 |
|
 |
|
 |
|
|
|
 |
|
About the Author
Dan Greenberg is a longtime observer of science policy and politics. He is the author of "Science for Sale: The Perils, Rewards, and Delusions of Campus Capitalism" (University of Chicago Press), as well as other books, and has published widely in newspapers and popular and professional magazines.
About Brainstorm
Mark Bauerlein
Marc Bousquet
Laurie Fendrich
Dan Greenberg
John L. Jackson Jr.
Stan Katz
Michael Nelson, guest blogger
Stephen Joel Trachtenberg
Brainstorm Alumni
Marty Nemko, guest bloggerBob Zemsky
Recent Posts
Nobel Hype | Comment
McCain and Obama Differ Little on Science Policy | Comment [5]
A Name on a Mailing List Is Forever | Comment
Stodgy NIH Announces New Program for Risky, Far-Out Reseach | Comment [2]
Obama Outlines Changes for Science Policy | Comment [4]
Featured Articles By the Author
Detecting Scientific Fraud
Fraud, fakery, or larceny is what ordinary people would call it. But in the sciences’ refined venues the proper term is “misconduct,” and there’s a lot more of it than official figures show, according to a report in Nature (19 June), “Repairing research integrity.”
Perhaps it’s nostalgia for my journalistic apprenticeship as a police reporter that draws me to such publications. But as much as I relish a crackdown on miscreants in lab coats, I’m wary of this report, though it has impressive authorship: a University of Wisconsin official responsible for research policy, and a current and a retired official of the U.S. Office of Research Integrity (ORI), which monitors scientific purity for the National Institutes of Health and other parts of the Department of Health and Human Services.
Under ORI’s procedures, grantee institutions are the first line of defense against misconduct, officially defined as “fabrication, falsification or plagiarism in proposing, performing, or reviewing research, or in reporting research results.” Upon receiving an allegation of scientific wrongdoing on their premises, research institutions are required to hold an inquiry — i.e., a quick review. If the allegation holds up, next comes a presumably deep investigation. Though locally conducted, the investigation must be reported to ORI.
The authors report that, on average in recent years, the number of investigations annually reported to ORI was merely 24! In 2007, a typical year, NIH supported 155,000 researchers — which leads the trio to conclude that they’ve come upon “an alarming picture of under-reporting,” or, as they put it, “the tip of a much larger iceberg.”
To determine whether there’s more going on than the official reports indicate, the authors sent out questionnaires that asked 4,298 NIH grantees at 605 institutions whether they had observed “possible research misconduct” during three prior years. Responses arrived from 2,212 researchers, of whom 164 reported 201 instances of “likely misconduct” during that time frame. Extrapolating from that response to the full pool of NIH researchers, the authors calculate that “more than 2,300 observations of potential misconduct are made every year. Not all are being reported to universities and few of these are being reported to the ORI.”
Now, given the rivalrous, sometimes venomous, atmosphere that prevails in departments and laboratories, especially in the sciences’ scarcity economy of recent times, claimed observations of misdeeds invite some skepticism. Only 11 percent of the misdeeds were “directly observed,” while nearly 30 percent came under the heading of “told first, then observed.”
Are we encountering here a whiff of hysteria reminiscent of ancient witch hunts?
Scientific misconduct probably exists, in excess of the official count. But the survey method is no way to determine incidence or to instill virtue in the tempted.
The Nature authors briefly refer to the most useful suggestion yet made in the long-running concerns over scientific misconduct: Occasional audits of laboratory records to match up publications with the underlying research. Knowing that the auditors may drop in unannounced could have a wondrously elevating influence on scientific integrity.
Posted at 02:37:04 PM on June 23, 2008 | All postings by Dan GreenbergComments
Commenting is closed for this article.
Previous: Countering Big Pharma's Misuse of Academic Prestige
Next: Prizes Won't Beat the Energy Crunch ...
Here is another interesting example of fraud.
http://www.sciam.com/article.cfm?id=glenn-mcgee
— Jean · Jun 24, 08:38 AM · #
During my scientific career of more than 40 years, from the time I entered graduate school and started doing original scientific research, to the present, I have encountered very few instances of misconduct. The most disturbing to me were incidents that occurred long ago (25-40 years ago) involving issues of laboratory animal care. One was a case where a professor holed up in his home to avoid being served with a warrant, and made no provision for the care of his research animals. Another was when an animal care staff person who had been fired, released various research animals into a space where some killed others. A third was a case where an investigator left his laboratory animals in the care of a volunteer who staged photos of some of the animals in horrendous postures suggestive of abuse, published and distributed the photos, and used the entire incident to establish what became a notorious anti-research NGO. Other cases involved selecting some human subjects for a drug trial who did not really meet the criteria for inclusion in the study (in fairness, the drug was already approved for safety and in wide use for one purpose; the study was to evaluate efficacy for an different use, and the additional subjects were people whose diagnoses suggested that they might also benefit from the drug—although the prospect of therapeutic benefit was less than for the real target population, of which there were too few available). Also problematic were cases where an investigator submitted several nearly identical manuscripts simultaneously (we journal editors caught this and all rejected the mss.) or sequentially submitted manuscripts that did not include sufficiently incremental results (also caught by our reviewers). We also ran across a couple of cases where scientists employed people in developing countries to go do field work and then reported the work as if it was their own, without adequate coauthorship or acknowledgement of those who actually did the work. This might be more common than I have experienced, in situations where lab techs or grad students do the work without sufficient credit—and there are allegations of this sort of thing in some pretty high profile cases. But, actual fabrication of data? I think that is pretty rare. It is too easy to get caught, and, after all, if the work is important, someone is likely to attempt work that involves and/or requires replication. One of the great strengths of scientific method is the degree to which it is self-correcting. Perhaps the greatest prospect of fraud or deception has to do with selective reporting of the results (many failures to reject the null hypothesis are not reported) and designing studies in ways that minimize the prospect of finding something one does not wish to find (e.g., no significant evidence of harmful effects of a drug, when too few subjects were tested to be able to detect harm if it existed). Some subtle deceptive practices probably do not cross the threshold of demonstrable fraud. I think these tend to be in applied science settings where significant incentives are contingent on specific outcomes. Now, a survey asking questions about specific practices and one’s experiences regarding them could be informative, as one of several assessment methods. The audits you suggest might be another appropriate method, but conducting them fairly and thoroughly with real validity, would, I think, be a very challenging task.
— Joe Erwin · Jun 24, 08:38 AM · #
“Falsification of data” should not be the only “gold standard” of misconduct. So should plagiarism, inappropriate “attributions” of credit, and mis-use of grant monies by institutions, etc., etc.
Indeed, Federal Justice Department investigations have uncovered such financial abuses in the area of Medicare-funding even at major Ivies, etc. However, Federal research-granting agencies by-and-large feel rather comfortable with letting the individual institutional foxes guard their chicken coops when it comes to “misconduct”. Oh, and where are is the ORI for humanities research, BTW? Nonexistent – even in the case of Federally-funded research.
Whistleblowers have everything to lose, of course, by complaining at their own institutions – unless they are rival researchers of significant standing. And anonymous “tips” to Federal agencies seem to get no attention at all (for fear they may be accurate?). One would think that such information could trigger an “audit” at the very least.
Again, one need only review the fascinating vagaries of the Philip Felig affair to see just what happens even when “serious” investigations occur. The lesser-positioned player in the scandal, a female researcher, received the most severe punishment; Felig had his endowed-chair and chairmanship offer from Columbia rescinded – primarily because of the negative publicity – but he remained a tenured (and, I believe, endowed-chaired) professor at Yale.
— Anti-hypocrisy advocate · Jun 25, 08:23 AM · #
Dear Mr. Greenberg, you “relish a crackdown on miscreants in lab coats” and you “published widely in newspapers and popular and professional magazines”. You should really report on University of Toronto fraud, see documents at http://ca.geocities.com/uoftfraud/
The problem is not with detecting scientific fraud, but with the media that helps to obstruct justice and lets the “miscreants in lab coats” get away with the crime.
The conspiracy of silence in the media is destroying the life of honest scientists and it makes their research and discoveries remain unknown! You should be the first to report the University of Toronto fraud!
— Michael Pyshnov · Jun 26, 10:32 AM · #