I. False Statements Purporting to be Fact

The first element of a defamation claim is that the defendant has made a “false statement purporting to be fact.” ²² Statements of truth and statements of opinion cannot be false statements in the context of defamation law. ²³ This nuance was central to the disposition of the defamation lawsuits arising from research misconduct allegations against Dr. Carlos Croce, a cancer researcher at Ohio State University. ²⁴ Dr. David Sanders, an independent research misconduct sleuth, found evidence of “manipulated data and plagiarized text” in a subset of Dr. Croce’s many scientific publications. ²⁵ Dr. Sanders brought his evidence to multiple scientific journals and to the attention of The New York Times, which in turn brought the allegations to Dr. Croce and Ohio State University after The New York Times published an article about the alleged misconduct.Reference Oransky ²⁶ Dr. Croce brought defamation charges against Dr. Sanders and The New York Times. ²⁷

In Dr. Croce’s case against Dr. Sanders, Dr. Croce identified several allegedly defamatory statements of Dr. Sanders, such as, “It’s a reckless disregard for the truth,” “image fabrication, duplication and mishandling, and plagiarism in Dr. Croce’s papers is routine,” and “Dr. Croce is knowingly engaging in scientific misconduct and fraud.” ²⁸ The U.S. Sixth Circuit Court of Appeals (“Sixth Circuit”) held that that the descriptors “reckless,” “routine,” and “knowingly” were “imprecise” value judgments that indicated that the statements were in fact protected opinion statements, particularly considering that they were framed in The New York Times as “Dr. Sanders argues…” ²⁹

Two other allegedly defamatory quotes were dismissed as true: that the Croce lab was “violating scientific norms” and “they were able to get away with this relatively simple manipulation … over and over again.” ³⁰ For the former, the Sixth Circuit held that because scientific journals had decided to alert the public to issues in certain of Dr. Croce’s papers, the statement regarding “violating scientific norms” was “substantially” true. ³¹ The latter was also found to be true; journals had found and corrected instances of image manipulation in several of Dr. Croce’s publications, and the authors had, in fact, “got away with it (for a time).” ³² The Sixth Circuit found in favor of Dr. Sanders, ruling his statements to be non-defamatory truth and opinion. ³³

Dr. Croce’s case against The New York Times was also resolved in favor of the defense, with the Sixth Circuit finding that The New York Times’ article “present[ed] two sides of the controversy,” including both the allegations against Dr. Croce and making clear that they are “allegations,” “criticisms,” and “complaints,” while also presenting Dr. Croce’s professional pedigree. ³⁴ The Court held that a neutral publication on the details of a dispute is not defamatory. ³⁵ Despite these findings for the defendants, Dr. Croce’s suit against Dr. Sanders required almost four years to resolve, from March 2017 to February 2021, and the suit against The New York Times more than two years, from May 2017 to July 2019. ³⁶

A finding of opinion was also central to the dismissal of certain of the defamation claims in Gino v. President and Fellows of Harvard College. ³⁷ In this case, Harvard Business School researcher Dr. Francesca Gino sued Harvard for defamation, among other claims, allegedly arising from the institution’s notices sent to journals and coauthors after a final finding of research misconduct by Dr. Gino. ³⁸ The notices stated that the institution “believe[s]” that data are invalid but did not discuss research misconduct nor accuse Dr. Gino of altering the data. ³⁹ The U.S. District Court of the District of Massachusetts (“Massachusetts District Court”) dismissed the defamation claims related to the notices, finding that the institution’s language “amount[ed] only to a statement of … subjective view or interpretation of its investigation.” ⁴⁰ The Massachusetts District Court additionally praised the “measured and professional … tone” of the notices and the absence of accusatory language. ⁴¹

Even if an institutional inquiry has concluded that a researcher has not been dishonest, courts have held that an expression of concern by a journal about the researcher’s work can be a non-defamatory opinion statement. ⁴² For example, Professor Mario Saad sued the American Diabetes Association (ADA) for its having posted an online “expression of concern” regarding his publications in the ADA’s journal, which the journal posted as a result of allegations of fabrication and falsification of data in those publications. ⁴³ Dr. Saad sued for defamation on the grounds that his university had conducted an inquiry and had found “mistakes,” but no dishonesty. ⁴⁴ The ADA argued that its expression of concern was protected opinion, but the Massachusetts District Court noted that “couching a statement in opinion will not automatically protect the speaker from liability” if it “implies the existence of underlying, undisclosed defamatory facts.” ⁴⁵ Nevertheless, the Massachusetts District Court found the expression of concern to be a non-defamatory opinion statement, as it expressed the ADA’s “subjective view” of the results of the university’s inquiry (which did find inaccuracies in the publications) and did not “accuse Dr. Saad of dishonesty or … misconduct.” ⁴⁶

II. Publication or Communication of an Allegedly Defamatory Statement to a Third Person

Under Part 93, institutions have a duty during the pendency of research misconduct proceedings to maintain the confidentiality of the accused and of witnesses, but also to “protect public health” and “the integrity of the PHS supported research process,” the latter of which could be interpreted as a duty to disclose to journals findings of invalid research results. ⁴⁷ In sub-regulatory guidance, the Office of Research Integrity within the US Department of Health and Human Services (ORI) states that Part 93 does not create an “absolute bar to disclosure.” ⁴⁸ This guidance notes that institutions may disclose information about actions “taken under the institution’s internal procedures,” as long as the institution does not disclose “the PHS component of the investigation. ⁴⁹ In practice, this is quite a narrow exception — an institution could, for instance, disclose that a researcher is on administrative leave, which would constitute an action taken under internal procedures, but could not explain that the reason the researcher is on administrative leave is due to a pending research misconduct investigation, which would constitute a disclosure of the PHS component of the investigation. Notably, a public disclosure of a research misconduct proceeding or result, even if allowed by ORI, may still be subject to state-law defamation claims, and this threat may have a chilling effect on institutional responses to allegations of research misconduct.

The second element of a defamation claim is that the alleged defamer has published or communicated the allegedly defamatory statement to a third person. ⁵⁰ This does not necessarily imply widespread disclosure, and could, for example, be triggered by an institution’s alerting even one scientific journal to possible issues with published research. ⁵¹ This adds defamation to the calculus of how an institution balances its duty to protect a respondent’s confidentiality against its duty to disclose issues with the research record to those who “need to know,” as alerting a journal to issues with a researcher’s work may constitute a de facto disclosure of that researcher’s identity. ⁵² In one case, Anversa v. Partners Healthcare System, Inc., the Massachusetts District Court implied that the institution had overstepped the permissible bounds of public disclosure of suspicions of research misconduct by having alerted journals to concerns regarding data integrity in certain papers. ⁵³

In that case, Drs. Piero Anversa and Annarosa Leri, cardiac researchers at Brigham and Women’s Hospital and Harvard Medical School, brought suit against their employer regarding how research misconduct proceedings had been conducted against them. ⁵⁴ After initiating but before completing the research misconduct process regarding allegations that Drs. Anversa and Leri had been involved in falsification or fabrication of data published in the journal Circulation and in unpublished manuscripts submitted to Science and The Lancet, administrators at the institutions alerted those journals to the ongoing proceedings and recommended that the papers be retracted. ⁵⁵ Drs. Anversa and Leri argued that through these disclosures the institutions had violated their regulatory duty under Part 93 to limit disclosure of the identities of respondents to “those who need to know.” ⁵⁶

Although the Massachusetts District Court did not reach a judgment on the merits in Anversa v. Partners, the wording of the decision seemed to evidence sympathy toward the plaintiffs. ⁵⁷ The Massachusetts District Court wrote that “[i]n this case … members of the scientific community and the media, who did not need to know, learned about the inquiry and the investigation due to the disclosures,” and wrote that one defendant “encouraged Circulation’s editor-in-chief to retract the paper and implied to him that Anversa and Leri had personally committed research misconduct.” ⁵⁸ The Massachusetts District Court decision identified alleged damage to the Plaintiffs’ reputations, finances, and professional prospects. ⁵⁹

Historically, ambiguity in the “need to know” confidentiality standard under Part 93 has created significant complexities for institutions in navigating what disclosures are appropriate. However, the Department of Health and Human Services (HHS) has recently promulgated revisions to Part 93, published September 17, 2024, effective on January 1, 2025 and applicable on January 1, 2026, which, among many other changes, clarify the circumstances in which an institution may disclose to a third party concerns about the validity of the research record. ⁶⁰ The revised rule provides that “those who need to know” may include “journals, editors, publishers, co-authors, and collaborating institutions.” ⁶¹ The rule also clarifies that the duty to maintain the confidentiality of the respondent terminates after the institution has made a final determination of whether the respondent did or did not commit research misconduct, and the institution may at that point publicly disclose the identity of the respondent. ⁶²

This revision to Part 93 only applies to research misconduct proceedings; it does not override state defamation law and will not eliminate the threat of defamation suits for institutions making disclosures to journals. Nevertheless, this feature of the new rule may persuade courts that such disclosures should be afforded the additional privilege granted to statements that are of public interest or public concern, which is discussed in greater detail below in Section IV, Immunity and Privilege. ⁶³ The new revisions to Part 93 countenance such disclosures and therefore imply that such disclosures are in the public interest and should receive this protection, although the impact of these revisions on defamation litigation remains to be seen. ⁶⁴

III. Fault

The third element of a defamation claim is “fault amounting to at least negligence.” ⁶⁵ Negligence is a legal standard considering whether an individual has acted as a theoretical “reasonable person” would in the same or similar circumstances. ⁶⁶ Therefore, an individual or institution that had reviewed the evidence and reasonably concluded that a respondent had committed research misconduct would not be defaming that respondent by publicizing the finding of research misconduct as a statement of fact. Nevertheless, it can be litigated whether such a conclusion was reasonable.

One consideration related to the fault element of defamation law that arises in some research misconduct-related cases is the “public figure” rule. ⁶⁷ While a false statement against a private individual may be defamatory if the speaker is negligent in making the statement, a false statement against a public figure must be made with malice in order to be defamatory. ⁶⁸ This higher threshold favors defendants: it means that a statement against a public figure must be made with “knowledge that the statements were false or with reckless disregard of whether they were false or not” in order to be defamatory. ⁶⁹ For instance, in Croce v. Sanders, discussed above, Dr. Sanders argued that Dr. Croce was a public figure, and therefore subject to the higher threshold for defamation. ⁷⁰ Dr. Sanders quoted statements of Dr. Croce such as “I am considered like the Pope of the genetics of leukemias,” and “[I am] one of the world’s most distinguished scientific researchers in the area of cancer genetics,” citing also to Dr. Croce’s vast scientific publication record. ⁷¹ The Sixth Circuit did not agree, finding instead that Dr. Croce was not a public figure and had only been injected into notoriety by Dr. Sanders and fellow critics, and therefore that the lower negligence standard applied. ⁷²

Conversely, in Jacobson v. Clack, Dr. Mark Jacobson, Professor of Civil and Environmental Engineering at Stanford University, was held to be a public figure, and therefore the higher malice standard applied to his allegations of defamation against him by critics. ⁷³ These disparate rulings demonstrate the unpredictable nature of judicial proceedings; Dr. Croce, who has over 400 scientific publications and an h-index of 249 (indicating a high number of publications citing to his work), was ruled not to be a public figure, whereas Dr. Jacobson, who has 185 scientific publications on his biographic webpage and has an h-index of 92, but who has other indicia of fame, such as several books and an appearance on the David Letterman television show, was held to be a public figure. ⁷⁴ Likewise, in Gino v. President and Fellows of Harvard College, discussed above, the Massachusetts District Court found Dr. Gino to be a public figure, citing her “countless journal publications … numerous honors and awards … [and coverage] in numerous media outlets.” ⁷⁵ Considering the conflicting precedents for Dr. Croce compared to Drs. Jacobson and Gino, one notable scientific researcher might be legally considered a public figure while another similarly accomplished one might not, rendering the standard of fault in defamation cases related to research misconduct allegations somewhat unpredictable.

IV. Immunity and Privilege

A defendant may enjoy legal protection against a defamation claim, even if the allegedly defamatory statements that the defendant made are false, if the defendant has a recognized immunity or “privilege.” Such a protection is separate from the protections on truth and opinion discussed above, and can, for instance, protect an unsubstantiated or even an untrue statement that a person has engaged in research misconduct. Such an immunity may arise in the context of a defamation claim against a state employee, such as an officer of a state university, and privilege may arise to protect a complainant, such as a colleague or internet commenter. ⁷⁶

Specifically, sovereign and governmental immunities are legal protections for governments and government actors against legal liability in certain circumstances in which a similarly situated non-state actor could face liability. ⁷⁷ In the context of research misconduct proceedings, such protections may aid state university officials accused of defamation. This was the case in Kreipke v. Wayne State University, in which neuroscientist Dr. Christian Kreipke sued the university’s president for defamation arising from the president’s published editorial in which he stated that Dr. Kreipke had been terminated for research misconduct. ⁷⁸ Rather than considering whether the statement was defamatory, the Sixth Circuit found that the President was acting in the scope of his authority as a state employee when he wrote the editorial, “us[ing] the pronoun ‘we’ to refer to WSU,” and therefore was protected from tort claims, including defamation, by the tort immunity of the state and its institutions. ⁷⁹

Similar to immunities, privileges may protect even a false or unsubstantiated statement from a claim of defamation, but the basis of such a privilege is most often related to the nature of the act rather than the identity of the actor. Privileges have long played a role in defamation cases, with the seminal Supreme Court case New York Times Co. v. Sullivan holding that there is a general privilege afforded to everyone by the First Amendment to make statements on issues of public interest or public concern free of claims of defamation as long as such statements do not rise to the level of malice. ⁸⁰ Such a privilege is referred to as “conditional,” as it may be lost if the privilege-holder does not meet the necessary conditions — i.e., if the speaker acts with malicious intent. ⁸¹

The conditional privilege to make statements in the public interest applied to one of the defamation claims in Gino v. President and Fellows of Harvard College. ⁸² In one of her claims, Dr. Gino alleged that Harvard Business School’s posting of a notice of on Dr. Gino’s faculty page that she was on administrative leave was defamatory. ⁸³ The Massachusetts District Court rejected this claim, holding that Harvard Business School’s notice was a protected statement of public interest. ⁸⁴ The Court noted that “the relevant community need not be very large” and in fact may be “a relatively small segment of the general public,” and that a statement may be one of public concern if it is “relating to any matter of political, social, or other concern to the community.” The Court therefore found that the institution’s notice was a protected statement of public concern to the relevant community — the Harvard Business School community and Dr. Gino’s colleagues — and therefore was afforded a conditional privilege against a claim of defamation. ⁸⁵

Although ORI lacks any direct jurisdiction over defamation claims, ORI notes in guidance that it considers research misconduct complainants to hold a similar “conditional privilege to disclose, in good faith … allegations of scientific misconduct to the proper institutional or ORI officials.” ⁸⁶ ORI’s guidance provides, “[s]uch a conditional privilege would protect whistleblowers from defamation claims even where the allegations ultimately prove to be untrue.” ⁸⁷ It also notes that “whistleblowers who abuse the privilege by making bad faith allegations or by intentionally violating the confidentiality of accused parties may not be protected from defamation claims.” ⁸⁸ These sorts of statements by ORI, though not binding on courts, nevertheless have an oblique protective effect, as they indicate that, from the perspective of the leading federal agency in this area, well-founded allegations of research misconduct are a matter of vital public interest.

Abusive behaviors, however, can negate a complainant’s conditional privilege to disclose allegations of scientific misconduct, as in Rossi v. Dudek. ⁸⁹ In that case, University of Utah doctoral student Christina Rossi’s thesis advisor Dr. F. Edward Dudek alleged that Rossi had falsified data in her dissertation. ⁹⁰ The U.S. District Court found that Dr. Dudek had acted in bad faith in that that he failed to produce, or apparently even search for, evidence of his allegations, despite having had access to Rossi’s data. ⁹¹ The Court also found that Dr. Dudek intentionally violated Rossi’s confidentiality by expressing his allegations to his colleagues, to Rossi’s thesis committee, to members of Dr. Dudek’s lab, and to Dr. Dudek’s daughter, many of whom had no “legitimate role” in the dispute. ⁹² The Court therefore found that Dr. Dudek had lost his conditional privilege as a complainant and was not protected against Rossi’s defamation claims. ⁹³ Dr. Dudek also could have enjoyed immunity as an employee of the University of Utah, but the Court found that Dr. Dudek lost this protection as well because he had engaged in willful misconduct by having made injurious statements with knowledge of their falsity. ⁹⁴ A jury awarded Rossi $160,000. ⁹⁵

Although that court faulted Dr. Dudek for having disclosed allegations of research misconduct to parties that had no need to hear them, other courts have protected anonymous complainants who have disclosed allegations of research misconduct to the world via the internet. ⁹⁶ Anonymous online complaints have become an increasingly common source of allegations in research misconduct cases, such as through the website PubPeer.Reference Caron ⁹⁷ In Sarkar v. Doe, cancer pathologist Dr. Fazlul Sarkar sought to subpoena from PubPeer the identities of anonymous commenters who had accused him of data fabrication and falsification in his scientific publications. ⁹⁸ Dr. Sarkar alleged that the commenters had defamed him, and that this cost him professional and financial opportunities, and he sought a judicial order mandating that the website disclose the identities of his critics. ⁹⁹ The Court of Appeals of Michigan cited to a prior case holding that anonymous commenters who may be unaware of a defamation case filed against them are subject to two additional conditions to protect their anonymity and First Amendment rights; first, the plaintiff must make “reasonable efforts” to alert them that they are subject to subpoena seeking their identities, and second, the plaintiff must be able to show that his or her defamation claims are sufficiently strong to survive summary disposition. ¹⁰⁰ The Court noted that the case itself had been publicized on the PubPeer website, thus satisfying the notice requirement, but that Dr. Sarkar had a minimal and deficient claim that the online comments were in fact defamatory. ¹⁰¹ The Court noted that online message boards are “generally regarded as containing statements of pure opinion,” that the comments “reflect[ed] the speaker’s opinion based on underlying facts that are available to the reader,” and that Dr. Sarkar had failed to point to particular language on PubPeer that he considered defamatory. ¹⁰² Therefore, the Michigan Appeals Court held that Dr. Sarkar was “not entitled to unmask” the anonymous PubPeer critics. ¹⁰³ This case is notable in that it suggests that even though ORI guidance asserts, and Rossi v. Dudek holds, that complainants may lose conditional privilege if they intentionally disclose to unnecessary parties accusations of research misconduct, anonymous complainants may nonetheless be protected from having their identities revealed if the respondent has articulated only a weak defamation claim. ¹⁰⁴

An additional form of protection for defendants in defamation cases, which is not a form of immunity or privilege but serves a similar function, is law combating “strategic lawsuits against public participation (“anti-SLAPP laws”). ¹⁰⁵ Some state legislatures have instituted anti-SLAPP laws to combat the use of defamation claims as a tool to silence protected First Amendment speech. ¹⁰⁶ Generally, anti-SLAPP laws allow a defendant to seek that a court dismiss a defamation claim early in the judicial process if that defendant can demonstrate that the claim is intended to prevent the protected exercise of First Amendment rights. ¹⁰⁷ For instance, in California, a defendant who shows that he or she is being sued for free speech in connection with a public issue can have the case dismissed within 30 days of a motion, and the court may in some instances make the plaintiff pay the defendant’s legal fees. ¹⁰⁸

In the context of disputes over data integrity in published research, commentators have noted that defamation suits against research misconduct complainants may have a chilling effect on open scientific debate, which is a significant source of self-correction in the sciences. ¹⁰⁹ Anti-SLAPP laws have helped to correct this problem in at least one case: in Jacobson v. Clack, climate scientist Dr. Mark Jacobson sued Dr. Christopher Clack and the National Academy of Sciences for alleged defamation in the form of a paper that Dr. Clack published through the National Academy of Sciences criticizing the methods and conclusions of a paper published by Dr. Jacobson. ¹¹⁰ The Superior Court of the District of Columbia noted that “criticizing ideas is not defamation,” and the appellate court in that case likewise found that the paper in question constituted “reasoned scientific disagreement” and not defamation. ¹¹¹ Nevertheless, anti-SLAPP laws have not been instituted in every state, and have significant differences in scope in the states that have instituted them, so these provisions cannot necessarily be relied upon to protect complainants and institutions involved in research misconduct disputes.Reference Piper ¹¹²