Scientific Misconduct Author(s): David Goodstein Source: Academe, Vol. 88, No. 1 (Jan. - Feb., 2002), pp. 28-31 Published by: American Association of University Professors Stable URL: http://www.jstor.org/stable/40252116 . Accessed: 23/03/2013 10:15

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp

. JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

.

American Association of University Professors is collaborating with JSTOR to digitize, preserve and extend access to Academe.

http://www.jstor.org

This content downloaded from 128.252.15.196 on Sat, 23 Mar 2013 10:15:34 AM All use subject to JSTOR Terms and Conditions /';-=09 )(8* =-0/'] i|||P|lfpElp|IHP^| HBlHHIIIHW

By David Goodstein

careerin scientificmisconduct began lectures,written articles,and taughtcourses about it. I have more than a decade ago. That's when I also draftedregulations, seen them adoptedby my institution realizedthat federalregulations would (the CaliforniaInstitute of Technology), copied by other uni- soon make it necessaryfor all universi- versities,and, much to my dismay,put into action in a high- ties to develop formalrules about what profile case at Caltech. to do if the unthinkablewere to hap- During that case, I had the remarkableexperience of seeing pen: that scientistsat their institutionswould be suspectedof a skilledlawyer, with a copy of my regulationshighlighted and fraudulentlymisrepresenting the resultsof an investigationor underlinedin four colors, guide participantsin following every the proceduresneeded to replicatethose results. word I had written, whether I had meant it or not. Through Since then, scientificmisconduct has become a virtualaca- all of that, I have learnedthings about conduct and misconduct demic subspecialtyfor me. I have given in that I would like to sharewith you. Let me begin by statingright up front what I have come to David Goodsteinis viceprovost and professorof believe. Seriousmisconduct, such as fakingdata, is rare.When physics and appliedphysics at the California it does occur, it is almostalways in the biomédicalsciences, Instituteof Technology,where he has been on not in fields like physics,astronomy, or geology, although thefaculty for morethan thirtyyears. In 1995 other kinds of misconductdo happenin these fields. Science is he was named the FrankJ. Gilloon self-correcting,in the sense that a falsehoodinjected into the Distinguished Teachingand ServiceProfessor. body of scientificknowledge will eventuallybe discoveredand

28 ACADEME

This content downloaded from 128.252.15.196 on Sat, 23 Mar 2013 10:15:34 AM All use subject to JSTOR Terms and Conditions rejected.For just that reason,dissemination of falsehoodsis some twenty-six casesof allegedmisconduct that surfaced never the purposeof those who perpetratescientific . between 1980 and 1986 revealedthat twenty-one came from Still, active measuresto protect science are needed, becauseif the biomédicalsciences, two from chemistryand biochem- the recordbecame badly contaminatedby fraudulentresults, it istry, one from physiology, and two from psychology.I don't would no longer be self-correcting. know of any more recent studies,but one cannot help notic- For a long time, the governmentmade a mess of trying to ing that the U.S. Public Health Service'sOffice of protect science. Governmentagencies performed poorly in Integrity,which investigatesmisconduct in biomédical this areapartly because they mistakenlytried to obscurethe research,seems constantlyto be embroiledin controversy.By importantdistinction between real fraudand lesserforms of contrast,the National Science Foundation,which supportsall misconduct. of the , includingbiology, has an inspectorgeneral's In additionto these observations,I have also concluded that office that conductsits businessin relativeanonymity, unmo- we scientistsare complicit in presentingto the public a false lested by seriousattention from the press. image of how science works, which can sometimesmake nor- Undoubtedly, multiple reasonsexist for this curious state mal behaviorby scientistsappear suspect. Let me try to of affairs.For example, many of the cases that have arisen explain what I mean by all of have involved M.D.'s, ratherthan Ph.D.'s. To this. an M.D., the welfare of the patient may be more First, a word about terminol- important than scientific truth. In a recent case, ogy. People are touchy about We scientists are a physician in Montreal was found to have falsi- words in this business. a When complicit in fied the records of participantsin a large-scale philosopher colleague and I study of breastcancer. Asked why he did it, he decided to offer a course in this presenting to the said his goal was to get better medical care for subject, we wanted to call it his patients. Most cases, however, arisefrom "Scientific Fraud."But the fac- public a false image more self-interestedmotives. ulty board, in its wisdom, didn't of how science want us teaching thatto our stu- dents, so we had to call it works, which can "Research ." The federal government, in all its gyrations, sometimes make has to this avoid- day studiously normal behavior by ed using the word "fraud" in connection with scientific mis- scientists appear conduct, because in civil law that word has a specific mean- suspect. ing. I, however, am not afraidto call a frauda fraud. Intent to Deceive Fraudmeans serious misconduct with intent to deceive. Intent to deceive is the very antithesisof ethicalbehavior in science. When you reada scientificpaper, you can agree or disagree with its conclusions,but you must be able to trustits account of the proceduresused and the resultsproduced by those procedures. To be sure, minor deceptionsarise in virtuallyall scientific papers,as they do in other aspectsof human life. For example, scientificpapers typically describe investigations as they logi- callyshould have been done ratherthan as they actuallywere done. Falsesteps, blind alleys, and outrightmistakes are usual- ly omitted once the resultsare in and the whole experiment can be seen in properperspective. Also, the list of authorsmay not revealwho deservesmost of the credit (or blame) for the Among the incidentsof scientificfraud that I have looked at, work. Such behaviormay or may not be correct or laudable, three motives, or risk factors,have been present.In all the but it does not amount to fraud.Real fraudoccurs only if the cases,the perpetrators(1) were under careerpressure, (2) proceduresneeded to replicatethe resultsof the work or the knew, or thought they knew, what the resultwould be if they resultsthemselves are in some way knowingly misrepresented. went to all the trouble of doing the work properly,and (3) As I have noted, this kind of misbehaviorseems to be were in a field in which individualexperiments are not restrictedlargely to the biomédicaland closely relatedsci- expected to be precisely reproducible.Simple monetary gain ences. A studyby Princetonsociologist PatriciaWoolf of is seldom, if ever, a factorin scientificfraud.

JANUARY-FEBRUARY2002 29

This content downloaded from 128.252.15.196 on Sat, 23 Mar 2013 10:15:34 AM All use subject to JSTOR Terms and Conditions It is, however, by no means true that Foundation decided to do something fraud always takes place when these about scientific misconduct. Their three factors are present; quite the view of what constituted serious mis- opposite, they are often present, and conduct in science, published in the fraud is rare. But they do seem to be FederalRegister, differed greatly from there whenever fraud occurs. Let's take what I have outlined in this article. them one at a time. Both agencies defined scientific mis- The first, career pressure, is clearly a conduct to be "fabrication, falsifica- motivating factor, but all scientists, at all tion, , or other practicesthat levels, from fame to obscurity, are pret- seriously deviate from those that are com- ty much always under career pressure. monly accepted within the scientific commu- Regarding the second risk factor, if sci- nity for proposing, conducting, and entific fraud meant knowingly inserting reporting research." (Emphasis an untruth into the body of knowledge, added.) it would be nonexistent or of little con- Controversy swirled around the cern, because science is self-correcting. statement from the moment it Scientific fraud always involves a trans- appeared. Yet both agencies issued it gression against the methods of science, as a "final rule" in 1990. No one took never against the body of knowledge. issue with their inclusion of "fabrica- Perpetrators think they know how an tion, falsification, [and] plagiarism" experiment would come out if they did it (referred to as "flp" by the cognoscenti) properly, and they decide against going to in the definition of misconduct. The all the trouble of doing it right. controversial part was the catch-all phrase The most obvious seeming counter- "practices that seriously deviate from example to this assertion is Piltdown Still, the belief that those commonly accepted." To many Man, a human skull and ape jaw planted someone else can scientists and other observers, it raised in a gravel pit in England around 1908. the horrifying specter of the government If ever a fraudulent physical artifact was repeat an experiment forcing scientists into some preconceived inserted into the scientific record, this mold of orthodox thought. was it. Yet the perpetrator was possibly and get the same In the rules developed at Caltech, we trying only to help along what was result can be a fearlessly called scientific misconduct known or thought to be the truth. "research fraud" and defined it in a way Prehistoric remains had been discovered powerful deterrent to essentially equivalent to "rrp"with no in France and Germany, and rumors catch-all phrase. One day, we received a even pointed to findings in Africa. cheating. letter from the Public Health Service Surely, human life could not have started informing us that we were not in com- in those uncivilized places. As it turned pliance with the agencies' definition, and out, the artifact was rejected by the body asking us to submit new rules within of scientific knowledge. Long before ninety days. Eighty-nine days later, I 1954, when modern dating methods showed it to be a , wrote back, explaining that revising rules in a university was growing evidence that our ancestors had ape skulls and not so simple, that changes had to be approved by various human jaws made an embarrassment at the bodies, and so on. The Public Health Service told me to send fringes of anthropology . the revised rules when we finished. The heat thus being off, I Regarding the third risk factor, , experiments referred the matter to the faculty's Academic Policies are seldom repeated by others in science. When a result is Committee without expressing any sense of urgency. A few found to be wrong, it is almost always because new work years later, before the committee had a chance to act, the gov- based on the result doesn't proceed as expected. Still, the belief ernment definition changed. that someone else can repeat an experiment and get the same Holding on to our own definition turned out to be of genuine result can be a powerful deterrent to cheating. This possibility importance. Twice during the. intervening period, tenure review appears to be the chief difference between biology and the committees at Caltech received letters accusing the person under - other sciences. Biological variability the fact that the same review of committing scientific misconduct. Letters like that can procedure performed on two organisms as nearly identical as touch off panic in a tenure review committee. What to do? In - possible is not expected to give exactly the same result may both cases, the committee chair came to me, as the resident provide cover for a biologist who is tempted to cheat. This last expert, to ask exactly that. I was able to say that even if the accu- point, I think, explains why scientific fraud is found mainly in sation were proven, it would not amount to serious misconduct the biomédical area. under the Caltech definition. Both candidateswere promoted Around 1988 the Public Health Service (the parent of the and are today among our brightest stars.I could not have done as National Institutes of Health) and the National Science much under the original government definition.

30 ACADEME

This content downloaded from 128.252.15.196 on Sat, 23 Mar 2013 10:15:34 AM All use subject to JSTOR Terms and Conditions In the wider academicworld, that definition causedmuch mous to authors. Funding agencies sometimes do the same, discontent.Commissions and committeeswere formed, hear- especially for small projects. ings were held, alternativedefinitions were proposed,all with- Peer review is a good way to identify valid science. It was out seriousconsequence. Then, nearlyten yearsafter the final especially useful when the only limit to scientific progress rule was issued,a taskforce under the directionof President was the number of good ideas available.It is not, however, Clinton's science adviserproposed a new formulationto be well suited to adjudicatingan intense competition for scarce applieduniformly by all federalagencies. resources, such as researchfunds or pages in prestigious Like the original rule, the new rule, adopted in 2000, journals. The reason is obvious enough. The referee, who is defined scientific misconduct to be fabrication,falsification, usually among the few genuine experts in a field, is a or plagiarism,but with the "f," "f," and "p" each carefully competitor for those same scarce resources.Most scientists defined. In addition, a practice would have to deviate seri- have high ethical standardsand try not to let their self- ously from what is generally accepted within the scientific interest interfere with their scientificjudgment. But every community to constitute misconduct, thus turning the "seri- scientist I know has war stories of having been mistreated ously deviate" phraseon its head. Now, ratherthan a catch- by anonymous referees. Because the refereesperform a all phrase,we have an additionalbarrier that must be over- professionalservice, almost always without pay, they are come to prove misconduct. Moreover, misconduct would never held to account for their actions. The temptation to have to be committed knowingly or with recklessdisregard find fault with a rival's efforts must sometimes be irresistible. for the truth- an essential element of the forbidden term Misconductof this kind is, I fear,rampant in all fieldsof sci- "fraud"- and proved by a preponderance of evidence. ence, not only biomédicalscience. Recently, in a presentation Within weeks, Caltech adopted revised to a largeaudience of mostlyyoung rules in precise compliance with the new researchersat a prestigiousuniversity, I government rule. Scientists are not outlinedthe crisisin peer review. The moderator,a famoussenior scientist, was New Pressures disinterested truth incredulous.He askedthe audiencehow Unfortunately, instancesof scientific mis- many disagreedwith my heresy.No one conduct may not remain as rare as they seekers; they are responded.Then he askedhow many have been in the past. Throughout most more like in agreed.Every hand in the house went up. of its , science was constrainedonly players Many of us in my generationwant to by the limits of its participants'imagina- an intense, winner- believe thatnothing importanthas tion and creativity. In the past few changedin the way we conductscience. decades, however, that state of affairshas take-all competition We arewrong. Businessas usualis no changed dramatically.Science is now held for scientific longer a realoption. back mainly by the number of research prestige As I mentioned earlier,I believe we posts and the amount of research funds and the resources scientistsare guilty of promoting, or at available.What had been a purely intel- least tolerating,a falsepopular image of lectual competition has become an intense that follow from that ourselves.I like to call it the Myth of struggle for scarce resources. In the long the Noble Scientist.It arises,I think, run, this change, which is permanentand prestige. out of the long-discreditedBaconian irreversible,will probably have an unde- view of the scientistas disinterested sirableettect on ethical behavior among sci- seeker of truth who gathersfacts with mind cleansedof preju- entists. Instances of scientific fraud will dices and preconceptions.The ideal scientist,in this view, almost surely become more common, would be more honest than ordinarymortals, certain- as will other forms of scientific ly immune to such common human failingsas pride misbehavior. or personalambition. When people find out, as they The institution of peer review, for invariablydo, that scientistsare not at all like that, they example, is now in danger. Scientific may react with understandableanger or disappointment. journals rely on peer review to Most scientistsare rigorouslyhonest about what reallymat- decide what to publish, and ters to them, like the accuratereporting of proceduresand grantingagencies depend on it data.In other areas,however, such as disputesover priorityor to decide what research to credit, they tend to behave like the ordinarymortals they are. support. Obviously, sound Scientistsare not disinterestedtruth seekers; they are more decisions in these areas are like playersin an intense, winner-take-allcompetition for crucial to the proper func- scientificprestige and the resourcesthat follow from that tioning of science. prestige.The sooner we admit to these factsand learn Journal editors usually to distinguishbetween seriousscientific misconduct send manuscript sub- and common human conduct by scientists,the better off missions to referees we'll be. who remain anony-

JANUARY-FEBRUARY2002 3 1

This content downloaded from 128.252.15.196 on Sat, 23 Mar 2013 10:15:34 AM All use subject to JSTOR Terms and Conditions