Research Integrity Principles

RESEARCH MISCONDUCT (formal definition used by NIH/ORI): intentional or reckless fabrication, falsification, or plagiarism in proposing, conducting, or reporting research. Key element is intent or recklessness; accidental error is not misconduct. Examples: fabricating data points, falsifying results, plagiarizing text. Misconduct is serious; proven misconduct can result in retraction, funding ban, job loss. RESEARCH INTEGRITY VIOLATION: broader category including both misconduct AND negligence, careless practices, or unintentional errors. Examples: undisclosed conflict of interest, inadequate methods description, failure to disclose funding. Integrity violations may not involve intentional deception but reflect inadequate research practices. Response differs: misconduct triggers formal investigation; integrity violations may result in corrective action (revision, retraction, or educational feedback) without formal misconduct finding.

Institutional response depends on severity and your institution's misconduct policies. Typical process: (1) Your concern is forwarded to research integrity officer or ombudsperson. (2) If allegation involves federal funding, institutional official may notify Office of Research Integrity (ORI) and agency funder. (3) Preliminary inquiry: institution gathers information to determine if allegation is credible. This is informal, quick (days to weeks), to determine if formal investigation warranted. (4) If inquiry indicates credible allegation, formal investigation begins. Investigation is confidential; involves fact-finding, interviews, documentation review. (5) Investigative report issued with findings and recommended actions. (6) Subject of investigation may appeal findings. (7) If misconduct found, consequences range from written warning (minor violations) to funding ban, journal retraction, termination (serious violations). Whistleblower protections exist; retaliation against person reporting misconduct is prohibited. Your identity may be protected depending on institutional policy. Note: investigation can take months or years; expect confidentiality constraints limiting what you learn about outcomes.

Reporting only positive findings is problematic and contributes to publication bias—overrepresentation of positive findings in literature. However, context matters. IF YOU CONDUCTED MANY TESTS, FOUND FEW SIGNIFICANT: This is research misconduct (selective reporting) if you fail to report the null findings. Prevent by preregistering hypotheses and promising to report all primary/secondary outcomes. IF YOUR RESEARCH DESIGN AIMED AT DETECTING POSITIVE EFFECT AND FOUND NONE: Not unethical to report negative findings; this is legitimate science contributing knowledge that intervention does not work. IF YOU FOUND POSITIVE FINDINGS IN ONE DOMAIN AND NEGATIVE IN ANOTHER: Ethical to focus manuscript on positive findings but note negative findings exist; do not misrepresent by omitting entire domains of investigation. Best practice: be transparent about what you tested and what you found, whether positive or negative. Negative and null findings are scientifically valuable; literature needs both to be credible.

Not necessarily. Transparency principles require making data available when possible, but legitimate reasons may prevent open data: (1) SUBJECT PRIVACY: If data contain sensitive personal information (genetic, health, financial) that cannot be adequately de-identified, sharing may violate subject privacy despite best efforts. In this case, data can be maintained securely with restricted access; researchers may request access through formal process. (2) PROPRIETARY AGREEMENTS: If you have contractual agreement with company/funder preventing data sharing, you may not be able to release data. In this case, disclose restriction in manuscript; provide as much detail as possible in methods to enable replication with other datasets; provide code/analysis scripts so methods are transparent. (3) LEGAL RESTRICTIONS: Some data are subject to legal restrictions (national security, export controls). Comply with law while maximizing transparency within constraints. Disclose restrictions honestly. In all cases, make explicit what is shared and what is restricted + why. Provide sufficient methods detail that independent researcher could replicate analyses with different dataset. This maintains transparency principle even when full data cannot be shared.

No. Having a financial interest is a POTENTIAL conflict of interest (COI), not automatically a disqualification. Many researchers have COI: oncologists who treat patients with disease they study; researchers with patents on interventions; investigators employed by companies producing study interventions. COI itself is not unethical; non-disclosure and failure to manage COI is unethical. Manage COI by: (1) DISCLOSING: Inform your institution (on grant application), journal (in manuscript), collaborators (in writing) about all financial interests. (2) MANAGING: Develop plan to minimize bias—examples: (a) independent researchers conduct data analysis (researcher with COI does not touch data); (b) data are locked before unblinding analyses (preventing selective reporting); (c) independent data monitoring committee oversees safety/efficacy; (d) pre-registration of hypotheses and analysis plan. (3) RECUSAL: In some cases (very large financial interest, extreme bias concern), you may be asked to recuse from certain decisions (e.g., data analysis) while remaining study leader. After full disclosure and appropriate management, research can proceed ethically.