01 · The Question
What Behavior Actually Qualifies as Research Misconduct?
"Research misconduct" is a serious label. It is sometimes used casually for anything from sloppy recordkeeping to unethical authorship, weak methodology, statistical mistakes, or outright invention of data. Formal policies are usually more precise.
Fabrication, falsification, and plagiarism, often abbreviated as FFP, form the core of many influential definitions of research misconduct. Yet identifying one of those categories is not always enough to establish a formal finding. The applicable definition, circumstances, state of mind, evidentiary standard, and institutional or regulatory jurisdiction can matter.
Understanding those distinctions helps researchers recognize serious wrongdoing without incorrectly treating every error or poor research decision as misconduct.
03 · What You Need to Know
What Fabrication, Falsification, and Plagiarism Actually Mean
Research misconduct is a defined category, not a synonym for anything wrong in research
The term should be used carefully. Under the U.S. Public Health Service framework administered by the Office of Research Integrity, research misconduct means fabrication, falsification, or plagiarism in proposing, performing, or reviewing research, or in reporting research results.
Other countries, institutions, funders, and professional bodies may define the scope differently. The National Academies has noted broad international agreement that fabrication, falsification, and plagiarism belong within research misconduct, while also recognizing that some systems include additional behaviors.
So FFP provides an important conceptual foundation, but a researcher dealing with an actual allegation should consult the policy governing that particular research rather than assuming one definition applies everywhere.
Fabrication means inventing data or results
Under the U.S. federal definition, fabrication is making up data or results and recording or reporting them.
The central problem is invention. The data or result is represented as part of the research even though it was not actually obtained as represented.
Hypothetical Example
A survey response that never existed
A researcher needs 300 completed questionnaires but has obtained only 276. Instead of reporting the actual sample or collecting additional participants appropriately, the researcher creates 24 fictitious responses and enters them into the dataset as though real participants had submitted them. Those invented observations illustrate fabrication.
Fabrication is not limited to published journal articles. Depending on the governing policy, misconduct can occur during proposals, data collection, research performance, review, or reporting. A fabricated result does not need to survive all the way to publication before it becomes potentially relevant.
Falsification means making the research record inaccurately represent what happened
Falsification is different from fabrication. Something real may exist, but it is manipulated, changed, omitted, or otherwise represented in a way that makes the research record inaccurate.
Under the U.S. federal definition, falsification includes manipulating research materials, equipment, or processes, or changing or omitting data or results so that the research is not accurately represented in the research record.
Hypothetical Example
Removing inconvenient observations without a defensible basis
A researcher obtains 50 legitimate observations. Five substantially weaken the expected result. After seeing the outcome, the researcher deletes those observations, not because they satisfy a legitimate exclusion criterion, but because removing them produces the desired conclusion. The researcher then reports the reduced dataset as though it represented the study as conducted. Depending on the circumstances and applicable standard, this could constitute falsification.
Not every alteration, correction, exclusion, transformation, or cleaning decision is falsification. Researchers routinely remove demonstrably invalid measurements, correct coding errors, process raw signals, exclude observations according to justified criteria, and transform data for legitimate analytical reasons.
The critical question is not simply, "Were the data changed?" It is whether the manipulation or omission caused the research to be inaccurately represented.
Plagiarism is not limited to copying sentences
Under the U.S. federal definition, plagiarism is the appropriation of another person's ideas, processes, results, or words without giving appropriate credit.
This makes plagiarism broader than verbatim copying. Taking distinctive ideas, research processes, or results and presenting them without appropriate attribution can also fall within the concept.
At the same time, not every similarity of wording establishes research misconduct. Questions involving common terminology, conventional descriptions, appropriate paraphrasing, attribution practices, or authorship disputes may require careful contextual assessment.
| Category |
Core problem |
Hypothetical illustration |
| Fabrication |
Something is invented and represented as research data or results. |
Creating participant responses that were never collected. |
| Falsification |
Research materials, processes, data, or results are manipulated or omitted so that the research record becomes inaccurate. |
Removing inconvenient observations solely to create the desired finding and concealing the removal. |
| Plagiarism |
Another person's ideas, processes, results, or words are appropriated without appropriate credit. |
Presenting another researcher's distinctive analysis or text as one's own without attribution. |
A formal finding can require more than showing that FFP occurred
Formal research misconduct frameworks may impose additional requirements. Under the U.S. federal policy, a finding requires a significant departure from accepted practices of the relevant research community. The conduct must also have been committed intentionally, knowingly, or recklessly, and the allegation must be proven by a preponderance of the evidence.
This distinction matters. An allegation can involve inaccurate data without automatically establishing misconduct. Investigators may need to determine what happened, whether the conduct falls within fabrication, falsification, or plagiarism, what the researcher knew or intended, whether the conduct was reckless, and whether the required evidentiary threshold has been satisfied.
The role of intent in research misconduct and whether recklessness can satisfy the required mental state therefore deserve separate attention.
Honest error is not the same thing as misconduct
Researchers make mistakes. Data can be entered incorrectly, formulas can contain errors, samples can be mislabeled, code can malfunction, and researchers can reach interpretations later shown to be wrong.
Under the U.S. Public Health Service definition, research misconduct does not include honest error or differences of opinion. That exclusion is essential because a research system in which every genuine mistake is branded misconduct would confuse fallibility with wrongdoing.
But calling something an error does not settle the question. The circumstances still matter. Determining the boundary between honest error and research misconduct may require examining what the researcher knew, what actions were taken, and whether the explanation fits the evidence.
Serious research problems can exist outside formal misconduct
FFP should not become a loophole. A practice does not become responsible merely because it falls outside a particular misconduct definition.
Researchers may engage in poor supervision, inadequate recordkeeping, irresponsible statistical practices, inappropriate authorship behavior, selective reporting, or other conduct that damages research without necessarily satisfying the formal definition of fabrication, falsification, or plagiarism.
The distinction between formal misconduct and questionable research practices is therefore important. The appropriate institutional response may differ, but the underlying integrity concern can remain substantial.
Watch Out
"It isn't FFP" does not mean "it is acceptable." Formal misconduct definitions establish boundaries for particular proceedings. Institutions, journals, funders, employers, and professional bodies may separately address other breaches of research integrity or professional standards.
06 · What This Means for You
How to Think About a Possible Misconduct Concern
If you encounter suspicious research, avoid beginning with the conclusion that someone "committed misconduct." Start with the observable facts.
Identify what appears to have happened. Were data invented? Were real data, materials, processes, or results manipulated or omitted in a way that misrepresented the research? Was another person's work appropriated without appropriate credit? Or is the concern actually about another kind of research practice?
A simple decision framework
If data or results appear to have been invented
Consider whether the conduct may fit the applicable definition of fabrication.
If genuine research was manipulated, changed, or selectively omitted
Ask whether the resulting research record inaccurately represents what occurred and whether the applicable definition of falsification may be engaged.
If another person's ideas, processes, results, or words were taken without appropriate credit
Consider the applicable plagiarism standard and the context in which the material was used.
If the behavior is troubling but does not clearly fit FFP
Do not assume it is acceptable. Determine whether it is a questionable, detrimental, negligent, unethical, or otherwise prohibited research practice under the relevant standards.
If a formal allegation may be necessary
Consult the institution's current research misconduct policy and appropriate research integrity office or equivalent authority rather than attempting to adjudicate the allegation yourself.
Precise language matters here. "I found an unexplained discrepancy between these records" is an observation. "The researcher fabricated the data" is a conclusion requiring evidence. Keeping those two statements separate is particularly important when allegations could affect careers, publications, funding, and the reliability of the scholarly record.
07 · A Quick Checklist
Before Calling Something Research Misconduct, Check
When assessing a possible misconduct concern, check:
What research misconduct definition actually applies to the research, institution, funder, or jurisdiction involved?
What specific conduct occurred, rather than simply what outcome appears suspicious?
Does the conduct potentially involve fabrication, falsification, plagiarism, or another category recognized by the applicable policy?
Could the discrepancy plausibly result from honest error, methodological disagreement, or another non-misconduct explanation?
What does the applicable framework require concerning intention, knowledge, recklessness, or other mental-state requirements?
Are original records, files, correspondence, code, or other relevant evidence being preserved?
Have you avoided making accusations more definitive than the available evidence supports?
Do you know the appropriate institutional channel for raising the concern if further action is warranted?
