In the wake of the Bell Labs scandal, Steve Farrar asks just how prevalent research fraud really is
One-third of the editors of the world's leading academic journals have encountered research fraud. Yet many do not know how the alleged miscreants were ultimately dealt with.
A survey by The THES in the wake of the Bell Labs scandal revealed a large degree of inconsistency in the way serious misconduct was handled once suspicions were raised.
Some journals prompted investigations at the author's institution that led to the individual's sacking. Others admitted that their concerns merely resulted in the withdrawal of a paper before publication and that they did not pursue the matter further.
Richard Horton, editor of The Lancet and a member of the Committee on Publication Ethics, a body set up by leading biomedical journals to deal with fraud and related issues, said many people did not want to create a fuss.
"I'm sure the vast majority of allegations are swept under the carpet in the hope that somebody else will deal with them," he said.
Dr Horton observed that many of the cases of misconduct investigated by Cope revealed that there had been suspicions about previous papers but that no action had been taken.
But Charles Jennings, editor of Nature Neuroscience and executive editor of Nature research journals, said that journals lacked the investigative resources or disciplinary authority to get to the bottom of fraud allegations.
"The responsibility for this must lie with institutions that have such powers - for instance, funding organisations, universities and research institutes," he said.
The survey, however, suggested that sometimes these bodies were not informed of what their employees might be up to.
The THES sent a questionnaire to the editors of 50 academic journals and received replies from 30. Of these, 16 handled primary research, while 14 were review journals.
Eleven admitted to coming across incidents of research fraud, but most said they had not had to deal with it.
Dr Jennings said he doubted whether fraud was as widespread in science as in many other professions. "Nevertheless, I suspect it is common enough that most scientists can expect to encounter a case at some point in their careers," he said.
One editor, who is waiting for the results of an investigation of a case concerning research submitted to her journal, said: "Unfortunately even a few cases make the entire community look bad."
Donald Kennedy, editor of Science , said: "It is relatively rare, but nevertheless occurs enough to make us uncomfortable."
The frequency of fraud was suspected to be higher in some fields, such as molecular biology and biomedicine, than in others, such as physics.
Fast-moving areas were believed to be less prone as competing researchers would soon try to replicate results.
"Lesser types of fraud, where the conclusions are in fact largely correct but the data have been distorted or altered to make them more convincing, may be more common," one editor said.
He added that obscure and less competitive fields might also be more prone to misconduct.
But another editor observed: "If fraudulent results are of any significance, the work will be repeated by others and the fraud will be uncovered. If they elicit no interest, the fraud may never be uncovered, but then it hardly matters."
While just three journal editors believed that the academic community was complacent about the problem, six felt more could still be done to tackle the problem.
"The temptations to commit fraud are prevalent, so the community is really quite vigilant at creating an atmosphere that is highly discouraging of such behaviour," one editor said.
Dr Horton was more scathing and warned that greater care on the part of journals was needed to reduce the likelihood of scandals that damaged public trust in science. "The scientific establishment has not taken the issue seriously and therefore has been reluctant to find out how common fraud is," he said.
At The Lancet , allegations of fraud that appeared to have substance were first referred to all of the authors and then to the head of the institution. If that did not resolve the situation, Dr Horton said the journal took action itself to raise the alarm.
Many editors agreed that peer review was not designed to provide a defence against misconduct, but they did not believe tougher reviewing was necessarily the answer.
Dr Kennedy said: "Detecting fraud is difficult and maybe not possible without doing more harm than good to science."
Among the measures suggested by respondents to reduce the problem were to:
* Publicise incidents of research misconduct when they come to light to show how rapidly most were uncovered, the damage they caused and the consequences for the miscreant
* Include ethics and even fraud-spotting tuition as part of a student's training
* Develop systematic bioinformatics methods to search the literature for duplicated images
* Appoint an ombudsman to keep track of research fraud incidents and protect the anonymity of whistleblowers
* Check references both on the phone and in writing before hiring anyone.
Dr Jennings remarked: "Unfortunately, it sometimes happens that senior scientists are eager to expel a problem researcher from their own lab but are reluctant to mention the problem to future employers."
The THES is conducting an investigation into research fraud. Do you have any concerns or examples? Contact in confidence: Phil Baty (020 7782 3298) or Steve Farrar (020 7782 3299).