Responsibilities for ensuring the integrity of the research record rests with a number of players – funding agencies, governments, publishers, journal editors, institutions who conduct research and the researchers themselves. Our responsibilities for providing research that is honest and trustworthy are extant at the very beginning of a research project and ever present thereafter. If one of the players in the research ecosystem finds that research isn’t honest or can’t or shouldn’t be trusted then we have to take steps to remove it from the research record or stop it from getting there. We don’t need a definition of research misconduct in order to do that.
In fact, there isn’t a definition of research misconduct, and this is part of the problem. Resnik et.al. describe this in their 2015 paper that reviewed and categorized misconduct definitions from 22 out of the top 40 research and development funding countries. They claim that the variation in research misconduct definitions might make it harder for potential complainants to raise a concern because they can’t work out whether something might be misconduct in any particular jurisdiction. Similar research by Resnik et.al. also looked at research misconduct definitions in US universities, and found that the majority go beyond the definition provided in US law, perhaps indicating that these universities recognise that there is more than falsification, fabrication and plagiarism that can impact on the honesty and trustworthiness of the research record. A ‘back of the envelope’ review of Australian research misconduct policies paints a similar picture with two broad clades – one that centres on research misconduct as a serious deviation from accepted practice and the other that requires misrepresentation. All of this means that saying Professor Y committed research misconduct doesn’t really mean much, and doesn’t tell us how the research is dishonest or untrustworthy. It stops us from making our own assessment of the trustworthiness of the research.
Many definitions also require that it can be shown that the researcher responding to the allegation committed the act of research misconduct, however defined, deliberately or intentionally or with recklessness or negligence. This ‘mental fault’ element is used to distinguish those lapses in responsible research that are honest mistakes or accidental from deliberate, mischievous attempts to deceive the users of the research output, whether that is a journal article, lab meeting presentation or grant application. The inclusion of this mental fault element also focusses the attention of those considering complaints or serving on investigation panels on the minds of the ‘accused’ – the investigations very much become concentrated on whether Professor Y was really trying to be evil and not whether the research should be trusted and allowed to have impact.
We believe that this is the fundamental question a research integrity investigation should be considering – can we trust the research and would we be happy for it to have impact?
Consideration of mental fault (mens rea if you’re a lawyer) is important when considering what disciplinary action to take, but is best not part of the rubric when considering trustworthiness, accuracy or honesty of research.
Research conduct occurs on a spectrum – from excellent research conduct at one end to research misconduct at the other. It is not only those deliberate or grossly negligent acts that cause us to question the honesty or trustworthiness of research. There are a range of behaviours that impact on the integrity of research and many of these are neither deliberate nor FFP. Some of these are described in the seminal paper by Martinson et.al. that reports on results of a survey of biomedical researchers. The most frequent ‘questionable research practices’ described in this paper include inadequate record keeping related to research projects (27.5% of researchers), ‘dropping observations or data based on gut feeling’ (15.3%) and ‘using inadequate or inappropriate research designs’ (13.5%). It is clear that these three QRPs will impact on the trustworthiness and accuracy of research findings, and the incidence of these QRPs is much greater than the 0.3% reported for ‘falsifying or cooking research data’. These and other QRPs fall outside of many definitions of research misconduct, and so can be overlooked by institutions forced or who choose to focus on research misconduct as defined. This leaves a broad range of activities potentially unchecked, and research on the record that perhaps really shouldn’t be.
Removing the definition of research misconduct simplifies the landscape. Investigations won’t need to consider the motivation for a departure from accepted practice or breach, but only if the research can be trusted or should be allowed to have impact. Disciplinary action can still happen through other misconduct related processes and this is where deliberation and intent can and should be considered. A system like this already exists. The Canadian Tri-agency Framework for Responsible Conduct of Research does not define research misconduct but instead sets out very clearly articulated principles for research integrity. A breach of these principles can trigger an investigation and consideration of deliberation or intent is not part of the framework. The absence of a definition has not stopped Canadian funding agencies taking appropriate action. Recently, the first disclosure of an investigation was made. It names the researcher responsible and provides detail about the nature of the breach and the action taken by the funding agency involved.
Research misconduct is not a well-defined term, but a better definition is not needed and is not the solution. We need to take action to protect the integrity of the research record and stop untrustworthy or dishonest research from reaching it. We can do that just as well or even better without narrowing the scope of these considerations.
References
David B. Resnik J.D.,Ph.D., Lisa M. Rasmussen Ph.D. & Grace E. Kissling Ph.D. (2015) An International Study of Research Misconduct Policies, Accountability in Research, 22:5, 249-266, DOI: 10.1080/08989621.2014.958218
David B. Resnik J.D., Ph.D., Talicia Neal M.A., Austin Raymond B.A. & Grace E. Kissling Ph.D. (2015) Research Misconduct Definitions Adopted by U.S. Research Institutions, Accountability in Research, 22:1, 14-21, DOI: 10.1080/08989621.2014.891943
Nature 435, 737-738 (9 June 2005) | doi:10.1038/435737a
Contributors
Paul M Taylor, RMIT University (bio) – paul.taylor@rmit.edu.au
Daniel P Barr, University of Melbourne (bio)- dpbarr@unimelb.edu.au
This post may be cited as:
Taylor P and Barr DP. (2016, 25 October) We don’t need a definition of research misconduct. Research Ethics Monthly. Retrieved from: https://ahrecs.com/research-integrity/dont-need-definition-research-misconduct
2 thoughts on “We don’t need a definition of research misconduct”
Of course, this raises the question of which principles we adopt and how we help researchers understand the implications of the principles. When reaching a decision on a case we would also need to outline clearly what we have determined to be research misconduct.
excellent research conduct need not always ensure excellent research impact. Research impact may not be visible immediately and sometimes create confusions and food for arguments. Good conduct research can ignite enthusiasm for further research. The problem is the abstract nature of the term misconduct. This is a complex construct and different people mean differently in different contexts. So attempting for a universally agreeable and standard definition is the need of the hour because as number of researchers are increasing the misconduct also is expected to increase.
Rajamohanan K Pillai , former president IndiaCLEN