THENEEDFORPSYCHOLOGICALSCIENCEtherealpurposeofthescientificmethodistomakesureUNITZ Naturehasn'tmisleadyouintothinkingyouknowsomethingyouactuallydon'tknowresearchmethods thescientificstudyofpsychology DIDWEKNOWITALLALONG HINDSIGHTBiashindsightbias thetendencytobelieveafterlearninganoutcomethatonewouldhaveforseenitIknewitallalongt thetendencytobelieveweknowmorethanwe outofsightoutofmindOVERCONFIDENCEabsencemakestheneartgrowfonderactuallydoit trustingourintuition memoryofthepastthatisstimulatedinusbysomethinginthepresentt overconfidenceintheaccuracyofourknowledge THREEMAINCOMPONENTSIT asandoverconadencethttthENFthfftruffytcuriouseagernessleadustooverestimateourintuition ggtskepticallyscrutinizecompetingideasscientificinquiryhelpssiftreality openmindedhumilitybeforenatureFuttearENFFIGAEHORadCRITICALTHINKING THEORYmerehunchthatintegratesasetofprinciplesthatorganizesobservations predictsbehaviorsoreventstiffhIndanyang HYPOTHESISatexaminesassumptionsthjjjjjnon atestablepredictionimpliedbytheorythatcanbeconfirmeditevaluatesevidence orrefutedOPERATIONALDEFINITIONstatementoftheexactprocedureusedinaresearchstudyREPLICATIONrepeatingtheessenceofaresearchstudytoseelfresultscanbegeneratedtootherrartionscircumstancesI0 AGOODTHEORY ISUSEFULIFITeffectivelyorganizesarangeofselfreportsandobservationsImpliesclearpredictionsthatanyonecanusetocheckthetheory03TYPESOFMETHODSTOTESTTHEORYFiretationalexperimentation DESCRIPTIONE NATURALISTICILIASESTUDY THESurveyC examinesoneindividual Ekatmanyeaseatonce OBSERVATIONC describesbehaviorC wordeffectscanaffect doesnotexplainInattrianempiricalwaysmuggestfurtherstudy Tinionsnomeasurementevidencecc Igytdiscerngeneral randomsamplingCC thepluralofanecdote balancestheinfluenceCdddcriticalthinkingofvividcasesorC anecdotesspaceistowardC GpopulationC randomsampleaC beinginoneplacecantdoforyouCORRELATIONgG ameasureoftheextenttowhichtwofactorsvarytogetherandthusofhowwelleitherfactorpredictstheotherHowwelldoesApredictsC WM TGELATIONC astatisticalindexoftherelationshipbetween2things Cautionf t strengthofthecorrelationC I to 1a tscatterplotC correlationhelpspredictbutdoesnotimplycause effectC EXAMPLE depression selfesteemarecorrelatedC IllusorycorrelationsC perceivednonexistantcorrelationC tarandomcoincidencebaseduponafocuson'conC formingevidenceonlyC PERCEIVINGORDERINRANDOMEVENTSC ttcomesfromourneedtomakesenseoutofworldGli.e.comflippokerhandE EI EIMATIONC RANDOMASSIGNMENTC eliminatesaltexplanations experimentC nfscoatauseettedC blinduninformed controlfactorsmanimanmortnetactorissotinterestC singleblindproceduresubjects nonconstanticontroningstacinsoC doubleblindprocedureresearchersC placeboeffect p n g p I lC receivestnetreatmentandependentindependentvariable confoundingvariableC controlgroupC doesnotreceivetnerreatment effectofrandomassignmentonconfoundingvariablesC dependentvariableC Gwhatisbeingmeasureda8Gallatimethnemen a phenomenon munchsome peopleexperience a2benefitaftertheadminofsymptomsisobserveddespite of an inactive look alikeusinganonactivetreatment substance on treatmentoccursduringfactorslikeexpectations classicalconditioningardcaneaseItypes ofplacebospainorfatiguepureorinactiveplacebossuchasspecifically sugarpillsorsaline injectionsused incontrolgroups impureoractiveplacebossuchasandexperimentation prescribinganantibioticfoaviralunfectionoravitamineventhoughthepep pygmy ymode occursthemostmean arithmeticaveragemedian middlescoremeasureofvariabilitynormalcurve bellshapedpositiveskewlongtailisonthepositivesidemedianistotherigntofthemedian modeorsomesaythedistributionisskewedtotherightINFERENTIALSTATISTICStryingtoreachconclusionsthatextendbeyondjustdescribingthedata usedtoleshypothesisaboutsamplesnegativeskew MAKINGINFERENCESFROMDATAlongtallisonthenegativeside Gresearchgoalmeanistotheleftofthemedianormodeor otomakeastatisticalstatementofhowsomesaythedistributionisskewedtotheleft frequentlyanobtainedresultoccuredbyexperimentalmanipulationorbuchancywhenisadifferencesignificantostatisticalsignificanceisastat AIsticalstatementofnowlikelyitisthatanobtainedtheaveragesarereliablethedifferencesbetweenaveragesisrelativelylargedoesimplytheimportanceoftheresultsrange 0 whensampleaveragesarereliableandthedifferenceStandarddeviation acomputedmeasureof betweenthemisrelativelylargewesaythediffhashowmuchscoresvaryaroundthemeanscore statisticalsignificance Itisprobablynotduetochancevariation forpsychologiststhisdifferencestandarddeviationfsumofldeviations ismeasuredthrualpnalevelisset 51cops.osnumberofscoresCONFIDENCELEVEL INTERVALSconfidenceinterval instatistics aconfidenceintervalis CONTROLaparticularkindofintervalestimateofapopulationparameter Ghatchanceeventsaresubjecttopersonalinsteadofestimatingtheparameterbuasinglevalue aninterval controlisanillusionofcontrolfedbylikelytoincludetheparameterisgiven Lex4012or40151 1illusorycorrelation theperceptionofarelationshipo confidencelevelakaconfidencecoefficientrepresentsthepossibilitymmmmwherenorelationshipactuallyexiststhattheconfidenceintervalistocontaintheparameter ex951 2 regressiontowardthemean thetendencyforconfidencelevel extremesofunusualscoreseventstoregress0 populationsize instatistics populationistheentireentitiesconcerningwhich towardtheaveragestatisticalinferencesaretobedrawnThepopulationsizeisthetotal oftheETHICS INRESEARCHwhenesariobserveddifferencerelvable I octniesingamsonstiisinganimaisinresearch0 representativesamples biasedsamples0lessvariableobservations morereliablethanmorevariable safeguardsforanimaluse0 morecases fewercases ethicsinhumanresearchTESTINGFORDIFFERENCES informedconsentIfwehaveresultsmeansfromagroupsbeforeweinfercausation protectedfromharm discomfortwemustasktheQuestion maintainconfidentialitytoanswerthis debriefingÉÉ mustbetweenthemeansofagroupsgruna TtestordidithappenbychanceOo
