Scientific integrity is a widely accepted institutional priority. Beyond detecting and addressing scientific misconduct, most higher education and research institutions now share the common goal of promoting a culture of good research practices.
It is tempting to place this culture in a historical perspective: the priority given to integrity today is only one moment in a longer history, that of the regulations governing scientific conduct. Montgomery and Oliver, for example, proposed distinguishing three major periods in the history of these regulations : the period before the 1970s characterized by the self-regulation of the scientific community, the period from the 1970s to the 1990s characterized by the importance given to the detection and prevention of scientific fraud, and, since the 1990s and the creation of the Office of Research Integrity , the first American federal agency responsible for investigating and preventing breaches of integrity in publicly funded biomedical research, the period of scientific integrity and the promotion of good practices.
This historical narrative of integrity may well have heuristic value, but like all grand narratives, it clashes with the facts. It suggests that the scientific community, up until the 1970s, was capable of defining and enforcing research norms and values on its own. However, the regulation that governs, for example, biomedical research predates the 1970s by a considerable margin, having taken shape in the post-World War II era with the Nuremberg Code (1947), been strengthened by the Declaration of Helsinki (1964), and gradually become institutionalized through ethics committees and legal mechanisms for the protection of individuals.
She then suggests that detecting scientific fraud and promoting best practices would represent setbacks for the autonomy of the scientific community. However, upon closer examination, institutional transformations—the adoption of the national code of ethics for research professions in 2015 , the creation of the French Office for Scientific Integrity in 2017, and the inclusion of integrity in law in 2020—are most often driven by representatives of the scientific community. And what may appear, from a distance, to be a political directive to scientists frequently stems from the scientific community’s own initiative, whether direct or indirect. The politicization of scientific integrity demonstrates the ability of a limited segment of the scientific community to seize political opportunities. Clearly, for France, the period 2015-2017 was one of those windows where, thanks to a high-profile case of scientific misconduct at the CNRS in 2015, the report by Pierre Corvol in 2016, and the circular letter from Thierry Mandon in 2017, these issues moved from a professional debate to a structured public policy issue.
Finally, this historical account of integrity seems to suggest that the institutional culture of detecting scientific fraud, characteristic of the 1970s-1990s, is now being replaced by a more positive culture of good practices and responsible research. Certainly, innovation and responsible research are now essential keywords for major scientific institutions, but the question of detecting misconduct remains more relevant than ever, especially given the shift in recent years from detecting scientific fraud to detecting questionable practices. The question of how to define and measure scientific misconduct has never been more pertinent.
While sociologists can only benefit from not attempting to become historians of science, they cannot remain blind to the major transformations in science and technology. In particular, the scientific community of the 21st century is clearly different from that of the last century. In its 2021 report, UNESCO noted that between 2014 and 2018, the number of researchers increased three times faster than the world’s population , reaching a total of more than eight million full-time equivalent researchers worldwide.
This densification of the scientific community presents significant challenges for scientific integrity. With an ever-expanding community, it is essential not only to speak with one voice and standardize guidelines and charters, but also to ensure that this voice is heard by all. While the institutionalization of integrity has accompanied a form of internationalization, not all countries are able to create the structures and roles we have described.
Furthermore, the growth of the scientific community is inevitably accompanied by an increase in the volume of scientific publications, which challenges traditional control mechanisms. Hence the development of alternatives to peer review. Much has been said about the surge in publications related to the Covid-19 crisis, but the underlying surge, perhaps less noticeable to the general public, is more structural: according to UNESCO data, between 2015 and 2019, global scientific publication output increased by 21%. Who today can guarantee the reliability of such a volume of publications?
The increasing size of the scientific community also potentially leads to a densification of international collaboration networks, within which researchers and research teams are often rival partners. The proliferation of international collaborations requires the ability to pool efforts and ensure the quality and recognition of each participant’s contribution. We have already highlighted the importance of considering the feelings of justice and injustice experienced by scientists when scientific misconduct occurs: an ignored contribution, a competitor rewarded for deviating from best practices, scientific institutions that prohibit in some cases what they reward in others—all situations that fuel these feelings.
Furthermore, the multiplication of collaborations, in a context where resources (funding, positions, distinctions, etc.) are by principle constrained, also implies logics of competition which can be experienced as so many incentives to take shortcuts.
Beyond the demographic transformation of the scientific community, the sociologist of science cannot ignore the evolution of how scientific work is carried out in the digital age. The “datafication” of science, underway since the end of the 20th century, corresponds as much to the increase in the volume of digital research data as to the growing importance of the technological infrastructures essential for its processing, storage, and dissemination. The accessibility and rapid dissemination of research data, in the name of open science, open up unprecedented perspectives for scientific collaborations. They create a redistribution of roles and expertise within these collaborations, with increasing emphasis placed on data engineering.
But here again, this technological evolution creates major challenges in terms of scientific integrity. Online scientific publishing has given rise to a predatory online market of journals that accept articles for a fee without properly evaluating them. While artificial intelligence tools are being added to the traditional tools of research teams, clandestine commercial structures, the “paper mills,” are misusing these digital tools to manufacture and sell fraudulent scientific articles to authors seeking publication. While the immateriality of digital data allows for the accelerated circulation of research results, as seen during the Covid-19 crisis, it also enables increasingly sophisticated manipulation. These structural transformations create an unprecedented demand for vigilance, which is being met, each in its own way, by post-publication evaluation, as well as by those sometimes called the new detectives of science who hunt online for anomalies in published articles (duplicate images, inconsistent data, plagiarism) and in doing so contribute to highlighting fraud or dubious practices.
More fundamentally, this surge in digital research data is generating unprecedented tensions between scientific integrity and research ethics. In an interview, a physicist, working daily with the terabytes of data generated by a particle accelerator, remarked, with a touch of irony, that he readily understood the institutional imperative to archive all his data on servers in the name of scientific integrity, but that this systematic storage of an ever-increasing volume of data directly contradicted the environmental ethics of research advocated by his institution. Scientific integrity or research ethics? Must we choose?
This highlights the diverse dilemmas that scientists face daily in their work. Science in the digital age operates within a delicate balance between scientific integrity and research ethics. One of the current challenges is undoubtedly how to maximize the benefits of digital technologies while minimizing their negative impacts. Hence the need for research institutions to promote a culture of integrity that goes beyond simply adhering to best practices, by incorporating the values of social responsibility to which, as our surveys show, a growing segment of the scientific community aspires.
Author Bios: Catherine Guaspare is a Sociologist, Research Engineer at the National Centre for Scientific Research (CNRS) and Michel Dub is a Sociologist and CNRS Research Director at Sorbonne University
