The publishing delay in scholarly peer‐reviewed journals · issue. Most electronic open access...
Transcript of The publishing delay in scholarly peer‐reviewed journals · issue. Most electronic open access...
1
Thepublishingdelayinscholarlypeer‐reviewedjournalsBo‐ChristerBjörkProfessor,InformationSystemsScienceHankenSchoolofEconomicsP.B.47900101Helsinki,Finlandbo‐[email protected](CorrespondingAuthor)Professor,DepartmentofMedicineandOMERADA‐202EFeeHallMichiganStateUniversityELansing,[email protected]+1517339‐0720
Abstract: Publishinginscholarlypeerreviewedjournalsusuallyentailslongdelaysfromsubmissiontopublication.Inpartthisisduetothelengthofthepeerreviewprocessandinpartbecauseofthedominatingtraditionofpublicationinissues,earlieranecessityofpaper‐basedpublishing,whichcreatesbacklogsofmanuscriptswaitinginline.Thedelaysslowthedisseminationofscholarshipandcanprovideasignificantburdenontheacademiccareersofauthors.Usingastratifiedrandomsamplewestudiedaveragepublishingdelaysin2700paperspublishedin135journalssampledfromtheScopuscitationindex.Theshortestoveralldelaysoccurinsciencetechnologyandmedical(STM)fieldsandthelongestinsocialscience,arts/humanitiesandbusiness/economics.Business/economicswithadelayof18monthstooktwiceaslongaschemistrywitha9monthaveragedelay.Analysisofthevarianceindicatedthatbyfarthelargestamountofvarianceinthetimebetweensubmissionandacceptancewasamongarticleswithinajournalascomparedwithjournals,disciplinesorthesizeofthejournal.Forthetimebetweenacceptanceandpublicationmostofthevariationindelaycanbeaccountedforbydifferencesbetweenspecificjournals.Keywords:ScholarlyPublishing;ReviewTime;ProcessingTime
2
1 Introduction ScholarlyjournalpublishinghasalonghistorygoingbacktoHenryOldenburg’sPhilosophicalTransactionoftheRoyalSocietyfoundedin1665.Forthepasttwocenturiesthevolumeofpeerreviewedarticlespublishedperyearhasincreasedbyarelativesteady3,5%peryear,withacurrentnumberofarticlesofaround1,8–1,9million,publishedinanestimated28’000journals(WareandMabe2012).Overtheyearsthescientificjournalasaninstitutionhasevolvedinmanywaysandafterthesecondworldwarandtheensuingrapidgrowthinsciencecommercialpublishershaveincreasinglyenteredthismarket,whichearlierwasdominatedbyscientificsocieties.Thedisseminationmediumhasveryrapidlychangedfromprintedissuestopredominantlydigitallydistributedpublishing(VanOrsdelandBorn,2002).Atthesametimethishastriggeredtheemergenceofnewbusinessmodelsfordigitalpublishing,includingbundlede‐licenses,pay‐per‐viewandopenaccesspublishing.ScholarlyjournalpublishinginitscurrentformhasbeentheobjectofincreasedcritiquesincetheadventoftheWorldWideWebandtheopportunitiesitoffersforprocessinnovation,Thedebatehasinparticularconcernedthreeaspects.Firstlythatthereachofthedisseminationthatthetraditionalsubscriptionmodelachievesissuboptimal.Secondlythatthepeerreviewprocessisflawedandfrequentlyleadstoarbitrarydecisions.Thirdlythattherearesignificantdelaysinpublishingarticles.Traditionalpaperpublishinginparticularcreatessignificantdelaysbothduetotheneedtobundlearticlesintoissuesandbacklogscreatedbypagelimitsresultingfromthehighperpagecostofthistypeofpublishing.ThesolutionproposedtothelimiteddisseminationisOpenAccess(OA),whichcanbeachievedeitherthroughpublishinginopenaccessjournals(“goldOA”)orthroughauthor’suploadingmanuscriptversionsoftheirarticles(“greenOA”)tosubjectorinstitutionalrepositories(Suber2012).OAjournalshaveincreasedtheiroutputby20‐30%peryearforoveradecadeandnowpublisharound12%ofallpeerreviewedarticles(LaaksoandBjörk2012).Theopenaccessibilitycanbeachievedviaanumberofbusinessmodelsofwhichthepublishingfeevariantisrapidlyincreasingitsmarketshare.Thecritiqueofthepeerreviewprocesshasledtoanumberofexperimentswithalternativemodels.Thewebmediumlendsitselftodifferentformsofopenreview,wheremanuscriptscanbe“published”priortorevieworwithminimalreviewandsubsequentlyevaluatedbyreadercommentsandelevatedtofullarticlestatusviapostpublicationfeedback.(Björk2011).Openreviewwastriedanddeemedafailureinawell‐knownexperimentbyNature(2006).MoresuccessfulthanopenreviewexperimentsisanalternativepeerreviewmodelpracticedbyanincreasingnumberofOA“megajournals”inthewakeofPLoSONE,whichcurrentlypublishesaround20,000articlesperyear.Inthisformofpeerreviewonlythescientificvalidityoftheresultsischecked,thedecisionconcerningthepotentialcontributionisleftforthereaderstodecide.
3
AnimportantreasonforthesuccessofPLoSONEisalsothatisoffersaveryattractivealternativetoauthorswhoaretiredofthelongdelaysinvolvedinpublishingintraditionaljournalsandrejectiononwhatarefelttobearbitraryandorbiasedopinionsofreviewersand/oreditor.Thedelaywasanecessaryfacetofthepublishingprocesspriortotheturnofthemillennium,whenjournalswerealmostexclusivelypublishedinpaperform,andwherejournalpagelimitswereaneconomicnecessity.Sincethenelectroniconlyjournalshaveshownthatthedelaycanbeconsiderablyshortened.Alsothetraditionaljournalshaveacknowledgedtheexistenceoftheproblembystartingtopost“inpress”orcompletelycopyeditedandformatted“aheadofprint”versionsofacceptedmanuscriptsevenbeforetheybecomepartofanissueandreceivepagenumbers.Arecentsurveywithauthorsshowedthatthespeedofpublicationwasthethirdmostimportantfactoraffectingauthors’choiceofjournal,aftertopicalfitandthequalityofthejournal(SolomonandBjörk2012).Insomefieldsofscienceauthorshavetriedtopartlybypassthesystembypublishingtheirmanuscriptsinopenwebrepositoriespriortosubmissionasworkingpapers(economics)orpreprints(physics),inordertospeedupthedisseminationoftheresults.Inothercasesexperimentshavebeenmadewithnewtypesofpeerreviewjournals,inwhichonlylightlyscreenedmanuscriptshavebeenopenlypublishedonthejournalwebsites,andthebetteroneshavelaterbeenelevatedtofulljournalarticlestatus(Björk2011),provingthesealofquality.Itisourbeliefthatthelengthofthedelayisnotconstantacrossdifferentfieldsofscience,butdependsonthereviewandpublishingculturesthathaveevolvedindifferentsciences.Forexampleadelayoftwoyears,commonineconomicsandmanagement,wouldbedifficulttoacceptforacademicsinthebiomedicalsciences.
1.1 The Life‐cycle stages of a peer‐reviewed article Duringitslife‐cycleascholarlyarticleundergoesanumberofstages,someofwhichareinfocusinthisstudy.Duringthewritingandfinalizingofamanuscriptmostauthorstendtoshowittoafewtrustedcolleagues,fromwhomtheyreceivefeedbackandsuggestionsforimprovement.Inmanydisciplinesit’salsocommontopublishversionsasconferencepapersandinafewdisciplines,inparticularphysicsandeconomics,atraditionofpublishingworkingpapershasevolved.Atsomestagetheauthor(orauthors)formallysubmitsthemanuscripttoaparticularjournal.Mostjournalsrequirethatamanuscripthasn’tbeenpublishedelsewhereandthatisnotunderconsiderationforpublishingbyanotherjournal.Inmedicinethisrulecanbeevenstricterinthatauthorsarealsorestrictedfromdiscussingtheresultswiththepopularmedia,theso‐calledInglefingerrule.Fromtheviewpointofthewholescholarlycommunitytheruleexcludingparallelsubmissionisunderstandableintermsofavoidingunnecessaryreplicationoftheunpaidrefereeworkdonebytheeditorandother
4
scholars.Ontheotherhandthiscausespublishingdelaysforauthorswhoseworkisrejectedinthefirstandevensecondjournaltowhichtheysubmit.Thequalityandextentofthepeerreviewthatamanuscriptundergoesvariesconsiderablyacrossjournalsanddisciplines.Theeditorsofmanyjournalsscreensubmissionsandquicklyrejectmanuscriptsthatareclearlyunsuitablewithoutsendingthemoutforexternalpeerreview.Thereviewprocesscanalsoinvolveseveralcyclesofreviewandrevision,apracticecommoninmoreselectivejournalsparticularlyinspecificdisciplinessuchasbusinessandmanagement.Manuscriptsatsomepointareaccepted,rejectedorinsomecaseswithdrawnbytheauthorwhomayfindtherequestedrevisionsortherevisionprocessunacceptable.Ifacceptedmanuscriptsaregenerallycopyeditedandtypesetbythepublisherorcontractor,afterwhichtheauthorisusuallyaskedtocheckthefinalversion.Intraditionalprintpublishingthefinalizedmanuscriptisthenputinthequeueforpublishing,awaitingitsturn,usuallythoughnotalwaysaccordingtoitspositioninthequeue.Articlessubmittedtoaspecialissuearetreatedabitdifferently.Thequeuingcantakeaslongasayearormoreifthejournalhasasignificantback‐log.Ifthejournalalsopublishesanelectronicversionmanuscriptsareoftenpublishedearlieronthejournalwebsiteunderheadingslike“in‐press”usuallywithoutexactpagenumbersandassignmentofissue.Mostelectronicopenaccessjournalspublisharticlesdirectlywhentheyarereadyratherthaninissues,thusspeedinguptheprocess.Ifwewouldtakeamanuscriptandnotjournal‐centricviewthetotaldelaywouldoftenbeevenlongersincemanymanuscriptsarerejected,andinsomecasesseveraltimesbeforepublication.Thistimefromsubmissiontorejection,insomecasesfrommultiplejournals,needstobeaddedtothedelayofthejournalthatfinallypublishesthearticle.Azar(2004)discussesthisforthecaseofeconomicsjournalsandpointsouttheimportanceoffirst‐responsedelays,sinceitisoftenatthisstagethatauthorsneedtofindalternativejournalsforsubmittingtheirmanuscripts.Inthisstudywetakethejournal‐centricviewlookinginparticularatthedelayfromsubmissiontoacceptanceandthedelayfromacceptancetofinalpublication,aswellasthetotaldelaytime.Althoughitmightbepossibletogetdataforotherstagesintheoverallprocessforsomejournalsthesethreepointsintimearecommonforallpeerreviewedjournals.
1.2 Previous Research Thereareanumberofpossiblesourcesofinformationaboutpublicationdelays.Ideallypublisherswouldtrackandmakethisdataavailable.Thisishoweverrare,perhapsbecausepublishersandeditorsmaybehesitanttodiscloselongdelays.Sometimestheinformationcanbefoundineditorialsinjournals,whichoftenalsoprovideinformationabouttheacceptanceratesofjournals.Anotheroptionistogatherarticledataaboutsubmissionandacceptancedateswhichisoftenpublishedindividuallyineacharticleoronthearticles’facepageonthe
5
publisher’swebsite.Thisisaverylabor‐intensiveprocessbutprovidesprecisestatisticsforthearticlessampled.Afinaloptionistogatherthedatafromauthorswhichwouldbedifficultandlikelytobefairlyinaccurate.Earlierstudieshavemostlycollectedthedataincludedinpublishedarticles.OneofthefewstudiesusingstatisticssolicitedfrompublisherswastheearlystudyofeconomicsjournalbyYohe(1980),whoobtainedstatisticsfromtheeditorsof20journalsandextractedarticleleveldatafor5more.Trivedi(1993)foundthattheaveragetotalpublicationdelayforeconometricsarticlesinsevenstudiedjournalswas22.8months,consistingof13.4monthsfromsubmissiontoacceptanceand9.4monthsfromacceptancetopublication.Ellison(2002)concentratedhisstudyonthereviewtimesonly(submissiontoacceptance)andfoundanaverageof16.5monthsin1999foraselectionof25journalsineconomicsandrelatedfields.HewasalsoabletodoalongitudinalanalysisforasubsetofthejournalsusingdatabothfromYohe(1980)andCoeandWeinstock(1967)andfoundthatthereviewtimeshadmorethandoubledinthreedecades(1970‐1999),forfiveleadingeconomicsjournalsfrom8.7to20.7months.Themainreasonforthisseemstobetheincreasingnumberofiterativeroundsinthereviewprocess.Healsofoundthattheaveragereviewtimesvarybetweendifferentsub‐specialtiesofeconomics,evenforarticlespublishedinthesamejournalswithbroaderscopes,andsuggestthattheexpectationsforthetypeandlengthofthereviewshavebeensociallyshapedwithinnarrowscholarlycommunities.AlsoHartmann(1997)reportsonadramaticincreaseinsubmissiontopublicationdelays.ForarticlesintheJournalofAtmosphericSciencesthetotaltimeincreasedfrom5.9to15.2monthsbetween1970and1997andwhiletheacceptancetopublicationlagincreasedsomewhat(4.4to6.6months)theincreasewasmainlyattributabletotheincreaseinthetimerequiredbythereviewprocess(1.5to8.5months).KlingandSwygart‐Hobaugh(2002)comparedtheevolutionofpublicationdelaysforthreenaturalscienceandthreesocialsciencejournalsbetween1970/1980and2000,inanattempttoseeiftheemailcommunicationwidelyinusein2000hadreducedaveragedelays.Theyfoundthatthedelaysinchemistryandphysicsjournalshaddecreasedfrom6.5monthsto5.8(andevenmoresoforaminorityofarticlespublishedelectronicallybeforepaperpublication)butthatthedelaysintheeconomics,managementandpsychologyjournalshadincreasedfrom9.0to23.8months.Diospatonyietal(2001)studiedtheevolutionofpublicationdelaysintenchemistryjournalsintheperiod1985‐1999,andcouldnotfindanycleardevelopmenttoshorterorlongerperiods,withtheyearlyaveragesrangingbetween6.7and7.5months.Thepapercontainsdetailedbreakdownsofthespreadofdelaywithinjournalsaswellasananalysisofthebreakdownbetweensubmissiontoacceptancevsacceptancetopublication.
6
Carroll(2001)comparedpublicationdelaysforsixstatisticsjournalsandfoundaslightdecreasefrom25.2monthsin1994to22.3in1999.Hesuggestthatthedeclinemightbeduetoelectronicpublishingbecomingmorecommoninthefiveyearinterval.Amat(2008)studied14journalsinfoodscienceandfoundanaveragepublicationdelayof11.8months(forarangeof6.2‐17.2months).ThedelaysofthreecivilengineeringjournalsreportedbyBjörkandTurk(2006)variedbetween6.7months(foranOAjournal)comparedto18.0and18.9fortwoconventionaljournals.ThestudybyLuwelandMoed(1998)differedfromtheabovebecauseitincludedjournalsfromdifferentsubjectareas.ThestudywastriggeredbyclaimsofDutchresearchersthatarticlesintechnicalsciencesandmathematicshavemuchlongerdelaysthanarticlesinphysicsandchemistry,andthatresearchersintheformerfieldsaredisadvantagedinshorttermbibliometriccomparisons,oftenusedwhencomparingcandidatesforpromotionetc.Inaselectionof15leadinginternationaljournalsintheabovefields,therangeofdelayswasbetween2.5and17.5monthswithmathematicsandengineeringjournalstendingtobetowardsthehigherend.Anotherstudywithjournalsfromdifferentdisciplineswasthestudyof26IranianjournalspublishinginthePersianlanguage(Khosrowjerdietal2011).Thedelayrangeforthesepredominantlysocialscienceandhumanitiesjournalswasverywide(5.8to34.6months)withanaverageof17.3months.ThestudybyDongetal(2006)istheonlystudythattriedtoanalyseifthedelaytimesforOAjournalsdifferfromsubscriptionjournalsinbiomedicine.TheycomparedsixOAjournalsfromtheleadingOApublisherBioMedCentral(BMC)withsixjournalsoncorrespondingtopicsfromNaturePublishingGroup(NPG)aswellassixotherBMCjournalswithelevensocietyjournals.TheresultsdemonstratedthattheNPGjournalswereequaltotheBMCjournalsinoverallpublicationdelay(4.5months)butmarginallyfasteriftheelectronicpublicationdateswerecompared.TheBMCjournalsclearlyoutperformedthesocietyjournals(4.8vs8.9months).Itisnoteworthythattheinthesubscriptionjournalstheprintversionstrailedtheelectronicversionsbyonlyshortperiodsofbetween0.5to1.5months.Yuetal(2004),aspartofthebuildingofamathematicalmodelofthedelayprocess,collecteddelaydataforsevenjournals.Scientometrics,aninformationsciencejournalhadadelayof5.5monthsandtheJournalofMathematicalphysicsadelayof9.0butthefiveotherjournals,fourofwhichwereindifferentengineeringfieldsandoneinthesocialsciences,haddelaysintherange16.4–20.0months.Tortetal(2011)studiedthedelaysbetweenelectronicandprintpublishinginneurosciencejournal,andfoundasignificantincreasebetween2003and2011.Theywerealsoabletodemonstratethatincreasingthedelayincreasesaparticularjournal’simpactfactor,duetothetimewindowsusedbytheISIincalculatingtheimpactfactor!
7
Table1aboutHerePreviousstudiespointtotwothings.Firstlythattherearesubstantialdifferencesinpublicationdelayswithleadingbiomedicalandchemistryjournalsachievingdelaysofroughlyhalfayearandattheotherendofthespectrumeconomicsandstatisticsjournalstypicallyhavingaveragedelaysofclosetotwoyears.Secondlythatthedelayshaveincreasedsubstantiallyinsomedisciplinesoverthepastdecades,partlyduetoanincreaseinthelengthofthereviewprocess.Twofactorswhichhavenotbeenexplicitlystudiedaretheeffectsofjournalsizeandscientificqualitylevelonthedelays.Mostofthepreviousstudieshavebeenbenchmarkingstudieswithinnarrowdisciplinesofrelativelyhomogeneous,highlycitedjournals.Sizecouldinparticulareffectthedelayafteracceptancesincesmallerjournalsmayappearonlyfourtimesayearoreventwiceayear,whichmeansthatarticlesmighthavetowaitinaqueueforquitesometimebeforepublication.Qualitymightlengthenthesubmissiontoacceptancetimessincearticlesmightgothroughseveraliterationsinthereviewprocess.Ontheotherhandthemosthighlycitedjournalsintheirfieldsmightfinditeasiertorecruitreviewersandaremorelikelytohavealargereditorialstaffandprocesssubmissionsmorequickly.
1.3 Aims Basedonthepreviouslypublisheddata,alotofanecdotalevidenceandpersonalexperiencesasauthorstheaimsofthisstudyweredefinedasfollows.Tostudypublicationdelaysinscholarlypeer‐reviewedjournalsacrossdisciplines,journalsizeandjournalquality.Weexplicitlyruledoutdoingalongitudinalanalysis,duetotheverytime‐consumingworkofdatacollection.
2. Method
2.1 Pilot study Beforestartingdatacollectionwedida“feasibilitystudy”thataddressedtwoissues.FirstlywecheckedourabilitytoobtaincopiesofarticlesfromjournalsindexedinScopusoratleasttheabstractsiftheyhappentocontainthenecessaryinformation.Secondlywecheckedwhetherthejournalsortheirfreelyavailableabstractsincludedsufficientinformationonthepublicationtimeframe.Itwasnecessarytocheckaccesstoelectroniccopiesofthejournalsthroughourlibraries’electronicholdingsaswefeltitwouldnotbefeasibletogatherthedatafrompapercopiesofagivenjournalorgetthenecessarycopiesviainterlibraryloan.AccesstothejournalswascheckedviathelibrariesofHankenSchoolof
8
EconomicsandMichiganStateUniversity.Forthispilotstudywerandomlyselected100journalsindexedinScopus.Atotalof66%ofthesejournalswereavailablethrougheithertheelectronicholdingsofourlibrariesortheywerefreelyavailableonlineandweredeemedtobeappropriateforanalysis.ThemajorityofjournalswecouldnotfindorgainaccessweresmallerjournalspublishedinothercountriesthantheUS,UK,NetherlandsandGermany.Sixty‐fourpercentoftheavailablejournalscontainedatleastthesubmissionandacceptancedatesanditwaspossibletodeterminethedateofpublicationeitheraslistedorbythedateoftheissueinwhichanarticlewaspublished.Wealsofoundthatjournalstypicallypublishedthedatesofuptofivedifferentkeypointsinthepublicationprocess.Theseincluded,submission,revisionbasedonfeedback,acceptance,publicationaheadofprintinanelectronicformat,andfinalpublicationaspartofanissue.Thefirstfourwereusuallyincludedasdates,whereasthelastitemcouldoftenonlybedeterminedbythemonthoftheissuewhichcontainedthearticle.Theresultsofthepilotstudyconfirmedthatthereisenoughdataavailabletomakethestudyfeasible.
2.2 Main study ThemainsourcedatabaseforthestudywastheScopuscitationindex,whichcontainsinformationaboutsome19,500scholarlyjournals,includingtheyearlyarticleandcitationcounts.TheSCImagoJournal&CountryRankwebsite(SCImago,2013)providesfreelyaccessibleScopusdataatthejournallevelwhichwasthedatasourceforthisstudy.Elsevier,thepublisherofScopusprovidesafreelydownloadablespreadsheetontheirwebsite(Scopus,2013)thatamongotherinformationprovidesahierarchicalclassificationofeachjournal’sdiscipline.Thehighestclassificationincludedonly4categoriesandwasfelttobetoobroad.Thesecondlevelincludes27categoriesandwasfelttobetoospecific.Wedecidedtomergesomeoftheselattergroupsbasedonoursubjectiveassumptionofsimilarityinreviewingcultureandpublicationspeedresultinginninegroups.Theseincludearts/humanities,biomedicine,business/economics,chemistry,earthscience,engineering,mathematics,physics,andsocialsciences.Wehypothesizedthatthereweredifferencesinthepublicationtimeassociatedwithjournalsize.Westratifiedbysizeinsuchawaytoensureeacharticlewithinadisciplinecategoryhadanequalchanceofinclusioninthestudy.ThejournalswereorderedbysizebasedonScopusarticlecountsin2010.Thejournalscontainingthefirstthirdofthearticlesinadisciplinemadeupthesmallestjournalstrata,thejournalscontainingthemiddlethirdmadeupthemiddlejournalstrataandthelastthirdofthearticlesthelargejournalstrata.Thisresultedinamuchsmallernumberofjournalsinthelargestjournalstratathoughanequalnumberofarticlesperstrata.
9
Werandomlyorderedthejournalsineachdiscipline/sizestrataandwentthroughthejournalsinordercheckingtoseeiftheywereavailablefromeitherofourtwolibraries,HankenSchoolofEconomicsandMichiganStateUniversityoratleasttheabstractorjournalwasfreelyavailableandcontainedthenecessarydates.Forthosejournalswewereabletoaccess,wecheckedfirstwhethertheyappearedtobepeer‐reviewedscholarlyjournalsandcontainedatleastthedatesofsubmissionandacceptance.Whenanappropriatejournalwasfoundweselected20articlesworkingbackwardfromthelastarticlepublishedin2012.Specialissues,invitedarticlesandeditorialswhereskipped.ForeacharticlewerecordedtheISSN,DOI,orifnoteasilyobtained,title,submissionandacceptancedates.Ifavailablewealsorecordedthedatearevisionrequestwasmadeandthedatethearticlewaspublishedelectronicallyaheadofprint.Publicationdateunlessstatedspecificallywasbasedonthemidpointofthepublicationperiod.Soifajournalwaspublishedmonthly,itwasthe15thofthemonththeissuewaspublished.Ifitwasquarterly,thedatewasthemiddleofthequarter,forexampleFebruary15thforthefirstquarter.AhandfulOAjournalscontainedexactdateoffinalpublication,whichwasusedinplaceofanestimateddate.Whenourmethodofdeterminingthepublicationdateresultedinanegativenumberofdaybetweenacceptanceandpublication,wesetthenumberofdaysfromacceptancetopublicationtozero.Whileweoriginallycalculatedthetimebetweensubmissionandacceptanceandthetimebetweenacceptanceandpublicationindays,forthepurposesofanalyzingandpresentingthedata,weconverteddaysintomonthsbydividingby30.44.Fivejournalswereincludedforeachsizegroupforeachofthe9disciplinecategoriesresultingindatafor135journalsand2,700articles.Forthepurposesofthisstudy,thetimefromsubmissiontoacceptanceandacceptancetopublicationmeasuredinmonthswasusedasthemainoutcomevariables.SourceNormalizedImpactperPaper(SNIP)version2citationmeasureswereobtainedfromtheJournalM3tricswebsite(2013).WealsoobtainedinformationonwhetherajournalwasintheDirectoryofOpenAccessJournals(DOAJ).DatamanagementandmostoftheanalyseswereconductedusingtheStatisticalPackagefortheSocialSciences(SPSS).Mostanalyseswereconductedatthelevelofindividualarticles.SinceSNIPvaluesareassignedtojournals,weaveragedthetimefromsubmissiontoacceptanceandfromacceptancetopublicationforassessingtherelationshipbetweenthesetimesandeachjournal’sSNIP.Thedatacollectedformedabalanceddesignandhenceitwaspossibleusinganalysisofvariance(ANOVA)topartitionthevarianceassociatedwitheachfactorinthedesign.Disciplinewascrossedwithsizegroup.Journalswerenestedinbothdisciplineandsizegroupandarticleswerenestedinajournal.Disciplineandsizegroupwereconsideredtobefixedeffectswhilejournalsandarticleswithinajournalwereconsideredtoberandomeffectsthatweresampled.BasedonthisdesignweestimatedthevariancecomponentsforthetimebetweensubmissionandacceptanceaswellasacceptanceandpublicationusingGENOVA(Brennan,2001).Thisanalysiswasusedtoassessthepercentageofthevarianceinthetimesfromsubmissiontoacceptanceandfromacceptancetopublication
10
thatcouldbeattributedtoeachsource,discipline,size,theirinteraction,journalsandarticleswithinjournals.
3. Results Althoughmoredetaileddatawereavailableforsomejournals,wefocusedthereportingonthetimefromoriginalsubmissiontoacceptanceandfromacceptancetofinalpublicationaswefeltthesewerethekeytimepointsandwewereabletoobtaincompletedataacrossalldisciplinesandsizegroups.Thefirsttimeperiodreflectsthedelayduetothepeerreviewandrevisionprocessusedbyajournalandthesecondthelengthofthepublishingprocess,backlogduetopublicationpagelimitsandpotentiallyotherfactors.Table2presentssummarystatisticsforsubmissiontoacceptance,acceptancetopublicationandtotaltimesubmissiontopublication.Figures1and2presentthisinformationingraphicformforthe9disciplines(Figure1)and3journalsizegroupings(Figure2).DetailedsummarystatisticsforthebreakdownbydisciplinesandsizegroupsarecontainedintheAppendix.AscanbeseeninFigure1,totaltimefromsubmissiontopublicationvariessignificantlybydisciplinewithbusinessatjustunder18monthshavingpublicationtimesnearlytwicethatofchemistryatabout9months.Largerjournalsappeartohavetheshortestpublicationtimeswithmid‐sizedjournalsthelongest.BasedoninclusionintheDOAJtherewere19OpenAccess(OA)journalsorapproximately14%ofthesample.Ofthese,7weredeterminedtobeOAfromtheirinceptionand12weredeterminedtohavebeenconvertedtoOAatsomepoint.Thelatterusuallymeansthatthejournalmaystillpublishaparallelpaperversionandalsoittypicallybundlesthearticlesinissues.Themethodologyfordeterminingbornversusconvertedaredescribedelsewhere(Solomon,Laakso&Björk,2013).Table3presentstheaveragetimeinmonthssubmissiontoacceptanceandacceptancetopublicationforjournalscreatedasOAandthosethatconvertedtoOA.SubmissiontopublicationtimesappeartobeconsiderablyshorterforOAjournals,particularlythosethatwerecreatedasOAjournals.Thedifferenceswerereflectedinbothreceivedtoacceptedandacceptedtopublishedbutthegreatestdifferences,particularlyforthejournalscreatedOAwereinacceptedtopublishedtimes.Thesedifferencesshouldbeconsideredwithcautionasthesamplesizesarefairlysmallandthepercentagesofjournalswithineachdisciplinearenotbalanced.Wefelttheremaybeacorrelationbetweenpublicationtimesandthecitationrateofthejournal.Sincecitationrateisatthelevelofthejournalratherthanthearticle,weaggregatedtothelevelofajournalusingaveragesforthetimesfromsubmissiontoacceptanceandfromacceptancetopublication.WeusedSNIPasthecitationmeasurebecausethesestatisticsarenormalizedtoaccountfordifferencesincitationratesacrossdisciplines.ThePearsonproductmomentcorrelationbetweenSNIPandsubmissiontoacceptanceandacceptancetopublicationwere0.20and‐0.09respectively.Thecorrelationforthetimefrom
11
submissiontoacceptancewithSNIPwassignificantlydifferentfromzerop<0.02.Table4containstheestimatedvariancecomponents1fordiscipline,sizegroup,journalswithindiscipline/sizegroupandarticleswithinjournals.Forsubmissiontoacceptance,thevariationamongjournalsandarticlesaccountedforthebulkofthevariation,mostlyintermsofarticleswithinjournals.Foracceptancetopublication,againthevariationwasalmostentirelyamongjournalsandarticlesnestedinjournals.Forthiscomponenthoweverthevariationamongjournalsaccountedforthebulkofthevariation.
4. Discussion Theresultsofthisstudyhavetobeinterpretedwithsomecaution.Themaincaveatisthatwewereonlyabletoincludedatafromjournalsthatpublishedthesubmissionandacceptancedateswhileinmostcasesthepublicationdatewasinferredfromtheissueandestimatedasthemid‐pointofthepublicationperiodfortheissue.Sincethedecisiontopublishthisinformationwasgenerallyconsistentacrossallthejournalsofaparticularpublisher,onlythosepublishersthatchoosetopublishsubmissionandacceptancedateareincludedinthestudy.Thisresultedin54%ofthesamplebeingpublishedbythetwobiggestpublishersElsevierandSpringer/Kluwer.Thiswasnotourintentionbutwastheresultofthelimitationnotedabove.AlistofthepublishersincludedinthestudythenumberofjournalsfromeachpublisherincludediscontainedintheAppendix.Therewerestrikingdifferencesbetweendisciplineswithbusiness/economicshavingaroundtwicethetotaldelaysubmissiontopublicationcomparedtochemistry.Differenceswerealsofoundintermsofthesizeofthejournalthoughtheywerefairlymodestwiththelargerjournalsappearingtobethemostefficientbothintermsofthetimefromsubmissiontoacceptanceandinpublishingarticlesonceaccepted.Openaccessjournals,particularlythosewhichwerecreatedasOAjournalsratherthanwereconvertedfromsubscriptionappeartobeabletopublisharticlesconsiderablymorequicklythansubscriptionjournals.Thisinpartmayreflectthefacttheyareelectroniconlyandtendtopublisharticlesastheyarereadyratherthanbundlingthemintoissues.GiventhesmallnumbersandthefacttheOAjournalsarenotevenlydistributedacrossdisciplinesthesefindingshouldbeinterpretedwithagreatdealofcaution.Theanalysisofvarianceindicatesmostofthevariationinpublicationtimesisatthelevelofindividualjournalsandarticles.Forthetimefromsubmissionto
1Thecomponentsforthefixedfactors,discipline,journalsizelevelandtheirinteractionarenottruevariancecomponents.Sincetheyarefixedeffectstheyarenotstatisticalexpectationsbutquadraticformsthatareaveragessimilarinnaturetoavariancecomponent.(Brennan,2001)Sincethedistinctionisirrelevantforthepurposesofthisstudy,wewillrefertothesequadraticformsasvariancecomponentsinthediscussionoftheresults.
12
acceptance,thebulkisamongarticles.Thisisnotsurprising.Therearemanyidiosyncraticfactorsthatinfluencethelengthofindividualarticlereviews.Editorsmoreoftenthannotacceptmanuscriptspendingrevisionsandauthorsvarygreatlyinhowquicklytheycompletetherevisions.Hencethelengthofthereviewprocessforaparticulararticlemayreflecttheactionsoftheauthorratherthantheeditororreviewers.Alltheseandotherfactorsresultinsignificantdifferencesinreviewtimesamongsubmissionsforaspecificjournal.Therewasalsoconsiderablymorevariationamongjournalswithinadisciplineandsizegroupthanamongdisciplinesandsizegroups.Thisindicatestherearerealdifferencesinthisimportantaspectofpublishingthatarenotexplainedbyeithertheanomaliesofindividualreviewsorthecultureofreviewofdifferentfields.Somejournalsjustappeartobefasterinconductingthereviewprocess.Thislikelyinpartreflectsthelevelandnumberofcyclesofrevisionstypicallyrequiredbytheeditor.Italsomayreflecthowquicklymanuscriptsgooutforreviewandwhatexpectationtheeditororeditorialteamhasforhowlongareviewershouldtakeinreviewingamanuscript.Forthetimefromacceptancetopublicationthevastmajorityofthevariationisamongjournals.Again,thisdoesnotseemsurprising.Thebacklogsinprocessingmanuscriptsthroughtypesettingandcopyediting,frequencyofpublicationandthebacklogduetopagelimitsiftheyexistwouldalllargelyimpactonpublicationtimesatthejournallevel.
5. Conclusions Webelievethistobethefirstbroadstudyofpublishingdelays,coveringallfieldsofscience.Ourstudyalsodiffersfromallearlierstudiesbyouruseofarandomsamplecoveringjournalsofallqualitylevels.Previousstudies,haveusuallyusedsmallconveniencesamplesoftypicallytopjournalsintheirfields,whichintroducesastrongbiastowardsjournalsthatmayincludelongreviewprocesses.Ourresultsare,nevertheless,notinconflictwiththeearlierstudies,butinstead,addtothem.Themethodologywasverylaborintensiveanditwouldbeveryusefulforfuturestudiesifpublishersincludedthedateofsubmissionandacceptanceasastandardpartoftheirarticleinformation.Thiswouldprovidealeveloftransparencyforpotentialauthorsastothedelaystheycouldexpectinreviewandpublicationprocesseswhenconsideringwheretosubmittheirmanuscripts.Itwouldalsoprovideastrongincentiveforjournalstospeeduptheseprocesses.Theaimofourstudywastoprovideoveralldataonreviewandpublishingtimesacrossvariousfieldsofscience.Wedidnotattempttodeterminehowdelayshaveevolvedovertime.Someoftheearlierstudieshavedonethis,butwemadeaconsciouschoicetoconcentrateonthedifferencesbetweendisciplines,duetotheresourceintensivenessofourmethod.Alongitudinalstudywouldbeagoodtopicforafollow‐upstudy,andshouldideallygobackaround25years,tothetimebeforeemail,websubmissionsystemsandelectronicpublishing.That
13
wouldontheotherhandalsoimplychallengesinfindingthedatawiththearticles.Itwouldbeveryusefultomakeamoredetailedstudyofwhydelaysdiffersomuchbetweendisciplines,Ourimpressionisthatthecleardifferencesamongfieldshaveevolvedoverdecadesthroughthedevelopmentofintra‐disciplinarysocialnormsforwhatisexpectedfromascholarlyjournalinthefield.Thisincludeswhatisanacceptabledelayforinformingauthorsofreviewresultsandacceptanceorrejectiondecisionsaswellastheprocessingandqueuingtimeonceamanuscriptisaccepted.ThisisinlinewiththeconclusionsofforinstanceEllison(2000).Thesedifferencesinreviewandpublicationtimesmayalsoreflectthenatureofthedisciplines.Forexampleinrapidlydevelopingfieldswhereseparategroupsofresearchersmayberacingtoachieveaparticularbreakthrough,thespeedofthepublicationprocesscandeterminewhichgroupgainscreditforthebreakthroughaspublicationhasbecomethedefactodeterminerofwhogetsthecreditforamajorfinding.Otherinterestingtopicsforfurtherstudieswouldbethedifferencesbetweenjournalswithinadisciplineandarticleswithinjournals.Forsomeindividualarticlesthedelaytimescanbeexcessivelylong.Thedelayscanbeduetotheauthorstakingexcessivelylongtimesmakingrevisionsaftertheoriginalreviewcycle.Theycanalsobeduetoexcessivelylongreviewperiodsordelaysinthepublicationprocess.Asfoundinthisstudy,mostofthevariationinsubmissiontoacceptancetimesisamongindividualmanuscriptswithinajournalwhilemostofthevariationinacceptancetopublicationtimeisamongjournalswithinadiscipline/sizegroup.Sincepublicationdelaysarebothdetrimentaltothecareersofindividualscholarsandretardtherateatwhichscientificfieldsadvance,understandingandattemptingtominimizeunnecessarydelaysinthepeer‐reviewandpublicationprocessisineveryone’sbestinterest.OneofthereasonsforthepopularityofOAjournals,inadditiontothewiderdissemination,isthebeliefthattheyhavemuchfastersubmissiontopublicationtimes.ThisperceptionisoftenhighlightedinthepromotionalmaterialforfullyelectronicOAjournals.ItappearsfromourverylimitedsampleofOAjournalsthatjournalswhichareonlydisseminatedindigitalformandpublisharticlesindividuallyastheyarereadytendtohaveconsiderablyshortersubmissiontopublicationperiodswithmostofthedifferenceduetoshorteracceptancetopublicationtimes.Afollow‐upstudycomparingsubscriptionjournalswithOA‐journalswouldneedtofurthersplitupOA‐journalsintoanumberofsubgroups,suchasmegajournals(PloSONEandcloses),journalsfromso‐calledpredatoryjournalswithspamacademicswithemailspromisingveryrapidpublicationandhighqualityOAjournals.Somecriticsofthecurrentsystemhavediscussedthealmostdefactostandardjournalpolicyofnotallowingauthorsthepossibilityofsubmittingtheirmanuscriptstootherjournalsinparallel(Torgersonetal2005),aslongasthearticlehasnotbeendefinitelyrejected(Piron2001).Thispolicycanresultinlongdelaysinthepublicationprocessofarticlesrejectedinthefirst‐choicejournalpotentiallyrenderingtheresultsoftheresearchoutdatedandoflittleuse
14
bythetimeitisfinallypublished.Thepolicyisoftenjustifiedbysayingthatitwouldbeveryinefficientandunfairtoeditorsandrefereesifthesamearticleswouldberefereedinseveraljournalsatthesametime.Ontheotherhandexactlythesamethinghappenswhenarticlesafterrejectionorauthorwithdrawalareresubmittedtootherjournalsandnewreviewersgetinvolved.Interestinglythereisonejournalcategorywherethisruleisnotenforced,lawjournalspublishedbyleadingUSuniversities,whichallowauthorstosubmittocompetingjournalssimultaneously.Althoughnoempiricalstudiescouldbefoundofthepublishingdelayintheselawjournals,severalauthorsforexample(Posner,2008)havepointedoutthatthedelaysaremuchshorterthaninotherfields.Ifpublishersaregoingtosticktothedemandthatauthorsrefrainfrommultiplesimultaneoussubmissionsofamanuscriptthenitseemstous,thattheyhaveanobligationtomakethepublicationprocessasfastandefficientaspossible.Electronicpublicationoffersarealpotentialforspeedingupthescholarlyjournalpublishingprocess,butinordertoachievethisjournalshavetostoppublishingaparallelpaperversionandneedtoconverttopublishingarticlesinanissue‐lessmodeastheybecomeavailable.Thisisexactlywhatmostbornelectronicjournalsdo,andastheirshareofpublishingincreases,averagepublishingdelayswilltendtodecrease.
Acknowledgements: WeareverygratefultoCharlottaBjörk,AleksiAaltonenandPatrikWellingforhelpingoutwiththetediousworkofgatheringthedata.
References: Amat,C.B.(2008)Editorialandpublicationdelayofpaperssubmittedto14selectedFoodResearchjournals.Influenceofonlineposting,Scientometrics74,3,379Azar,OferH.,Rejectionsandtheimportanceoffirstresponsetimes,InternationalJournalofSocialEconomicsVol.31No.3,2004pp.259‐274Björk,B‐C.andTurk,Z.(2006).TheElectronicJournalofInformationTechnologyinConstruction(ITcon):anopenaccessjournalusinganun‐paid,volunteer‐basedorganization.InformationResearch,11(3)paper255,Availableathttp://InformationR.net/ir/11‐3/paper255.html]
15
Björk,B‐C.(2011)InnovationsinScholarlyPublishing–AmultipleCaseStudyofOpenAccessJournals,JournalofMedicalInternetResearch,13(4):e115Retrievedfromhttp://www.jmir.org/2011/4/e115/,doi:10.2196/jmir.1802Brennan,R.L.(2001)GeneralizabilityTheoryStatisticsforSocialScienceandPublicPolicy(series)2001SpringerNewYork.CarrollR.(2001)ReviewTimesinStatisticalJournals:TiltingattheWindmills?Biometrics,Vol57,1‐6.CoeR.K.,WeinstockI.(1967):“EditorialPoliciesoftheMajorEconomicsJournals,”QuarterlyReviewofEconomicsandBusiness7,37‐43.I.Diospatonyi,†G.Horvai,†andT.Braun*,‡PublicationSpeedinAnalyticalChemistryJournals,J.Chem.Inf.Comput.Sci.2001,41,1452‐1456Dong,Peng,Loh,Marie,Mondry,AdrianPublicationlaginbiomedicaljournalsvariesduetotheperiodical'spublishingmodel,Scientometrics,2006,69(2),271‐286Ellison,G.(2002)TheslowdownoftheEconomicsPublishingProcess,JournalofPoliticalEconomy,110,947‐993Hartmann,D.L.“AreAMSJournalsTooSlowtoBeUseful?AnOpenLettertoAMSMembers,”http://www.atmos.washington.edu/~dennis/AMS_Publ_Paper.970610.psJournalM3tricsResearchAnalyticsRedefined.(2013).http://www.journalmetrics.com/Khosrowjerdi,M.,Zeraatkarb,N.,Varac,N.(2011)PublicationDelayinIranianScholarlyJournals,SerialsReview37(4)262–266.Kling,R.,Swygart‐Hobaugh,A.J.(2002).TheInternetandthevelocityofscholarlyjournalpublishing.WorkingPaperNo.WP‐02‐12,RobKlingCenterforSocialInformatics,SchoolofLibraryandInformationScience,IndianaUniversity,Bloomington,IN.https://scholarworks.iu.edu/dspace/handle/2022/148Laakso,M.,Björk,B‐C.(2012).Anatomyofopenaccesspublishing:astudyoflongitudinaldevelopmentandinternalstructure.BMCMedicine,10:124.doi:10.1186/1741‐7015‐10‐124LuwelM.,MoedH.F.(1998)PublicationDelaysintheScienceFieldsandTheirRelationshiptotheAgeingofScientificLiterature,”Scientometrics41(1–2)29–40.Nature.Overview:Nature'speerreviewtrial(2006)|doi:10.1038/nature05535.
16
Piron,Robert(2001)TheyHavetheWorldonaQueue,Challenge,200144(5)95‐101.Posner,Richard(2006)LawReviews,WashburnLawReview,46(1)http://washburnlaw.edu/wlj/46‐1/articles/posner‐richard.pdfRaney,K.(1998).Intoaglassdarkly.JournalofElectronicPublishing,4(2).Retrieved3March,2006fromhttp://www.press.umich.edu/jep/04‐02/raney.htmlSCIMago.(2013).SJR–SCImagoJournal&CountryRank.Retrievedfromhttp://www.scimagojr.comSolomon,D.J.,Björk,B‐C.(2012)PublicationFeesinOpenAccessPublishing:SourcesofFundingandFactorsInfluencingChoiceofJournal,JournaloftheAmericanSocietyforInformationScienceandTechnology,63(1):98–107,2012,DOI:10.1002/asi.21660Solomon,David,Laakso,Mikael,Björk,Bo‐Christer,2013,Alongitudinalcomparisonofcitationratesandgrowthamongopenaccessandsubscriptionjournals,JournalofInformetrics,Vol7,No3,642‐650.Suber,P.(2012).OpenAccess.MITPress.230p.TorgersonD.J.,AdamsonJ.,CockayneS,DumvilleJ.,PetherickBritishE.(2005)Submissiontomultiplejournals:amethodofreducingtimetopublication?BMJ,330,305–307.TrivediP.K.,(1993)AnAnalysisofPublicationLagsinEconometrics,”JournalofAppliedEconometrics893–100.VanOrsdel,L.andBorn,K.(2002)PeriodicalsPriceSurvey2002:DoingtheDigitalFlip.LibraryJournal,http://www.libraryjournal.com/article/CA206383.html.)Ware,M.,Mabe,M.(2012)TheSTMReport:AnoverviewofScientificandScholarlyJournalPublishing,InternationalAssociationofScientific,TechnicalandMedicalPublishers,TheHague,Netherlands,http://www.stm‐assoc.org/2012_12_11_STM_Report_2012.pdfYohe,G.W.,(1980)CurrentPublicationLagsinEconomicsJournals,"JournalofEconomicLiterature18,1050‐55YuG.,YuD.R.,,LiY.J.(2004)TheuniversalexpressionofperiodicalaveragepublicationdelayatsteadystateScientometrics60(2)121–129.
17
Figure1:AveragePublicationTimesinMonthsbyDisciplineFigureisbasedon15journalsperdiscipline,5foreachsizegroup,20articlesperjournalresultinginatotalof300articlesperdiscipline.
18
Figure2:AveragePublicationTimesinMonthsbyJournalSizeGroupFigureisbasedon45journalspersizegroup,15foreachdiscipline,20articlesperjournalresultinginatotalof900articlespersizegroup.
19
Table1.Previousstudiesconcerningpublicationdelaysinscholarlyjournals.Study
Includedjournals
Periodstudied
Discipline Delay(months)AverageRange
Yohe1980
25journals 1980 Economics 18.9
4.9–28.7
Trivedi1993 7journals
1986‐1990 Econometrics 22.8 19.7–31.4
Carroll2001
6journals 1994,1999 Statistics 22.3 15.0‐26.0
KlingandSwygart‐Hobaugh2002
3socialsciencejournals
1970/1980,2000
Econ.,management
23.8 17.0–29.4
KlingandSwygart‐Hobaugh2002
3naturalsciencejournals
1970/1980,2000
Physics,Chemistry
5.8 4.0–7.4
Hartmann1997
Onejournal 1970,1997 AtmosphericSciences
15.4
LuwelandMoed1998
15journals 1992 Physicalsciences,Eng.
9.4
2.5‐17.0
Diospatonyietal2001
10journals 1985‐1999 Analyticalchemistry
7.1 3.5–12.5
Raney1998
Onejournal 1997 Geoscience 21.8 11.5‐36.5
Yuetal2004
7journals 2002 Mainlyengineering
15.1 5.5–20.0
Amat2008 14journals 2004 Agriculture 11.8 6.2–17.2
Dongetal2006 28commercial,SocietyandOA
2004 Biomedicine 6.3 3.0‐11.0
BjörkandTurk2006
OneOAandtwoconventional
2005 CivilEngineering
14.5
6.7–18.9
Khosrowjerdietal2011
26Iranianjournals 2009 Cross‐disciplinary
17.3
5.8–34.6
20
Table2 TimeSubmissiontoPublicationTotals
Months Submitted to
Accepted
Months Accepted to Published
Months Submitted to
Published
Mean 6.41 5.78 12.18
Std. Deviation 5.35 4.21 7.17
Std. Error of Mean* 0.10 0.08 0.14
Statisticsbasedon135Journals/2,700Articles*TheStandarderrorsofthemeansareapproximateduetothelackofindependencebetweenarticlesinthesamejournal.
21
Table3 TimeSubmissiontoPublicationforOAJournals
Created Open Access
Months Submitted to
Accepted
Months Accepted to Published
Months Submitted to
Published
Yes
Mean 4.17 1.80 5.97
Std. Deviation 3.08 1.56 3.77
Std. Error of Mean 0.26 0.13 0.32
Number 7 Journals / 140 Articles
No Mean 5.12 4.76 9.88
Std. Deviation 4.37 5.17 7.90
Std. Error of Mean 0.28 0.33 0.51
Number 12 Journals / 240 Articles
Thestandarderrorsofthemeanareapproximateduetolackofindependenceamongarticlesinthesamejournal.
22
Table 4 Estimated Variance Components
Submit to Accept Accept to Publish
Variance Percent Variance Percent
Discipline 3.44 12% Discipline 0.83 5%
Journal Size 0.52 2% Journal Size 0.31 2%
Journal 8.49 29% Journal 12.88 71%
Article 16.46 56% Article 4.20 23%
Size x Discipline 0.49 2% Size x Discipline 0.00 0%
Total 29.41 Total 18.23
23
Publication Time in Months by Discipline
Discipline Months
Submitted to Accepted
Months Accepted to Published
Months Received to Published
Chemistry Mean 4.73 4.18 8.91
Std. Deviation 5.46 3.60 7.30
Std. Error of Mean 0.32 0.21 0.42
Engineering Mean 5.00 4.30 9.30
Std. Deviation 3.68 3.06 5.29
Std. Error of Mean 0.21 0.18 0.31
Biomedicine Mean 4.65 4.82 9.47
Std. Deviation 3.47 4.11 5.18
Std. Error of Mean 0.20 0.24 0.30
Physics Mean 5.21 5.72 10.93
Std. Deviation 3.26 2.66 4.41
Std. Error of Mean 0.19 0.15 0.25
Earth Science Mean 5.74 5.96 11.70
Std. Deviation 4.80 4.66 7.24
Std. Error of Mean 0.28 0.27 0.42
Mathematics Mean 8.20 5.11 13.30
Std. Deviation 6.21 2.45 6.87
Std. Error of Mean 0.36 0.14 0.40
Social Science Mean 6.17 7.93 14.10
Std. Deviation 4.36 5.73 7.32
Std. Error of Mean 0.25 0.33 0.42
Arts and Letters Mean 7.21 7.00 14.21
Std. Deviation 5.26 5.38 7.71
Std. Error of Mean 0.30 0.31 0.44
Business/Economics Mean 10.75 6.96 17.70
Std. Deviation 7.15 3.19 7.52
Std. Error of Mean 0.41 0.18 0.43
All Journals Mean 6.41 5.78 12.18
Std. Deviation 5.35 4.21 7.17
Std. Error of Mean 0.10 0.08 0.14
There are 15 journals, 5 per each size group and 20 articles per journal
Standard error of the mean is approximate due to lack of independence among articles in a journal.
24
Number of Journals from Each Publisher Included in the Study
Publisher Number Percent American Chemical Society 1 0.8%
American Dairy Science Association 1 0.8%
American Physiological Society 1 0.8%
American Psychiatric Publishing, Inc. 1 0.8%
American Psychological Association 1 0.8%
American Society of Civil Engineers 1 0.8%
American Vacuum Society 1 0.8%
Arizona State University 1 0.8%
Australasian Association of Psychology and Philosophy 1 0.8%
Bentham Science Publishers 1 0.8%
BioMed Central 1 0.8%
Blackwell Publishing Inc. 3 2.3%
Butterworth Scientific Ltd. 1 0.8%
Cell Press 1 0.8%
Central Fisheries Research Institute 1 0.8%
Cognizant Communication Corp. 1 0.8%
Consejo Superior De Investigaciones Cientificas 2 1.5%
Copernicus Gesellschaften 1 0.8%
Electrochemical Society, Inc. 1 0.8%
Elsevier 61 46.6%
European Respiratory Society 1 0.8%
Geophysical Society of Finland 1 0.8%
Hindawi Publishing Corporation 1 0.8%
Institute for Ionics 1 0.8%
Institute of Physics Publishing 1 0.8%
Istituti Editoriali e Poligrafici Internazionali 1 0.8%
JAI Press 1 0.8%
Maik Nauka/Interperiodica Publishing 1 0.8%
Marcel Dekker Inc. 3 2.3%
Molecular Diversity Preservation International 1 0.8%
Oldenbourg Wissenschaftsverlag GmbH 1 0.8%
Opragen Publications 1 0.8%
Oxford University Press 1 0.8%
Pan American Health Organization/Organizacion Panamericana de la Salud
1 0.8%
Prolegomena: Journal of Philosophy 1 0.8%
Public Library of Science 1 0.8%
Raptor Research Foundation, Inc. 1 0.8%
Royal Society of Chemistry 2 1.5%
Royal Society of London 1 0.8%
SAGE Publications 1 0.8%
Springer Pub. Co., 14 10.7%
Taylor & Francis 1 0.8%
25
Universidad de los Andes 1 0.8%
Universidad de Murcia 1 0.8%
Universidad Nacional de Colombia 1 0.8%
Universidade Estadual Paulista 1 0.8%
Universidade Federal do Rio de Janeiro 1 0.8%
University of the Aegean 1 0.8%
Versita (Central European Science Publishers) 1 0.8%
Wayne State University Press 1 0.8%
Wiley-Blackwell 2 1.5%
Publisher names retrieved from SCImago Journal & Country Rank website. http://www.scimagojr.com/