调查模式对数据质量的影响：尼日利亚的实验证据.docxVIP

TheEfectofSurveyModeonDataQuality

ExperimentalEvidencefromNigeria

y

2

1Introduction

High-qualitysocioeconomicdataisfundamentalforadvancingresearchandforguiding,targeting,andmonitoringhumanitariananddevelopmentinterventions.Inlow-andmiddle-incomecountries(LMICs),surveysremainacriticalsourceofrepresentativedataandakeytoolfor“groundtruthing”socioeconomicoutcomes.Inrecentyears,phonesurveyshavebecomeanincreasinglyimportantmodeofdatacollectioninthesesettings,particularlyfollowingtheCOVID-19pandemic,whichtemporarilyforcedashiftfromin-persontoremoteinterviews(Gourlayetal.,2021a;Zezzaetal.,2023).Theirappealisclear:phonesurveysarelesscostlythantraditionalface-to-facesurveysandcanbeimplementedquicklyandflexiblytoprovidetimelyandhigher-frequencydata.Assuch,phonesurveysofferapragmaticresponsetotwoconvergingpressures.Ontheonehand,thegrowingnumberandfrequencyofcrisesandshocksinLMICshasincreaseddemandfortimely,high-frequencydatatoinformrapidresponseandtoanalyzehouseholdresilienceandrecovery(HeadeyandBarrett,2015).Ontheotherhand,constrainedbudgetsinbothresearchandinternationaldevelopmenthavereinforcedthepushforcost-effectiveapproachestodatacollection.Againstthisbackdrop,phonesurveyshavebecomeaprominentandlikelyenduringfeatureofthedatalandscapeinLMICs.

Yet,asphonesurveyshaveproliferated,concernsabouttheirdataqualityremain.Likealldatasources,phone(andin-person)surveysaresubjecttoerrorsfrommultiplesources,butthesurveymodeitselfisincreasinglyrecognizedasanimportantsourceofmeasurementerrorandbias.Despitegrowingevidenceonphonesurveydataquality,importantgapsremaininunderstandingthemagnitude,direction,anddriversofmodeeffects,systematicdifferencesinresponsesattributabletothedatacollectionmoderather