PROCEDURESAND TECHNIQUESINMARKETINGRESEARCH 59
1.Thedeclineinmarketshareistheresultofincreasedcompetitiveadvertisingin
thearea.
- Thedeclineinmarketshareistheresultofthete::t marketingofa newproduct
bya majorcompetitor.
3.Thedeclineinmarketshareis theresultof"stockouts"attheretaillevelcaused
bya truckingstrikeinthearea.
Inattemptingtoverifyoneormoreofthesehypotheticalstatements,theresearcherexam-
inescompanyrecordstouncovernewsourcesofinformationortodiscoverrelationships
inolddatawithbearingonthecurrentproblem.Interv iewswithcompanyexecutivesand
operatingpersonnelareoftenconducted.Interviews are alsoconductedwithvariousper-
sonsoutsidethecompanywhoseopinionsmightbeexpectedto havesomerelevancetothe
problem.Thepreliminarysearchis alwayslimitedtoobtaininganinsightintotheproblem
andintopossible.solutionsforit.
inthefinalphaseofthepreliminaryinvestigation,thei'esea~cheranalyzestheresults
hehassofarobtainedandrestatesthemintheformofresearchquestionstobetestedin
thesubsequentresearchsteps.
CreatingtheResearchDesign
Thedesignofa marketingresearchprojectis theplanproposedfort~stingtheresearchques-
tionsaswel1ascollectingandprocessinginfoITiation.TheadministratioDoftheproject
accordingtothedesigninsuresthatthefact-findingprocesswiIlbeadequatelycontrol1ed.
" Design"meansmorethansimplyusinggoodmarket researchprocedures.Everyresearch
projectshouldbeindividuallydesignedtoproducethekindsofinformationneededtosolve
a particularproblem. Forthisreason,nctwomarket researchprojectsareeverexactlyalike.^3
Sixstepsareinvolvedincreatinga researchdesign:choosingtheapproach,deter-
miningtypesofdataneeded,locatingdatasources,choosinga methodofcollectingdata,
selectingth e sample,andanticipating/col1ectingtheresults.
ChoosingtheApproach Threealternativeapproachesarepossibleincreatinga
researchdesign.Theyarenotmutuallyexclusive,butinmostcases,thedesignofaresearch
planislimitedtotheuseofoneofthethree.
Thefirstapproachis theexperimentalapproach.This approaChrequiresthatcertain
proceduralrulesmustbefollowed. Essentially,thevariableofinterest---c.g.,price,message-
mustbemanipulatedandeveryoneparticipating intheexperiment musthavea knownand
equalchanceofbeingselected.
Ina marketexperiment,informationrelatingtothebasic problemis obtainedthrough
theuseofa small-scalesimulatedprogramdesignedtotestasp~cificresearchhypothesis.
Suppose,forexample,thatwewishtotestthequestionthatfamiliesofsimilarsizeand
economiccharacteristicslivinginthreedifferentcitiespurchasedifferentamountsofapar-
ticularformulaofa softdrink,suchasDr.Pepper.Thefirststepwouldbetoestablishthe
researchquestion:"Fora givent;meperiod, t heaveragefluidouncesofa FormulaA , B,
orC purchaseineachcitywerethesame."Next,a sampleofthefamiliesineachcitywould
beselectedandrandomlyassignedeitherA, B,orC.Next,a surveywouldbetakentodeter-
minethenumberofouncespurchasedbyeachfamily.Once this wasdone,a statisticaltest
wouldbeusedtotesttheresearchquestion.Ifstatisticallysignificantdifferencesinpur-
chasesofFormulaA,B,orCofDr.Pepperwere noted, it couldbeconcludedthattaste
doesinfluencetheamountofthissoftdrinkpurchasedby familieswiththesamesocialand
economiccharacteristics.Ofcourse, otherhypothesesaboutsoftdrinkpurchasingcould