(Notes) 2: Research Methods & Scientific Study
THENEEDFORPSYCHOLOGICALSCIENCE
therealpurposeofthescientificmethodistomakesure
UNITZ Naturehasn'tmislead
youintothinkingyouknowsomething
youactuallydon'tknow
researchmethods the
scientificstudyofpsychology DIDWEKNOWITALLALONG HINDSIGHTBias
hindsightbias thetendencytobelieveafterlearning
anoutcomethatonewouldhaveforseenit
Iknewitallalong
t thetendencytobelieveweknowmorethanwe outofsightoutofmind
OVERCONFIDENCE
absencemakestheneartgrowfonderactuallydo
it trustingourintuition memoryofthepast
thatisstimulatedinusbysomethinginthe
present
t overconfidenceintheaccuracyofourknowledge THREEMAINCOMPONENTS
IT asandoverconadence
thttthENFthfftruffy
tcuriouseagerness
leadustooverestimateourintuition ggtskepticallyscrutinizecompetingideas
scientificinquiryhelpssiftreality openmindedhumilitybeforenature
FuttearENFFIGAEHORadCRITICAL
THINKING THEORY
merehunchthatintegratesasetofprinciplesthatorganizes
observations predictsbehaviorsorevents
tiffhIndanyang HYPOTHESISatexaminesassumptions
thjjjjjnon atestablepredictionimplied
bytheorythatcanbeconfirmed
itevaluatesevidence orrefuted
OPERATIONALDEFINITION
statementoftheexactprocedureusedinaresearchstudy
REPLICATION
repeatingtheessenceofaresearchstudytoseelfresultscan
begeneratedtootherrartions
circumstances
I
0 AGOODTHEORY ISUSEFULIFIT
effectivelyorganizesarangeofselfreports
andobservations
Impliesclearpredictionsthatanyonecanuseto
checkthetheory
03TYPESOFMETHODSTOTESTTHEORY
Firetational
experimentation
DESCRIPTION
E NATURALISTIC
ILIAS
ESTUDY THESurvey
C examinesoneindividual Ekatmanyeaseatonce OBSERVATION
C describesbehavior
C wordeffectscanaffect doesnotexplainInattrianempiricalway
smuggestfurtherstudy Tinions
nomeasurementevidencec
c Igytdiscerngeneral randomsampling
C
C thepluralofanecdote balancestheinfluence
C
ddd
criticalthinking
ofvividcasesor
C anecdotes
spaceistoward
C Gpopulation
C randomsample
a
C beinginoneplacecantdoforyou
CORRELATION
gG ameasureoftheextenttowhichtwofactorsvarytogetherandthusofhowwelleitherfactorpredictstheother
HowwelldoesApredicts
C WM T
GELATIONC astatisticalindexoftherelationshipbetween2things Cautionf t strengthofthecorrelationC I to 1
a tscatterplot
C correlationhelpspredictbutdoesnotimplycause effect
C EXAMPLE depression selfesteemarecorrelated
C Illusorycorrelations
C perceivednonexistantcorrelation
C tarandomcoincidencebaseduponafocuson'con
C formingevidenceonly
C PERCEIVINGORDERINRANDOMEVENTS
C ttcomesfromourneedtomakesenseoutofworld
Gli.e.comflippokerhand
E EI EIMATIONC RANDOMASSIGNMENT
C eliminatesaltexplanations experiment
C n
fscoatauseettedC blinduninformed controlfactorsmanimanmortnetactorissotinterest
C singleblindproceduresubjects nonconstanticontroningstacins
o
C doubleblindprocedureresearchers
C placeboeffect
p n g p I l
C receivestnetreatmentandependent
independentvariable confoundingvariableC controlgroup
C doesnotreceivetnerreatment effectofrandomassignmentonconfoundingvariables
C dependentvariable
C Gwhatisbeingmeasured
a
8
Gallatimethnemen a phenomenon munch
some peopleexperience a
2benefitaftertheadmin
ofsymptomsisobserveddespite of an inactive look alike
usinganonactivetreatment substance on treatment
occursduringfactorslike
expectations classical
conditioningardcanease
Itypes ofplacebospainorfatigue
pureorinactiveplacebossuchas
specifically sugarpillsorsaline injections
used in
controlgroups impureoractiveplacebossuchas
andexperimentation prescribinganantibioticfoaviral
unfectionoravitamineventhoughthe
pep pygmy y
mode occursthemost
mean arithmeticaverage
median middlescore
measureofvariability
normalcurve bellshaped
positiveskew
longtailisonthepositiveside
medianistotherigntofthemedian modeor
somesaythedistributionisskewedtotheright
INFERENTIAL
STATISTICS
tryingtoreachconclusionsthat
extendbeyondjustdescribingthe
data
usedtoleshypothesisaboutsamples
negativeskew MAKINGINFERENCESFROMDATA
longtallisonthenegativeside Gresearchgoal
meanistotheleftofthemedianormodeor otomakeastatisticalstatementofhow
somesaythedistributionisskewedtotheleft frequentlyanobtainedresultoccuredby
experimentalmanipulation
orbuchancy
whenisadifferencesignificant
ostatisticalsignificanceisastat A
Isticalstatementofnowlikelyitisthatanobtained
theaveragesarereliable
thedifferencesbetweenaveragesisrelativelylarge
doesimplytheimportanceoftheresults
range 0 whensampleaveragesarereliableandthedifference
Standarddeviation acomputedmeasureof betweenthemisrelativelylargewesaythediffhas
howmuchscoresvaryaroundthemeanscore statisticalsignificance Itisprobablynotdueto
chancevariation forpsychologiststhisdifference
standarddeviation
fsumofldeviations ismeasuredthrualpnalevelisset 51cops.os
numberofscores
CONFIDENCELEVEL INTERVALS
confidenceinterval instatistics aconfidenceintervalis CONTROL
aparticularkindofintervalestimateofapopulationparameter Ghatchanceeventsaresubjecttopersonal
insteadofestimatingtheparameterbuasinglevalue aninterval controlisanillusionofcontrolfedbylikelytoincludetheparameterisgiven Lex4012or40151 1illusorycorrelation theperceptionofarelationship
o confidencelevelakaconfidencecoefficientrepresentsthepossibility
mmmm
wherenorelationshipactuallyexists
thattheconfidenceintervalistocontaintheparameter ex951 2 regressiontowardthemean thetendencyfor
confidencelevel extremesofunusualscoreseventstoregress
0 populationsize instatistics populationistheentireentitiesconcerningwhich towardtheaverage
statisticalinferencesaretobedrawnThepopulationsizeisthetotal ofthe
ETHICS INRESEARCH
whenesariobserveddifferencerelvable I octniesingamsonstiisinganimaisinresearch0 representativesamples biasedsamples
0lessvariableobservations morereliablethanmorevariable safeguardsforanimaluse
0 morecases fewercases ethicsinhumanresearch
TESTINGFORDIFFERENCES informedconsent
Ifwehaveresultsmeansfromagroupsbeforeweinfercausation protectedfromharm discomfort
wemustasktheQuestion maintainconfidentiality
toanswerthis debriefing
ÉÉ must
betweenthemeansofagroups
gruna Ttest
ordidithappenbychance
Oo