Research integrity

Research integrity or scientific integrity is an aspect of research ethics that deals with best practice or rules of professional practice of scientists.

First introduced in the 19th century by Charles Babbage, the concept of research integrity came to the fore in the late 1970s. A series of publicized scandals in the United States led to heightened debate on the ethical norms of sciences and the limitations of the self-regulation processes implemented by scientific communities and institutions. Formalized definitions of scientific misconduct, and codes of conduct, became the main policy response after 1990. In the 21st century, codes of conduct or ethics codes for research integrity are widespread. Along with codes of conduct at institutional and national levels, major international texts include the European Charter for Researchers (2005), the Singapore statement on research integrity (2010), the European Code of Conduct for Research Integrity (2011 & 2017) and the Hong Kong principles for assessing researchers (2020).

Scientific literature on research integrity falls mostly into two categories: first, mapping of the definitions and categories, especially in regard to scientific misconduct, and second, empirical surveys of the attitudes and practices of scientists. Following the development of codes of conduct, taxonomies of non-ethical uses have been significantly expanded, beyond the long-established forms of scientific fraud (plagiarism, falsification and fabrication of results). Definitions of "questionable research practices" and the debate over reproducibility also target a grey area of dubious scientific results, which may not be the outcome of voluntary manipulations.

The concrete impact of codes of conduct and other measures put in place to ensure research integrity remain uncertain. Several case studies have highlighted that while the principles of typical codes of conduct adhere to common scientific ideals, they are seen as remote from actual work practices and their efficiency is criticized.

After 2010, debates on research integrity have been increasingly linked to open science. International codes of conduct and national legislation on research integrity have officially endorsed open sharing of scientific output (publications, data, and code used to perform statistical analyses on the data) as ways to limit questionable research practices and to enhance reproducibility. Having both the data and the actual code enables others to reproduce the results for themselves (or to uncover problems in the analyses when trying to do so). The European Code of Conduct for Research Integrity 2023 states, for example, the principles that, "Researchers, research institutions, and organisations ensure that access to data is as open as possible, as closed as necessary, and where appropriate in line with the FAIR Principles (Findable, Accessible, Interoperable and Reusable)
for data management" and that "Researchers, research institutions, and organisations are transparent about how to access and gain permission to use data,
metadata, protocols, code, software, and other research materials". References to open science have incidentally opened up the debate over scientific integrity beyond academic communities, as it increasingly concerns a wider audience of scientific readers.

Definition and history

''Research integrity'' or ''scientific integrity'' became an autonomous concept within scientific ethics in the late 1970s. In contrast with other forms of ethical misconducts, the debate over research integrity is focused on "victimless offence" that only hurts "the robustness of scientific record and public trust in science". Infractions to research integrity include chiefly "data fabrication, falsification, or plagiarism". In that sense, research integrity mostly deal with the internal process of science. It can be treated as community issue, that should not involve external observers: "research integrity is more autonomously defined and regulated by the community, while research ethics (again, a narrow definition) has closer links to legislation".

Emergence of the issue (1970–1980)

Before the 1970s, ethical issues were largely focused on the conduct of medical experiments, especially in regard to tests on humans. In 1803, the "code" of Thomas Percival created a moral foundation for experimental treatments that "was built upon fairly regularly" throughout the next two centuries, notably by Walter Reed in 1898 and by the Berlin code in 1900. After the Second World War, the Nazi human experimentations motivated the development of international, widely acknowledged codes of research ethics, such as the Nuremberg code (1947) and the World Medical Association Declaration of Helsinki.

According to Kenneth Pimple, Charles Babbage was the first author to set aside the specific issue of scientific integrity. In the ''Reflections on the Decline of Science in England, and on Some of its Causes'', first published in 1830, Babbage identified four classes of scientific frauds, from outright forgery to varied degrees of arrangements and ''cooking'' of the data or the methods.

Research integrity became a major debated topic in biological sciences after 1970, due to a combination of factors: the development of advanced data analysis methods, the growing commercial relevancy of fundamental research, and the increased focus of federal funding agencies in the context of big science. In 1974, the "painted mouse incident" attracted unprecedented media attention: William Summerlin inked a black dot on a mouse to claim a treatment has been a success. Between 1979 and 1981, several major cases of scientific fraud and plagiarism drew a greater focus on the issue from researchers and policymakers in the United States: as many as four important frauds occurred in the summer of 1980.

At the time, the "scientific community responded to reports of 'scientific fraud' (as it was often called) by asserting that such cases are rare and that neither errors nor deception can be hidden for long because of science's self-correcting nature". A journalist of ''Science'', William Brad, took the opposite position and made an influential contribution to the issue of research integrity. In an answer to the US House of Representatives Science and Technology subcommittee, he highlighted that "cheating in science was nothing new" but, until recently, "had been handled as an internal affair". In a detailed investigation co-signed with Nicholas Wade, ''Betrayers of Science'', Brad described scientific fraud as a structural problem: "As more cases of frauds broke into public view (…) we wondered if fraud wasn't a quite regular minor feature of the scientific landscape (…) Logic, replication, peer review — all had been successfully defied by scientific forgers, often for extended periods of time." Other early assessments of the systematicity of scientific frauds presented a more nuanced picture. For Patricia Wolff, along with a few obvious manipulations, there were a wide range of grey areas, which were due to the complexity of fundamental research: "the boundaries between egregious self-deception, culpable carelessness, fraud, and just plain error, can be very blurred indeed". Characteristically, the debate led to a re-evaluation of past scientific practices. In 1913, a well-known scientific experiment on electron charge by Robert Millikan was explicitly based on discarding some results that would not agree with the underlying theory: while well received at the time, by the 1980s this work had come to be considered as a textbook example of scientific manipulation.

Formalization of research integrity (1990–2020)

By the end of the 1980s, the amplification of misconduct scandals and the heightened political and public scrutiny put scientists in a difficult position in the United States and elsewhere: "The tone of the 1988 US congressional oversight hearings, chaired by Rep. John Dingell (D-MI), that investigated how research institutions were responding to misconduct allegations reinforced many scientists’ view that both they and scientific research itself were under siege." The main answer was procedural: research integrity has "been codified into numerous codes of conduct field specific, national, and international alike." This policy response largely stemmed from research communities, funders and scientific administrators. In the United States, the United States Public Health Service and the National Science Foundation adopted "similar definitions of misconduct in science" in 1989 and 1991. The concepts of ''research integrity'' and its reverse, ''scientific misconduct'' were especially relevant from the perspective funding bodies, since it made it possible to "delineate the research-related practices that merit intervention": lack of integrity led not only to unethical but inefficient research and funds have better to be allocated elsewhere.

After 1990, there was a "veritable explosion of scientific codes of conduct". In 2007 the OECD published a report on best practices for promoting scientific integrity and preventing misconduct in science (Global Science Forum). Such international texts include:
* European Charter for Researchers (2005)
* the Singapore statement on research integrity (2010)
* European Code of Conduct for Research Integrity of All European Academies (ALLEA) and the European Science Foundation (ESF) (2011 revised in 2017).

There are no global estimates of the total number of codes of conduct related to research integrity. A UNESCO project, the ''Global Ethics Observatory'' (no longer accessible after 2021), referenced 155 codes of conduct but "this is probably just a fraction of the total number of codes produced in recent years." Codes have been created in highly diverse settings and show a wide variation in scale and ambition. Along with national-scale codes, there are codes for scientific societies, institutions and R&D services. While these normative texts may frequently share a core of common principles, there has been growing concern "over fragmentation, lack of interoperability and varying understandings of central terms can be sensed".

Taxonomy and classification

In codes of conduct, the definition of research integrity is usually negative: the collection of norms aims to single out different forms of unethical research and scientific misconduct with varying degrees of gravity.

The multiplication of codes of conduct has also corresponded with an expansion of scope. While the initial debate was focused on "three deadly sins of scientific and scholarly research: fabrication, falsification and plagiarism", attention has later shifted "to the lesser breaches of research integrity". In 1830, Charles Babbage introduced the first taxonomy of ''scientific frauds'' that already encover some forms of questionable research practices : hoaxing (a voluntary fraud "far from justifiable"), forging ("whereas the forger is one who, wishing to acquire a reputation for science, records observations which he has never made"), trimming (which "consists in clipping off little bits here and there from those observations which differ most in excess from the mean" and cooking. Cooking is the main focus of Babbage as an "art of various forms, the object of which is to give to ordinary observations the appearance and character of those of the highest degree of accuracy". It falls done under several sub-cases such as data selection ("if a hundred observations are made, the cook must be very unlucky if he cannot pick out fifty or twenty to do the serving up", model/algorithm selection ("another approved receipt (…) is to calculate them by two different formulae") or use of different constants.

In the late 20th century, this classification has been greatly expanded and have come to encompass a wider range of deficiencies than intentional frauds. The formalization of research integrity entailed a structural change in the vocabularies and the concept associated with it. By the end of the 1990s, use of the expression "scientific fraud" was discouraged in the United States, in favor a "semi-legal term": ''scientific misconducts''. The scope of scientific misconducts is expansive: along with data fabrication, falsification and plagiarism it includes "other serious deviations" that are demonstrably done in bad faith. The associated concept of questional research practice, first incepted in a 1992 report of the Committee on Science, Engineering, and Public Policy, has an even broader scope, as it also encompass potentially non-intentional research failures (such as inadequacies in the research data management process). In 2016, a study identified as much as 34 questionable research practices or "degree of freedom", that can occur at all the steps of the project (the initial hypothesis, the design of the study, collection of the data, the analysis and the reporting).

After 2005, research integrity has been additionally redefined through the perspective of research reproducibility and, more specifically, of the "reproducibility crisis". Studies of reproducibility suggest that there is continuum between irreproducibility, questionable research practices and scientific misconducts: "Reproducibility is not just a scientific issue; it is also an ethical one. When scientists cannot reproduce a research result, they may suspect data fabrication or falsification." In this context, ethical debates are less focused on a few highly publicized scandals and more on the suspicion that the standard scientific process is broken and fails to meet its own standard.

Current landscape and issues

Prevalence of ethical issues

In 2009, a meta-analysis of 18 surveys estimated that less than 2% of scientists "admitted to have fabricated, falsified or modified data or results at least once". Real prevalence may be under-estimated due to self-reporting: regarding "the behaviour of colleagues admission rates were 14.12%". Questionable research practices are more widespread as more than one third of the respondents admit to have done it once. A large 2021 survey of 6,813 respondents in the Netherlands found significantly higher estimate, with 4% of the respondents engaging in data fabrication and more than half of the respondents engaging in questionable research practices. Higher rates can be either attributed to a deterioration of ethic norms or to "the increased awareness of research integrity in recent years". The higher rates of self-declared scientific misconducts are found in the medical and life science, with at much as 10.4% respondents surveyed in the Nerthelands admitting a scientific fraud (either fabrication of falsification of the data).

Other forms or scientific misconducts or ''questionable research practices