Recent headlines have been glaring. “Errors found in dozens of papers by top scientists at Dana-Farber Cancer Institute.” ¹ “Stanford president will resign after report found flaws in his research.”Reference Saul ² “Honesty researcher committed research misconduct, according to newly unsealed Harvard report.”Reference O’Grady ³ “A top government scientist engaged in research misconduct, NIH finds.” ⁴

Clearly, something is very wrong, affecting even top researchers and institutions, across fields ranging from basic science to clinical research to social science. Meanwhile, public trust in science is on shaky ground following the COVID-19 pandemic,Reference Achenbach ⁵ with a second Trump administration upending government and academic science through communication freezes,Reference Wadman and Kaiser ⁶ data suppression,Reference Stobbe and Schneider ⁷ grant holds,Reference Kozlov and Garisto ⁸ and unprecedented funding cuts.Reference Jewett, Stolberg, Diamond, Keating and Kaiser ⁹ The stakes for ensuring good, trustworthy science couldn’t be higher. Yet research misconduct wastes resources and hobbles progress as scientists pursue false leads.Reference Piller ¹⁰ It can also more directly harm patients as scientific misdirection influences their care.Reference Piper ¹¹

Research misconduct — typically defined as fabrication, falsification, or plagiarism — is not a new concern. Representative Al Gore chaired the first Congressional hearing on research misconduct in 1981. ¹² In 1985, Congress passed a law requiring the Department of Health and Human Services (HHS) to issue a regulation mandating awardee and applicant institutions to establish processes to review scientific fraud and report “substantial” fraud to the Secretary, while requiring the National Institutes of Health (NIH) to establish a process for receiving and responding to institutional reports. ¹³ Final regulations were published in 1989, the same year the Public Health Service (PHS) created the precursor offices to today’s Office of Research Integrity (ORI). ¹⁴ New regulations, the PHS Policies on Research Misconduct, became effective in 2005 and were recently revised through ORI’s 2024 Final Rule. ¹⁵ At least for now, the government’s approach continues to rely on research institutions to “self-police” by investigating and responding to misconduct perpetrated by their own researchers. ¹⁶ NIH has also taken recent steps to protect integrity in intramural research carried out by the agency’s own scientists, ¹⁷ although the 2024 NIH Final Scientific Integrity Policy was rescinded on March 28, 2025, “to ensure alignment with the [Trump] Administration’s priorities.” ¹⁸ Complementing these approaches to publicly-funded research, the Food and Drug Administration (FDA) plays a central role in vetting industry research submitted in support of product marketing applications. Across the government’s scientific agencies, however, major staffing cuts imposed by the Trump administration are likely to hobble research integrity efforts.Reference Stolberg and Jewett ¹⁹

Despite decades of effort to address research misconduct concerns, misconduct has been frustratingly persistent, driven by a variety of structural factors including the “publish or perish” mentality of academia, bias against publishing “negative” studies, accolades heaped on scientists who are first to report splashy high-impact results, limited resources and incentives for scientific replication, insufficient coordination between relevant stakeholders, and reputational concerns, among other challenges. While these issues are not new, there are several novel factors influencing today’s research misconduct landscape. Instances of misconduct are increasingly identified thanks to a growing community of online “sleuths” committed to rooting them out, alongside proliferating platforms for discussing and publicizing concerns.Reference Wapner ²⁰ There are also a variety of new screening tools that can be employed by journals, research institutions, and others, although with imperfect results.Reference Chen and Wosen ²¹ At the same time, generative artificial intelligence (genAI) capable of doing everything from drafting manuscripts to fabricating datasets and images, which can be especially challenging to spot, has facilitated misconduct.Reference Gerhard ²² Misconduct has also been allowed to flourish through the growth of so-called “paper mills” that sell papers, authorship slots, citations, and peer reviews.Reference Marcus and Oransky ²³

With these changes and new examples of research misconduct seemingly reported every day, we are at an opportune moment to reassess how institutional stakeholders — including universities, publishers, funders, and regulators — should address persistent and emerging concerns about research misconduct, from prevention to response. On October 10, 2024, under the auspices of the University of Pennsylvania Perelman School of Medicine’s Institute for Translational Medicine and Therapeutics (ITMAT), we held a public symposium to bring together experts across sectors and disciplines to reflect on key challenges and solutions. In this special issue of The Journal of Law, Medicine, and Ethics (JLME), symposium participants — leading institutional research integrity officers, science sleuths, researchers, academic leaders, journal editors, and experts in ethics and law — share their views on how we’ve arrived at the current research misconduct crisis and what should be done about it.

Before summarizing the contributions of these participants, we must acknowledge a conspicuous absence from this special issue. ORI Director Sheila Garrity participated in our October symposium, making public remarks about ORI’s new regulations and efforts to enhance partnership with research institutions — an approach that contrasts with alternatives favored by some commenters that would instead have ORI play a stronger direct role identifying and investigating research misconduct and imposing penalties. With her colleagues, Loc Nguyen-Khoa (also of ORI) and Jason Weixelbaum (of Rose Li and Associates), Garrity contributed an article for this issue based on her symposium presentation, which was then peer reviewed and accepted for publication in December 2024 by the special issue editors. After formal submission to the journal, JLME’s editor-in-chief, Aaron Kesselheim, accepted the article for publication on January 19, 2025.

On January 20, 2025, President Trump took office and on January 21, Acting HHS Secretary Dorothy Fink issued a memo imposing a widespread communication freeze, directing HHS staff to “[r]efrain from publicly issuing any document … or communication … until it has been reviewed and approved by a Presidential appointee.” ²⁴ The same day, we received a message from ORI’s communications director indicating that the ORI manuscript accepted for this special issue could not proceed further and would “need to be reviewed by OASH [Office of the Assistant Secretary for Health]’s Comms Team to ensure it aligns with HHS’ priorities moving forward.” She subsequently explained that this was part of the new administration’s directive “that anything going out has to undergo review irrespective of what happened before inauguration.” As the special issue goes to press, we have received no updates on the status of this review.

Government employees lack the privilege of academic freedom and, across administrations, traditionally have been required to have their publications approved prior to submission. Although purely scientific publications should be reviewed exclusively for accuracy, it is reasonable for the government to review its employees’ policy publications to allow the government to advance consistent messaging. However, the Trump administration’s blanket approach and suppression of an article that had already been internally reviewed, peer reviewed, formally submitted, and accepted for publication — about regulations that have been final for months — raises concerns about how government regulators and scientists will be allowed to communicate with the public going forward and by what standard their communications will be tested for alignment. We hope to be able to include ORI’s article in a future issue of JLME, although we have no timeline, no guarantees, and no indication of whether the article’s contents will be required to change as a condition of publication. This is a loss for the regulated community, for which the article would have provided deeper insight into ORI’s thinking and approach regarding the new regulations.

With that challenge aside, the special issue begins from the researcher’s perspective, with an article from Max H. Bazerman describing a high-profile episode of presumed research fraud in which he was involved as a collaborator and coauthor. In his article, “An Insider’s Perspective on How to Reduce Fraud in the Social Sciences,” Bazerman outlines changes in social science research practices, initiated since the paper in which he was involved was published, that have made such events less likely. These changes include greater transparency about research hypotheses and methods, preregistration of research plans, and data sharing. He highlights exemplary social science journals, particularly in psychology, that have instituted practices to enhance rigor and reproducibility. These include public sharing of data, original research materials, and analysis scripts; “registered reports” in which study plans are reviewed by journals in advance and, if the methods are rigorous and the question sufficiently important, the journal pre-commits to publish results; and incorporation of statistical review into the editorial process. Finally, Bazerman emphasizes that there is much more that universities can do to promote integrity, including robust investigations of alleged misconduct, transparency about outcomes, and support for coauthors (especially junior collaborators) whose careers are threatened by association with episodes of misconduct.

In the next article, “Misconduct Detection — Evolving Methods and Lessons from 15 Years of Scientific Image Sleuthing,” Paul S. Brookes draws on his extensive experience as a scientific image sleuth to provide what he calls “a toolbox for discovery and reporting of problematic scientific images and other data.” He begins by describing hardware and software requirements for image analysis before detailing a range of methods for spotting irregularities in scientific images and data. Next, he shares best practices for documenting workflow and assesses the pros and cons of different avenues for reporting problematic images and data. He also provides advice for dealing with the legal ramifications that can come with reporting suspected scientific misconduct. While his article is addressed primarily to investigators currently engaged — or seeking to become engaged — in identifying instances of scientific misconduct, Brookes closes with recommendations about how journals and universities should respond to reports of potential misconduct.

Moving on to the realm of science publication, Ellie Rose Mattoon, Arturo Casadevall, and Ferric C. Fang use COVID-19 article retractions to highlight the concerns that allow problematic science to make its way into journals and often stay there. In their article, “Retractions of COVID-19-related Research Publications During and After the Pandemic,” Mattoon et al. describe the characteristics of 244 articles related to COVID-19 that were indexed in PubMed and retracted between 2020 and early 2024. Corresponding authors on these articles were most likely to be from China, the United States, India, and Saudi Arabia. Almost all were published in open-access journals and their mean time from publication to retraction was about 13 months. The most common reasons for retraction were error and manipulation of the peer review system. Among retraction notices that provided a reason for the retraction, fraud was seen less often than in prior reviews. Mattoon et al. suggest that lower quality standards during the pandemic, combined with urgency of the public health need and the move to open-access publishing, help explain the characteristics of the retractions that they observed. Finally, they suggest that many of the problems that led to retraction, such as “compromise of the peer review process, plagiarism, authorship issues, [and] lack of ethics approvals,” could have been identified by better journal editorial processes.

In her article, “Systematic Reviewers Have an Obligation to Promote Research Integrity,” Lisa Bero identifies authors of systematic reviews, such as those working under the auspices of the Cochrane Collaboration, as ideally placed to identify and draw attention to potentially problematic studies. These systematic reviewers have the advantage of assessing individual studies in the context of other studies addressing the same question. They also routinely use risk-of-bias instruments and exercise an overarching commitment to improving the quality of the scientific literature. Bero describes Cochrane’s ongoing effort to develop and validate a tool, “Investigating Problematic Clinical Trials in Systematic Reviews,” or INSPECT-SR, to identify studies that may be compromised by integrity problems. In addition, she describes Cochrane’s policy for how authors should proceed when they identify potential problems, starting with attempts to contact authors and proceeding to notification of journal editors if concerns remain unresolved. Initiatives and policies such as Cochrane’s can improve how the scientific community handles potentially problematic studies and, more importantly, may reduce the likelihood that such studies enter the literature in the first place.

Next, Howard Bauchner, Robert Steinbrook, and Rita F. Redberg share their experience dealing with research misconduct as editors of prominent medical journals. In their article, “Research Misconduct and Medical Journals,” these authors outline the important role journals play disseminating research results and — by extension — their responsibility for addressing allegations of research misconduct. They argue that all journals ought to have a consistent, clearly articulated process for dealing with allegations of misconduct, which they then outline. The authors also name possible remedies, including corrections or exchanges of letters to the editor — for simple errors — or, in more serious cases, simple retraction or retraction and replacement. Ultimately, such processes and remedies are intended to promote public confidence in the research enterprise.

The special issue next moves from journals to research institutions. The ORI article was slated to come next but had to be pulled for reasons described above, leaving a gap regarding important regulatory perspectives. Instead, we move to several articles describing the practical challenges faced by research institutions as they continue to play a critical role leading misconduct investigations under the new ORI regulations.

In their article, “Managing Multi-Institutional Jurisdiction in Cases of Research Misconduct,” Leslie Thornton, Devin Cohen, Bailey Jenks, Mark Barnes, and Barbara Bierer take on an important challenge in the modern research environment: research misconduct concerns that span institutional collaborations. As more research is conducted across research sites, sometimes including collaborations between academic centers and industry partners, the question arises: how should these institutions handle misconduct allegations and determine their respective jurisdictional roles? Thornton et al. propose a “reasonableness test” to answer this question, encouraging collaborators to determine “jurisdiction by examining the nexus between the research at issue and the various individual entities involved, including the acuity of any actual or potential institutional conflicts of interest.” More specifically, they suggest considering the location of the research activities in question, the respective capacity and expertise of each collaborating institution, the position and seniority of the respondents, and the implications of external funding and associated legal obligations. Ultimately, the coauthors suggest that collaborators make these decisions about their respective roles through contractual arrangements before the need arises. If disagreements arise, however, regulators may need to weigh in.

Next, Nathaniel Jaffe, Minal Caron, Lauren Walsh, Barbara Bierer, and Mark Barnes address the potential for defamation claims to arise out of research misconduct allegations, investigations, and responsive disclosures, in their article “Defamation Claims Arising from Research Misconduct Cases: Best Practices for Institutions.” Emphasizing that defamation lawsuits can be used to chill appropriate efforts to identify and address misconduct, these authors clarify the relevant legal standard, noting that true statements and opinions are safe but might still face the burdens of litigation. They also describe key areas of legal uncertainty at the intersection of defamation and institutional regulatory obligations to protect the public health in the face of alleged research misconduct, including disclosures to journals and other parties within the limitations of a “need to know” standard. Although some states have passed laws to combat “strategic lawsuits against public participation” (anti-SLAPP laws), which are intended to address the problem of defamation suits being brought to scare or punish critics, the authors note these have not always succeeded in protecting those raising concerns about research misconduct. They conclude with a series of recommendations for institutions, journals, and complainants to minimize the risk of facing or losing a defamation case, including careful tailoring of public statements and clarification of institutional policies regarding disclosure of research misconduct allegations.

Beyond the practical challenges facing those responsible for conducting research misconduct investigations, the next several articles offer ideas for how to more fundamentally reshape them. In their article, “Academic Research Integrity Investigations Must be Independent, Fair, and Timely,” Matthew Schrag, Kevin Patrick, and Elisabeth Bik observe that research institutions face powerful incentives to avoid the reputational and financial harm that can come from a finding of research misconduct. The authors argue that these incentives create “an inherent conflict of interest” that can lead to “delayed and anemic” institutional investigations of alleged misconduct. To overcome this structural challenge, the authors propose that “credible reports of a violation or potential violation of research integrity of substantial magnitude should be investigated by an independent investigative body.” While such investigations are not the norm, they are not unprecedented. As the authors note, the Stanford Board of Trustees recently convened an independent expert body to investigate allegations of research misconduct from the laboratory of Marc Tessier Lavigne, who was then the university’s president. The article concludes with a number of recommendations on the composition and operational guidelines for independent investigative bodies.

Following on the idea of stronger investigations, in their article, “Why Universities Should Make Misconduct Reports Public,” Ivan Oransky and Adam Marcus make the case for why universities ought to disclose reports of their research misconduct investigations. Such reports can provide a critical source of evidence for evaluating and ultimately strengthening universities’ research integrity programs and policies. Yet, as Oransky and Marcus note, the “vast majority of misconduct investigation reports … remain hidden from view.” In defense of the status quo, universities have cited the need to maintain confidentiality in order to protect the privacy of those accused of research misconduct and those who serve on investigation committees. The authors acknowledge that there is some basis for these concerns but contend that “the preservation of confidentiality is not in itself a societal good on par with, say, safeguarding the public trust in science, protecting the public health, or the effective stewardship of taxpayer resources.” Additionally, they propose practical measures, such as redacting committee member names and sensitive witness statements, that could mitigate privacy risks.

Next, Garret A. FitzGerald advocates for academic institutions to clearly distinguish between their responsibility to ensure scientific validity when investigating research integrity concerns, which demands speed and transparency, and their obligations to address suspected misconduct, which demands due process, fairness to all parties, and a high burden of proof. In his article, “Segregate Assessment of Data Validity from the More Complex Issue of Fraud,” FitzGerald emphasizes the importance of replication to trustworthy science, noting that while the biopharmaceutical industry has strong incentives to replicate academic studies before relying on them to inform business decisions about drug development, such replications are rarely published to inform the broader scientific community. In contrast to industry, FitzGerald argues that academic investigators have few incentives or resources to conduct replication studies and that their institutions are “also not motivated to address the issue of scientific validity in a transparent way.” This lack of motivation is particularly problematic, he argues, because other entities that are central to the scientific enterprise — particularly funders and journals — rely on academic institutions to ensure rigor and investigate when concerns are raised.

The special issue closes with two foundational ideas for reshaping current approaches to research integrity. First, Julia L. Briskin and C.K. Gunsalus argue that research misconduct can best be avoided — and research integrity promoted — by improving institutional culture and holding institutions accountable. In their article, “Fostering Accountability: How Institutions Can Promote Research Integrity with Practical Tools and Knowledge,” Briskin and Gunsalus note that the existing flawed reward structure around scientific success focuses on outcomes (what research is done) without regard to process (how research is done), leading to tolerance for toxic work environments that can undermine research integrity. In addition, complacency around Responsible Conduct of Research training allows it to become tedious, insufficiently tailored, and rote, not fit for purpose of meeting the needs of scientific researchers across career stages. The antidote, according to Briskin and Gunsalus, will come through tailored professional development programs geared toward helping scientists address interpersonal challenges and team dynamics, as well as efforts to create transparent institutional processes for raising and responding to concerns about institutional culture. They argue that institutions must measure and assess institutional “temperature” through periodic surveys of the research climate, including comparisons to relevant benchmarks such as prior performance and performance of peer units and institutions, with the results used to tailor specific responses. Ultimately, Briskin and Gunsalus conclude that “[c]ultivating a culture of accountability through attending to the environments in which people work, and equipping them with the tools to interact effectively within those environments will lead to better research[.]”

Finally, in her article “The Need for Prospective Integrity Standards for the Use of Generative AI in Research,” Kayte Spector-Bagdady questions the current inconsistency in approaches to regulating human subjects research versus regulating research misconduct. She explains that human subjects have been appropriately protected through prospective oversight requirements, including prior review by Institutional Review Boards, grounded in an understanding that the stakes are too high to rely on trust in researchers combined with the imposition of backward-looking penalties only if harm in fact occurs. In contrast, she argues that approaches to research misconduct have been persistently and problematically retrospective, with overly heavy emphasis on trust and professional self-regulation. Now, with new data-driven research technologies, including genAI, the risks of research misconduct have become even greater. Rather than relying on retrospective approaches, Spector-Bagdady argues that researchers should be held accountable for following prospective requirements for research integrity — as they are for protecting human subjects — without the need for a third party to later establish that falsification, fabrication, or plagiarism in fact occurred. Violation of the rules would become the key criterion for research misconduct penalties, demanding establishment of clear expectations around appropriate research conduct and use of AI technologies.

Together, these articles provide a range of ideas for how scientists, sleuths, journal editors, research institutions, and regulators can prevent and respond to research misconduct, ranging from ideas that are presently actionable with sufficient institutional will and commitment to those that would require new regulatory changes — which are unfortunately unlikely given recent revisions and the current political climate. Considering the frequency and impact of research misconduct, every stakeholder should take the steps they can to improve the integrity of science, with special obligations on the institutions that fund, oversee, and publish research to ensure that scientists prioritize the pursuit of truth and that their work is trustworthy.