Skip to content
PlagiatScanner.de
Research integrity · assessment

How to distinguish research misconduct from an honest mistake

Not every research error is misconduct. A responsible assessment considers the act, the person’s knowledge, the standard of care, the significance of the problem and the response after discovery. Only the competent procedure at the relevant institution can make a formal finding; a similarity score and this guide cannot.

Page ID INT-02Sources checked 4 September 2026

Use three levels, not instant labels

An honest error can occur despite reasonable care: a transposed column label, a transcription mistake or a misunderstood technical term. Negligence describes an avoidable failure of care, such as skipping a basic check or keeping source notes so poorly that ownership becomes unclear. Research misconduct is a formal category. DFG procedural guidance focuses in particular on intentional or grossly negligent misrepresentation, misappropriation of another person’s work and interference with another person’s research. The institution’s applicable policy ultimately defines its own process and findings.

The distinction matters because one visible problem may have several causes. A missing citation might result from a footnote deleted during formatting, careless note-taking or deliberate concealment. The finished paragraph rarely proves the writer’s state of mind. A fair preliminary review therefore starts with observable facts, records credible alternatives and keeps suspicion, evaluation and decision separate.

Decision tree: error, negligence or a matter for formal review?

  1. What exactly happened? Describe the observable discrepancy without assigning motive: “three paragraphs closely resemble source X and contain no citation”, not “the writer stole them”.
  2. Which rule or professional expectation may apply? Identify the assessment regulation, integrity policy, citation guidance or methodological standard. If no relevant rule is evident, seek disciplinary clarification first.
  3. How material is the discrepancy? Consider scale, effect on conclusions, affected people and ease of correction. A typographical mistake is not equivalent to an invented dataset.
  4. Which explanation is supported by records? Examine notes, raw materials, version history, correspondence and dates. Deliberately test a benign explanation as well as a concerning one.
  5. Was the expected care reasonable in context? An avoidable lapse may be negligence. Repetition, clear prior warnings or disregard of elementary duties can make the concern more serious, but none automatically proves intent.
  6. Are there indicators of intent or gross negligence? Stop making private findings here. Preserve relevant evidence and consult the confidential route; a trained and authorised panel should assess contested intent.
  7. What happened after discovery? Prompt disclosure, correction and cooperation do not erase an original breach, but they matter to the complete facts. Concealment or further alteration of records should be recorded without speculation.

The same symptom can have different meanings

Examples and counterexamples for an initial, non-binding assessment
ObservationPossible errorPossible negligencePossible formal concern
Citation missingFootnote lost in formatting; drafts retain attributionSources mixed in notes and final check omittedAnother person’s prose knowingly presented as original
Data value changedClear entry error corrected under a stated ruleCleaning performed without a log, but original retainedValue altered to manufacture the desired result
Figure reusedLicence credit accidentally incompleteRights and provenance never checkedAuthorship knowingly represented falsely
Contrary study absentMissed during an otherwise defensible searchSearch stopped prematurelyKnown evidence concealed to create a misleading account

These examples are prompts, not automatic offences. Context may change an assessment: experience, training received, access to information, disciplinary convention and the consequence of the discrepancy all matter. For student work, it is also relevant whether a requirement was communicated clearly. Lack of knowledge does not excuse every failure, but it can be material when distinguishing mistake, ordinary carelessness and more serious culpability.

Responding when the error may be yours

Do not keep editing the only original files. Preserve a copy and record when you noticed the issue, which material is affected and what consequences may follow. Separate established facts from assumptions. Before submission, many problems can be corrected openly. After submission, contact the supervisor, examinations office or named integrity contact in accordance with local rules. A useful report gives the location, evidence, apparent reach and proposed correction; it avoids an amateur legal diagnosis or a dramatic confession that outruns the facts.

If you notice a possible problem in someone else’s work, maintain confidentiality. Do not accuse publicly, recruit outsiders or investigate protected data on your own. Retain only evidence you may lawfully access, then ask a neutral adviser about the correct route. Germany’s Research Ombudsman provides confidential advice on integrity concerns, including authorship and supervision disputes; an institutional ombudsperson may be best placed for local processes.

Why a proper procedure protects everyone

Procedures modelled on DFG guidance distinguish preliminary assessment from formal investigation, protect confidentiality and allow affected people to be heard. An initial suspicion is not a conclusion. Records are examined, statements requested and jurisdiction established. Consequences should never follow solely from an online text comparison or a software alert. A plagiarism check may identify passages for expert review; it cannot establish intention, culpability or research misconduct.

Not legal advice: This page explains general assessment principles. A real examination, employment or legal matter must follow the rules and competent bodies of the institution involved.

A fair five-sentence preliminary record

Write one sentence for each of the following: the observation, the potentially relevant rule, the supported circumstances, the unresolved alternative explanations and the next competent step. This small discipline prevents a suspicion from silently becoming “fact”. It also prevents a serious issue being dismissed as a mere accident without examination. Research integrity requires both: determined clarification and restraint until the evidence has been tested.

Official sources

  1. German Research Foundation: procedural guidance on good research practice, official publication.
  2. DFG Code of Conduct: Guidelines for Safeguarding Good Research Practice, official English portal.
  3. German Research Ombudsman, independent confidential advice.