知識(shí)產(chǎn)權(quán)調(diào)查設(shè)計(jì)指南

Guidelines

fordesigninganIPsurvey

Whycarryoutan

IPsurvey?

Asurveycanbeusefulforunderstanding

factorsbehindandbeyondaggregated

statistics.Surveysbasedonintellectual

property(IP)canthereforebeavaluable

toolinguidinginnovationandIPpolicies.

Thisshortguideoutlinesbestpracticesfor

designingIP-relatedsurveys,withtheaimof

promotingtheiradoptionbygovernments

andresearcherskeentounderstandthe

economicbehaviorofstakeholdersinthe

IPsystemanddesignpoliciestoassist

itsdevelopment.

1

2

roadmap

RoadmapfordesigninganIPsurvey

APlanningthesurvey

Startbyhavingaclearpictureofyourfinaloutputandthemainresearchquestion(s)youneedtoanswer.

BDesigningthesurvey

Usingdescriptivestatisticalanalysis,analyzeyourtarget

populationtobetterunderstandit,andusethisanalysistohelpyoudesignthesurvey.

CCollectingthedata

Draftastep-by-stepplanforthesurveyimplementationanddatacollectionstage,basedonatimeline.

DCheckingandanalyzingthedata

Assess,analyzeandvalidatetheaccuracyofthedatayouhavecollected.

GuidelinesfordesigninganIPsurvey

3

planningthesurvey

APlanningthesurvey

Thinkaboutthefinaloutputfirst.Thisisthemostimportant

step,anditwillhelpyouplanforthelaterstages:design,validity

checkingandcollecting.

Tounderstandtheoutput,considerthefollowing:

?Whatisthepurposeofthesurvey?

?Whatisthemainresearchquestion(s)?

?Surveyscantakealongtimeandusealotofresources.Iscarryingoutasurveyreallythebestwaytoanswerthe

researchquestion?

?Dependingonyourresearchquestion(s),youmightneed

tounderstandbothusersandnon-usersoftheIPsystem.

Whoarethetargetrespondentsofthesurvey?Thetarget

respondentmaydifferdependentoncountry,typeofcompany

orinstitution.Forexample,IPusersurveyscanbetargetedat:

?applicants(individuals,companies,universities,etc.)

?applications(patents,industrialdesigns,etc.)

?inventors/creators/designers,etc.

Smarttip

Whendealingwithmultiple

applicants/inventors/designersperapplication,useonlythe

firstapplicantforthesake

ofsimplicity.

GuidelinesfordesigninganIPsurvey

planningthesurvey

?Whatistheidealsurveymethodforyourresearchquestion(s)?Forexample:

?Panel–asequenceofdatapointscollectedatintervals,suitablefortrackingchangesovertime.

?Cross-sectional–datacollectedbyobservingmanysubjects

(suchasindividuals,firms,territoriesorregions)atonepointoroveralongerperiod.

?Mixeddesign–integrationofquantitativeandqualitativedatawithinasingleinvestigation.

Themethodyouchoosewillaffectthedatayoucancollect,aswellastheresourcesandtimerequired.

?Whatisthebesttimeframeforthesurvey,basedonyour

surveymethod?

?Forexample,patentapplicationsfiledbetween2017and2019.

Choosingaveryrecenttimeframemaymeancertaindataisnotreadytoberecorded.However,amorehistoricaltimeframe

couldresultininaccuraterecallbytherespondents.

?Whatisthebestcollection(distribution)methodforyourtarget

4

respondents?

?Forexample,online,email,post,face-to-face,etc.

Identifywhetheryouhavesufficientcontactinformationfor

thetargetrespondents,suchasemail,phonenumberorpostaladdress.Contactmethodsdifferintheiruseofresourcesand

theirresponserate.

GuidelinesfordesigninganIPsurvey

planningthesurvey

?Determinetheidealtimelineforeachstageofthesurveyandsetclearmilestones.Rushingthroughthestagescanberisky,butalongcollectionandanalysisprocesscanmeansurveyresultsbecomeirrelevant.Allowenoughtimeforeachstage,withsomecontingency.Youmightneedtosendreminders,forexample.

?Determinetheidealfinaloutput.Forexample:

?datasetforresearchers

?reportforpolicymakers.

?Havecontingencyplansinplace.Forexample:

?second-bestrespondents

?second-bestdistributionmethod.

Smarttip

Differentinfrastructureand

culturalpreferenceswillimpactthesuccessrateofthesurvey.Identifythebestcollection

methodsforyourlocalsetting.

5

GuidelinesfordesigninganIPsurvey

designingthesurvey

BDesigningthesurvey

Sampledesign

6

?Conductadescriptivestatisticalanalysisofyourtarget

populationtounderstanditsnuancesanddemographics,usingunit-recordinformation(acomprehensivecollectionofdata

relatingtoindividuals).Thiswillhelpyoutoidentifytheidealrespondentsandpotentialbiasesintheresults.

?Ifthetargetpopulationistoolargetosurveyinitsentirety,youwillneedtosurveyarepresentativesample.Thisisasubsetofapopulationthataccuratelyreflectsthecharacteristicsofthelargergroup.

?Forexample,apopulationof1,000(300non-residentsand700residents)couldgeneratearepresentativesampleof200(60non-residentsand140residents).

?Itmightbedifficulttoachieve,butaimforasamplethatmimicstheconsistencyanddiversityofthelargerpopulation.

?Thereareseveralpotentialbarrierstoachievinga

representativesample.Inthiscase,betransparentaboutthelimitationsofyoursamplingprocedure.Anunrepresentativesamplewillresultinbiasedresultsand,ultimately,biased

conclusions.

GuidelinesfordesigninganIPsurvey

designingthesurvey

Questionnairedesign

?ProvidecleardefinitionsofIPinstrumentsandanyother

technicaltermsupfront.

?Definetermssuchas“patent,”“filingdate,”etc.

?Designthequestionnaireaccordingtotheresearchquestion(s)andthedescriptivestatisticalanalysis.

?Includequestionsthatwillallowyoutovalidatetheresults(e.g.,non-responsebias).

?Ensurethelanguageyouuseisclearandconsistent.

?Arrangeyourquestionslogically.

?Groupquestionsbytopic.

?Startwiththemostimportantquestionsandtopics.

?Withineachtopic,startwithgeneralquestionsandgraduallymakethemmorespecific.

?Organizethesurveyinsuchawaythattherespondentisnotrequiredtorecallearlierquestions.

?Avoidlengthyquestionnairesasitreducestheresponserate.

7

GuidelinesfordesigninganIPsurvey

8

designingthesurvey

Questionwording

Althoughitisimpossibletoensurethateachquestionmeansexactlythesamethingtoalltherespondents,aimtominimizemisunderstandingasfollows:

?Besimple,coherentandconcrete.

?Avoidabstract,generalandambiguousquestions.

?Avoidaskingformorethanonepieceofinformationinasinglequestion.

?Minimizebiasesbyavoiding:

?leadingquestions

?sociallydesirableresponses

?jargonandname-dropping.

?Allowforallpossibleresponses,butensuretheyaremutuallyexclusive.

?Allowforratedresponseswherenecessary.

?Forexample,theLikertscalewithfivetosevenpossibleresponses.

GuidelinesfordesigninganIPsurvey

9

designingthesurvey

Pre-testing

?Pre-testingthequestionnaireonbothexpertsandnon-experts

iscrucialtohelpyouunderstandsourcesofconfusioninthe

textofthequestionsandchoices,andeliminateredundancy

andtypos.

?Pre-testthequestionnaireonexpertstogettechnicalfeedback.

?Forexample,targetrespondents,IPexperts,examiners.

?Pre-testthequestionnaireonnon-expertstogetfeedbackaboutitsclarityandeaseofuse.

?Conductinterviewswiththepre-testrespondentstogettheirfeedback.

?Forexample,askthemaboutthelengthofthequestionnaire,theclarityofthequestionsandtheirthoughtprocesstoreach

ananswer.

?Reviseandupdatethequestionnairebasedonthefeedbackyoureceive.

?Allowenoughtimeformultipleroundsofrevisions.

GuidelinesfordesigninganIPsurvey

10

collectingthedata

CCollectingthedata

?Draftastep-by-stepprocessforimplementation,andsetatimeline.

?Havethenecessarycontingencyplansinplace.Forexample,identifypotentialalternative(second-best)respondents.

?e.g.CEO,thedesigner,etc.

?Haveastrategyforreassuringrespondentsaboutanonymityandconfidentiality.

?Draftacoverletter:

?statingthepurposeofthesurvey

?includingendorsementsfromauthoritiesandinstitutions

?guaranteeingtheanonymityandconfidentialityofthedata.

?Prepareacentralizedandharmonizeddatarecordingstrategytominimizethepotentialforhumanerror.

?Traintheresearchersinthefieldtomaintainaunivocalapproachandminimizebiases.

GuidelinesfordesigninganIPsurvey

collectingthedata

?Pre-testthedatarecordingstrategytoensureitworks.

?Distributethequestionnairewithanaccompanyingcoverletter.

?Sendoutremindersandfollow-upsaccordingtotheagreedtimelineandmilestones.

?Ateachstage,keeptrackoftheresponserateandpotentialbiases.

?Identifysegmentsofyourpopulationwithalowresponse

rateandmobilizecontingencyplanstominimizeanypotential

biases.

?Forexample,consideralternativerespondents,orchangethecollectionmethod.

?Documentanychangeofstrategyduetounforeseenproblems.

?Recordandcollectthedata.

11

checkingthedata

GuidelinesfordesigninganIPsurvey

DCheckingandanalyzingthedata

?Assesswhetherthecollecteddataaccuratelyrepresentsthetargetsampleandlargerpopulation.

?Ifnecessary,treatandcleantherecordeddatatomakeadatasetthatiscompatiblewithstatisticalsoftware:

?Removeanyduplicates.

?Removeemptyorincompletequestionnaires(e.g.,basedonathreshold).

?Calculatetheresponserate.

?Conductadescriptivestatisticalanalysisonthecollecteddata.

?ThereisoftenadelayinpublicationofIPapplications.Whenanalyzingt