Research Integrity & Compliance

Research integrity and compliance is the set of standards, obligations, and oversight
mechanisms that keep the research record trustworthy — from how a study is approved before
it starts, to how misconduct is defined and investigated, to how the published record gets
corrected when something goes wrong. It spans three overlapping layers: ethics
oversight (approvals required before data collection begins, such as IRB/REC
review), conduct standards (training and disclosure obligations that apply
throughout a research career), and publication integrity (the editorial and
post-publication mechanisms — COPE, retraction, correction — that catch problems after
work is submitted or published).

This page is a hub, not a single guide: it frames the cluster’s scope and links out to the
in-depth guides and dictionary terms that cover each piece. Coverage here is honest about
what exists today and what’s still a gap — CASRAI is actively building out this cluster, and
not every sub-topic below has a dedicated guide yet.

Ethics oversight: approval before research begins

In the United States, research involving human subjects that is conducted or funded by a
federal agency is generally governed by the Common Rule (45 CFR 46, most
recently revised effective January 21, 2019), which requires review and approval by an
Institutional Review Board (IRB)
before human-subjects research can proceed, and sets the baseline requirements for IRB
composition, review criteria, and informed consent. Institutions receiving federal support
file a Federalwide Assurance (FWA) with the HHS Office for Human Research Protections
committing to follow Part 46. Outside the US, the equivalent body is usually called a
Research Ethics Committee (REC)
— same underlying function, different national framework.

Ethics oversight extends beyond human-subjects approval. Research involving shared
biological or research materials typically runs on a
Material Transfer Agreement (MTA);
research involving personal data collected on or transferred from EU/UK subjects has to
account for GDPR; and
some international collaborations are constrained by
export-controlled research rules
(EAR/ITAR in the US) that limit what can be shared with foreign nationals or institutions
regardless of funding source.

Conduct standards: training, disclosure, and the definition of misconduct

The globally-cited reference point for what “responsible conduct” means in principle is
the Singapore
Statement on Research Integrity (2010) — four principles (honesty, accountability,
professionalism, stewardship) and fourteen responsibilities, produced at the 2nd World
Conference on Research Integrity and not itself a binding regulation, but widely adopted as
a reference framework by institutions and funders since.

In the US, that principle gets a binding definition through the
Office of Research Integrity
(ORI), which administers 42 CFR Part 93 for Public Health Service–funded research. ORI’s
definition of
research misconduct is specific and
narrower than “bad research practice” — it covers
fabrication (making up data or results),
falsification (manipulating research materials,
equipment, processes, or data so the record doesn’t accurately reflect what was done), and
plagiarism (appropriating another’s ideas,
processes, results, or words without credit) in proposing, performing, or reviewing research
or in reporting results — collectively referred to as FFP. It explicitly excludes honest
error and legitimate differences of scientific opinion. A closely related but distinct
concept, self-plagiarism (reusing one’s own
previously published text or data without disclosure), isn’t misconduct under the ORI
definition but is treated as an ethics violation by most journals and COPE guidance.

Funder-mandated Responsible
Conduct of Research (RCR) training is the practical mechanism institutions use to keep
researchers current on these standards — NIH requires a minimum of 8 contact hours with
substantial face-to-face discussion, NSF requires equivalent training (expanded to cover
faculty and senior personnel by the 2022 CHIPS and Science Act), and NSF separately now
requires Research Security Training as an active, distinct obligation. The full guide
covers exactly who needs it, how often, and — the harder operational question — how
institutions actually track completion across a career and across collaborating
institutions, a gap CASRAI’s own analysis found no standardized machine-readable solution
for yet.

Alongside conduct training sits disclosure: most funders and journals require
conflict of interest
disclosure — financial or other relationships that could plausibly bias study design,
conduct, or reporting — using instruments such as the ICMJE’s standard COI disclosure
form.

Clinical trial registration and reporting

Clinical research carries its own specific compliance layer. In the US, the Food and Drug
Administration Amendments Act of 2007 (FDAAA 801) requires registration of “applicable
clinical trials” — broadly, controlled trials of FDA-regulated drugs, biologics, and
devices, other than Phase I — on ClinicalTrials.gov within 21 days of enrolling the first
subject, with a further requirement to submit summary results; the 2016 HHS Final Rule
(42 CFR Part 11) clarified and expanded these obligations. Separately, the ICMJE requires
registration in a public trials registry at or before first patient enrollment as a
condition of considering a trial for publication in any ICMJE member journal — a
publication-side requirement that applies more broadly than FDAAA 801’s regulatory scope,
since it covers any prospective assignment of human subjects to compare a health
intervention’s effect, not only FDA-regulated products.

Publication integrity: catching problems after submission or publication

Editorial and post-publication integrity is the layer most directly relevant to CASRAI’s
publishing-side work, and where the cluster currently has its deepest coverage. The
Committee on Publication
Ethics (COPE), a nonprofit founded in 1997, sets the reference standards most scholarly
publishers use for handling ethics questions — plagiarism suspicions, authorship disputes,
duplicate submission, and retraction — through its published guidance to editors and
journals. When something does need correcting after publication, CASRAI’s guide on
post-publication peer review covers how a
concern moves from initial flag (commonly via PubPeer, an independent nonprofit not
affiliated with CASRAI or any publisher) through to a formal
retraction or correction, and how Crossref’s
Crossmark and update-to metadata field propagate that correction to anyone
holding the original DOI. Retraction Watch, also independently operated, maintains the most
comprehensive public retraction database and has had its dataset integrated into Crossref’s
infrastructure since September 2023.

A related, ongoing concern in this space is publication-venue quality itself — how to
recognize a predatory journal, one that
solicits and charges publication fees without providing the peer review or editorial
services associated with legitimate scholarly publishing. This overlaps with, but is
distinct from, the misconduct and correction mechanisms above: a predatory venue is a
structural problem with the publisher, not necessarily with the individual researcher’s
conduct.

Reporting integrity in a broader sense — whether research assessment itself rewards the
right things — connects to DORA
(the San Francisco Declaration on Research Assessment) and
CoARA (the Coalition for Advancing Research Assessment),
both of which argue that over-reliance on journal-level metrics like impact factor distorts
incentives away from research quality and integrity; institutions and funders can become
DORA signatories or
CoARA Agreement signatories as a formal
commitment to reform.

What’s covered here today, and what’s still a gap

Two full guides are live in this cluster as of publication: RCR training requirements and
tracking, and how post-publication peer review actually moves a concern to a correction.
The dictionary terms linked throughout this page — misconduct, IRB, COPE, retraction,
self-plagiarism, conflict of interest, and the rest — give operational definitions for the
underlying vocabulary. Several adjacent topics that genuinely belong in this cluster don’t
yet have dedicated guide-length treatment on CASRAI: a practical walkthrough of the
IRB/REC approval process itself, misconduct investigation procedure as distinct from
definition, and a dedicated look at COPE’s editorial flowcharts are active gaps, not
oversights — this page will update as that coverage is built out.