July 8, 2022
Research integrity: An evolving challenge for scholarly publishers
Scholarly publishing has seen a steep rise in research integrity cases over recent years. Motives for this activity are usually profit-driven but sometimes accidental. Either way, integrity breaches threaten the trustworthiness of scientific research and the publishers' reputation.
Several leading publishers have approached us for help after retracting numerous proceedings volumes. The retractions are due to various forms of research misconduct discovered by the publishers. Since large-scale retractions are a big concern in the industry (and are likely to receive unwanted publicity from groups such as Retraction Watch), they are eager to take steps to improve their research integrity efforts. High retraction rates damage reputations and negatively impact the wider scientific community.
Publishers are understandably keen on quick and simple ways to solve this problem. The challenge boils down to the need to quickly identify articles in a proceedings volume with potential integrity issues so that they can be excluded, preferably in an automated way. The intention is that this will minimize delays to the publishing process and ensure all published materials will protect research integrity.
Challenges faced in ensuring integrity and quality of research
There are three main trends that publishers in the industry face when assessing the authenticity and trustworthiness of scientific research. Tortured Phrases, Citation Manipulation and Salami Publishing are often symptoms of a broader phenomenon known as Paper Mills. These are commercial companies that create fake articles and sell citations for their customers to help them get published. When these issues are missed during peer-review and pre-publication, they contribute to a high rate of retractions.
These known issues are widespread and continue to grow.
Nature covered the issue of tortured phrases in August 2021, when a group of computer scientists discovered some unusual phrasing in a research piece. They could not understand why the terms ‘counterfeit consciousness’, ‘profound neural organization’ and ‘colossal information’ were used instead of the more widely recognized terms ‘artificial intelligence,’ ‘deep neural network’, and ‘big data.’
To understand the potential scale of the problem, the researchers searched so-called tortured phrases in journal articles listed in Dimensions, uncovering over 800 publications that included at least one of the phrases, many of which came from a single article.
“It harms science. You cannot trust these papers, so we need to find them and retract them,” said Guillaume Cabanac, a computer scientist at the University of Toulouse who worked on the study.
Authors, editors, publishers, and institutions see citations as an important indicator of productivity. So what is citation manipulation? It is a situation where guest editors, authors, or related author groups are overly cited in more than 5% of the reference list in a paper. Researchers are rewarded with promotions and additional funding.
Citation Manipulation can occur in three ways (according to COPE):
- Excessive citation of an author’s research to build the author’s number of citations - the author’s own research (self-citation)
- Excessive citations of articles from published journals in which the author contributed, increasing the number of journal citations.
- Excessive citation of another author’s work or journal.
Originally conceived as a threat to research integrity in 1955, this is a complex issue to resolve, not least because of the interpretation of multiple co-authors collaborating on one paper.
Salami publishing or salami slicing can be defined as a publication of two or more articles derived from a single study. It is usually characterized by similarity of hypothesis, methodology, or results, but not similarity in text. Existing software applications do not objectively detect these aspects of publications.
The segmentation of a large study into two or more publications differs from reporting the same data in two publications, but it remains an unacceptable scientific practice. Salami slicing can lead to a distortion of the literature by leading unsuspecting readers to believe that data presented in each salami slice (i.e., journal article) is derived from a different subject sample. Consider the examples provided by Kassirer and Angell (1995), former editors of The New England Journal of Medicine:
“Several months ago, for example, we received a manuscript describing a controlled intervention in a birthing center. The authors sent the results on the mothers to us and the results on the infants to another journal. The two outcomes would have more appropriately been reported together.”
According to COPE, this term describes the process of submitting manufactured manuscripts to a journal on behalf of researchers for a stipulated fee. This makes it easier for the researcher to be published or offered as an ‘author for sale.’
The concerns with these submissions include faked or manipulated data and images, the use of stock images, substantial authorship changes, and plagiarism. These are not detected because they come from a translated version of another article.
Paper mills can sometimes only be detected after a pattern of submission behavior across several submissions and journals. Even original data can be convincingly fabricated and must be independently verified, a challenge for a busy journal with busy peer reviewers and editors, who are usually not in a position to spot such problems.
The scholarly publishing world is facing unprecedented pressure from rogue elements out to profit from these hard-to-detect practices. As the issue of research integrity evolves, peer reviewers and editors will need more innovative tools to combat these challenges and manage research proceedings better. Morressier takes this threat to research extremely seriously and constantly innovates to create solutions for a more seamless review process.