Honest Error vs. Research Misconduct: Where Is the Line?

When Is a Research Mistake More Than an Honest Error?

A spreadsheet formula is wrong. Samples were mislabeled. A researcher accidentally analyzes the wrong version of a dataset. An image contains an unexplained duplication. Several observations disappear during data processing. Any of these discoveries can be serious, especially if they affect published findings.

But a serious error is not necessarily research misconduct.

The difficult question is often not whether something went wrong. That may already be obvious. The harder question is how the error arose and whether the evidence supports honest error, negligence, recklessness, knowing conduct, or intentional wrongdoing. Those distinctions matter because an inaccurate research record and researcher culpability are related, but they are not the same thing.

An Error Becomes Misconduct Because of the Conduct Behind It, Not Simply Its Consequences

The Difference Between Being Wrong and Committing Misconduct

Research misconduct is not defined by the size of the error

A common instinct is to judge misconduct from the consequence: if the mistake changes a major result, invalidates a paper, or requires a correction or retraction, it must have been misconduct.

That reasoning is unreliable.

Under the U.S. Public Health Service framework, research misconduct means fabrication, falsification, or plagiarism in proposing, performing, reviewing, or reporting research. Honest error and differences of opinion are expressly excluded. A formal finding also requires a significant departure from accepted practices of the relevant research community, the required state of mind, and proof by a preponderance of the evidence.

A substantial error can therefore remain an honest error. Conversely, conduct potentially satisfying a misconduct definition does not become acceptable merely because its effect on the final results happens to be small.

Honest error means more than saying "I made a mistake"

Calling something an honest error does not make it one. The explanation must be assessed against the evidence.

Current Office of Research Integrity guidance recommends examining the research record and the circumstances surrounding the questioned conduct. Relevant materials can include raw and processed data, laboratory or study records, proposals, manuscripts, correspondence, computer files, and other records that help reconstruct what occurred.

The practical question is whether the evidence is consistent with a genuine mistake rather than fabrication, falsification, or plagiarism committed with the mental state required by the governing framework.

The research record can reveal how an error arose

Suppose an incorrect value appears in a published table. The number alone tells you very little about culpability.

Perhaps the researcher mistyped 0.18 as 0.81. Perhaps a spreadsheet reference shifted when a row was inserted. Perhaps an older dataset was analyzed by mistake. Or perhaps the researcher deliberately changed the number because the correct value contradicted the hypothesis.

The published error may look identical in each case. The history behind it is not.

That is why contemporaneous records matter. Version histories, analysis scripts, raw data, laboratory notebooks, audit trails, emails, file metadata, documented procedures, and repeated patterns can help distinguish accidental events from deliberate or reckless conduct.

A pattern may be informative, but repetition alone does not automatically prove misconduct

One isolated discrepancy may have a plausible accidental explanation. A repeated pattern of changes all favoring the same hypothesis may be harder to explain that way.

ORI's current state-of-mind guidance identifies evidence such as patterns of falsification, affirmative manipulation, motive, and falsifications consistently supporting a hypothesis as potentially relevant to determining intentional or knowing conduct.

Still, pattern evidence should be evaluated rather than converted into an automatic rule. Repeated mistakes can also arise from a persistent software bug, misunderstood procedure, systematic instrument problem, inadequate training, or a flawed workflow. The task is to determine which explanation is supported by the complete evidence.

Negligence complicates the boundary

Between an ordinary mistake and deliberate deception lies a difficult territory of carelessness, inadequate checking, poor documentation, and failure to follow appropriate procedures.

Negligent research can create serious integrity problems without necessarily satisfying a formal misconduct definition. Under the current U.S. Public Health Service framework, for example, a finding requires intentional, knowing, or reckless conduct. Mere proof that a researcher should have exercised greater care does not by itself establish recklessness.

This distinction is why negligent research can be an integrity problem even when it is not formal misconduct. Poor practice may require correction, retraining, supervision, or institutional action even if a misconduct finding is unwarranted.

Recklessness can cross a line that ordinary carelessness does not

Under the current U.S. Public Health Service framework, recklessness can satisfy the state-of-mind requirement for research misconduct. ORI's 2026 guidance explains that evidence may support recklessness where a researcher knew of a risk of fabrication, falsification, or plagiarism and acted with indifference to that risk.

That is more demanding than simply showing that the researcher should have been more careful. A departure from accepted practice alone is not enough to establish a reckless state of mind under that guidance.

The difference between negligence and recklessness in a misconduct analysis can therefore become decisive.

What a researcher does after discovering an error is relevant, but it does not rewrite the past

A researcher who promptly preserves records, informs collaborators, rechecks the analysis, and seeks an appropriate correction behaves differently from someone who discovers the problem and tries to conceal it.

That response can provide useful evidence about the circumstances and is important for protecting the research record. But subsequent correction does not automatically prove that the original act was innocent, just as an initially defensive response does not by itself prove misconduct.

The original conduct and the later response should be evaluated separately.

An honest error can still make research unreliable

There is a crucial distinction between culpability and reliability.

Suppose a researcher honestly uses the wrong statistical code and publishes a conclusion that disappears when the analysis is corrected. If the mistake was genuinely accidental, it may not be misconduct. The published conclusion can nevertheless be wrong.

The research record may therefore need correction regardless of whether anyone committed misconduct. A researcher who discovers a consequential problem in their work should focus not only on classification but also on what must be done about the error itself.

One Wrong Figure, Three Very Different Explanations

Common Mistakes When Distinguishing Error From Misconduct

Assess the Evidence Before Assigning the Label

If you discover a serious discrepancy, resist two premature conclusions: "someone committed misconduct" and "it was obviously just a mistake." Both may outrun the evidence.

The safest conceptual separation is simple: first ask what happened to the research; then ask why it happened; finally ask what the applicable policy calls that conduct. Those questions interact, but collapsing them into one judgment is where many integrity discussions go wrong.

Questions to Ask When an Error Might Be Misconduct

Frequently Asked Questions About Honest Error and Misconduct

A Wrong Result Does Not Tell You Why It Became Wrong

How to Cite This Guide

This guide is intended to be read, shared, and used in research, teaching, and academic work. If you draw on its ideas, explanations, or other content, please acknowledge the source by citing the guide. Doing so gives appropriate credit and helps your readers locate the original resource.