The Case for Revising the Definition of Research Misconduct

By Elizabeth McEvoy, Esq. and Melissa Hogan

Twenty years after the Office of Research Integrity (ORI) adopted “falsification, fabrication, or plagiarism” as the black-letter definition of research misconduct, at least one member of the scientific community is challenging what it means to commit research misconduct. In a recent article, bioethicist David Resnick revisits the longstanding federal definition of research misconduct – fabrication of data, falsification of data, or plagiarism, or “FFP,” – which has remained unchanged since its enactment in 2000 – and suggests ways in which the scientific community can expand the definition that has long served as a “floor” for institutional definitions of research misconduct. Professional organizations, including research institutions, universities, and professional societies, retain great discretion when it comes to defining misconduct but often adopt, almost verbatim, FFP for the sake of consistency with the federal regulations.

New areas of “misconduct”

Resnick’s article examines four courses of conduct – sexual harassment, sabotage, deceptive use of statistics, and failure to disclose a significant conflict of interest – that do not fall squarely within the definition of research misconduct but raise serious ethical concerns that call into question the integrity of the underlying science.

Sexual harassment

Sexual harassment has recently emerged as an ever-growing issue facing all areas of our daily lives. Of late, it has warranted sharp response from employers. In the lab, complaints of sexual harassment remain a grey area for institutions looking to take a hard line. While recent policies published by two of the federal oversight agencies charged with policing research misconduct, the National Science Foundation (NSF) and the National Institutes of Health (NIH) have taken a “zero tolerance” stance against any form of sexual harassment in research, the scientific community is split when it comes to codifying sexual harassment as research misconduct as a matter black-letter law. In fact, NSF, which has adopted a more aggressive stance to than its NIH, requires that institutions comply with federal law and procedure but need only notify NSF when a Principal Investigator (PI) or co-PI is put on administrative leave, or when findings of sexual harassment have been returned. This protocol imposes requirements outside the research misconduct framework and leaves open the question of whether sexual harassment is also a form of fraud.

Sabotage

Research sabotage is another area of concern that does not fit neatly within FFP. Sabotage in the lab is no different than it is outside the lab – referring to acts that deliberately destroy, damage, or obstruct. In the lab context, that destruction is frequently to another’s data or experiments. Resnik points out that research sabotage does not fit neatly into the existing laws or policies, including those prohibiting misconduct that constitutes FFP. The 2009 case of a University of Michigan researcher who was punished both criminally by the state of Michigan and civilly by ORI for contaminating a colleague’s cell culture samples with ethyl alcohol is one egregious example of sabotage that was addressed by ORI, which asserted jurisdiction over the matter as an instance of data falsification. Considering the destruction of another’s data to be “data falsification,” however, generally runs contrary to the strongly held belief in the scientific community that one can only fabricate or falsify one’s own data, and not those of another. Under this premise, there is a strong argument to be made that reprehensible acts of sabotage would remain beyond the scope of misconduct as currently defined.

Deceptive use of statistics

Although manipulation and omission are traditionally considered instances of fabrication, selective or “deceptive” use of statistics is not addressed within the guidelines of FFP. Resnik aptly recognizes that researchers can achieve favored outcomes by selecting a statistical model or test that, while presenting the desired hypothesis in the most favorable light, obscures certain assumptions inherent in the interpretive frameworks and has the effect of misleading the reader as to the significance of the data. This type of deception flies below the radar and is often only identified by diligent members of the scientific community or professional watchdog organizations that make it their business to test such hypotheses for reproducibility and sound the alarm when the tests do not yield the reported outcome. Although it may draw negative attention among the scientific community, deceptive use of statistics is not a recognized form of misconduct and could arguably apply to a wide variety of research studies, triggering a slippery slope with regard to policing such activities. Today, researchers are not expressly barred from such practices, regardless of how unethical they are.

Conflicts of interest

A fourth area that draws concern from Resnick is the failure of researchers to adequately disclose conflicts of interest that could affect their research. Researchers are generally required to complete a financial disclosure form as a precondition for receiving funding, but articles are rarely retracted for failing to disclose a potential or actual conflict of interest. Failing to inform the publisher or funding agency of these affiliations could compromise the integrity of the science reported and could be grounds for retraction or at the very least, a published correction. To some, failure to disclose a conflict of interest is tantamount to falsifying the research record. But to others, like Resnick, this type of inaccuracy is an omission that does not involve manipulation of materials, equipment, processes, or changing or omitting data, and thus does not rise to the level of formal misconduct.

Institutional action: what you can do.

Given the limitations of the traditional FFP framework, institutions looking to protect the integrity of their research should consider revising their existing policies to expand upon the traditional concepts of FFP and address sabotage, sexual harassment, deceptive statistics, and conflicts of interest, as well as other, ill-advised research tactics that call into question the validity of the underlying data. Rewriting institutional policies, of course, is a significant undertaking. Institutions and scientific organizations considering an update to current policies should engage in a careful evaluation of the research landscape as well as alternative practices.

To weigh the need for potentially expanding the scope of misconduct, Resnick suggests applying a four-part test: (1) Is the behavior widely regarded as highly unethical? (2) Does the behavior significantly threaten the integrity of the science? (3) Can the behavior be clearly defined? And (4) Are the existing methods of enforcement or deterrence inadequate? Although a helpful guidepost, Resnick’s approach only underscores the extensive grey areas that remain in detecting and preventing fraud in research. Any expansion of the research misconduct doctrine must be balanced against the considerable deterrent posed by vague and broadly defined concepts – generalities that invite collateral disputes over their meaning in the first place.

There are also practical concerns with codifying new rules governing ethical conduct in science. For example, it is virtually impossible without extensive inquiry to assign a judgment to just how conflicted an author is or may be. At the end of the day, professional associations, journals, private research sponsors, and universities are free to adjudicate unethical behaviors as they see fit as “they do not need to meet the same standards of legal due process that apply to state actors.”

And Resnik himself, a nationally recognized ethicist, does not recommend expanding the federal definition to address any of the four behaviors he highlights. Perhaps it is the role of institutions and professional organizations to pave the way for a new brand of ethical and responsible publishing by redefining what is truly fraudulent or what constitutes misconduct. Either way, as Resnick’s article makes clear, it is critical that the research and scientific community continue this dialogue and continue to promote transparency as well as a firm dedication to combating research misconduct.

For more, visit. David B. Resnik (2019) Is it time to revise the definition of research misconduct?, Accountability in Research, 26:2, 123-137, DOI: 10.1080/08989621.2019.1570156