Mendeley使用手册-最新最全版

1Networksofreaderandcountrystatus:AnanalysisofMendeleyreaderstatisticsRobinHaunschild*,LutzBornmann**,&LoetLeydesdorff****MaxPlanckInstituteforSolidStateResearchHeisenbergstr.1,70569Stuttgart,Germany.Email:[e-mail]**DivisionforScienceandInnovationStudiesAdministrativeHeadquartersoftheMaxPlanckSocietyHofgartenstr.8,80539Munich,Germany.Email:[e-mail]***AmsterdamSchoolofCommunicationResearch(ASCoR)UniversityofAmsterdamPOBox157931001NGAmsterdam,TheNetherlands.Email:[e-mail]2AbstractThenumberofpaperspublishedinjournalsindexedbytheWebofSciencecorecollectionissteadilyincreasing.Inrecentyears,nearlytwomillionnewpaperswerepublishedeachyear;somewhatmorethanonemillionpaperswhenprimaryresearcharticlesareconsideredonly.Sophisticatedandcompactbibliometricmethodshavetobeappliedinordertoobtainanoverview.Onepopularmethodisanetwork-basedanalysis.Inthisstudy,weanalyzeMendeleyreadershipdataofasetof1,133,224articlesand64,960reviewswithpublicationyear2012togeneratethreenetworks:(1)Thenetworkbasedondisciplinaryaffiliationspointsoutsimilaritiesofanddifferencesinreadershipsofpapers.(2)Thestatusgroupnetworkshowswhichstatusgroups(e.g.students,lecturers,orprofessors)commonlyreadandbookmarkpapers.(3)Thecountrynetworkfocussesonglobalreadershippatterns:Itvisualizessimilaranddifferentreadingpatternsofpapersatthecountrylevel.Withthesenetworksweexploretheusefulnessofreadershipdatafornetworking.KeywordsMendeley,network,Pajek,VOSViewer,bibliometrics,altmetrics3IntroductionBibliometricsisnotonlyamatureresearchfield,whichdevelopsadvancedindicatorsforresearchevaluationpurposes,butalsoaresearchfield,whichstudiespatternsinscience.Thebestmethodforstudyingthesepatternsisbibliometricnetworkingorsciencemapping.Here,bibliometricdataareusedtogeneratenetworksofcitationrelations(e.g.betweenscholarlyjournals),networksofco-authorships(e.g.betweenhighly-citedresearchersininformationscience),ornetworksofco-occurrencerelationsbetweenkeywords,wordsinabstractsand/orwordsintitles(e.g.co-occurrencerelationsbetweenwordsinabstractsofpaperspublishedininformationscience)(vanEck&Waltman,2014).Powerfulcomputershaveledtotheanalysisoflargenetworks(whichmayincludethewholeWebofScience,WoS,ThomsonReuters,database)(Milojević,2014).Today,thesenetworksarenotonlyofinterestforspecialistsinbibliometricsornetworking,butalsoforstakeholdersfrompublishers,researchinstitutions,andfundingagencies.AccordingtoMartin,Nightingale,andRafols(2014)“networkandscience-mappingvisualisationshaveconsiderablyenhancedthecapacitytoconveycomplexinformationtousers.Thesetoolsarenowsufficientlymaturetobeusednotonlyavailableinacademiabutalsoinconsultancyandfundingorganisations”(p.4).Overviewsofpublicationsdealingwithnetworkingandmappinghavebeenpublished,forexample,byandBörner,Sanyal,andVespignani(2007)andLeydesdorff(2014).Sincerecentyears,altmetricshasdevelopedtoapopularresearchfieldinbibliometrics(Bornmann,2014).Altmetricscountsandanalyzesviews,downloads,clicks,notes,saves,tweets,shares,likes,recommends,tags,posts,trackbacks,discussions,bookmarks,andcommentstoscholarlypapers.Becauseitisnotclear,whatthesecountsreallymeasure,mostof4thestudiesinthisfieldhavecalculatedthecorrelationbetweenaltmetriccountsandcitationcounts(Bornmann,inpress).Asubstantialpositivecorrelationpointstoacertain,butotherwiseundefinedmeaningofaltmetricsinascientificcontext.Similartobibliometricdata,altmetricdatacannotonlybeusedforresearchevaluationpurposes,butalsofornetworkingorsciencemapping.Kraker,Schlögl,Jack,andLindstaedt(2014)presentedamethodologyandprototypeforcreatingknowledgedomainvisualizationsbasedonreadershipstatistics(fromMendeley).HaunschildandBornmann(2015)generatedareadershipnetworkwhichisbasedonMendeleyreadersper(sub-)disciplineforalargedatasetofbiomedicalpapers.Inthisstudy,weuseMendeleyreadershipdataforallpapers(articlesandreviewswhereaDOIwasavailable)from2012togeneratethreenetworks:(1)Thenetworkbasedondisciplinaryaffiliationswillpointoutsimilaritiesofanddifferencesinreadershipsofpapers.(2)Thestatusgroupnetworkwillshowwhichstatusgroups(e.g.students,lecturers,orprofessors)commonlyreadpapers(ornot).(3)Thecountrynetworkfocusesonglobalreadershippatterns:Itwillvisualizesimilaranddifferentreadingofpapersatthecountrylevel.Usingthesenetworks,weexploretheusefulnessofreadershipdatafornetworkinganalysis.MethodsDatasetusedBetweenthe11thand23rdofDecember2014,MendeleyreadershipstatisticsfornA=1,133,224articlesandnR=64,960reviewswereretrievedviatheApplicationProgrammingInterface(API),whichwasmadeavailablein2014,usingHTTPGETrequestsfromR().AnexampleRscriptisavailableat(MPG)basedontheWoSandadministeredbytheMaxPlanckDigitalLibrary(MPDL).TheDOIwasusedtoidentifythepapersintheMendeleyAPI.1,074,407articles(94.8%)and62,771reviews(96.6%)werefoundatMendeley.Intotal,werecorded9,352,424readercountsforarticlesand1,335,764readerc