廣告調(diào)查（第二版）英文 指導(dǎo)手冊(cè) Davis-Instructors-Manual-ch10 Data Collection Through Surveys

AdvertisingResearch:Instructor’sManual

Copyright?2012PearsonEducation,Inc.publishingasPrenticeHall

PAGE

PAGE

AdvertisingResearch:Instructor’sManual

Copyright?2012PearsonEducation,Inc.publishingasPrenticeHall

10.DataCollectionThroughSurveys

ChapterGoals

Thechaptershouldprovidestudentswithabetterunderstandingof:

? thedifferentoptionsforcollectingsurveydata,andthestrengthsandweaknessesofeachoption.

? howtoselectthemostappropriatedatacollectiontechniquefordifferentinformationneedsandresearchsituations.

? whyresponseandnonresponserates,andnonresponseerrorareimportantconsiderationsinevaluatingthequalityofsurveydata.

? howtoincreasesurveyresponserates.

NotestotheInstructor

TheChapterLectureprovidesaguidetokeytopicsandcontent.ThePowerPointslidesarenamed:davis_adresearch_ch10.ppt.

ChapterLecture

Slide10-

2

Slide10-

3

Surveysprovide“pointintime”snapshotlookatrespondents’attitudes,beliefsandbehaviors.

Foursurveydatacollectionmethods:(1)personal(facetoface)interviews,(2)

telephoneinterviews,(3)mailandpaperself-completedsurveys,and(4)online.

I.PersonalInterviews

Occurwhenanintervieweradministersasurveytoarespondentinaface-to-facesetting.Twomaintypes:interceptandpre-recruited.

A.InterceptInterview

Recruitsrespondents"onthespot,"forexample,asanindividualiswalking

throughashoppingmall,grocerystore,airport,ortrainstation.Mallinterceptisprobablymostcommonsettingforinterceptinterviews.

Mallandotherinterceptapproachesworkwelltoobtainhighincidenceindividualsorthosewhoarelikelytohavelocation-specificcharacteristics.

Slide10-

4

Slide10-

5

Slide10-

6

Slide10-

7

Slide10-

8

Slide10-

9

B.Pre-recruitedInterviews

Interceptsareaninefficientwayoflocatingindividualswithlesscommoncharacteristics.Inthesecases,personalprescreeningandpre-recruitmentworksbetter.Individualsareinitiallycontactedbymailortelephone,screened,andthenarrangementsforpersonalinterviewaremade.

C.AdvantagesandDisadvantages

Generallyprovidethehighestlevelofdataqualitydueto(1)personaladministrationofsurveybytrainedinterviewerand(2)face-to-facecontactbetweeninterviewerandinterviewee.Additionaladvantagesinclude:

? Workextremelywellwhenquestionnaireisverylongand/orcomplex

? Permituseofvisualstimuli

? Permitinterviewertomakecertainthatrespondentunderstandsindividualsurveyquestionsorcomplicatedinstructions

Primarydisadvantagesofpersonalinterviewsare:(1)highcostand(2)largeamountoftimerequiredforcompletion.

II.TelephoneInterviews

Typicallyconductedbyteamoftrainedinterviewerstelephoningfromacentrallocation.AlmostalltelephoneinterviewsuseCATI(ComputerAssistedTelephoneInterviewing)whereatelephoneinterviewerreadsthesurveyquestionsoffacomputerandtypesintheresponsesastheyareprovided.

A.Advantages

Provideopportunityforinterviewertoexplaincomplicatedinstructionsandquestions.Additionaladvantagesinclude:

? AbletoadministerinterviewswithverycomplexskippatternsbecauseCATIsystemsautomaticallycallsupnextappropriatequestionbasedontherespondent'sanswer.

? Lowermarginalcostsbecauseindividualscanbeinexpensivelycontactedseveraltimesiftheyarenotavailableduringthefirstcontact.

? Canmoreefficientlyobtainatruerandomsample

Slide10-

10

Slide10-

11

Slide10-

12

B.Disadvantages

Severallimitationsversusotherformsofinterviewing.

? Questionnairelengthmustberelativelyshort.Itisdifficulttokeeprespondentsonthetelephoneforextendedperiodoftime.

? Questioncomplexitymustberelativelysimple.Complexordetailedquestionsandscalesaredifficultforrespondentstorememberandansweraccurately.

? Difficulttocollectsensitivedata

? Refusalratesarehigh.Manyindividualsdonotliketobedisturbedathome.

? Extremelydifficulttousevisualstimuli

III.MailSurveys

Entailmailingeachrespondentapackagecontainingacoverletter,thesurveyquestionnaire,instructionsforcompletionandreturn,andastampedenvelopeaddressedtotheresearchcompanyconductingtheresearch.Thereisnopersonalinteractionbetweentherespondentandtheinterviewer.

? Anincentive,suchasamoneyorasmallgift,maybeincludedinthepackage

A.Approaches

Maybeconductedinoneoftwoways.

? Mailingstoselectedsampleofpopulation.Similartotelephonesurveys,listofnamesandaddressescanbegeneratedinternallyorpurchased.

? Mailpanel.Mailpanelisacontinuinggroupofindividualsthathaveagreedtoparticipateinsurveyresearch.Panelsarecreatedandmaintainedbyindependentmarketingresearchorganizations.

B.Advantages

Slide10-

13

Slide10-

14

Slide10-

15

Advantagesofmailpanelsinclude:

?Responseratestendtobesignificantlyhigher.Panelmembershaveagreedtoinadvancetoanswerquestionnairesmailedtothem.

?Costefficienciestendtobegreater.Higherlevelsofresponselowerperinterviewcost.

?Theselectionofindividualswithcertaindemographicorproductusagecharacteristicscanoftenbemadewithoutprescreening.

?Prescreeningoflow-incidencerespondentscanbeefficientlyconducted.

?Costefficiency-Thecostpercompletedinterviewcanbeconsiderablylowerthancomparablepersonalortelephoneinterviews.

?Respondentconvenience-Respondentsmaybemorewillingtoparticipatebecausetheycancompletethequestionnaireattheirconvenience.

C.Disadvantages

Weaknessesofmailsurveysinclude:

? Lowresponserate

? Limitedquestionnairelengthandcomplexity

? Extendedtiming

D.Self-AdministeredSurveys

Closelyrelatedtomailsurveysareself-administeredsurveysthataredistributedin-personbutarereturnedviathemailortocentrallocation.Commonlyusedtomeasurecustomersatisfactionorqualityofcustomer’sbrand-relatedexperience.

IV.OnlineSurveys

ThesesurveysaretypicallyadministeredovertheInternet.Theyentailthree

components:(1)invitation,(2)questionnairedesignandposting,and(3)datacollection.Invitationsmostcommonlymadeviaemail,whichcontainsalinktowebsitehostingquestionnaire.

Slide10-

16

Slide10-

17

Slide10-

18

A.Advantages

Abilitytoexecutecomplexskippatterns.

Manyonlinesurveysitesallowforrandomassignmentofindividualstodifferentgroupsandfortherandompresentationofstimulusmaterials.

Costsaretypicallylowerversusotherformsofdatacollection.Candrawfromaworld-widesample.

B.Disadvantages

Theprimarycautionregardingonlinesurveysrelatestosamplecomposition.SinceInternetaccessisnotyetuniversalandcomfortlevelsofInternetusagediffersacrossagegroups,thereisalwaysquestionofsamplerepresentativness.

V.Hybrid(Mixed-Mode)Surveys

Forsomesituations,the“best”datacollectiontechniqueisonethatisactuallymulti-modal,thatis,onethatusesmorethanoneapproachtodatacollection:

“Multi-modesamplingistheprocessofaccessingrespondentsusingmorethanonemodeforthesamestudy.Thiscouldhappentwoways:Differentmodesmaybeusedonlyatthecontactstage,whilethesampleitselfremainsononemode-for

example,contactingsomeoneviae-mailtorequestthattheydoastudyviaphoneorrecruitingpeopleviaphoneforanonlineormailstudy.Multi-modemayalsoapplytothequestionnaire,wheresomeportionofthesampleanswersonlineandanotherportionmayanswerviatelephone.”

Hybriddesignshavepotentialtoincreaseresponserateandasaconsequence,increasedataquality.

A.Limitations

Complextoputintopractice,withdifferentsampleframesrequiredforeachmodeofdatacollection.

Assumesthatthesameperson(oraverysimilarone)wouldgivethesame

responsetothesamequestionirrespectiveofmode,samplingframeandsamplingmethod.

VI.CriteriaforSelectingaDataCollectionMethod

Slide10-

19

Slide10-

20

Slide10-

21

Sixfactorsshouldbeconsideredwhendeterminingwhichsurveymethodisthebestforaparticularresearchstudy:

?cost

?timingrequirements

?sample,interview,andadministrativecontrol

?informationneeds

?topicandquestionnairecomplexity

?interviewlength

A.Cost

Costofresearchstudyreflectsquestionnairelength,requiredresponserate,geographiccoverage,andsamplecharacteristics.However,whenfacedwithcomparableparameters,personalinterviewstendtobeconsiderablymoreexpensivethantheothermethodsofdatacollection.

? Onlinesurveystendtobetheleastexpensive.

B.TimingRequirements

Telephoneandonlinesurveysgenerallyrequiretheleastamountoftime

?Telephonesurveysarerelativelyquickbecausealltrainingandcoordinationcanbeperformedatacentralfacilityandteamsofinterviewerscanworksimultaneously.

?Onlinesurveyscollectdataquicklybecause(1)thereislittletonolagbetweentimeofrespondentcontactandsurveyadministrationand(2)multiplequestionnairescanbecompletedsimultaneously.

Mailsurveystendtotakerelativelylonger.Questionnairesmustbeputintomail,processedanddeliveredbypostoffice,completedbyrespondentsattheirconvenience,placedbackintothemail,thenprocessedandreturnedbythepostoffice.

Personalinterviewsarealmostalwaystheslowestformofdatacollection,especiallyassamplesizesincrease.

C.Sample,Interview,andAdministrativeControl

Slide10-

22

Slide10-

23

Slide10-

24

Slide10-

25

Slide10-

26

Samplecontrolreferstoextentaresearchercancontrolwhorespondstothesurveyquestionnaire.

? Personalandtelephoneinterviewsprovidegreatestdegreeofsamplecontrol.Mailandonlineinterviewsprovideconsiderablylesscontrolasthereisnoassurancethatpersontowhomquestionnaireisdirectedisinfactpersonwhocompletesthequestionnaire.

Interviewcontrolreferstotheextenttowhicharesearcherhascontroloverthecircumstancesinwhicharespondentprovideshisorheranswerstothesurvey.

? Personalandtelephoneinterviewsprovidegreatestdegreeofinterviewcontrol.Onlineprovidesarelativelylowerdegreeofinterviewercontrol.Mailsurveysprovidetheleastamountofinterviewcontrolasthereisnoassurancethattherespondentwillreadnecessarymaterials,orcorrecterrorsinresponse.

Administrativecontrolreferstothedegreetowhicharesearcherisabletomonitorinterviewerquality.

? Thiscanonlybedoneintelephoneorpersonalinterviews.Personalinterviews,particularlythoseconductedinthefield,providetheleastamountofadministrativecontrol.Aresearchercannotpossiblymonitoreachpersonalinterview.

D.InformationNeeds

Somesurveytechniquesworkbetterthanothersforsensitivetopics.Respondentstendtobemorehonestandcandidintelephoneandmailinterviewsversuspersonalinterviews.Telephoneinterviewstendtoelicitfewer"sociallyacceptable"responsesorresponsesrelatedtosensitivebehaviorssuchasdruguse.

Personalinterviewsappeartobebettermeansforcollectingcompleteandconsideredresponses.

E.ComplexityoftheTopicandQuestionnaire

Personalinterviewsaremostappropriatewhenboththetopicandquestionnairearecomplex.

Telephoneinterviewsworkwellwhentheissueiscomplexandquestionnaire

complexityisentirelyduetoskippatterns.Telephoneinterviewsarenotappropriatewhenthequestionnairecomplexityisduetoverbalorquestioncomplexity.

Onlineinterviewsareappropriateforrelativelysimpletopicswithorwithoutquestionnairecomplexity.Respondent'sabilitytoseequestionsandbeguidedbythecomputerpermitsrelativelyhighlevelsofquestionnairecomplexity.

Mailinterviewsareappropriateforrelativelysimpletopicsandrelativelysimplequestionnairesdueto:(1)lackofanexternalcheckonresponses,(2)respondent'sinabilitytoclarifyquestions,and(3)inconvenienceofhavingtowritedownresponsestoopen-endedquestions.

Slide10-

27

Slide10-

28

Slide10-

29

Slide10-

30

Slide10-

31

F.InterviewLength

Personalcontactisessentialforlongersurveys.Permitsinterviewertokeeptherespondentinterested,attentive,involved,andacontinuingparticipantintheresearch.Mail,telephone,andinteractivedatacollectiontechniquesaregenerallyappropriateonlyforshorterinterviews.

VII.ResponseRate,NonresponseRateandNonresponseError

Responseratereferstothepercentageofavalidsamplewhoparticipateintheresearchbycompletinganinterview.

? Responserateiscalculatedbydividingthenumberofrespondentscompletingthesurveybythenumberofrespondentsinthevalidsample.

? Nonresponserateisthecomplementofresponserateandiscalculatedbydividingthenumberofindividualsinthevalidsamplenotcompletingthesurveybythetotalnumberofrespondentsinthevalidsample.

Slides10-29to10-31illustratecalculationofresponseandnonresponserate.

Highresponserate,andthusalownonresponserate,generallyindicatesthattherearenomeaningfuldifferencesbetweenthosewhorespondedtothesurveyandthosewhodidnot.Alowresponserategenerallyindicatesthatthereisprobablybiasinthedata,reflectingthefactthatthecharacteristicsofthosewhoparticipatedintheresearcharedifferentfromthosewhodidnotparticipate.

Example:

Slide10-

32

Slide10-

33

Slide10-

34

Slide10-

35

Slide10-

36

Slide10-

37

? Datapresentsresultsofanopinionsurveytakenbeforethestartofantismokingcampaign.Responseratesformenandwomenarehighandequivalent.

? Thechartcomparespre-campaigntopost-campaignaveragesfortotalsample.Outcomesappeartobegood.Totalsample’saveragesroseforallthreemeasures.

? Theconclusionthatthecampaignhasbeeneffective,however,isincorrectandisdirectlyduetotheeffectsofthenonresponseerror.Noticehowmenandwomen'sopinionsareunchangedfromthefirstmeasure.However,becauseofthenonresponseerror(i.e.,lowproportionofresponsesfrommen)theoverallaveragesincreasedduetooverrepresentationofwomeninthefinalsample.

A.SourcesofNonresponseError

Mainsources:refusalsandnot-at-homes.

? Refusal:whenaparticipantdeclinestoparticipateintheresearchstudyorfailstocompleteasurveyonceithasbegun.

? Not-at-homeoccurswhenarespondentisunavailableatthetimeofpotentialcontact.

B.ImprovingResponseRateToReduceNonresponseError

Primaryapproachesare:(1)advancenotification,(2)incentives,(3)callbacks,and(4)recontacts.Implementedsomewhatdifferentlyinpersonal,telephone,mail,andonlinesurveys.

1.ReducingNonresponseinPersonalInterviewsandTelephoneSurveys

Stepscanbetakentoensurethecooperationofthepotentialrespondent,forexample:

? Explicitlymentiontheinterviewtopicanddescribewhythetopicisimportantandrelevanttorespondent.

? Effectivelyandaccuratelydescriberesearchpurposeandgoal,notingtheimportanceoftheindividual'sresponsestoachievingthegoal.

? Makecertainthatrespondentsdonotfeelthreatenedbyparticipationintheresearchorbyhowtheirresponseswillbe

Copyright?2012PearsonEducation,Inc.publishingasPrenticeHall

AdvertisingResearch:Instructor’sManual

Copyright?2012PearsonEducation,Inc.publishingasPrenticeHall

111

used.

? Useincentivesatlevelsappropriatetothedemandsofthesurvey.

? Useadvancenotification.

Inadditiontopriorrespondent-centeredtechniques,carefultrainingandmonitoringofinterviewerscanreducerefusalrates.Importanttomakecertainthatinterviewersare

? successfullytrainedtoestablishquickrapportwithrespondents.

? carefullymonitoredwithregardtotheirindividualrefusalrates.

Interviewerswhohaveexceptionallyhighrefusalratesshouldbe

immediatelyretrainedorreplaced.

Nonresponseerrorduetorespondentunavailabilitycanbesignificantlyreducedwhencallbacks/multipleattemptstoreachaspecificrespondentareused.

Slide10-

38

2.ImprovingResponseRateinMailSurveys

Causesofnonresponseinmailsurveysaredifferentthanthatofpersonalinterviewsandtelephonesurveys.Majorprobleminmailsurveysrelatestorefusals;manysimplydonottakethetimetocompleteandreturnthesurvey.

Advancenotificationiseffectiveforinducingrespondentcooperationandstudyparticipation.Advancenotice:(1)brieflyinformsarespondentofthenature

andimportanceofthestudy,(2)thefactthattheyhavebeenselectedtoparticipate,and(3)thedatearoundwhichtheycanexpecttoreceivethesurvey.

Othertechniqueseffectivetechniquesinclude:

? Useacoverletterthatclearlystatestheinterviewtopic,describeswhythetopicisimportantandrelevanttotherespondent,effectivelyandaccuratelydescribesthepurposeandgoaloftheresearch,andexplainstheimportanceoftheindividual'sresponses.

? Includepostage-paidreturnenvelope

? Preparequestionnai