Fear of getting caught could reduce scientific misconduct

A system of audits could put a stop to research fraud

Regular auditing of university research labs could be an effective deterrent against misconduct, the vice-chair of the UK Research Integrity Office has said.

Speaking at the 3rd World Conference on Research Integrity, Michael Farthing, also vice-chancellor of the University of Sussex, likened the potential effect of auditing to the way in which speed cameras slow drivers down.

Although discussions around the responsible conduct of research usually centre on appealing to people’s honesty, “I know personally that I’m encouraged to do things when there’s a chance that I might actually be caught”, Professor Farthing told the conference, held in Montreal from 5 to 8 May.

Universities’ finances and teaching are regularly audited but, outside clinical trials, no equivalent exists for research, he said, despite some institutions receiving more than 50 per cent of their income through public research funding.

“We have ways to assess quality…but no one asks the question ‘is this high-quality work also work of high integrity?’,” he added.

Speaking to Times Higher Education after the event, Professor Farthing said he had discussed introducing a pilot audit system at his own institution. However, he stressed that the idea was in its early stages and that cost implications had not been worked out.

Under such an initiative, a small percentage of research would be randomly picked for auditing, either before or after publication, he said. Measures could include checking that raw data support conclusions, running manuscripts through plagiarism software and checking images for manipulation.

Rather than give the impression that Sussex had a particular problem with misconduct, such a move would help the university to provide assurance as to the integrity of its work and show a “strong institution, not a weak one”, he added.

Professor Farthing said that he believed auditing formed part of a “toolbox” of efforts that could be used to prevent and detect research misconduct.

Other measures included registering research studies as they begin, in the manner of clinical trials, and better tracking of researchers as they moved from institution to institution.

Communication between universities over potential misconduct remained a problem, he said.

He highlighted the case of researcher Jatinder Ahluwalia, whose dismissal from the University of Cambridge’s doctoral programme in 1998 for suspected misconduct was highlighted to other institutions only after a 2010 University College London investigation concluded he was likely to have falsified results while at UCL.

“It is a good example of how you have got a 15-year history of misconduct, yet nobody was able to talk to each other and perhaps truncate the activity,” Professor Farthing said.

Elizabeth Iorns, chief executive of Science Exchange, a Silicon Valley company that calls itself a “marketplace for scientific services”, said that she was pleased to see systematic auditing of research being highlighted at the conference.

She cited data suggesting that more than 70 per cent of biomedical research results cannot be replicated. Her company is hoping to run a pilot for a system in which biomedical papers are randomly audited for their reproducibility in non-academic labs, potentially acting as a deterrent to misconduct.

“The current system we are relying on is clearly not very effective,” she said. “If 73 per cent of papers are not reproducible and only 1 per cent are retracted, there’s a huge gap there.”

elizabeth.gibney@tsleducation.com