Science Police

For ten years, two zealous, self-appointed investigators of scientific fraud made headlines--and enemies. Now they have been reined in, yet the questions remain: what were they after? And what did they gain?

By Edward Dolnick|Tuesday, February 01, 1994
Some people have perfect pitch. Others can shuffle cards with one hand or turn a somersault in midair. Walter Stewart is a virtuoso of a different sort: he has an extraordinary talent for annoying others. "I can't think of anyone who has as many people angry at him as I do," he says matter-of-factly.

Stewart professes to find the anger surprising. Not likely. He and his longtime colleague Ned Feder, both employed as researchers at the National Institutes of Health, have made a career of pursuing fraud in science. No scientist welcomes the message that science--which is, after all, a search for truth--is itself tainted. Coming from these messengers, however, the message is that much less welcome.

Stewart and Feder are, in the words of one representative critic, "zealots" and "vigilantes" and "self-appointed policemen." Of course fraud is wrong and should be punished, skeptics say. The problem isn't Stewart and Feder's crusading but their recklessness and self-righteousness. Washing dirty linen in public is one thing. Gleefully ferreting out laundry from under the neighbor's bed is another. Besides, they ask, if such investigation is necessary, why should these two freelance sleuths do it? Why not leave the chore to the Department of Health and Human Services' Office of Research Integrity, which is charged with the responsibility? "It seems to me the entity constituted to look into these things ought to look into them," says J. Edward Rall, an NIH biologist and its former director of intramural research, "and people hired to do research ought to do research."

Stewart and Feder pounce on such objections. They take every opportunity to embrace the doctrine that a free and open debate is essential to the health of science. "I find it preposterous that people ask what I have to do with these cases," Stewart sputters. "It's like my asking what they have to do with the star they're studying. It's a scientist's job to evaluate information, form opinions, and then tender them to the public."

The case in which their opinions were tendered loudest and longest was the David Baltimore affair, a tangle unresolved to this day. The matter centered on allegations of fraud in a paper of which Baltimore-- winner of a Nobel Prize in 1975 for his discovery of the reverse transcriptase enzyme--was a coauthor. Stewart and Feder played a key role in publicizing those charges.

Had not John Dingell, the formidable congressman from Michigan, decided to get involved, the whole to-do might have remained a scientific family squabble. But in 1988 Dingell held hearings to air the anti- Baltimore charges. Those hearings culminated a year later in a clash between Dingell and Baltimore, powerhouse politician against Nobel laureate, King Kong versus Godzilla. Although ultimately Baltimore would be the big loser--forced by the scandal to resign his presidency of Rockefeller University--he initially won the public relations battle at the hearings by painting Dingell's investigation as an indefensible intrusion of politics into science. Stewart and Feder, who (with NIH approval) had spent two years working with Dingell, were damned by many of their colleagues as traitors to the cause of science.

More recently Stewart and Feder have been in the news because of their "plagiarism machine," a computer program that gobbles down texts by various authors and then points an accusatory finger when it finds suspicious matches. In February 1993, Stewart and Feder sent the American Historical Association 2,100 pages of documentation alleging plagiarism in three works by biographer Stephen Oates. They detailed, for example, 175 passages they said were lifted from a 1952 work by historian Benjamin P. Thomas for use in Oates's 1977 biography of Abraham Lincoln. Oates protested that he was not a plagiarist and that Stewart and Feder had no business laboring to show that he was; as a historian, not a scientist, he received not a penny in government support.

Last April, apparently in response to Oates's complaints, the NIH put an end to Stewart and Feder's full-time dedication to detection. The two researchers were reassigned to other duties, their offices and files locked. Stewart retaliated by going on a hunger strike, abandoned only after 33 days and considerable media attention. The fast was originally someone else's idea, Stewart says, "but the moment I heard about it, it seemed completely natural." He grows indignant at the suggestion that a hunger strike in response to a mere job squabble seems a bit much. "It's a moral issue, not an administrative one. You don't try to silence scientists." At this writing, Stewart and Feder are still protesting their reassignment; but for now they have been split up like a pair of naughty fifth graders, and they are forbidden to investigate misconduct during work hours.

For such notorious figures, Stewart and Feder were slow in finding their ultimate calling. From 1968 to 1983 the two were conventional scientists tucked into the bowels of the NIH. Stewart spent about half of those 15 years developing a dye to stain neurons, while Feder studied cell behavior in mice. Then for five years Stewart and Feder worked together studying nerve cells in snails. They labored in a lab that looked like a mad apothecary shop, with shelf after shelf lined with jar upon jar of snail upon snail. For all the attention this work received, the two men might have been enrolled in a federal witness protection program.

But then ten years ago, with the grudging approval of the NIH (whose philosophy is to let its scientists wander where their curiosity leads), Stewart and Feder began looking into fraud. Their first venture quickly made them famous in the science community. The case involved John Darsee, a Harvard cardiologist who used fraudulent data in roughly 20 papers.

Stewart and Feder's research didn't expose Darsee; that had already been done by Darsee's colleagues. Instead they focused on the work of Darsee's coauthors, and it seemed that everywhere they looked they found improprieties. The duo's paper barely made it to print. For four years it wandered from journal to journal looking for a home, while libel lawyers fought to keep it from being published. Undaunted, Stewart and Feder rewrote their exposé a remarkable 100 times. Finally, in 1987, their chapter-and-verse accounting of "lapses from generally accepted standards of research" appeared in the influential British journal Nature. Some "lapses" were simply errors, ranging from silly to glaring. Others were more serious: authors reported data they knew to be false, packaged old work under new titles to pad their résumés, or listed "honorary coauthors" whose names granted credibility to work they had barely participated in. Everyone knew such things were done, but no one before Stewart and Feder had made the case in public and in detail.

As a consequence of their newly won fame, other cases began coming Stewart and Feder's way, and the sleuthing that they had pictured as a onetime diversion from their real work became their real work. "I used to paddle my canoe on an absolutely flat lake," Stewart says. "The work was fascinating but nothing changed from day to day. Then it was like white water and the landscape changed every single hour."

Stewart, at 48, is nearly a generation younger than the 65-year- old Feder, but he is the conspicuous member of the pair. They met at Harvard in 1966, when Stewart took an undergraduate biology course taught by Feder. Except for a one-year stretch when Stewart was in graduate school at Rockefeller, they have worked together ever since--that is, until their recent reassignment. While they don't complete each other's sentences, as the cliché would have it, their conversational style does bear telltale marks of countless hours spent together. When one talks, the other continually echoes and interrupts and endorses his thoughts. It is an odd effect, the academic equivalent of a church congregation's calling out "Amen!" at critical junctures. "The answer is not to build a wall to keep out Congress," says Feder. "No!" cries Stewart. "The answer is to fix the problem--" "Absolutely right!" "--that interested them in the first place."

The two partners appear sharply different. Stewart has managed to carry a teenage gawkiness well into middle age. Padding across the floor in mismatched blue socks, with large iron-on patches on his pants and an elastic strap holding on to his glasses, he still looks and sounds like the brightest boy in high school. Ask a question and he all but waves his hand in the air and calls out, "Me, I know, ask me, I know!" Take too long posing the question and he emits a staccato burst--"right, right, right, right, right"--in a desperate attempt at hurrying you along.

It is not only Stewart's eagerness and energy that make him seem boyish. When he talks about the scientific establishment and its self- betrayal and abandonment of its own high standards, he sounds eerily like an idealistic freshman lambasting his parents for their flabby and compromised lives.

Feder is a quieter, calmer man. Tall and thin, with a head of thick white hair, he looks like someone who might be found puttering with tomatoes in a backyard garden or teaching a grandchild how to fish. Where Stewart occasionally stammers in excitement, as if he has more thoughts than he can express, Feder weighs each word as if checking it one last time before sending it out into the world. "I don't like being disliked by a very high proportion of my colleagues," he will say, and then a moment later will reconsider. "In order to avoid the unpleasant echo of 'like' and 'dislike,' I should say, 'I don't enjoy being disliked.' "

Feder's motives for pursuing fraud seem straightforward-- scientists should tell the truth and liars shouldn't get away with their deceit--but Stewart is a more complex character. His manner is almost painfully intense. "Walter is one strange duck," a Dingell staffer recalls fondly. "I think he wore the same pair of pants every day for a year and a half, four sizes too big, with this big belt hanging down. Every single time he went through the metal detector he set it off."

As Stewart talks, he jiggles his legs, clicks his pen, tugs his hair. In midsentence he will hop up on the sofa he has been sitting on and squat on his heels like a catcher waiting for a fastball. Occasionally he scrunches his eyes tightly shut, as if he has just rubbed lemon juice into them. In his wife's succinct description, he is "submanic."

Stewart fends off queries about motivation, his own and other people's. "I used to do nothing but play with test tubes and ask nature questions," he says. "I was a very un-people person, and I'm still not people-oriented in the conventional sense. People are quite a mystery to me. I don't understand them and I never have. Certainly one of the reasons I went into science, and I think it is one of the classic reasons, is that it's about things you can understand--like why the sky is blue. Things like that have an enormous beauty and clarity."

Early on, Stewart's urge to explore the world, and his considerable intelligence, seemed to mark him for stardom. Indeed, one of the few points of general agreement in the fraud debate is that Feder is a bright man and Stewart is a brilliant one. But there is hardly any agreement at all on whether Stewart and Feder have harnessed their brainpower to a worthwhile goal. Part of the problem is that despite countless interviews with reporters they have done a poor job of spelling out their views. Feder tends to defer to Stewart, and Stewart, like Stephen Leacock's fictional Lord Ronald, flings himself upon his horse and rides madly off in all directions. He starts a reply headed one way, remembers a different point and reverses course, and then interrupts himself to change tack yet again, all at top speed. It is hard to extract a coherent philosophy from this torrent of words, like trying to get a sip of water while standing under Niagara Falls.

As a result, and without much resistance on their part, Stewart and Feder have been assigned to a pigeonhole that others have chosen for them. Article after article--Fraud Busters seems to be the preferred headline--has recounted their activities. (In addition to the Baltimore, Oates, and Darsee cases, Stewart and Feder launched two separate investigations involving researchers at the University of Wisconsin: neurophysiologist James Abbs, who allegedly misused the work of a graduate student, and biologist Hector DeLuca, who tried to patent a vitamin D derivative that was first synthesized by someone else. This last case involved the university's then chancellor, who refused to turn over key material regarding DeLuca. That chancellor was Donna Shalala, now, ironically, secretary of the Department of Health and Human Services, which oversees the NIH.)

Newspaper and magazine stories recount other recent instances of scientific misconduct too. Over the past ten years a fair number of cases of scientific fraud--ones not involving Stewart and Feder--have made headlines. There was the case of Barry Garfinkel, a prominent psychiatrist at the University of Minnesota, who was convicted in August of falsifying data in a $250,000 study of antidepressants; Stephen Breuning, a University of Pittsburgh psychologist specializing in drug treatment for the mentally retarded, who was shown to have invented many of his findings; and Robert Slutsky, a University of San Diego cardiologist who produced a remarkable 137 papers over seven years but falsified data. Perhaps the most publicized case of all was that of Robert Gallo, a top NIH researcher who was accused of misrepresenting his role in discovering the AIDS virus; ultimately the Office of Research Integrity dropped the case, but only after it had dragged on for years and generated untold pages of news. Read a stack of such stories and there seems no escaping the conclusion that science is just another corrupt enterprise, as rife with scandal as Congress or the used-car business or the savings and loan industry.

In this atmosphere of science bashing, Stewart and Feder have become icons. Science's other whistle-blowers have tended to focus on one particular scandal that they happened to witness. Only Stewart and Feder have taken on case after case, spearheading half a dozen investigations and playing a supporting role in dozens more. And they have learned to publicize their findings, not only in countless interviews and speeches but also in testimony before Congress.

The unsightly picture they paint supposedly undercuts science's standard self-defense that in science, more than in any other human activity, the truth will out. Businessmen may steal millions and live out their days in Caribbean luxury, gunmen may get away with murder, but fraud in science cannot remain undetected. Not because scientists are more virtuous than other people, but because important and intriguing claims will be poked and prodded from every angle by researchers eager to extend the new findings. Announce a startling but false discovery--the moon is made of green cheese--and today's publicity will give way to tomorrow's retraction.

That is the conventional wisdom. Here's the surprise: Stewart and Feder agree. "First of all," Stewart says, "nobody sensible is alleging that the foundations of science are based on fraud. I don't believe there are any important ideas in science today that are due to fraud. I don't think there ever were for very long, because they don't stand up to the test of replication."

So why all the fuss? Because, says Stewart, even if fraud does not damage the edifice of science, it does damage the profession of science. "The profession is different from the body of received knowledge that we call science," he argues. "The problem is not that fraud is going to retrofit bad ideas into science. The problem is that it makes the profession a bad one."

Scientists can be robbed of credit they deserve or hailed for work they didn't do. Patients can be hurt if their treatments are based on fraudulent findings. And in instance after instance whistle-blowers are threatened or ostracized for telling the truth. "A single case in which a person is treated awfully is a problem," Stewart says, "and lots of cases are a big problem. So no one has to believe that the foundations of science are threatened in order to believe that we have to have appropriate ethical responses to people who allege misconduct."

Stewart then recasts the question of why the subject of fraud preoccupies him. "We all have this responsibility," he says. "The question is not why I take it seriously, but how soon the rest of my colleagues will join me in taking it seriously." The echo of Thoreau is so strong that it is hard to imagine it is accidental. "Henry, why are you here?" Ralph Waldo Emerson is supposed to have asked Thoreau, who was in jail for refusing to pay war taxes. "Waldo, why are you not here?" Thoreau replied.

This is a view that concedes Stewart and Feder the moral high ground, and it is as far as either man likes to follow questions of motivation. Indeed, their dislike of such talk calls to mind Mr. Gradgrind of the Dickens novel Hard Times, who proclaimed, "In this life, we want nothing but Facts, sir; nothing but Facts!" "Eventually people have to address the merits of our argument," Stewart says impatiently. "Whether they decide we're vindictive, nasty, mean, self-interested, glory-seeking, or whatever, they still have to say, 'But is their evidence correct?' "

This is a bit coy. Stewart and Feder are as guilty as any of their critics of blurring the line between a person's career and his character. John Maddox, the editor of Nature, is a lukewarm supporter of Stewart and Feder, but he noted in a 1987 editorial that "they have not understood that the unfettered right to publish scientific data does not equate with a right to denigrate others' characters." Admittedly, the distinction is subtle. Is attacking scientists' work as shoddily conceived and sloppily carried out different from attacking the scientists themselves? How can we know the dancer from the dance?

Stewart and Feder's worship at the altar of "facts" carries with it a deeper problem: it promises objectivity but cannot deliver it. Take their latest venture, the "plagiarism machine." Programmed by Stewart (a self-taught programmer), this computer chops entire texts into 30- character-long stretches and then sets to work looking for matches. The explicit aim is to avoid the messiness and ambiguity of the whistle-blower cases in favor of a neater, document-based approach.

Stewart and Feder hurry to concede that the phrase "plagiarism machine" is a misnomer. Their computer can detect overlaps between texts but not plagiarism; it lacks the judgment to know that two authors utterly unaware of each other might both have quoted the same phrase from, say, the preamble to the Constitution. But they gloss over the real issue. In all but the simplest cases, facts do not speak for themselves. Except in cases of outright copying, evaluating even as homely a fraud as plagiarism requires judging and interpreting and sifting facts rather than simply stacking them to the heavens like a child's blocks.

As an example of what Stewart calls their "fantastically strong" case against Oates, for instance, they cite his sentence: "The two Presidents said little to one another as the carriage bumped over the cobblestones of Pennsylvania Avenue, part of a gala parade." They compare that with the earlier work by Benjamin P. Thomas: "As the open carriage jounced over the cobblestones of Pennsylvania Avenue, Lincoln looked into the faces of the crowd that jammed the sidewalks." Is this really a compelling example of theft?

How to account for Stewart and Feder's zeal? Why devote thousands of pages of documentation to indicting Oates, not a scientist but a historian? Why rewrite their first joint paper on scientific misconduct 100 times in order finally to find a publisher?

Stewart and Feder's disdain of such questions leaves a vacuum, and armchair analysts abhor a vacuum. The mildest theory those analysts propose is sour grapes. Stewart and Feder hardly publish any work of their own, the argument runs, so they spend their time trying to drag down their more successful rivals.

The premise, at least, is correct. Five years' work on snails yielded a few papers on techniques but essentially no reports at all on the snails themselves. "We fell on our faces there," Stewart concedes. In 1988, then NIH director James Wyngaarden called Stewart and Feder's productivity as scientists "extraordinarily low."

The sour grapes theory also comes in a variant form. Here the supposed spur is not envy but lust for publicity. Stewart and Feder do plainly enjoy their notoriety; they are on a first-name basis with reporters at the New York Times and the Washington Post, among many others, and they carefully save their press clippings, which now make a stack nearly a foot thick. (Both say, however, that the clippings are used to educate people about science fraud.) Better yet, many of those stories raise Stewart and Feder to dizzying heights. There, duking it out on the front page on roughly equal terms, are the onetime snail tenders from the basement of the NIH and such titans of science as David Baltimore.

Still another explanation of Stewart and Feder's fraud-hunting passion comes from one of Stewart's old mentors at Harvard, Gerald Holton. Stewart calls Holton "a man I dearly respect and love," and Holton, a physicist and distinguished historian of science, professes "considerable fondness" for his former student. Holton sees Stewart's career as "a tragedy in which a very worthy, very clever, highly motivated, highly moral person went far beyond what was necessary and sacrificed his own talent.

"Remember Savonarola?" asks Holton. "He was the fifteenth-century Florentine priest who wanted to rescue the human race by putting up large bonfires of unwanted books and unwanted people. I think there is a little bit of that feeling in Stewart, that he has to rescue science from its own misdeeds, and unfortunately he has just gone too far."

This is too harsh. Stewart and Feder's criticisms of science never come near the fury of Savonarola's condemnation of the church. ("O prostituted Church," Savonarola thundered, "you have unveiled your abuse before the eyes of the entire world and your poisoned breath rises to the heavens.") On the contrary, both Stewart and Feder seem genuinely enamored of science. When Stewart talks about a scientific hero--Kepler, for example--he sounds almost giddy: "Kepler wanted to understand the universe, and he burst into tears as he was teaching a class because he suddenly understood how the five planets were laid out--they were laid out on the five regular solids. It makes us smile nowadays, but what he actually found out about the equal areas and the ellipses is absolutely wonderful. He was an enormous figure, just gigantic--how did I get onto this?"

The theory that Stewart and Feder would do anything to get their names in the headlines doesn't ring true, either. Even those who paint the two as fanatics don't see them as petty. Moreover, at least in Stewart's case, the skepticism and relentless hunt for other people's errors are not some midlife reaction to disappointed hopes. At the beginning of his career, at a time when his résumé still glowed with promise, Stewart launched his first fraud investigation. It involved a report on a mysterious substance called scotophobin. Extracted from one rat's brain and injected into another's, it somehow conveyed learned skills from donor to recipient. This "memory molecule," its excited discoverers told the newspapers, meant hope for the retarded and the senile, and possible cures for alcoholics and drug addicts.

The research was published in Nature in 1972. Following the article was a second one, written by Stewart, who had been assigned to referee it. His article was twice as long as the one it critiqued. In it Stewart demonstrated in enormous and irrefutable detail that the researchers had no idea what they'd found, and he strongly implied that, whatever it was, it did not perform as claimed when injected into rats. Scotophobin hasn't been heard of since.

Yet even false theories about what drives Stewart and Feder have one merit: they illustrate the hostility the two inspire. Some measure of resentment is easy to understand. Stewart and Feder have not only made a career of sniping at others, they have done it while safe on the government dole at the NIH. Shielded from academia's pressure to "publish or perish," they have been able to pronounce on the rat race while watching it from padded seats on the sideline.

But the hurt goes deeper. Stewart and Feder's criticisms leave their fellow scientists feeling willfully misunderstood. Why don't they do science, their colleagues wonder, instead of complain about it? "There are far fewer cases of fraud in science than in journalism, banking, politics, or any other field of human enterprise," says Holton indignantly. "The amazing thing is that science can be done so well although it is done by mere human beings."

Hurt feelings are one thing, though, and a sense of betrayal is another. To hear many scientists tell it, Stewart and Feder crossed that line when they went to work for Congressman Dingell, to help his investigation of the Baltimore case.

As chairman of the House Subcommittee on Oversight and Investigations, Dingell can roam where he wants. He is six foot three, bald-headed, lantern-jawed, and widely regarded as a bully (his "default setting," according to one longtime observer, "is seething outrage") who exults in his power. In 1988 scientific fraud caught his eye. On the prowl for a good test case, he settled on the Baltimore affair. It involved big names at big institutions, it was timely, and it had already spurred investigations at Tufts and MIT, so science's self-policing could be put to the test. Stewart and Feder hadn't brought the case to Dingell directly, but they had spent two years, in Stewart's words, "hawking it to anyone who would listen." Dingell picked up on the buzz that Stewart and Feder had generated and recruited the two men as temporary investigators.

Scientists were outraged. Stewart and Feder had argued for years that only scientists should police science. Time and again they had proclaimed that they spoke only for themselves. They had no official standing, they said, and they wanted none. Now, suddenly, they had gone over to the enemy.

Dingell, as echoed by Stewart and Feder, insists that he is not an enemy of science. He consistently supports big budgets for biomedical research; his father, once a congressman himself, was a supporter of the NIH in its early days; and his brother is a high-ranking scientist there. "Dingell understands science quite well and simply wants it to police itself," Stewart says. "His good faith and good sense are shown by the fact that he hasn't passed any laws or regulations or anything--all he's done is ask scientists to fix the system. It's simply not possible to take a taxpayer-supported activity and argue that those who have an official oversight role shouldn't perform it just because scientists don't want them to."

It is a tough sell. Even Stewart concedes that "a politician less well inclined or less well informed than Dingell could use the fraud issue as an excuse either to disparage science or to set up some giant bureaucracy doing something inappropriate." If scientists don't police science, Stewart goes on, "somebody will, and I view that prospect with enormous alarm."

But if politicians pose such risks, why work with them at all? Stewart's argument seems disingenuous. Politicians in general are best kept away from science, he maintains, but this particular one, feared though he is, is all right. It is as if a hiker conceded that, yes, throwing leftovers from dinner into the woods might lure grizzlies into camp, but not to worry--the particular bear who's been coming around so far seems harmless enough.

Sportswriters are fond of saying of a particularly fierce competitor that "he doesn't know when to quit." In the world of science this is not necessarily a compliment. Stewart and Feder spent five fruitless years studying snails. They filled 600 notebooks with data on 62,562 jars of snails and in the end learned almost nothing worth publishing. Their fraud investigations have consumed years of their lives, often without a lot of clear resolutions. "There's a common theme," claims Bernard Davis, a retired Harvard microbiologist and a longtime critic of Stewart and Feder. "They didn't have enough sense of proportion to shift gears after a few years of snail work, and they don't have enough sense of proportion to say, 'Look, what constitutes really significant misconduct?' "

Davis's charge is true, more or less, but it is not the sort of accusation that fazes Stewart and Feder. They are crusaders, and crusaders don't go in much for cost-benefit analysis. And they have had an impact, though perhaps not in propor
Comment on this article
ADVERTISEMENT

Discover's Newsletter

Sign up to get the latest science news delivered weekly right to your inbox!

ADVERTISEMENT
ADVERTISEMENT
Collapse bottom bar
DSCNovCover
+

Log in to your account

X
Email address:
Password:
Remember me
Forgot your password?
No problem. Click here to have it emailed to you.

Not registered yet?

Register now for FREE. It takes only a few seconds to complete. Register now »