ACOMPARISON OFTWO DATACOLLECTING …

1 Hacettepe niversitesiE itimFak ltesiDergisi18: 1 -10[2000}ACOMPARISONOFTWODATACOLLECTINGM ETHODS:INTERVIEWSANDQUESTIONNAIRESB urcuAKBAYRAK* ZET:Bumakaledeveritoplamaarac olaraksosyalbi-limlerveegitimdeyap lanara t rmalardas k akullan lang r meveanketteknikleriincelenmi tir. nceherikitek-nigintan mlanvet rleri zerindedurulmu r melerveanketler e itliy nlerdenkar la t r lm t la t rmamaliyet,s re,planlama, rneklemb y kl - ve rnekleme,istenilenbilgiyeula ma,yanl l k,gizlilik,yan tlamaoran ,ge erlik,g venirlikveverilerinanalizi l- tlerineg reyap lm t ,herikitekniginavan-tajlanvedezavantajla r birtabloda zetlenmi ZC KLER:G r ,Ara t rmaY :Inthisartide, , ,time,schedule,samplesizeandsampling,acc esstoinformation,bias,anonimityandconfid entiaIty,responserate,validity, ,theadvantagesanddisadvantagesofbothtech niquesweresummarisedina :lnterviews, ,Inthisartiele, 'self-administered'and'mailed'questionna ires, kindofconversationandasHull(l)tellsus".. butofParticularkind,whereactorstalktoa specificandconsciouspurpose.]

2 "ChannelandKahn(1968)definedinterviewas" atwo-personconversationinitiatedbytheint erviewerforthespecificpurposeofobtaining research-relevantinformation,andfocusedb yhimoncontentspecifiedbyresearchobjectiv esofsystematicdescription,prediction,ore xpla-nation"[citedin(2) ].Atitssimplestform,a questionnaireisnomorethana ,toassurethatmis-understandingsorambigui tiesinthequestioningarereducedtoaminimum ,andtoenabledatatobecomparedacrossthemem bersofa sample,anumberofdifferentwaysofpresentin gques-tionshavebeendeveloped(3).Whencons tructingquestionsforinterviewsandquestio nnaires,Foddy(4)emphasizesthattheresearc hermustelearlydefinethetopicofeachquesti on,determinetheapplicabilityofthequestio ntoeachrespondent, , (5)emphasizethat"Thereisgenerallylittler oomforvariationinresponseexceptwhereanin frequentopen-endedquestionmavbeused".Als othevadd"Thereisverv* ,FacultyofEducation, [J. ofEdLSlittleflexibilityinthewayquestions areaskedoranswered".Theunstructuredinter views, ,theircontent, ,thisdoesnotmeanthattheunstructuredinter viewisa morecasualthing,andforinitsownwayitmustb ecarefullyplanned(2).]

3 ,namelythemail, ' (9 )saythat" ."Thepurposeofthequestionnaireisexplaine d,andthentherespondentsisleftalonetocomp letetheques-tionnaire, highresponserate,accu-ratesampling,andmi nimumofbias,providingnecessaryexplanatio nsandgiyingthebenefitofpersonalcontact(6 ). (10 )tellus"Withoutdoubt,themailquestionnair eisgener-allycheaperthanothermethods" "Questionnairescanbesentthroughthemail;i nterviewerscannoL"Therearetravelingandex istenceexpensestobeoccurred, m(11 )hasfoundthisfactorasadifficultyinhisres earchandheindicatesthat"Abouttwiceasmany teacherapplied,forwhichreasonselectionwa sbasedonpracticalfactors,suchastimetable sandtravelIinatoexpenses."Asa result, , ,suchasfollowingofthegraduatesofa doesnotrequireatrainedstafforthefieldwor kers,thepaymentstothemandtheirtravelexpe nsesdonotexisLFurthermore,allit requiresthecostofplanningandpiloting,pri ntingexpenses,sampling,andmailing,provid ingstamped, ,thereisadangeroflowresponserateinmailqu estion-nairethatitcanmakethesurveyexpens ive."..theresponsetoa mailsurveymaybesolowthatthecostpercomple tedquestionnaireishigherthanwithaninterv iewsample.

4 "(10 ). takesseveralinterviewswithAComparisonofT woDataCollectingMethods:lnterviewsandQue stionnairesdifferentpeoplebeforeasimilar ,a ; caneffectofreducingthenumberofpersonstop ar-ticipate, (12 )admitsthat"Stafffeelit isimportanttoallowtimeforstudentstofeela teaseduringtheinterview,sointerviewsaret imedtolastatleast30minutes."AIsoshequot- edfromHowardJones"Studentswhohavebeenfor interviewatothercollegesareof tensur-prisedatbeinggi yenthisamountoftime,butwefeelthatit isimportanttogivestaffandstudentstimetof indoutabouteachother",Additionally, (13)saythat"Interviewstook45minutestoi hourandwerecompletelyvolun-tary," it "Arrangementstovisit;securingnecessarype rmission;confirmingarrangements;reschedu Iingappointmentstocoverabsencesandcrises ; ;tapesifusedrequirewholeorpartialtranscr iption(onehouroftapetimerequirestenhours totranscribefully)"(7 ).AIso, ,Huse'nandPostlethwaite(14)suggestthatfo radults30minutesisabouttheupperlimitthat canbeexpectedinthewayofansweringtimewhen questionnairesareadmin-isteredingroupset ting.

5 "Whenquestionnairesaremailedtorespondent s, "(14:pA883).Therefore,thefullquestionnai reshould3requireslessthan30minutestocomp leteand,preferably,lessthan15or20minutes ."Questionnaireswerecompletedinabout20mi nutesandnopupilsrefusedtotakepart."(15 ).Alongandtime-consumingquestion-na re,aspointedoutbyHuse'nandPostlethwaite( 14:pA883)"..maycausearespondenttoceaseto cooperateaftera ,onewillreceiveanincompletequestionnaire and,atworst,thequestionnairewillnotberet urned."Therefore,thesefactorsinfluenceth eresponserate, ,self-completedquestionnaires,whichrespo ndentsfillinforthemselves, school,ortothehundredsofemployeesinfirm, canbecompletedbythemandreturnedtotherese archeraboutthesameamountoftimethatit takestocompletea , ,whichofferstoarespondentonlytwoalternat ives:'yes-no'or'agree-dis-agree'.Accordi ngtoKerlingerthechiefadvan-tageofthistyp equestionisthatithasgreateruniformityofm easurementandsohashigherreliability;maki ngrespondentsanswerina man-nerfittingtheresponsecategory; ,thepos-sibilityofirritatingrespondentsw hofindnoneofthealternativessuitable,andt hepossibilityofforeingresponsesthatarein appropriatecanbefoundtheirdisadvantages[ citedin(2)],Theopen-endedtypewhichsupply a frameofrefereneefortherespondents'answer sbutputaminimumcontrolontheanswersandthe ir4 BurcuAkbayrak[ ] ,theopen-endedquestionsobtainseveraladva ntages:theyareflexible;theyallowtheinter viewertoprobesothathe/shemaygointomorede pthifhe/shedecides,ortoelearupanymisunde rstandings;theyalsoenabletheinterviewert otestthelimitsoftherespondentsknowledge; theyencourageco-operationandhelptoestabl ishrapport;andfinallytheyallowtheintervi ewertomakea (16),(9),(13).

6 Clementexplainsthereasonofthisas"Thelast sectionofeachquestionnairewasleftopensot hatHeadteachersandassistantteacherscould addtheirpersonalcommentsinresponsetotheS tatutoryOrdersforart."(16 ). (intervieweesandrespondentsofquestionnai res)neededfroma (19)recom-mendsthat5 to10% , ,whereastheinterviewerhastofindtherighti ndividualathomeinaparticularhouseholdori naworkplace(20).Inaninterview,regardingi tstypethesam-ples ,Schenk(21)hasreducedhersamplegrouptoget depthinformation."Structuredinterviewswi th50experienceddesignerswereconductedand supplementedbyshorter,focusedinterviewsw ith20juniordesigners."(21:p. 74). ,TizardandHugheshadtoreducetheirsamplesi zebecauseoftheprocessofrecordingandtrans cribingofchildrenconversationswasmuchmor etime-consumingthantheyhadantic-ipated[c itedin22]. formationWhateverdatacollectingtechnique isusedwhetherquestionnaireorinterviewtec hnique, , ,af terreachingtheintervie-wee, , 'filter'questions.(20 )explainsthat"itstartsoffwitha broadquestionandthenpro-gressivelynarrow sdownthescopeofftheques-tionsuntilit comestosomeveryspecificpoint.

7 "Hecontinuesthat,afilterquestionisusedto excludea respondentfroma particularquestionsequenceif ,manyopen-endedquestionswillproducea gooddealofinformationthatis~ respondent' (3:p .57)pointsoutthat"theaimofaninterviewist odrawfromanintervieweea rangeofinforma- ,andtomakeit explicit."Apartfromthat,interviewerisabl etoanswerquestionsconcemingthepurposeoft heinterview,andtheintervieweemaybeputati na ,ina .al(1962)pointout,"inter-viewersarehuman beingsandnotmachines"[citedin(18) )]. (20 )pointsoutthattheinterviewhasseveralsour cesofbiases."Theinterviewermaygiveaninkl ingofheropinionorexpectationsbyhertoneof voice,thewayshereadsquestions,orsimplyby herappearance, ,byprobingwithleadingquestions, ;shemayshowsur-priseorboredomintoneoremp hasisorinotherwaysunconsciouslycommunica teherownattitudesandherexpectationsofthe responden!' 'delicate'issuesraisespecialproblems."Th emainsourcesofbi asarethecharacteris-ticsoftheinterviewer andthecharacteristicsoftherespondents, ,reli-gion, ,whicharethecarefulformulationofquestion ssothatthemeaningisveryclear;thor-oughtr ainingproceduressothataninterviewerismor eawareofpossibleproblems;probabiIitysamp Iingofrespondents;andsometimesbymatching interviewercharacteristicswiththoseofthe samplebeinginterviewed(2).

8 Someoftheproblemsmentionedabovecanbeexte ndedbysuitableselectionandtrainingofinte rviewers,bycarefulpreparationoftheinter- viewschedulecontrol, , (19 )indicatesthat"itismoreof ,anony-mousquestionnairesgeneratea higherandbet-terqualityresponse,especial lyontheincomeandsecondjobquestion.".This isveryimpor-tantincaseofaskingquestionsa boutone' , [ ,Boulton(23 )tofacilitatehonestanswersandtocomplywit hethicalstandardsfroma sampleof8-10year-olds,informedthechildre nthat"..theirresponseswouldnotbecommunic atedtoanybodyfromtheirschooL."andheconti nuesby"Attheoutsetoftheinter-view,it waspointedoutthatit wasnota test,astherewerenorightorwronganswers,an dthatitwasacceptableforthechildrentosayw hatevertheywantedinresponsetothequestion s." ,forexamplenotchild'sownclassroom,should beobtainedforinterviewing(23). , ,it shouldbenecessarytohavea recordofa respondent'sper-sonalknowledge,theseshou ldbekeptentire1yseparatefromthedata,andt hereshouldbenowaythata (3 )pointsoutthat"Beyondthisminimum,everyre searchershouldsatisfyhimselfthatallperso nalinformation,suchasplaceofresidence,oc cupation,andthelike,whichcouldleadtoanid entification,andwhichis notabsolutelyessen-tialtothetopicofinqui ry,haseitherbeenremovedoreffectivelydisg uised.]

9 "Forexample,HolroydandHarlen(9 )haveassuredtheconfidentialityofthequest ionnairestothesam-plegroup."Teacherswere fullyinformedoftheaimsoftheresearchandth eprocedurethatwasadoptedforthereturnofco mpletedquestion-nairesensuredconfidentia lity." , , (8 )emphasizesthat"Theinterviewerneedstohav eanethicalframeworkfordealingwithsuchiss ues." ,theyusuallypro-ducelowerresponserate(19 )."Dependingonthesubjectofthequestionnai re,fortheuninter-estedgroups , figuresof40percentto60percentaretypical; eveninsurveysofinterestedgroups,80percen tisseldom"(20 ).AstudywascarriedoutbyKrysanandetaL.(24 ) , (25)haveconductedanexperimentalstudytode terminetherelativeeffectsofquestionnaire length,respondent-friendlyquestionnaired esign, "theshorteningquestionnaireandrespondent -friendlyquestionnairedesignimproverespo nserate,whereasaskingapotentiallyquestio nsand/orobjectionablequestions,..,lowers AComparisonofTwoDataCollectingMethods:ln terviewsandQuestionnairesresponse."(25 ).Interestingly,Scott(26) (Brown,1960andSletto,1940).

10 , ,theresponserateswererespective-ly68perc ent,60percent, ,therewasnoclearevidenceofa cleardeclineinresponseregardingthequesti onnairelength(citedin(26) )].Unfortunately, (26) 'sresearchin1996(9).Theyexplainthat"With thevaluablecooperationofthedirectorsofAA Pscienceproject, ,thatis, 'esponseratesareunusuallyhighfora questionnairesur-vey."(9 ). goodplanningofthequestionnairescanincrea setheresponserateforthequestionnaires, ,whichcanseriouslyreduceanyclaimstogener alizability(7 ). ,Koenig,andBecker(27)toexam-inetherelati onshipbetweenageandresponse7ratestomails urveys(mailquestionnaires).Theyworkedona randomsampleof1,000retiredpublicemployee sina (1978) "..age,withinelderlypopulation,doeshavea significantimpactonresponseratetoamailqu estionnaireMailsurveyscanbeadministereds uccessfullytoelderlypopulationsanda reasonableoverallresponseratemaybeobtain ed-butonlyifappropriateprecautionsaretak en."(27 ). "Highvaliditydemands highreliability, , "(26 ).Althoughtherearedifferenttypeofvalidit y,inferencesaboutthevalidityofinterviews aremadeonthebasisoffacevalidity,thatis, ,it 'svalidityontheoralproficiencyofaninterv iewee.