386 Shaking Up Science Shaking UpScience and question theethostheir of profession Two journal editors takeahard look athonesty in science they saw unwelcome an as Dis- science. academic in transformation own achievements aside, the two had nagging worries about what nothingmore. intoeach other a handful oftimes, but bumped the Bronx, New York (Casadevall). They were acquaintances, having ton, Seattle, (Fang) and the Einstein Collegeof in Albert various and tenure securing University the at accolades of Washing- the ranks ofmicrobiology and , large labs and running get paid todoresearch.” science,” in acareer about Casadevall says. “Ididn’t know you could ald’s spending money. and as a bank teller toearn “I never thought New York fi the because yearrst was there worked and free McDon- at never left. He enrolled at Queens College ofthe CityUniversity of at was 11, reunited with his family inNew York, and tor, Harvard attended and University.Casadevall fl the for ed voice,” Fang admits.) Fang grew upinLos Angeles, the son ofa doc- were toimmediately a couple of times that I failed recognize his They live and barely a continent apart speak onthe phone. (“There FANGFERRICCASADEVALL ARTURO AND UNLIKELY AREAN DUO. NEWS Disenchantment brought them together almost 5 years ago. Their Despite their differences, theyrose onparallel tracksthrough LeoTolstoy A FAVORITE BOOK: HOBBY: EDITS JOURNAL: of Medicine inNew York Albert Einstein College HOME INSTITUTION: microbiologist Immunologist and OCCUPATION: AGE: Arturo Casadevall Anna Karenina 55 Reader ofhistory FOCUS

mBio

, 25 JANUARY25 2013 VOL 339 P u b l i s h e d SCIENCE b we can? And if not,what’s inour power tochange? oneself. than They wanted ask: to weAre way best the science doing fi ofsomething bigger rst place, the thrill of the chase, ofbeing part something about it—torecapture what brought them toscience inthe party.cocktail a at passions fessional focused on narrow fields and often couldn’t communicate their pro- was Funding in high-impact journals. scarcer than ever. Scientists covery for its own sake was being sidelined by a push to publish of bad scientific behavior, which Fang has been studying indepth. or the nextpressure to score the next publication, grant and highrates todiscuss something else: the toxicwhom mixof he met preferred oxide bacteria. and days, these happens often As with researchers the the University of Pennsylvania, where he’s just given a talk onnitric in a Philadelphia coffeemorning shop with music blaring, steps from says. Wonkish and graying at the temples, he’s sitting one November are really scientists, of majority the 99%, “The driven by fear,” Fang The partnership y

A A None ofthis is new. But Fang and Casadevall do to decided totry Fang and Casadevall came together in2008.Fangwas, and still A S

www.sciencemag.org happy towrite to [us].” they but were letterthe tojournal, “Very few people wanted writeto a back about these essays,” Fang says. “We gettingwonderful started feed- would(NIH) makedifference. much Healthof Institutes National the at proposed changes topeer review their salaries and questioned whether entists’pay to dependence ongrants or Lipstick a on Pig?” explored sci- Review WeReform—Change Need, in early 2009 and titled “NIH Peer fitheir published commentaries, rst of One area. the in leadership for ger was Arturo.”ward science. “The one who stepped for- editors for ideas” about the state of Fangsays. badgering my “I started privilege of writing opinionpieces,” tion, asked to head upa new ASM publica- well. as there was (He subsequently Casadevall(ASM). was editor an American Society for Immunity is, editor-in-chief of The two quickly recognized ahun- mBio , a journal published, a journal by the .) “I realized that I had this this had I that realized “I .) and

CREDIT: ALBERT EINSTEIN COLLEGE OF MEDICINE

Downloaded from www.sciencemag.org on January 24, 2013 CREDIT: PHOTO BY MATTHEW RYAN WILLIAMS Online (http://scim.ag/pod_6118). Jennifer Couzin-Frankel sciencemag.org their tenure. Why didthe Mori case drive Fang inan obsessive new hasn’t that there out science of corrected,” been now he believes. Suddenly he realized that probably wasn’t the case. “There’s a lot asmany science, that assumption the like say, to self-correcting. is lished, settingoff an inquiry. Until then, Fang had operated under happened to recognize some figures as having been previously pub- uncovered by almost reviewer peer chance—a journal the for university. the as clusion papers. the retract to agreed Mori by pixel analysis of the fi con- gures and came tothe same unfortunate of VirologyJournal Mori’s papers. Three more appeared in ASM’s lowing a tipfrom another journal. dozens of papers by virologist NaokiMori fol- to Fang, the university had been investigating of the Ryukyus inOkinawa, Japan. Unbeknownst manuscripts,”other University the at adean wrote manuscript as well as the use offi guresused in found repeated use of the same fi gures within the published inthe [sic] to the manuscript indicated below, which was would send the two ona different path. ning when he received an unsettling e-mail that poet and I’m the prose,” Fang says. Casadevall always defer toFerric. He’s always right.” Square. “Whenever there’s I a disagreement, eggs andtoast at a diner near New York’s Times arms,”in Casadevall says over of brunch arecent friendship developed. comrade “I found a great fltorials ew by coast to coast from adeep e-mail, Fang kept pushing forward. As draftsof their edi- applications since then had been successful. although she was awarded tenure, none of her grant quote from her letter, toldFang inNovember that tion is‘not just me.’ ” fi and affiall the rumors rmed that the situa- rmed down on my desk and cried. Your con- article after reading your I putmybut commentary head and Casadevall. “I am notan emotional person, notime topublishbut them,” she wrote toFang modest research award. “I have results … great in 3years and was subsisting number on a 11, applications She had submitted 10failed grants. ure was contingent onsecuring two major NIH versity inMorgantown. At the time, her ten- Downs, at researcher asleep West Virginia Uni- Still, many editors encounter research misconduct during journal Fang was shaken by the experience. The image manipulation was Infection andImmunity “We you are writing toinform that inregard “We really see eye toeye, but is the Arturo Inspired by letters like this one, Casadevall and Montgomery-Downs, whoallowed One note came from Hawley Montgomery- with author Podcast interview . Digital data experts at . Digital data experts a pixelASM performed Infection andImmunity his offihis ce early eve- one ized, Fang was working in how character- is research review, and science, basic essays onpeer ruminating jointseveral bers. After Fang the num- crunches massages words, the and had published six of www.sciencemag.org Science , we we , to to

SCIENCE A FAVORITE BOOK: HOBBY: EDITS JOURNAL: HOME INSTITUTION: OCCUPATION: AGE: Ferric Fang Blood P u b l i s 55 h e d Jazz musician b

1979, he was startled to learn that the stipend he would1979, he was to learn startled receive was University admitted himoff thewaitlist to its M.D./Ph.D.program in borders. national across transfer could that a trade NewWhen York practice law inthe United States, encouraged Casadevallto develop was and camp prison aCuban in time spent unqualifi deemed toed for his ownand gratitude Hisfather, good fortune. who an attorney Casadevall’s worldview is shaped by his experience asa Cuban exile The problem with Arturo.” “is back. tion “It’s not with me. The problem,” he says tongue-in-cheek, direction? “So what’s wrong with me?” he asks, echoing the ques- y

VOL 25JANUARY339 2013 A A A S Microbiologist

Infection and Immunity Infection

Four Quartets University of Washington,Seattle University , T. S.Eliot NEWS FOCUS 387

Downloaded from www.sciencemag.org on January 24, 2013 NEWSFOCUS

greater than what he’d been earning at four jobs combined. He asked that examined 21 surveys of research misconduct. Pooling the to enroll in June rather than September partly to start drawing that data, the author, an Italian researcher at the University of Edin- paycheck. The school agreed. burgh in the United Kingdom named Daniele Fanelli, concluded Like Fang, Casadevall began his scientifi c career full of idealism. that 14% of scientists said that they knew of a colleague who But as time passed, he grew troubled by the reluctance of scientists had falsified data. About 2% admitted committing misconduct to study how they perform their craft—the science of science, as it themselves. were. “We don’t look at our own belly button,” he says. There’s virtually no information about ques- The drivers tions he considers critical to running an effi cient Fang and Casadevall have given much thought research enterprise. Are prizes helpful or harmful to what’s behind bad behavior. Both now detect to science in general? What’s the optimal size of a ì signs of a system they consider fl awed. For exam- lab? How common is research misconduct? We donít look ple, faculty applicants to their respective depart- Casadevall and Fang chose to tackle the last at our own ments are invited for interviews only if they’ve question fi rst. And they hit upon a rich source of been first authors on a publication in a high- data to help them: the scientifi c literature, includ- belly button. profile journal. “We defer to the editors of ing a treasure trove of high-profi le papers and óArturo Casadevall Science and Nature to tell us what’s good,” retractions spanning decades. Casadevall laments. These days, “you get a fi nd- The two fi rst explored whether there was any ing and the whole discussion is not about the connection between a journal’s impact factor—a fi nding, it’s where you’re going to publish.” ranking based on citations of papers published there—and its retrac- An underlying issue is funding. Fang’s father, a physician-scientist tion rate. They speculated that the more prestigious a journal, the working in the 1960s, told him that back then, half or more of grants more likely scientists might be to cut corners, or even fudge data, were funded. “He always felt that the challenge in science was sci- to get their work published in it. They searched the biomedical lit- ence itself,” Fang says. “The level of competition has changed dra- erature database PubMed for retractions in journals with a range of matically. … If you talk to any student or postdoc, they’ll say the

impact factors and found a robust correlation. The pair published picture they’re getting, the name of the game, is to get money.” Sitting on January 24, 2013 their “retraction index” in in August 2011. on promotion and tenure committees, Fang has watched colleagues They weren’t the fi rst to uncover this connection, but the paper pay “lip service” to teaching and quality of science, but the real made a splash: The retraction index was republished in newspapers yardstick is the applicant’s funding levels. In part, this is because and magazines worldwide, including this one. In April 2012, The universities now depend heavily on “soft money”—grant funding— New York Times featured Fang and Casadevall’s work on retractions, to support their own infrastructure. further elevating their profi le. “The most productive scientists are still worried because they Their next project was more ambitious. The pair wanted to quan- have a lot of mouths to feed,” Fang says. Fear struck him a few years tify scientifi c misconduct as best they could in the published litera- ago when his own funding situation grew dire and half a dozen indi- ture. For assistance they recruited R. Grant Steen, a medical writer viduals in his lab were at risk of losing their jobs. They were saved— www.sciencemag.org in North Carolina. Snatching time in airports, on airplanes, and after for the moment—by support from the federal stimulus package. “It’s all about money,” Fang says. “How can you be sure that you get money?” The answer comes back to publications—and sometimes skirting the rules to get them. Fang began his journey with Casadevall thinking cheaters were The 99%, the majorityì of scientists, inherently different from the rest of us. Now, he appreciates how a

toxic environment can subtly encourage bad behavior. One infl uence Downloaded from are really driven by fear. has been the writings of Dan Ariely, a social scientist at Duke Uni- óFerric Fang versity in Durham, North Carolina, who studies cheating. “We have a tendency to point fi ngers, we have a tendency to identify some peo- ple and say, ‘These are bad people,’ ” Ariely says. But “it’s unrealistic to create a system that tempts people and expect them to behave very hours, they assembled an enormous Excel fi le of every retraction they well.” Ariely cites an example from his own history: As a teenager, could fi nd in PubMed, more than 2000 dating back to 1977. he was badly burned and spent years receiving treatment. One of They cross-referenced many retractions with other sources, such his favorite physicians pressured him to tattoo the right side of his as reports from the U.S. Offi ce of Research Integrity (ORI), which face to give the appearance of stubble, which had been erased by the investigates misconduct. The three attributed about 67% of all the burns. “And then I found out I was going to be the third patient in the retractions to scientifi c misconduct, including fraud and plagiarism. paper,” Ariely says, and that the doctor needed a critical mass of vol- The results were published in October 2012 in the Proceedings of the unteers in order to publish his research. “This was an amazing phy- National Academy of Sciences. sician who took great care of me for 3 years, … but at that moment, “We never anticipated that the problem was going to be so wide- he wanted the paper out.” spread, ever,” says Casadevall, who’d expected honest errors to This week, Fang and Casadevall published their latest missive in explain the vast majority of retractions. “We need to clean up our act.” mBio: With Joan Bennett, a prominent microbiologist at Rutgers Uni- Retractions remain rare, at about one out of every 10,000 papers. versity in New Brunswick, New Jersey, they analyzed ORI reports But “even a single retracted paper for fraud can be very damaging to determine whether men were found guilty of misconduct to a dis- to the credibility of science,” Fang says. The actual problem rate proportionate degree. At the senior level, the gender imbalance was is much higher.” Backing him up is a 2009 paper in PLOS ONE dramatic. The three reviewed the ORI fi les of 72 faculty members,

388 25 JANUARY 2013 VOL 339 SCIENCE www.sciencemag.org Published by AAAS NEWSFOCUS and only nine of them were women. That’s one-third of what one would predict based on female representation in the life sciences. Among trainees, the gap narrowed. Fang and Casadevall speculate that the inclination of male principal investigators to cheat tracks the We defer to the editorsì of Science increased likelihood of men to engage in risky behavior, well docu- mented in the social science literature. Younger scientists—male and and Nature to tell us whatís good. female—may cheat to please their boss, or because of pressure to óArturo Casadevall arrive at certain results. “Scientists love to think that they are totally objective,” when in fact they’re often not, Bennett says. “I think it’s very important to look at these questions.” He’s in discussions with Albert Einstein College of Medicine about Fang and Casadevall admit that they’re trapped in the system that “putting the Ph”—philosophy—“back in Ph.D.,” and launching a troubles them. “I think it’s crazy to focus so much on impact factors,” graduate track that includes training in epistemology and meta- Fang says. “But I have a postdoc right now who has a great story, and physics—or as Casadevall puts it, “How do you know what you we’re going to try to submit his paper to Nature. … I see the rules as know?” and “How much can you push your lab?” they are, and I’m not going to sacrifi ce his career.” At the same time, He’s also disillusioned by peer review, which he believes yields the two are trying to modulate how they run their labs and mentor endless demands to add data to a paper without necessarily improv- their students. Casadevall, ever the broad thinker, urges lab members ing it. At the open access journal Casadevall runs, mBio, the rule is to read widely outside their fi eld. Fang has postdocs working on mul- that papers are either accepted or rejected, period. tiple projects at once, with the hope that something will pan out and Fang and Casadevall know that they can’t come up with all of they’ll be under less pressure. Fang’s partnership with Casadevall has the answers. Rather, their goal is to start a conversation and hope also changed how he’d react to fraud in his own lab. If a trainee faked others identify solutions. “We have to somehow change the incen- data, Fang says, “I would question myself, that I had failed strategi- tives,” says economist Paula Stephan of Georgia State University cally and not created the right environment for them, and they felt in Atlanta. Her book, How Economics Shapes Science, examines

afraid of failure.” the ways in which scientists and institutions compete for resources on January 24, 2013 Despite the stresses they face, most scientists, of course, don’t and rewards. “Historically, a lot of the criticism [of the scientifi c cheat. Even as Fang looks inward and contemplates sweeping enterprise] … comes from people outside science,” says Stephan, changes to the system, he doesn’t absolve individuals who succumb who met Casadevall and Fang at a Health Research Alliance event to its temptations. At ASM, Fang worried about trusting Mori’s work in Washington, D.C., last year. Although she doesn’t agree with them in the future, and argued that he should be barred from publishing on everything, “it’s very exciting when you see people like the two of ever again in ASM journals. He was outvoted them, who are editors of journals, really begin- in favor of a 10-year ban. ning to question the system.” Casadevall and Fang are shifting gears now, The solutions moving away from misconduct and into other www.sciencemag.org With every joint publication on the state ì issues that may prove tougher to tackle quan- of science—they have 14 so far—Fang and Itís unrealistic to titatively. Casadevall hypothesizes that prizes Casadevall see more hunger in the com- create a system are detrimental to science, because they fos- munity to hash out these topics. Casadevall ter a “winner take all” system and reward cut- travels constantly. He spent parts of October that tempts people throat behavior, rather than cooperation that and November in Michigan, London, Paris, and expect them to might better advance knowledge. He is cur- and Chile. Everywhere, the conversation was behave very well. rently cataloguing all the Nobel prizes and Downloaded from the same. assessing which were said to have left out Scientists, especially younger ones, “feel óDan Ariely, potential awardees. Fang is considering using powerless,” he says. “The older group is wor- Duke University data from ASM journals to ask how often peer ried, surprised” by the misconduct fi ndings. review changes the substance of a paper. “We Fang and Casadevall find themselves in are going to continue to take on question after increasing demand. Fang participated in a question,” Casadevall says. roundtable on scientifi c integrity last month at the National Acad- Meanwhile, the two must stay abreast of their day jobs: Editing emy of Sciences in Washington, D.C. Together they’re writing an a journal each, running large labs, in Fang’s case directing a bus- article for Scientifi c American Mind on cheating. tling clinical laboratory, and in Casadevall’s sitting on a national bio- Fang and Casadevall have heard concerns from researchers that defense advisory post. When Casadevall worries that he’s stretched their work will be used to discredit science. It’s been cited on anti- too thin, his 89-year-old mother, who lives in Queens, urges him science blogs, like those questioning the safety of vaccines or the onward. “She assures me there will come a day when nobody’s going role of human activities in climate change. While they’ve considered to invite me anywhere. … She says to me, ‘Don’t turn down an invi- the risks, “I think we need to have this conversation to try and make tation.’ ” He grins. “If my mom tells me to do it, I’ll do it.” science better,” Fang says. Then Casadevall, his glasses folded neatly and hanging from his Casadevall favors a more generalized science education, shirt collar, turns serious. “I really do think that what Ferric and I rather than the extreme specialization that now occurs in gradu- try to do may be the most important thing I do in my life,” he says. ate school. An enthusiastic reader of history, he points out that in Others, he knows, will keep building the edifi ce of science. These the 19th century and before, scientists such as Isaac Newton and two want to shake its modern foundations. Gottfried Leibniz were philosophers fi rst and scientists second. –JENNIFER COUZIN-FRANKEL

www.sciencemag.org SCIENCE VOL 339 25 JANUARY 2013 389 Published by AAAS