Environment

52 views

Problems With Indexing and Citation of Articles With Group Authorship

Problems With Indexing and Citation of Articles With Group Authorship
of 3
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Share
Tags
Transcript
   AUTHORSHIP ANDCONTRIBUTORSHIP Problems With Indexing and Citationof Articles With Group Authorship Kay Dickersin, PhDRoberta Scherer, PhDEunike Sri Tyas Suci, MA Michelle Gil-Montero G ROUP AUTHORSHIP  ( ALSO known as corporate or col-lectiveauthorship),thelist-ing of the name of a groupin place of the names of individual au-thors,hasbeenrecognizedsinceatleast1841. 1 Modified group authorship, inwhich individual names are listed fol-lowedbythenameofthegroup,isalsoused. Group authorship is most oftenused by investigators associated withstudies involving many investigators(eg, multicenter clinical trials or ge-nomics) because it allows investiga-tors to share credit equally. 2 PubMed is the National Library of Medicine’s(NLM’s)bibliographicdata-base,withcitationsgenerallydatingbackto 1966. The NLM has not includedgroupauthorsintheMEDLINEauthorfield; rather, the group name has beenincludedasanadd-ontothetitle.Asof April 2001, group authors will appearin a new, separate “collective name”field. In the case of modified groupauthorship, named authors have beenincluded in the author field and the re-search group name in the title field.ScienceCitationIndex(SCI)isasci-entific and biomedical bibliographicdatabase used to track citations to in-dividual articles. Instead of listing theresearchgroupintheauthorfieldinitssource listing, SCI lists either all indi-vidual names in the group or the writ-ing committee members’ names in theorder they are listed in the article (eg,thefirstauthorlistedcouldbethefirstinvestigator name in an alphabeticallyordered list). Weexamineherehowreportsofcon-trolledtrialswithgroupauthorshipareindexed in PubMed and SCI and howcitations are listed in SCI. METHODS  Weidentifiedallcontrolledtrialsfundedby the National Eye Institute (NEI) aslistedontheirWebsiteinJanuary2000. We chose to study randomized clinicaltrials because they are associated witha strong tradition of group authorship. We selected NEI because it has a long-standingWeb-basedregisterofitsclini-cal trials and associated articles thatwould provide us with an unbiasedsample for comparisons of group andnamed authorship.PublicationsassociatedwitheachtrialwereidentifiedusingPubMed(searchesconductedbetweenJanuaryandAugust2000).Weexcludedallletters,editori-als,andabstracts.Identifiedrecordsweredownloaded directly into ProCite, pa-percopiesretrieved,andallsourcedata(eg,authorship,articletitle)enteredex-actly as listed on the publication.ArticlesarelistedinSCIin2ways:firstasa source article,whichcontainsthecita-tion,abstract,andreferencescitedinthatarticle,aswellasthenumberoftimesthatarticlewascitedinotherpublications;andsecond as a  cited reference  (all cases inwhich the article was cited as a refer-ence in other publications). The SCI Author Affiliations:  Departments of CommunityHealth (Dr Dickersin and Ms Gil-Montero) and Soci-ology(MsSuci),BrownUniversity,Providence,RI;andDepartment of Epidemiology and Preventive Medi-cine, University of Maryland, Baltimore (Dr Scherer). Corresponding Author and Reprints:  Kay Dickersin,PhD,DepartmentofCommunityHealth,BrownUni-versity,169AngellSt,BoxG-S2,Providence,RI02912(e-mail: kay_dickersin@brown.edu). Context  It is not known whether articles with group authorship (ie, with a researchgroup name listed as the author) are difficult to identify or whether use of group au-thorship may lead to problems with citation. Methods  To examine ways in which reports of controlled trials with group author-shipareindexedandcitationscountedinbibliographicdatabases,weconductedacross-sectional study in January 2000. We identified 47 controlled trials funded by the Na-tional Eye Institute and 285 associated articles. Between January and August 2000, wesearched PubMed and Science Citation Index (SCI) and recorded the citation practicesfor these articles. Our main outcome measures were ways in which trial reports werelistedinPubMedandSCIandnumberofcitationstoeachreportbytypeofauthorship. Results  Ofthe285publishedreportsidentified,126(44%)hadgroupauthorship,109(38%) had modified group authorship (listing individual names plus the name of theresearchgroup),and50(18%)hadnamedauthorsonly.InPubMed,nogroupauthorswerelistedintheauthorfield(perMEDLINErules);inSCI,group-authoredreportsgen-erally were incorrectly attributed (first name on investigator list [35.3%], first name onwriting committee [25.5%], contact name [16.7%], anonymous [16.7%], and other [5.9%]). Using the SCI general search, we identified citations to 16.7% of group-authoredreports,comparedwithcitationsto96.9%ofreportswithmodifiedgroupau-thorship and 93.9% of citations to reports with named authors only. Other systematicsearchmethodsfoundthatmorethan98%ofgroup-authoredreportsactuallyhadbeencited and that group-authored reports were cited more than other reports. Conclusions  Indexing systems are not optimally adapted to group authorship. Werecommend that indexing services change their practices to include group authors inthe author field to help correct the problem.  JAMA. 2002;287:2772-2774  www.jama.com 2772  JAMA,  June 5, 2002—Vol 287, No. 21  (Reprinted)  ©2002 American Medical Association. All rights reserved. Downloaded From: http://jama.jamanetwork.com/ on 02/25/2013  searchoptionsforfindingarticlesinthese2 ways are the general and cited refer-encesearches,respectively.WesearchedSCIduringtheperiodJanuarytoAugust2000;thedatabaseincludedarticlespub-lishedfrom1987onward.Toidentifytheactual SCI source record for each pub-lication,weusedtheSCIgeneralsearchoptionwithvariouskeywordsandauthornames until each record was found. We next searched for citations toeach publication (citation record) inSCI, first using the general search op-tion.AllsourcerecordsincludedinSCIare linked to all cases in which a pub-lication has been cited exactly in theway it is listed in the SCI source data-base. Thus, if a record is listed in theSCI source database incorrectly, a ci-tation to that record would have be toincorrectinexactlythesamewaytobecounted and linked. We also used the cited referencesearch to count the number of cita-tions to the articles in our sample. Asingle article may be cited many differ-entways,sowelookedforcitationsthatused named authors, when they ex-isted;thefullgroupname;anabbrevia-tionofthegroupname,usingSCIrulesaswellastheabbreviationsusedbytheresearch study group (eg, ISCH OPTNEURDECaswellasIONDTfortheIs-chemicOpticNeuropathyDecompres-sionTrial);thefirst4lettersofeachtermin the research group name; the firstnameonthewritingcommittee;andthefirst name on the list of investigators.Since it would be necessary for thesearchertoguessatthepotentialincor-rectspellingsandtypographicalerrorstocapture all citations to any article, ourstandardized citation search strategycould be an underestimate of the truenumber of citations. To obtain a betterestimate of the actual number of cita-tions,wealsoperformedanadhocsearchfor citations to 3 publications emanat-ing from the IONDT. We used the textsearch terms  ischemic optic neuropathy and opticnervedecompression ,aswellasthe authors’ names for all 24 IONDTclinical center directors and comparedtheresultsofthisadhocsearchwithourstandardized study searches. RESULTS  We identified 47 NEI-funded con-trolled trials associated with 306 re-ports,ofwhich285articlesareindexedin PubMed. All 126 articles with groupauthorship were included in PubMed,but no author was listed in the authorfield,perNLMrules.All109articleswithmodified group authorship had thenamedauthorslistedintheauthorfield,butthenameoftheresearchgroupwasnot listed there. All 50 articles withnamed authors had the named authorsincludedintheauthorfield.Eighty-onepercent of the articles listed in PubMed(232/285) were included in the SCIsourcedatabase.Excludedarticleswereall published before 1987.Using the general search, we identi-fiedeacharticleandrecordedthewayitwas listed in the SCI source index.Articles with group authors (n=102)wereneverlistedashavingagroupauthor(per SCI guidelines): 35.3% listed theinvestigatornames(sometimesnumber-ing in the hundreds); 25.5% listed thenamesofthewritingcommittee;16.7%listed the contact person named in thepublication; 16.7% listed “anony-mous”;and5.9%listedsomeotherindi-vidual(s). Articles with modified groupauthorship(n=97)werelistedusingtheauthors’namesastheyappearedonthepublication 99% of the time. In 1 case,thesameauthorswerelistedbuttheorderwasdifferentfromthatontheactualpub-lication. Articles with named authors(n=33) were listed 100% of the timeusingthesamefirstauthorwhoappearedon the srcinal publication.Foreacharticlelocatedinthesourcedatabase (ie, using the general search),weenumeratedthenumberofcitationslisted: articles with named and modi-fied group authorship were almost al-waysrecordedasbeingcitedatleastonce(31/33[93.9%]and94/97[96.9%]),re-spectively. Articles with group author-ship,however,werelistedasbeingcitedonly16.7%ofthetime(17/102).Thisisnotsurprising,giventhatauthorshipwasnever listed correctly in the source da-tabaseforarticleswithgroupauthors,soonlyincorrectcitationswouldhavebeencounted. When we searched for citations us-ingthecitedreferencesearch,wefoundthat98.1%ofthearticleswithgroupau-thorship (100/102) had actually beencitedatleastonce.Citationsweremadeusing the actual group author (37/102articles;117citations);theactualgroupauthorwiththeexceptionthatthecita-tion was abbreviated per SCI rules (99/ 102; 2671 citations); the first name onthewritingcommittee(10/102;19cita-tions); the first name on the investiga-tor list (6/102; 18 citations); or anotherway(7/102;15citations).Somearticleswerecitedinmorethan1wayandwerecounted more than once. Articles withmodified group authorship were citedmost of the time using the named au-thors(94/97articles;1790citations)butsometimesbyusingtheactualorabbre-viated group name (19/97; 54 cita-tions)oranotherway(1/97;5citations).Articleswithnamedauthorswerecitedusing the named author exclusively. When we considered all known ci-tationstoanindividualarticle,wefoundthatalmostallcitationstoarticleswithnamed and modified group authorswereidentifiedusingthegeneralsearch( F IGURE 1 ).Onlyasmallproportionof citationstoarticleswithgroupauthor-ship were identified using the generalsearch, however, and almost all werefound using the search algorithm ap-plied to the cited reference search. Figure 1.  Proportion of Citations FoundUsing General Search vs Cited ReferenceSearch 100604080200No. of CitationsNo. of Articles    A   l   l   A  r   t   i  c   l  e  s   F  o  u  n   d ,   % 28401021849 9778733Citations Found in General Search Additional Citations Found in Reference SearchNamed AuthorsOnly496ModifiedGroup Authorship496Group Authorship982 GROUP AUTHORSHIP AND ARTICLE CITATION ©2002 American Medical Association. All rights reserved.  (Reprinted) JAMA,  June 5, 2002—Vol 287, No. 21  2773 Downloaded From: http://jama.jamanetwork.com/ on 02/25/2013  Using information from the generalsearch, it appeared as if articles withgroupauthorswerenotcitedasoftenasother articles ( F IGURE  2 A). The meannumber of citations per report was 2.7forarticleswithgroupauthors,18.9forarticleswithmodifiedgroupauthors,and24.5forarticleswithnamedauthors.Us-ing the cited reference search, a differ-ent sort of picture emerged. On aver-age, there were 10 times as manycitationsperarticleaswehadfoundus-ing the general search for group-authoredarticles(27.8citesperarticle)butaboutthesamemeannumberforar-ticleswithmodifiedgroupauthors(19.1cites per article) and with named au-thors(23.8citesperarticle)(Figure2B).Thus,group-authoredarticlesthatwerecited had a higher mean number of ci-tations than articles with other types of authorship.Most likely, even these results rep-resent an underestimate of the num-berofcitationstogroup-authoredpub-lications. When we completed an adhoc cited reference search on the 3group-authored publications of theIONDT, 3-5 wefound22previouslyuni-dentified citations (68 total citations).None of the 22citationshad anyentrywhatsoever in the author field. COMMENT Both NLM and SCI indexing policiesmake searching for group authors dif-ficult for users, since most people areprobablyunawareofindexingrulesandexpect authors of all types to be listedintheauthorfield.Thesepoliciescounterthe effort authors have made to focusattention away from individual con-tributorsandtowardtheproductsofsuc-cessfulcollaboration.Inaddition,thefacttheneitherPubMednorSCIrecognizesgroupauthornamesintheauthorfieldofthedatabaseimpliesalesserstatusof these articles. University promotionscommittees do not always understandgroup authorship, and this problem iscompounded when PubMed and otherdatabases fail to acknowledge groupauthorship as a standard option. It isespecially disappointing for an investi-gatortofindacitationincorrectlylistedunderthenameofacoinvestigatorwhomayhaveearnedfirstauthorpositioninthe citation purely by his name’s placein the alphabet or even by being listedin the acknowledgments. 6 Failure to recognize group author-ship has also meant that there may bealargeundercountofcitationstothesearticles. 7 For example, using multiplemethodsofidentification,wefoundal-most 70 citations to 3 group-authoredarticlesfromtheIONDT.UsingtheSCIgeneralsearch,thetypethatwouldtypi-cally be performed by the averagesearcher,wefoundzero.Giventhefactthatthecostofthistrialwasinthemil-lions of dollars, this finding should bedisconcerting to the funding agency.Changes are needed immediately inbibliographicindexingsystemssothatappropriate credit can be accorded tothose electing group authorship. Ad-dition of a new field such as PubMed’s“collectivename”isnotsufficient.Wespecifically recommend that indexingservices change their practices to in-cludegroupauthorsintheauthorfield. Author Contributions:  Study concept and design: Dickersin, Scherer. Acquisition of data:  Scherer, Gil-Montero. Analysis and interpretation of data:  Dickersin,Scherer, Suci, Gil-Montero. Drafting of the manuscript:  Dickersin. Criticalrevisionofthemanuscriptforimportantintel-lectual content:  Dickersin, Scherer, Suci, Gil-Montero.  Statistical expertise:  Dickersin, Suci. Obtained funding:  Dickersin. Administrative, technical, or material support: Dickersin, Scherer.  Study supervision:  Dickersin, Scherer. Funding/Support:  This research was funded by Na-tional Eye Institute grant U10 EY09608. Acknowledgment:  We thank Richard Noble of theBrownUniversityLibraryforhishelpinidentifyingsourceson the history of group or corporate authorship. REFERENCES1.  TaitJA. AuthorsandTitles:AnAnalyticalStudyof the Author Concept in Codes of Cataloguing Rulesin the ENGLISH Language, From That of the British Museum in 1841, to the Anglo-American Catalogu-ingRules,1967  .Hamden,Conn:ArchonBooks;1969. 2.  Meinert CL. In defense of the corporate author for multicenter trials.  Control Clin Trials . 1993;14:255-260. 3.  IschemicOpticNeuropathyDecompressionTrialRe-searchGroup.Opticnervedecompressionsurgeryfor nonarteritic anterior ischemic optic neuropathy(NAION) is not effective and may be harmful: resultsoftheIschemicOpticNeuropathyDecompressionTrial.  JAMA . 1995;273:625-632. 4.  Ischemic Optic Neuropathy Decompression TrialResearch Group. The clinical profile of nonarteriticanterior ischemic optic neuropathy (NAION): experi-ence of the Ischemic Optic Neuropathy Decompres-sion Trial.  Arch Ophthalmol . 1996;114:1366-1374. 5.  IschemicOpticNeuropathyDecompressionTrialRe-search Group. The Ischemic Optic Neuropathy De-compressionTrial(IONDT):designandmethods. Con-trol Clin Trials . 1998;9:276-296. 6.  NewallML.Multicentrestudiesshouldbecitedbystudy name.  BMJ . 1995;311:632-633. 7.  Errorsincitationstatistics[editorial]. Nature. 2002;415:101. Figure 2.  Science Citation Index Search Results 01-1011-2021-3031-4041-5051-100  > 100100604080200 No. of Times Cited    A  r   t   i  c   l  e  s   C   i   t  e   d ,   % Science Citation Index General Search  A  01-1011-2021-3031-4041-5051-100  > 100 No. of Times Cited Science Citation Index Cited Reference Search B Group Authorship (n   =   102)Modified Group Authorship (n   =   97)Named (n   =   33) GROUP AUTHORSHIP AND ARTICLE CITATION 2774  JAMA,  June 5, 2002—Vol 287, No. 21  (Reprinted)  ©2002 American Medical Association. All rights reserved. Downloaded From: http://jama.jamanetwork.com/ on 02/25/2013
Advertisement
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks