Research Data Fabrication vs Falsification

Fabrication is making up data or results and reporting them as real; falsification is manipulating research materials, equipment or processes, or changing or omitting data or results so that the research is not accurately represented. Both are forms of research misconduct as defined by the US Office of Research Integrity, and both undermine the scientific record.

Differences with examples

What they are not

Honest error, differences in interpretation, and legitimate data cleaning that is documented and justified are not misconduct. Transparent documentation of exclusions and corrections is a protective practice.

Prevention and good practice

• Keep raw data, laboratory notebooks and audit trails.
• Follow the protocol and document deviations.
• Use data management plans and version control.
• Share data where possible.
• Create a culture in which people can ask questions.

If you suspect it

Do not accuse publicly. Gather basic facts, keep records, and raise the concern confidentially with a research integrity office, a senior colleague or the journal editor, following their procedures. COPE provides guidance for editors on handling suspected fabrication.

Illustrative example

Invented for teaching. A student notices that two "independent" datasets contain identical rows in a different order. She asks the data manager how they were created and, finding no explanation, discusses her concern confidentially with her supervisor rather than altering or deleting the data.

Common pitfalls

• Confusing honest error with misconduct.
• Discarding unexpected data without recording why.

Key takeaways

• Fabrication invents; falsification distorts.
• Keep records and use proper channels for concerns.

This page is general educational information about research methods. It is not medical, legal or regulatory advice; follow your institution's ethics board and your target journal's instructions.