McDaniel&Gates–MarketingResearch,12thEdition Instructor’sManual

Copyright?2021JohnWiley&Sons,Inc. 6-

CHAPTER6

TraditionalSurveyResearch

LEARNINGOBJECTIVES

1.Understandthereasonsforthepopularityofsurveyresearch.

2.Learnthetypesoferrorsinsurveyresearch.

3.Distinguishthetypesofsurveys.

4.Gaininsightintothefactorsthatdeterminesthechoiceofparticularsurveymethods.

KEYTERMS

Adhocmailsurveys

Callcentertelephoneinterviews

Chancevariation

Computer-assistedtelephoneinterviews(CATI)

Door-to-doorinterviews

Executiveinterviews

Frameerror

Inputerror

Interviewererror,orinterviewerbias

Longitudinalstudy

Mailpanels

Mall-interceptinterviews

Measurementerror

Measurementinstrumentbias

Nonresponsebias

Populationspecificationerror

Randomerror,orrandom

Refusalrate

Responsebias

Sampledesignerror

Samplingerror

Samplingframe

Selectionerror

Self-administeredquestionnaires

Surrogateinformationerror

Systematicerror,orbias

CHAPTERSUMMARY

Thischapterexploresthebasicsofsurveyresearch.Inunderstandingsurveyresearch,itisimportanttounderstandthetypesoferrorswithinit.Thereissamplingerror,systematicerror,andmeasurementerror.Ofthese,systematicerrorandmeasurementerrorcanbeaffectedbytheresearcher.Thedifferenttypesofsurveys,thereappropriateuseandtheadvantagesanddisadvantagesofeachtypearediscussed.Thedifferenttypesofsurveysaredoor-to-doorinterviewing,mallinterceptinterviewing,executiveinterviewing,telephoneinterviewing,callcentertelephoneinterviewing,computer-assistedtelephoneinterviewing,IVRautomatedtelephonesurveys,directcomputerinterview,self-administeredinterviews,andmailsurveys.

Therearemanyfactorsdeterminingthechoiceofaparticularsurveymethod.Manyofthesearewithinthecontroloftheresearcher,whileseveralarenot.

QuestionsforReview&CriticalThinking

TheownerofahardwarestoreinEureka,California,isinterestedindeterminingthedemographiccharacteristicsofpeoplewhoshopathisstoreversusthoseofpeoplewhoshopatcompetingstores.Healsowantstoknowwhathisimageisrelativetothecompetition.Hewouldliketohavetheinformationwithinthreeweeksandisworkingonalimitedbudget.Whichsurveymethodwouldyourecommend?Why?

Atelephonesurveyisfastandhasarelativelowcost.Thisseemstofittheneedsoftheclient.ThesampleframeusedforthesurveycouldbefocusedontelephoneexchangesintheareasurroundingthestoreinEureka.Chooseareasonableradius.

Anotherpossibilityare“mall-intercepts”wherecustomersareinterviewedleavinghisstoresandthestoresofhiscompetitors.Whileinterviewerswouldhavedifficultygainingaccesstocustomersleavingcompetitors’stand-alonestores,theywouldhavenothavethatproblemifthestoreswerelocatedinstripshoppingcenters.

Discussthisstatement:“Amall-interceptinterviewisrepresentativeonlyofpeoplewhoshopinthatparticularmall.Therefore,onlysurveysthatrelatetoshoppingpatternsofconsumerswithinthatmallshouldbeconductedinamallinterceptinterview.”

Ifallofthepersonsshoppingatamallwereverysimilar,thisstatementwouldbetrue.However,shoppingmallsarelocatedinurbanareasindenselypopulatedregions.Oneneedsonlytovisitamalltoseethatpersonsfromawiderangeofdemographicsshopthere.Ifthepeoplewhoshopatthemallarerepresentativeofthepopulationofthesurroundingareaofthecity,orevenofthenation,thenarandomsampleoftheshopperswouldberepresentativeofthatpopulation.Thefindingswouldbeabletobegeneralized.However,regionalbigmallscanhaveshoppersfromawidegeographicareaplustourists.Youneedtoscreenrespondents.

Acolleagueisarguingthatthebestwaytoconductastudyofattitudestowardcitygovernmentinyourcommunityisthroughamailsurveybecauseitischeapestmethod.Howwouldyourespondtoyourcolleague?Iftimewerenotacriticalfactorinyourdecision,wouldyourresponsechange?Why?

Iftimeisacriticalfactor,thetelephone(randomdigitdial)wouldbethemostviablesolution.Amailsurveyisnotalwaysthecheapestandalowresponseratewouldgivealowreliabilityresponse.Nonrespondentsandthehighcostofpostagecandriveupthecostofconductinganadequatesurvey.However,withtimenotbeingacriticalfactor,asecondarymailingcouldincreasetheresponserate,butalsoincreasethecost.IfthecityisinanareawithhighInternetaccess,anon-linesurveywouldalsobeapossibility.

Discussthevarioussourcesofsampledesignerrorsandgiveexamplesofeach.

Thefirsttypeofsamplingerrorisrandomerror.Thiserroroccursbecauseofchancevariationinthesample.Toreducetheinfluenceofrandomerror,alargersampleshouldbeemployed.Thelargerthesample,otherthingsbeingequal,themoreitrepresentsthepopulationfromwhichitcame.Also,thelargerthesamplethehigherthecost.Forexample,perhapsyouinterviewedfiftypeopleinapopulationof50,000onthemeritsofbuildingaprisonandgotresultsverycontrarytowhatyouexpected.Thiscouldbeduetochance,becausethesamplewassosmall.Givenalargersample,theresultsmaybemorecloselyalignedwithexpectations.

Thesecondtypeofsamplingerrorissystematicerrorwhichincludesthreetypesofbiases.Thefirstisframeerror,whichresultsbecauseofusinganincompleteorinaccuratesamplingframe.Anexampleofthisischoosingyourrespondentsfromaphonebookandmissingallthosewithunlistednumbers,peoplewhohavemoved,andpeoplewithnewlistings.Thesecondtypeofbiasispopulationspecificationerror.Thiserrorisaresultofnotcarefullydefiningthepopulationofinterestfromwhichasampleisdrawn.Forexample,ifyouwereinterestedintheattitudesofconsumerswhohadpurchasedaparticularbrandoftoothpaste,yetspecifiedthepopulationaspeoplewhoshopataparticulargrocerystore,thenyoursamplemaynotmatchyourpopulationofinterest.

Thethirdtypeofbiasisselectionerrorwhichoccursasaresultofnotchoosingaparticularrespondentforsomereasonoutsideoftheresearchproject.Forexample,aninterviewerusingthemallinterceptmethodtypicallyleavesouttheelderly,becauseheorshedoesn’twanttotalktoolderpeople.

Whyisitimportanttoconsidermeasurementerrorinsurveyresearch?Whyisittypicallynotdiscussedinprofessionalmarketresearchreports?

Measurementerroroccurswhenthereisavariationbetweentheinformationbeingsoughtandtheinformationobtainedbythemeasurementprocess.Measurementerrorisveryimportantinsurveyresearchbecauseifthequestionsbeingasked,orthewayinwhichtheyarebeingasked,arenotconsistentandaccurate,thenthedatareceivedareoffarlessvalue.Itisalsoassumedthattheconfidenceintervalselectedwilltakecareofmeasurementerrors,whichisnotalwaysthecase.Aspecifiedalphalevelforstatisticalanalysishelpstominimizetheeffectsofthesebiases.

Whattypesoferrormightbeassociatedwiththefollowingsituations?

Conductingasurveyaboutattitudestowardcitygovernmentusingthetelephonedirectoryasasampleframe.

Sampleframeerror.

Interviewingrespondentsonlybetween8:00a.m.and5:00p.m.onfeaturestheywouldliketoseeinanewcondominiumdevelopment.

Sampleselectionerror.

Askingpeopleiftheyhavevisitedthepubliclibraryinthepasttwomonths.

Responsebias.

Askingpeoplehowmanytubesoftoothpastetheyusedinthepastyear.

Responsebias.

Tellinginterviewerstheycanprobeusinganyparticularexampletheywishtomakeup.

Interviewererror.

REALLIFERESEARCH

Case6.1–DoConsumersLikeChatbots?

Summary

Mostcustomerspreferliveagentsupporttoself-serviceandchatbots.

Preferencesareevenstrongincertainsituations,suchastechnicalsupport,emergencies,andmakingacomplaint.

Youngercustomersaremoreopentousingchatbots.

Questions

Whatmightthequantificationsbetoparticipateinthissurvey?

Studentswilllikelysuggestthatthequalificationwouldbehavingtriedself-serviceand/orchatbotsinthepast.Thiswouldbeamistakeasitwouldmissconsumerswhosostronglyopposechatbotsthattheyrefusetoeventrythem.

Anotherlikelyqualificationthatstudentsmightgiveishavingshoppedonlinebefore.Thistoowouldbeamistakesinceself-serviceandchatbotsarebothusedontelephonecalls.

Basically,anyonewhohaseverinteractedwithafirmlargeenoughtousethesetechn