturn,psychologyisaffectedbyBigDataindifferentways•Psychologyisbecomingincreasinglyinterdisciplinary,requiringinformaticsskillsincludingexpertiseinprogramming,signalincludingexpertiseinprogramming,signalprocessing,patternrecognition,machinelearning,etc.TheAgeofSmallData:HippchronoscopeandCatellchronometer/IndustrialEngineering&InnovationSciencesComputersinPsychology/IndustrialEngineering&InnovationSciencesComputersinPsychology/IndustrialEngineering&InnovationSciences•InsertfMRI/IndustrialEngineering&InnovationSciences/IndustrialEngineering&InnovationSciencesTheSmartphoneLaboratory/IndustrialEngineering&InnovationSciencesMobileelectronicdevicesinpsychology•Aggregatedcallroutingrecords,providedbytelecomproviders•Bigsamples,butnotrichdata•Socialnetworkanalysis•CustomPDAsorelectronicallyactivatedrecorders(EARs)•Diarystudies,experiencesampling(ESM)studies•Dedicatedsmartphonewithpsychapp•ContextualisedESM•E.g.,ContextPhone,SoundSense,EmotionSense•Psychappsonly(i.e.,supposedtorunonyourownsmartphone)•Fewstudiesavailable•E.g.,TrackYourHappiness.org–with250,000happinessreportsfrom5000peopleacross83countriesSmartphonePsychology•Intimate,personal,trustedpiecesoftechnology•Onboardsensors,e.g.,•GPS:identifywhereparticipantslive,study,eat,andplay,andhowtheyareaffectedbylocalweather,traffic,socialevents,etc.•Microphone;bluetoothdensity:localsocialinteractions/activity•Accelerometers:walking,running,dancing,calmversusnervous,etc.•Accelerometers:walking,running,dancing,calmversusnervous,etc.•Futuresensorsforlight,humidityorpollution:studyhowenvironmentalvariablesinfluencebehavior•RFIDtagreader:loginteractionswithsystemsandproductsinshops,athome,ontheroad,etc.•Externalsensors,e.g.,•Psychophysiologicalsensors:e.g.,heartrateorsweat–indicatorsofexertion,stressoremotions•Biosensors:e.g.,bloodpressure,bloodglucose,bloodalcohol,immunesystemactivity,hormonelevels,etc.Influencinghumanbehavior/IndustrialEngineering&InnovationSciencesExperimentalcontrolvs.ecologicalvalidityvalidity“Field”/IndustrialEngineering&InnovationSciencesExperimentalcontrolEcologicalvalidity“Lab”Experimentalcontrolvs.ecologicalvalidityvalidity“SmartphoneLab”/IndustrialEngineering&InnovationSciencesExperimentalcontrolEcologicalvaliditySmartphonePsychologysomeethicalconcerns•Informedconsent•Mostusersdonotreadlicencingagreements•Difficulttoguaranteeanonimityindata-richsmartphonestudies•Potentialinferencesdrawnfromseeminglyharmlessdata•Potentialinferencesdrawnfromseeminglyharmlessdatamaybesurprisingtomanyusers•Participantconfidentialityisvulnerabletoauthorities•Allowretroactivedata-managementbyparticipants?•Global,multi-sitestudiesareproblematicinIRBterms•OneIRBapprovalforallsites?•Datastoredinthecloud?ChrisAnderson,Editor-in-Chief,Wired.com•“Thisisaworldwheremassiveamountsofdataandappliedmathematicsreplaceeveryothertoolthatmightbebroughttobear.Outwitheverytheoryofhumanbehavior,fromlinguisticstosociology.Forgettaxonomy,ontology,andpsychology.Whoknowswhypeopledowhattheydo?Thepointistheydoit,andwecantrackandmeasureit