誠信制度之危誠信制度之危誠信制度之危資料僅供參考文件編號：2022年4月誠信制度之危版本號：A修改號：1頁次：1.0審核：批準(zhǔn):發(fā)布日期：誠信制度之危[Dishonoring

the

Honor

Systemcientificreviewisnotdesignedtocatchcheats.ButtheSouthKoreancloningscandalsuggeststhatthejournalsNatureandSciencemaybetoopowerfulindecidingwhatresearchreachesthepublic.<xml:namespaceprefix=ons="urn:schemas-microsoft-com:office:office"/>科學(xué)評審的本意并非是要抓騙子。然而，韓國克隆丑聞?wù)f明，在決定將哪些科研成果公諸于眾方面，《自然》和《科學(xué)》雜志或許有些過于強(qiáng)大了。

JanHendrikSchon'ssuccessseemedtoogoodtobetrue,anditwas.InonlyfouryearsasaphysicistatBellLaboratories,Schon,32,hadco-authored90scientificpapers—oneevery16days—detailingnewdiscoveriesinsuperconductivity,lasers,nanotechnologyandquantumphysics.Thisoutputastonishedhiscolleagues,andmadethemsuspicious.Whenoneco-workernoticedthatthesametableofdataappearedintwoseparatepapers—whichalsohappenedtoappearinthetwomostprestigiousscientificjournalsintheworld,ScienceandNature-thejigwasup.InOctober2002,aBellLabsinvestigationfoundthatSchonhadfalsifiedandfabricateddata.Hiscareerasascientistwasfinished.

IfitsoundsalotlikethefallofHwangWooSuk—theSouthKoreanresearcherwhofabricatedhisevidenceaboutcloninghumancells—itis.Scientificscandals,whichareasoldasscienceitself,tendtofollowsimilarpatternsofhubrisandcomeuppance.Afterwards,colleagueswringtheirhandsandwonderhowsuchmalfeasancecanbeavoidedinthefuture.Butitneverisentirely.Scienceisbuiltonthehonorsystem;themethodofpeer-review,inwhichmanuscriptsareevaluatedbyexpertsinthefield,isnotmeanttocatchcheats.Inrecentyears,ofcourse,thepressureonscientiststopublishinthetopjournalshasincreased,makingthejournalsthatmuchmorecrucialtocareersuccess.ThequestionsraisedanewbyHwang'sfallarewhetherNatureandSciencehavebecometoopowerfulasarbitersofwhatsciencereachesthepublic,andwhetherthejournalsareuptotheirtaskasgatekeepers.

Eachscientificspecialtyhasitsownsetofjournals.PhysicistshavePhysicalReviewLetters,cellbiologistshaveCell,neuroscientistshaveNeuron,andsoforth.ScienceandNature,though,aretheonlytwomajorjournalsthatcoverthegamutofscientificdisciplines,frommeteorologyandzoologytoquantumphysicsandchemistry.Asaresult,journalistslooktothemeachweekforthecreamofthecropofnewsciencepapers.Andscientistslooktothejournalsinparttoreachjournalists.WhydotheycareCompetitionforgrantshasgottensofiercethatscientistshavesoughtpopularrenowntogainanedgeovertheirrivals.Publicationinspecializedjournalswillwintheaccoladesofacademicsandsatisfythepublish-or-perishimperative,butScienceandNaturecomewiththeaddedbonusofpotentiallygettingyourpaperwrittenupinTheNewYorkTimesandotherpublications.

ScientistsarealsotryingtoreachotherscientiststhroughScienceandNature,notjustthepublic.Thelinebetweenpopularandprofessionalnotorietyisnotdistinct.ScientiststendtopaymoreattentiontotheBigTwothantootherjournals.Whenmorescientistsknowaboutaparticularpaper,they'remoreapttociteitintheirownpapers.Beingoft-citedwillincreaseascientist's"ImpactFactor",ameasureofhowoftenpapersarecitedbypeers.FundingagenciesusetheImpactFactorasaroughmeasureoftheinfluenceofscientiststhey'reconsideringsupporting.BecauseNatureandSciencepapershavemorevisibility,thenumberofsubmissionsisgrowing,saytheeditors.Naturenowgets10,000manuscriptsayear,andthatfigureisrising,sayseditor-in-chiefPhiipCampbellviaemail."Thispartlyreflectstheincreaseinscientificactivityaroundtheworld,"hesays."Italsonodoubtreflectstheincreasingandsometimesexcessiveemphasisamongstfundingagenciesandgovernmentsonpublicationmeasures,suchasthetypicalratesofcitationofjournals."

Whateverthereasons,thewhimsoftheeditorsatScienceandNatureloomlargeformanyscientists.Wheneithermagazineisconsideringapaperforpublication,theauthorsaretoldnottospeaktothepresslesttheywanttoriskrejection."Everyscientistshatesthemandlovesthem,"saysaprominentscientistwhowouldnotspeakforattributionforfearofoffendingtheeditors."Wehatethembecauseit'ssopoliticaltogetanarticleinthem.FranklyI'mastonishedatsomeofthethingstheyaccept,andsomeofthethingstheyreject."

Whethertheclamortoappearinthesejournalshasanybearingontheirabilitytocatchfraudisanothermatter.Thefactis,fraudisterrificallyhardtospot.ConsidertheprocessScienceusedtoevaluateHwang's2005article.Scienceeditorsrecognizedthemanuscript'simportalmostassoonasitarrived.Aspartofthestandardprocedure,theysentittotwomembersofitsBoardofReviewingEditors,whorecommendedthatitgooutforpeerreview(about30percentofmanuscriptspassthistest).Thisrecommendationwasmadenotonthescientificvalidityofthepaper,butonits"novelty,originality,andtrendiness,"saysDenisDuboule,ageneticistatthe<xml:namespaceprefix=st1ns="urn:schemas-microsoft-com:office:smarttags"/>UniversityofGenevaandamemberofScience'sBoardofReviewingEditors,intheJanuary6issueofScience.(EditorswouldnotcommentforthisstoryaheadofthecompletionofSeoulNationalUniversity'sinvestigation,whichwasreleasedtoday.ThepanelfoundthatHwanghadfabricatedalloftheevidenceforresearchthatclaimedtohaveclonedhumancells,butthathehadsuccessfullyclonedthedogSnuppy.)

Afterthis,Sciencesentthepapertothreestem-cellexperts,whohadaweektolookitover.Theircommentswerefavorable.Howweretheytoknowthatthedatawasfraudulent"Youlookatthedataanddonotassumeit'sfraud,"saysonereviewer,anonymously,inScience.AtaDecembernewsconference,editor-in-chiefDonaldKennedymaintainedthatthepaper,despiteitsimportance,wasnotrushedtoprint."Anyimportantpapergetscarefulscrutiny,andIthinkourpeerreviewersgaveitthat,"hesaid."It'sverydifficultforapeer-reviewprocesstodetectmistakesthatarenotclearlyevidentoraredeliberatemisrepresentations."

Intheend,abigscandalnowandthenisn'tlikelytodomuchdamagetothebigscientificjournals.Whateditorsandscientistsworryaboutmorearethemyriadsmallerinfractionsthatoccurallthetime,andwhicharealmostimpossibletodetect.ANaturesurveyofscientistspublishedlastJunefoundthatone-thirdofallrespondentshadcommittedsomeformsofmisconduct.Theseincludedfalsifyingresearchdataandhaving"questionablerelationships"withstudentsandsubjects—bothchargesleveledagainstHwang.Nobodyreallyknowsifthiskindoffraudisontherise,butitisworrying.

Scienceeditorsdon'thaveanyplanstochangethebasiceditorialpeer-reviewprocessasaresultoftheHwangscandal.Theydohaveplantstoscrutinizephotographsmorecloselyinanefforttospotinstancesoffraud,butthatpolicychangehadalreadybeendecidedwhenthescandalstruck.Andevenifithadbeeninplace,itwouldnothaverevealedthatHwanghadmisrepresentedphotographsfromtwostemcellcoloniesascomingfrom11colonies.(Nature'sCampbellwouldnotanswerquestionsaboutreviewpolicyorthestatusofHwang's2005NaturepaperonthecloningofSnuppythedog,whichNatureisinvestigating.)Withthefinancialanddeadlinepressuresofthepublishingindustry,it'sunlikelythatthejournalsaregoingtotakemarkedlystrongermeasurestovetmanuscripts.Beyondreplicatingtheexperimentsthemselves,whichwouldbeimpractical,it'sdifficulttoseewhattheycoulddotomaketakesciencebeyondthehonorsystem.■

簡·亨德里克·舍恩的成功似乎令人難以置信，但它的確如此。３２歲的舍恩曾是貝爾實(shí)驗(yàn)室的物理學(xué)家，在短短４年間，他卻與人合作撰寫了９０篇科技論文——每１６天一篇——這些論文詳盡闡述了他在超導(dǎo)、激光、納米技術(shù)和量子物理學(xué)方面的最新發(fā)現(xiàn)。如此多產(chǎn)令他的同事大為驚訝并產(chǎn)生懷疑。當(dāng)一位同事發(fā)現(xiàn)同一個數(shù)據(jù)圖表出現(xiàn)在兩篇不同的論文之中——而且恰恰分別出現(xiàn)在世界上兩家最為知名的科學(xué)期刊《科學(xué)》與《自然》雜志上時——把戲被拆穿了。２００２年１０月，貝爾實(shí)驗(yàn)室調(diào)查發(fā)現(xiàn)，舍恩曾篡改并捏造數(shù)據(jù)。他的科學(xué)家生涯就此止步。

如果此事聽起來很像是黃禹錫的名譽(yù)掃地一事，那么事實(shí)的確如此。這位韓國研究者在克隆人體細(xì)胞上偽造了證據(jù)?？茖W(xué)丑聞與科學(xué)本身的發(fā)展亦步亦趨，并且往往遵循著類似的規(guī)律，即自以為是必然會受到懲罰。丑聞曝光之后，同行們都在痛定思痛，思考今后怎樣才能避免此類事件的發(fā)生。但是，要完全避免絕不可能?！犊茖W(xué)》雜志是建立在誠信制度之上的；同行評審的辦法，即由同一領(lǐng)域?qū)＜覍φ撐脑暹M(jìn)行評審的方法，并非旨在抓騙子。當(dāng)然，近年來要求科技工作者在頂級期刊上發(fā)表論文的壓力越來越大，使得這些期刊在他們事業(yè)成功上所起的作用越來越關(guān)鍵。黃禹錫事件讓這些問題再次浮出水面：《自然》和《科學(xué)》在決定將哪些科研結(jié)果公諸于眾方面的權(quán)力是不是太大了，這些刊物是否能夠勝任其守門人的工作。

每個科學(xué)專業(yè)領(lǐng)域都有自己的一系列刊物。物理學(xué)家有《物理評論快報》，細(xì)胞生物學(xué)家有《細(xì)胞》，神經(jīng)科學(xué)家有《神經(jīng)元》，等等。但是，《科學(xué)》和《自然》是僅有的兩家覆蓋全部科學(xué)學(xué)科的主要刊物，學(xué)科范圍從氣象學(xué)和動物學(xué)到量子物理學(xué)和化學(xué)等。因此，媒體記者每周都關(guān)注這兩家期刊，以便在林林總總的新科學(xué)論文中覓取精華。在某種程度上，科學(xué)工作者也期望通過這兩家期刊為新聞界所獲知。他們?yōu)楹稳绱嗽谝饽匾驗(yàn)榭蒲薪?jīng)費(fèi)的競爭異常激烈，要想比對手略勝一籌，科學(xué)工作者就需要尋求大眾知名度。在專業(yè)期刊上發(fā)表論文可以贏得學(xué)術(shù)界的贊譽(yù)，也可以滿足“不發(fā)表就完蛋”的迫切需要。但是，在《科學(xué)》和《自然》雜志上發(fā)表論文還會帶來不期之喜，那就是，你在上面發(fā)表的論文可能會被《紐約時報》及其他報刊所報道。

科學(xué)工作者不但希望通過《科學(xué)》和《自然》雜志成為知名人士，而且還希望為其他科學(xué)工作者所了解。公眾領(lǐng)域與職業(yè)領(lǐng)域內(nèi)的聲名狼藉兩者之間的界線并不分明。科學(xué)工作者對期刊“兩巨頭”的關(guān)注往往多于其他期刊。如果某篇論文為更多科學(xué)工作者所了解，他們在自己論文中引用它的可能性就更大。被經(jīng)常引用會增加科學(xué)工作者的“影響因子”，而這是衡量論文被同行引用次數(shù)的一個尺度。資助機(jī)構(gòu)用“影響因子”來粗略衡量他們考慮資助的科學(xué)工作者的影響力。兩家雜志的編輯稱，因?yàn)椤蹲匀弧泛汀犊茖W(xué)》雜志受注目的程度更高，提交給它們的論文數(shù)量與日俱增。目前《自然》雜志每年收到１萬份稿件，且這個數(shù)字還在增長，該雜志主編菲利普·坎貝爾通過電子郵件表示?！斑@一方面反映了全世界科學(xué)活動的增加，”他說，“同時無疑也反映了資助機(jī)構(gòu)和政府日益強(qiáng)調(diào)甚而有時過分地強(qiáng)調(diào)論文在發(fā)表方面的標(biāo)準(zhǔn)，如各期刊的平常引用率?！?/p>

無論原因如何，對許多科學(xué)工作者來說，《科學(xué)》和《自然》雜志的編輯們一時心血來潮就讓人誠惶誠恐。無論哪家雜志在考慮發(fā)表某篇論文的時候，都會告知作者不要將消息透漏給媒體，除非他們不怕退稿?！懊總€科學(xué)工作者都對它們愛恨交加，”一位擔(dān)心得罪雜志編輯而不愿透露自己身份的著名科學(xué)家說道。“我們恨它們是因?yàn)樵谶@些雜志上發(fā)表論文太過于倚重人際關(guān)系。說實(shí)話，對他們接受和退掉的一些稿件我都感到很吃驚?！?/p>

人們拼命想在這些雜志上露面是否會影響到雜志的打假能力是另一碼事。事實(shí)上，要發(fā)現(xiàn)造假極為困難。我們來看看《科學(xué)》雜志評估黃禹錫２００５年