"Long before it's in the papers"
January 18, 2011

RETURN TO THE WORLD SCIENCE HOME PAGE

Are U.S. scientists more likely to publish fake research?

Nov. 16, 2010
Courtesy of BMJ-British Medical Journal 
and World Science staff

U.S. sci­en­tists are sig­nif­i­cantly more likely to pub­lish fake re­search than sci­en­tists from else­where, sug­gests a study that tal­lied of­fi­cially re­tracted stud­ies from var­i­ous coun­tries.

Re­search­er Grant Steen, of Chap­el Hill, N.C.-based com­pa­ny Med­i­cal Com­mu­nica­t­ions Con­sul­tants, searched PubMed, a vast U.S. gov­ern­ment database of bi­o­log­i­cal stud­ies, for eve­ry re­search pa­per that had been with­drawn—and there­fore ex­punged from the pub­lic record—be­tween 2000 and 2010.

A to­tal of 243 pa­pers on the database were re­tracted due to pre­sumed fraud dur­ing this pe­ri­od, found Steen’s stu­dy, pub­lished on­line in the Jour­nal of Med­i­cal Eth­ics. In about 35 per­cent of these cases, the first name in the list of au­thors was that of a U.S. sci­ent­ist.

The many cases in which re­trac­tion was due to re­search er­ror rath­er than fraud weren’t counted, Steen said.

“Perhaps sur­pris­ingly, fraud occurs more often in the USA than the rest of the world,” he wrote. “And there was sig­ni­fi­cantly more fraud than error among re­tract­ed papers from the USA... com­pared with the rest of the world.”

Steen al­so found that the fakes were more likely to ap­pear in lead­ing pub­lica­t­ions with a high “im­pact fac­tor,” a meas­ure of how of­ten re­search is cit­ed in oth­er peer re­viewed jour­nals. Moreo­ver, 53 per­cent of the faked pa­pers had been writ­ten by a first au­thor who was a “re­peat of­fend­er,” and faked re­search pa­pers were sig­nif­i­cantly more likely to have mul­ti­ple au­thors.

Each first au­thor who was a re­peat fraud­ster had an av­er­age of six co-au­thors, each of whom had had anoth­er three re­trac­tions, the study found.

* * *

Send us a comment on this story, or send it to a friend

 

Sign up for
e-newsletter
    
 
subscribe
 cancel

On Home Page         

LATEST

  • B­irds may at­tract mates with “sca­ry mo­vie” ef­fect

  • W­hy do men use sil­ly pick­up lines?

EXCLUSIVES

  • From bra­in sci­ence, new ques­tions about fr­ee wi­ll

  • Scien­tists ex­plore whe­ther some ap­es shake he­ads for "no"

  • It se­ems we're all more hu­man than aver­age

  • Ba­by temper­ament found to pre­dict ad­ult bra­in struc­ture

MORE NEWS

  • What hit the Mo­on? New cra­ter makes a spl­ash

  • T­racks may tell tale of rep­tilian land con­quest

  • Plan­ets found shar­ing strange dan­ces

  • Newfound sta­rs seen sha­ttering known si­ze lim­its

U.S. scientists are significantly more likely to publish fake research than scientists from elsewhere, suggests a study that tallied officially retracted studies from various countries. Researcher Grant Steen, of Chapel Hill, N.C.-based company Medical Communications Consultants, searched PubMed, a vast U.S. government database of biological studies, for every research paper that had been withdrawn—and therefore expunged from the public record—between 2000 and 2010. A total of 243 papers on the database were retracted due to presumed fraud during this period, found Steen’s study, published online in the Journal of Medical Ethics. Of these, the first listed author was from the United States in slightly over one third of cases. The many cases in which retraction was due to research error rather than fraud weren’t counted, Steen said. Steen also found that the fakes were more likely to appear in leading publications with a high “impact factor,” a measure of how often research is cited in other peer reviewed journals. Moreover, 53% of the faked papers had been written by a first author who was a “repeat offender,” and faked research papers were significantly more likely to have multiple authors. Each first author who was a repeat fraudster had an average of six co-authors, each of whom had had another three retractions, the study found.