CASRAI Dictionary

Tag: ORI research misconduct

  • Research Misconduct Investigations: How ORI and UKRIO Procedures Compare

    Grant administrators rarely encounter research misconduct allegations often — but when one arrives, the clock starts immediately, and the procedural path depends entirely on which side of the Atlantic the funding sits. A US Public Health Service (PHS) grant triggers a federally regulated process overseen by the Office of Research Integrity (ORI); a UK Research England or UKRI grant triggers an institution-led process shaped, but not enforced, by the UK Research Integrity Office (UKRIO). Confusing the two — assuming ORI’s binding timelines apply to a UK case, or that UKRIO can compel an outcome the way ORI can — is a common and costly error for administrators managing cross-border collaborations.

    This guide sets out, side by side, what each body actually is, how each defines misconduct, and how the investigation stages differ — so administrators handling an allegation tied to a grant know which rulebook applies.

    How ORI and UKRIO define research misconduct

    Both frameworks agree on a common core — fabrication, falsification, and plagiarism (FFP) — but they diverge sharply in scope.

    ORI operates under 42 CFR Part 93, the PHS Policies on Research Misconduct. Its definition is deliberately narrow: research misconduct means fabrication, falsification, or plagiarism in proposing, performing, or reviewing research, or in reporting research results. A finding requires three elements to be met: a significant departure from accepted practices in the relevant research community; committed intentionally, knowingly, or recklessly; and proven by a preponderance of the evidence. Honest error and legitimate differences of scientific opinion are explicitly excluded.

    UKRIO, by contrast, works from a broader, non-statutory definition: “behaviours that deliberately or recklessly fall short of the standards expected in the conduct of research.” Its guidance, aligned with the UK Concordat to Support Research Integrity (Universities UK, 2019), extends beyond FFP to cover breaches of ethical approval, undeclared conflicts of interest, and mismanagement of research data — while also naming a distinct, lower-severity category, “questionable research practices” (QRPs), for avoidable errors that fall short of intentional misconduct.

    Regulator versus adviser: two different roles

    The most consequential difference is not definitional but structural: ORI is a federal oversight body; UKRIO is an independent charity with no regulatory power.

    • ORI sits within the US Department of Health and Human Services and directly oversees how institutions handle misconduct allegations tied to PHS-funded research (including NIH grants). It can conduct its own oversight review of an institution’s findings, recommend administrative actions, and refer findings to the HHS Departmental Appeals Board.
    • UKRIO was established as an independent advisory charity in 2006. It publishes a model investigation procedure that UK research organisations may adopt or adapt, offers case-by-case advice, and promotes good practice — but it does not investigate cases itself, does not mandate a single national procedure, and has no statutory sanctioning power.

    UK funders address this gap contractually rather than through a regulator. UKRI’s Guidance for Research Organisations on the Investigation of Research Misconduct (April 2025) requires any organisation receiving UKRI funding to investigate allegations against staff or students under its own Grant Terms and Conditions, with UKRI able to take funding action if an organisation fails to do so.

    How an investigation actually runs, stage by stage

    Both systems separate a preliminary triage stage from a full inquiry, but they name and time these stages differently.

    Stage ORI (US, PHS-funded research) UKRIO model (UK institutions)
    Trigger Allegation received by institution’s Research Integrity Officer Concern raised with a “Named Person” or responsible officer
    Triage Assessment: does the allegation meet the FFP definition and involve PHS funding? Initial assessment: does it fall within the misconduct procedure’s scope?
    Formal fact-finding Inquiry (institution-level, time-limited) Initial investigation to establish if there is a case to answer
    Full review Investigation, following a sufficient inquiry finding Full investigation by an academic panel, including external members
    Standard of proof Preponderance of the evidence Not codified nationally; set by each institution’s procedure
    External oversight ORI oversight review of institutional findings; report to PHS agency None mandatory; UKRIO offers advice only
    Appeal route HHS Departmental Appeals Board Institutional appeal, managed by someone other than the original Named Person

    Under the UKRIO model, once a panel reports its findings, the Named Person decides on next steps: referral to institutional disciplinary proceedings, correction of the published record, and notifying relevant funders. Where allegations are not upheld, the same procedure is meant to protect the reputation of the person accused — a feature both systems share in principle, though neither publishes comparable statistics on false-allegation rates.

    Answer-first: common questions on research misconduct

    What are the three types of research misconduct?

    Both ORI and most UK institutional policies converge on the same core triad: fabrication (inventing data or results), falsification (manipulating research materials, equipment, or data to misrepresent results), and plagiarism (using another person’s ideas, processes, or words without credit). This is often abbreviated FFP.

    What are some examples of research misconduct?

    Common examples include inventing patient consent records, selectively deleting inconvenient data points, copying text or images from another paper without attribution, and misrepresenting the outcome of a statistical test. UKRIO guidance also treats proceeding without required ethical approval as a form of misconduct, even without FFP intent.

    What counts as research misconduct?

    Conduct counts as misconduct when it represents a significant, intentional or reckless departure from accepted research standards — not an honest mistake or a genuine scientific disagreement. ORI requires proof by a preponderance of the evidence; UKRIO-aligned institutions apply a similar intent-based threshold under their own procedures.

    What this means for research administrators

    For administrators managing grants that cross jurisdictions — a common scenario in NIH-funded international collaborations or Horizon Europe partnerships involving UK institutions — three practical points follow from the comparison above:

    • Know which body has enforcement power. Only ORI can conduct oversight review and refer a case to a federal appeals process; UKRIO cannot compel an institutional outcome.
    • Check the funder’s own reporting clause. UKRI’s April 2025 guidance obliges the receiving institution — not UKRIO — to investigate and report; PHS grant terms impose parallel obligations that run through ORI.
    • Do not assume a single global timeline. ORI-regulated inquiries and investigations run to defined federal timeframes; UKRIO-aligned UK procedures are set institution by institution, so the applicable deadline sits in the local Code of Practice for Research, not in UKRIO’s own documents.

    Administrators supporting research administration functions across both systems should hold copies of both the relevant institutional misconduct procedure and the specific grant terms — the procedural detail, not the high-level definition, is where jurisdictional mismatches cause delay.

    Where the two systems are heading

    Both frameworks are converging on the same underlying principle even as their governance models remain distinct: misconduct findings should correct the scholarly record, not just discipline an individual. UKRI’s 2025 guidance tightened institutional reporting obligations, and ORI continues to publish case summaries and administrative actions as a transparency mechanism. Neither change closes the structural gap — one system regulates, the other advises — so for the foreseeable future, administrators handling cross-border allegations will need to work both playbooks rather than assume one substitutes for the other.

    As with contributor-role standards, where CASRAI originated the CRediT taxonomy in 2014 and the standard is now stewarded by NISO as ANSI/NISO Z39.104-2022, research integrity governance illustrates a broader pattern in research administration: originating bodies and enforcement bodies are frequently separate, and knowing which is which is a prerequisite for compliant practice.

  • Famous Cases of Research Misconduct: Procedural Lessons from Stanford and Harvard

    The most famous cases of research misconduct rarely turn on a single fabricated data point. They turn on how an institution responds once a concern surfaces: who is notified, how quickly a panel is convened, what gets published, and what stays confidential. The 2023 resignation of Stanford University’s president, Marc Tessier-Lavigne, and the parallel investigation into Harvard Business School professor Francesca Gino gave research integrity offices two live, closely watched case studies in exactly that process, separate from the underlying scientific disputes.

    Both cases ran on strikingly different tracks: one via an independent trustee-commissioned panel responding to press and PubPeer scrutiny, the other via a confidential internal Research Integrity Officer inquiry triggered by a data-forensics blog. Comparing them, and three earlier landmark cases, surfaces recurring procedural failure points that any research administration office can audit against its own protocols.

    Why these cases matter to research integrity offices

    Research misconduct investigations are procedurally distinct from routine peer-review corrections. Under the US Office of Research Integrity’s (ORI) definitions, misconduct comprises fabrication, falsification, and plagiarism — not honest error, and not scientific disagreement. Getting that distinction right, early, determines whether a case is handled as a correction, a retraction, or a formal misconduct finding with employment consequences.

    The Stanford and Harvard cases are instructive precisely because they show two institutions applying that distinction under public pressure, at speed, with reputational stakes attached to the outcome.

    The Stanford presidency: oversight, not fabrication

    Marc Tessier-Lavigne resigned as Stanford’s president in July 2023 after a Special Committee of the Board of Trustees, assisted by an outside scientific panel, reviewed allegations first amplified by student newspaper reporting and image-integrity comments on PubPeer. The panel’s report did not find that Tessier-Lavigne personally falsified data. It found manipulated images and unreliable data across several papers from labs he had run at different institutions, and — the procedurally decisive finding — that he had failed to act “decisively and forthrightly” to correct the record once concerns were raised over several years.

    • Investigation body: an independent Special Committee plus an outside scientific review panel, not the university’s standing research-integrity office, reflecting the conflict of interest inherent in investigating a sitting president.
    • Trigger: student journalism and public PubPeer commentary, not an internal whistle-blower report.
    • Outcome: resignation as president; five papers slated for retraction or correction; Tessier-Lavigne retained his tenured faculty position.

    For research administrators, the lesson is less about the science and more about escalation pathways: when the subject of a complaint holds institutional authority, governance structures must route the inquiry outside the normal chain of command from the outset.

    Francesca Gino at Harvard: whistle-blowers, investigation, litigation

    Harvard’s case followed a different route. In 2021, the behavioural-science bloggers behind Data Colada flagged anomalies in four of Francesca Gino’s papers to Harvard directly. The university’s Research Integrity Officer opened a confidential inquiry that produced a roughly 1,200-page report, concluding Gino had “committed research misconduct intentionally, knowingly, or recklessly.” She was placed on unpaid administrative leave in June 2023, shortly before Data Colada published its findings publicly.

    Gino has denied the findings and, in August 2023, filed a $25 million lawsuit against Harvard, its dean, and the three Data Colada researchers, alleging defamation and gender discrimination. A federal judge dismissed the claims against the Data Colada authors and several claims against Harvard in September 2024, though parts of the litigation continue.

    • Investigation body: Harvard’s internal Research Integrity Officer process, kept confidential during the inquiry itself.
    • Trigger: an external, methodologically detailed whistle-blower report from named researchers running a public blog.
    • Outcome: administrative leave, retractions, and multi-year litigation that has kept the case — and the institution’s internal report — in the public record long after the initial finding.

    The litigation risk here is the procedural lesson generic “list of cases” coverage tends to skip: a confidential process does not stay confidential once litigation compels disclosure, and institutions should draft misconduct reports with that eventual exposure in mind.

    Five cases compared: triggers, investigators, outcomes

    Set alongside three earlier landmark cases, the Stanford and Harvard investigations show a consistent pattern: the trigger for an inquiry is now more often external scrutiny — journalists, forensic bloggers, or peer-commentary platforms — than an internal audit.

    Case Institution What triggered the inquiry Investigating body Outcome
    Marc Tessier-Lavigne Stanford University Student journalism, PubPeer comments Independent trustee-commissioned panel Resigned as president, July 2023; retractions pending
    Francesca Gino Harvard Business School Data Colada whistle-blower report (2021) Harvard Research Integrity Officer Unpaid leave, June 2023; ongoing litigation
    Diederik Stapel Tilburg University Junior researchers raised concerns Levelt, Noort and Drenth Committees Resigned 2011; roughly 55 papers retracted
    Jan Hendrik Schön Bell Laboratories Peers noticed duplicated data across papers Internal Bell Labs investigation committee Dismissed 2002; PhD revoked 2004
    Hwang Woo-suk Seoul National University Investigative journalism and insider tips University investigation panel Dismissed 2006; later convicted on related charges

    Common questions about research misconduct cases

    What are some examples of research misconduct?

    Classic examples include fabricating data that was never collected, as in the Stapel case; falsifying results to match a hypothesis, as alleged in the Gino papers; and image manipulation across multiple publications, the pattern identified in the Tessier-Lavigne review. Plagiarism of text or ideas is the third recognised category under most institutional and federal definitions.

    What is considered the most serious form of research misconduct?

    Fabrication is generally treated as the most serious category, because it invents phenomena outright rather than distorting real data. Falsification is a close second, since it corrupts genuine findings. Both differ fundamentally from honest error or a good-faith methodological dispute, which do not meet the threshold for a misconduct finding.

    What are some examples of academic misconduct?

    Academic misconduct is a broader category than research misconduct, and includes duplicate publication, undisclosed conflicts of interest, ghost or guest authorship, and failure to secure ethics approval, alongside the fabrication, falsification, and plagiarism that define research misconduct specifically under ORI and most university policies.

    What research offices should take away

    Five recurring procedural patterns emerge across these cases:

    1. Escalation independence. When a complaint touches senior leadership, route the inquiry to a body outside the normal reporting line — Stanford’s use of an independent trustee committee is the model.
    2. External triggers now dominate. Four of the five cases here were surfaced by outsiders — journalists, junior researchers, or forensic bloggers — not internal audit. Offices should treat external tip lines as a primary detection channel, not a fallback.
    3. Draft for eventual disclosure. Gino’s litigation shows that confidential reports can become public exhibits; findings should be evidenced and worded accordingly.
    4. Corresponding-author accountability matters. Several of these cases turned on unclear responsibility for data supplied by co-authors or lab members — a governance gap that clearer authorship and contributorship practices help close, including explicit use of contributor-role frameworks for accountability mapping.
    5. Timelines are long. None of these cases resolved in under a year; Gino’s litigation is still active more than three years after Data Colada’s initial report to Harvard. Offices should plan communications and interim-measures policy for multi-year timelines, not single-semester ones.

    Institutions building or revising misconduct policy can benchmark their escalation, notification, and disclosure procedures against these patterns through established research administration frameworks, and can consult standard definitions of the terms involved via CASRAI’s open research vocabulary.

    Looking ahead: building resilient response processes

    The Stanford and Harvard cases will keep generating secondary coverage — retractions, court filings, follow-on reporting — for years, which is itself a lesson: a misconduct case is not closed when a leave notice or resignation is announced. Research integrity offices that treat these as multi-year, cross-functional processes, spanning HR, legal, communications, and the research office itself, are better positioned than those that treat a finding as the end state.

    As data-forensics blogs, image-integrity tools, and post-publication review platforms proliferate, external detection will keep outpacing internal audit capacity. The institutions that fare best will be the ones with pre-agreed escalation pathways, disclosure-ready documentation standards, and clear contributor-accountability frameworks already in place before the next high-profile case breaks.