Research misconduct refers to a breach of ethical or professional standards during the research or reporting processes. It is a direct threat to research integrity, which upholds the values and expectations of science conducted by professionals and national bodies, often for the benefit of the public.
Here, we shall look at the definition and types of research misconduct; explore the typical consequences for researchers, institutions, and the published record; cover the detection tools and processes used by journals and preprint servers; and demonstrate how preprints and Preprints.org support broader community feedback, scrutiny, and faster correction.
What is research misconduct?
To understand what accounts for research misconduct, it’s useful to refer to the ‘FFP’ (Fabrication, Falsification, Plagiarism) framework used by the US Office of Research Integrity and other national bodies. According to this framework, research misconduct is the evidence of fabrication, falsification, or plagiarism in the conducting of research and any subsequent reporting. Here are some examples of research misconduct:
- Fabrication: Making up of data or results and reporting them as if they were true.
- Falsification: Manipulating research materials, equipment, or processes, or purposefully altering data or results to falsify the research record.
- Plagiarism: Appropriating other people’s ideas, processes, results, or words without giving credit.
Research misconduct directly threatens the integrity of science. By spreading awareness about it, we help maintain public trust and ensure that progress is not based on falsities and ethically unsound methods.
Typical consequences of research misconduct
Research misconduct often results in typical consequences for researchers, institutions, and the public record.
Researchers
For researchers found guilty of research misconduct, the consequences are often severe and potentially career-ending. They include loss of employment, formal retraction of publications, reputational damage, loss of funding eligibility, and penalties for any collaborators.
Institutions
Consequences for institutions found guilty of research misconduct (or directly responsible for colleagues found guilty) include widespread reputation damage, significant losses to funding, administrative and legal costs for prolonged investigations, and diminished trust from the public and larger scientific community.
Published record
The published record (or version of record) is the final, peer-reviewed, citable, and digitally preserved version of the manuscript. If authors are found guilty of research misconduct, this can result in paper retractions and expressions of concern by journals. Additionally, misconduct can lead to the misdirection of future research and damage to the credibility of scientific output.
Detection tools and processes
Research misconduct is more prevalent than you might expect. A study found that one in 12 Dutch scientists admitted to committing fabrication or falsification. Of course, this study only investigates respondents from one country, but the findings show that misconduct will remain prevalent unless we take a firmer approach to maintaining research integrity.
Maintaining research integrity is a shared responsibility among researchers, institutions, publishers and the broader research community, supported by tools and detection processes.
But what detection tools and processes are being employed to reduce instances of misconduct?
Similarity Check
Similarity Check, a service provided by Crossref, produces immediate feedback regarding a manuscript’s similarity to other published academic and general web content, helping editors verify research originality.
Submitted manuscripts are cross-referenced against a large, automatically updated database containing vast numbers of web pages and scholarly items. The tool then generates a similarity report and score.
Proofig AI
The development of artificial intelligence (AI) tools means that image manipulation in research is a growing concern for science. But AI-powered image proofing tools such as Proofig AI can help ensure integrity.
Proofig AI uses advanced algorithms to identify patterns unique to AI-generated images, revealing image duplication and manipulation in the process.
Preprints.org: supporting community detection, faster correction, and screening
Preprints and servers like Preprints.org also serve as tools to be used to counter research misconduct.
Due to their rapid open availability through Creative Commons (CC BY) 4.0 licensing, preprints offer community scrutiny and feedback. Because this research is openly available via open access, communities can come together to verify methods and data. Authors can also receive public or private feedback on their work from their peers, meaning that if errors are honest and not malicious, issues can be addressed accordingly.
Preprints.org also runs a screening process that aims to ensure that content adheres to basic publishing ethics, and that the authors comply with international research ethics regulations. Preprints.org adheres to the Committee on Publication Ethics (COPE) Code of Conduct and Best Practice Guidelines. Within these guidelines, plagiarism, data fabrication, and image manipulation are all forbidden, and manuscripts that violate these ethical standards may be rejected or withdrawn.
Preprints are also more flexible than peer-reviewed articles when it comes to addressing ethical oversights and honest errors identified post-publication. Whereas peer-reviewed articles may need to be withdrawn completely, posted preprints are free to be revised and updated through a process called versioning. Each new version is assigned a unique DOI, preserving the integrity of the research record.
