Then versus now: keeping pace with research integrity threats
Long-term solutions must be dynamic, involve multiple sectors and hold all stakeholders to account, writes Sabina Alam
Research integrity and publishing ethics concerns have always impacted scholarly publications.
After all, where humans, research and the communication of research findings are involved there may indeed be issues around accuracy, debates around interpretation of data and results, authorship disputes, unaddressed competing interests and plagiarism. However, these were part and parcel of the research, authoring and editorial process, which until more recent times, were addressed primarily through established processes for editorial quality control, peer review and implementation of publishing ethics policies and practices.
The fact you’re reading this leads me to assume you know this is no longer the case! Issues involving authorship and data integrity concerns have not been uncommon, however the nature of the concerns 10 years ago versus now are quite different. For example, 10 years ago we wouldn’t have scrutinised identity verification markers of an author or reviewer quite as much as we need to do now. Now we ask questions such as are they who they say they are, do they even exist, are they a bot, has someone’s identity been hijacked?
Authorship issues used to mainly focus on disputes (including order of authorship), however now we are asking further details not just on identity, expertise and contribution, but also scrutinising any authorship changes, as buying and selling authorship is typically linked to paper mill activity. In fact, where authorship changes used to be comparatively rare, since paper mills entered the scene, these occur much more frequently which has led to necessary changes in checkpoints to detect unethical authorship behaviours.
Regarding data issues, where we used to mainly ask questions such as is there any reporting bias, are the conclusions in line with the data, are there any errors in the statistical analysis, now we ask questions such as is the data even real?
During my researcher days, one of the main questions we would debate when submitting a paper is whether we could pay the extra charges to publish our experiment images in colour rather than black-and-white, whereas now the questions we ask about images in submissions is are they real, cloned, manipulated, ‘stock’ images etc?
Even something as standard as the reference section has been impacted in a big way. Where we used to focus on whether the citations are relevant or a suitable source to back a claim, now due to irresponsible and careless use of AI tools in producing content, we’re asking questions such as are the references hallucinated?
And while we’re on the topic of AI tools… this has introduced a host of challenges including scaling up the concerns mentioned above, introducing major errors into content (including nonsense phrases and images), and making it easier for fake content to hide in plain sight.
So, the tools and editorial processes that are in place to detect research integrity concerns at the point of submission, through the peer review process and all the way through to publication cannot remain the same for very long. This is something that publishers continuously need to adapt to ensure accuracy, efficiency and scalability.
Dispelling myths
Part of the process of determining solutions involves dispelling myths, such as:
“This doesn’t happen in my area”: At Taylor and Francis we see research integrity issues arise in all types of content across STM and AHSS, from submissions coming in from across the world, in journals operating on open access, hybrid and subscription models. While the specific issues and the scale of these can vary, there are no areas which remain unaffected by research integrity and publishing ethics issues, and so solutions need to encompass all types of content.
“Paper mills are solely to blame for fraud”: In reality, systematic manipulation of the publishing process can happen in multiple ways and predates the large-scale operation of paper mills we often discuss today. It can manifest through small networks of people working together to increase their publications as well as citations to their work even when this is irrelevant. Therefore, solutions need to be adapted for small scale as well as large scale multi-faceted types of fraudulent activities.
“All you need is a research integrity tool”: Most research integrity tools are not yet ready to ‘plug and play’ independently. While the technical solutions these tools offer are critical, none of these can be used without involvement of humans in terms of analysis of the outputs and subsequent decision-making. Therefore, it cannot be assumed that deploying these important and necessary tools will immediately reduce overall cost (time as well as resources). Using these require significant investment from a technical as well as a people perspective. Specialist skills are required even more now to interpret the outputs of these tools to reliably inform decision-making.
What are we doing to address this?
A multi-pronged approach is the only way to address the many issues, balancing between reactive solutions (e.g. early detection of problematic content) and preventative solutions:
- Tools – Piloting and deploying various internal and third-party content validation and research integrity tools to screen different aspects is a very active area and continuously evolving to respond to emerging trends across several areas (e.g. author verifications, image integrity, citation accuracy, plagiarism, reviewer reports, and so on).
- Training and resources – Internal ethics and integrity due diligence checklists and training for colleagues across the business who may be dealing with third parties connected to potential submissions, including the establishment of our code of conduct for conference proceedings organisers.
- Cross-stakeholder collaborations – We are actively involved in key cross-industry efforts to tackle research integrity and emerging publishing ethics issues, including with United2Act, COPE and STM Integrity Hub, as well as establishing a joint research integrity lab with CAS Library which resulted in research to shed light on training gaps in research integrity and publishing ethics, and more recently led to the launch of a new advisory committee to ensure long term solutions are developed and implemented.
- Education and engagement – various activities across the world including author workshop programmes and more targeted engagement with institutions to establish long-term education and training support in publishing ethics.
Overall, the research integrity issues we are dealing with today won’t remain the same issues tomorrow. As discussed previously, these trends evolve and grow more complex over time and so long-term solutions need to be dynamic, involve multiple sectors and hold all stakeholders to account.
Dr Sabina Alam is Vice President of Publishing Ethics and Integrity at Taylor and Francis Group
- Want to understand the full scope of how publishers detect and investigate research misconduct at scale? Watch Dr. Alam’s expert-led session for library professionals, featuring real-world case studies and insights into the frameworks protecting the scholarly record. https://think.
taylorandfrancis.com/access- research-integrity-session/? utm_source=external_party_ad& utm_medium=display&utm_ campaign=P11853916650_EMENAB