You devote yourself to science. You tackle difficult questions. You win tenure. Grants start rolling in. And then someone in your lab starts fabricating data. Bogus numbers go from a lab book to a journal club to an abstract to a paper. Without realizing it, you've staked your reputation on a lie.
Blatant fraud in the lab is rare -- or, at least, rarely reported -- but it happens. And as a biologist at a large Southern university can attest, a few bad numbers can threaten an entire lab.
Phony numbers
The researcher, who didn't want to be named in this column, thought he was on the verge of a breakthrough. One of his research fellows was producing exciting data, the kind of stuff that could shake up a field. There was only one problem: The numbers couldn't be replicated. Suspicious, he took a careful look at the fellow's lab book. It soon became apparent that he was reading a work of fiction.
Following protocol, the senior researcher delivered the evidence to his department chairman. The fellow eventually confessed and was barred from receiving any federal grants for three years.
As it turned out, the fellow's hot new project was really a dead end. "We wasted a lot of time and effort on a series of papers that never got submitted," the biologist says. "We're not going to get funded by that grant again." He never had to retract any papers -- the ultimate humiliation in science -- but he was still deeply shaken. "The person who did this was by all appearances nice and completely honest," he says. "Everybody in the lab loved her. That was the biggest devastation. The rest of the lab just went into a depression."
Scientific misconduct is by no means limited to junior researchers. Primary investigators (PI's) are perfectly capable of making up numbers, and, as many young scientists have discovered, they can be all-too-willing to take credit for other people's work. (The latter issue will be the focus of next month's Catalyst column.)
Integrity at the Top
Unethical behavior can occur anywhere in the chain of command, but honesty should start at the top, says Chris Pascal, director of the Office of Research Integrity, a division of the U.S. Department of Health and Human Services. "If a PI cares about integrity, he or she can set the standard," he says. The ORI has handled many cases involving lab technicians, graduate students, and postdoctoral fellows. In Mr. Pascal's view, the young researchers weren't always entirely to blame. "You can have lab heads or PI's who are very aggressive and cut corners," he says. "They're sending a message: That's the way we do it."
Of course, even careful, conscientious scientists can get blindsided. The aforementioned biologist says he always stressed "100 percent integrity" in his lab. "I told everybody that we're here to generate data, publish papers, and get grants, but good science is still the most important thing." The list of other victims includes some of the most prestigious scientists in the world, such as Francis Collins, director of the National Center for Human Genome Research. In 1997, Mr. Collins had to retract all or parts of five papers after uncovering deceitful work by one of his graduate students.
Senior researchers may never be able to ensure absolute honesty in their labs, but they are in an ideal position to keep science as clean as possible, says Nicholas Steneck, a science historian with the University of Michigan and a consultant for the Office of Research Integrity. Phony data are only one potential land mine. "The number of people who go out and manipulate data is fairly low," Mr. Steneck says. "But the number of people who have done things that they wouldn't be enormously proud of if they were published on the front page of The Chronicle -- that's much higher."
Not surprisingly, the constant pressure to publish inspires many instances of dodgy science. "You find ways to multiply your publications," Mr. Steneck says. "You chop things into pieces, make subtle changes, and publish things twice. Is it misconduct? That's anybody's guess. But it's not good for science."
Very few people go into science with anything but the best intentions, says Michael Kalichman, an adjunct professor of pathology at the University of California at San Diego and director of a research ethics program there. "I start with the premise that most people want to do the right thing, but they can't do it if they don't know what's right."
Leading the Way
Mr. Kalichman offers several tips for helping young scientists do the right thing:
-
Set an example. "As a PI, you need to model the appropriate way to conduct yourself."
-
Be available. Visit the lab regularly and keep your office door open. Young researchers have a hard time emulating someone they rarely see. There's another advantage to staying accessible: If someone suspects something fishy in the lab, he or she will have ample opportunity to sound the alarm. "The most frightening thing for a scientist is not finding out about something until after a paper has been published."
-
Set firm, explicit rules. Students have to know what's expected of them. A few possible examples: All researchers must keep a detailed notebook, they must be prepared to defend their data, and they must gain approval from the PI before submitting any abstract or articles.
-
Talk about integrity. "People think they have to drive so hard on science that there's no time to discuss other issues," Mr. Kalichman says. But in his experience, young scientists are usually eager to talk about ethics. "It's not a waste of time, and it's not boring."
-
Encourage young researchers to take courses in research ethics. The courses, now offered at most institutions, help students prepare for the many dilemmas that arise in science.
Such safeguards may seem obvious, but many researchers fail to take even these basic steps. "We've seen many cases where there were no real integrity checks," says Larry Rhoades, director of the ORI's Division of Policy and Education. "Data are missing, or articles get submitted without anybody looking at them."
To get a better sense of the practices in the real world, the ORI will soon send surveys to 5,000 PI's who receive support from the National Institutes of Health. (The surveys will be sent either this spring or fall.) Mr. Rhoades hopes the survey will uncover some innovative strategies that could be shared with other institutions. He fully expects, however, that it will also document some appalling lapses in judgment.
The biologist at the Southern university has already changed his approach. He still tells his trainees that he expects absolute integrity. But he doesn't stop there. "I also say: 'If you [commit misconduct], I'll prosecute you to the full extent.'" Such dire warnings would have been unthinkable a couple of years ago, but now they seem essential. "I've changed," he says. "I don't trust anybody."
