PlagiatScanner.de
Integrity · grey areas

Questionable research practices: examples and boundaries

A questionable research practice is a methodological or ethical choice that weakens reliability or transparency without automatically constituting a formal finding of misconduct. Context, applicable rules, effects, care and disclosure matter more than a sweeping label.

INT-07 · Checked 4 September 2026

Why grey areas need precise language

Between exemplary practice and formally established misconduct lie choices that can distort research quality: unclear exclusions, selective reporting, incomplete documentation or inappropriate authorship. Naming the grey area allows improvement before a problem escalates. It is not a universal offence list. Institutions define misconduct through their own policies; depending on intention, seriousness and rules, one practice may be an honest error, negligence or a matter for formal investigation.

A fair assessment begins with the scientific function at risk. Can the choice be reconstructed? Are evidence and sources reliable? Are participants and colleagues respected? Is responsibility assigned accurately? These questions reflect ALLEA’s principles of reliability, honesty, respect and accountability. They support specific remedies without turning every imperfection into a judgement about character.

Case matrix with counterexamples

Prompts for a non-binding professional review
Practice Questionable when… Defensible counterexample Safeguard
Changing the question change after seeing results is concealed new question is labelled exploratory decision log
Excluding cases rules shift with desired outcome predefined quality exclusion raw state and exclusion list
Trying analyses only favourable variant is reported transparent sensitivity analysis analysis and variants log
Selecting literature known contrary evidence is systematically absent documented, relevant boundary search and screening record
Editing a figure processing changes meaning or comparison uniform, disclosed readability adjustment original and processing steps
Reusing text origin and novelty become unclear cited reuse with rights checked overlap declaration
Assigning authorship status replaces contribution roles reflect actual work early contribution matrix

Planning, collection and data

Problematic flexibility often begins without a plan to deceive. Time pressure, uncertain roles or poor version control make decisions impossible to reconstruct. Preserve the original state, date plan changes and explain departures. In qualitative work this includes codebooks, categories and case selection; in quantitative work it includes variables, outliers, missing values and models.

More data are not automatically better. Collection without adequate consent, unnecessary personal attributes or unauthorised access breaches responsibility even when analysis is competent. Conversely, privacy should not become a blanket excuse for withholding every methodological detail. Describe process and constraints while protecting confidential material.

Use analytical flexibility openly

Research needs exploration. Patterns can suggest hypotheses, models can respond to diagnostics and qualitative categories may evolve. Flexibility becomes questionable when the final report conceals that development and presents a retrospective choice as predetermined. Separate confirmatory and exploratory work, show relevant alternatives and explain why one analysis carries the principal claim.

Null, negative and conflicting findings are information. Concealing them distorts the record and may misdirect later work. Report limits of measurement and sample rather than searching endless subgroups or outcomes for a convenient result. Do not condemn an operation by its label alone; ask what was chosen, when, why and with what consequence.

Writing, citation and publication

Selective citation arises when only supportive sources receive attention or a source is made to claim more than its design permits. A for-and-against evidence matrix is a practical safeguard. Quotations, paraphrases, translations and borrowed structures need traceable origins. Similarity software can locate passages for inspection but cannot determine appropriate attribution by itself.

Publication grey areas include fragmented reporting, concealed overlap, undeserved authorship and incomplete competing-interest statements. Separate papers are not inherently improper when they answer genuinely different questions. The relationship must remain visible, prior findings should be cited, and readers must not be misled about novelty or the underlying dataset.

Turn a grey area into a correctable process

  1. Describe the observable decision without assigning motive.
  2. Name the affected function: reliability, traceability, fairness, rights or accountability.
  3. Check local policy and disciplinary standards.
  4. Preserve lawfully accessible originals and credible alternative explanations.
  5. Choose a proportionate remedy: document, reanalyse, narrow prose, clarify credit or consult the competent office.
  6. Record how the correction will be visible in the report.

Disclose a problem in your own work early. If it concerns someone else, protect confidentiality and use advisory routes rather than public accusation. A repeated, consequential pattern may require formal assessment; only the authorised institution can decide that.

Questions for supervision meetings

Ask whether the decision was made before or after seeing the relevant result, whether an alternative would materially change the interpretation, what evidence preserves the original state and where the final report discloses the choice. These questions are neutral enough for routine quality assurance and specific enough to reveal hidden flexibility. Recording the answer creates a useful prevention tool rather than a retrospective blame exercise.

At project close, compare those records with the methods and limitations sections. A change documented internally but absent from the report may alter how readers understand the evidence. Correct the account rather than erasing the history. This simple reconciliation turns supervision notes into a genuine integrity safeguard.

Test proportionality: Not every departure needs the same response. Ask whether it affects evidence, interpretation, rights or fair credit; how many conclusions depend on it; whether the original state survives; and whether transparent correction can contain the problem. A missing file note may be repaired through documentation, while outcome-dependent deletion may require reanalysis from the preserved original. Record a reasoned decision to take no further action as well. This prevents a warning sign becoming an automatic sanction, while ensuring consequential problems are not dismissed as style. If seriousness or jurisdiction remains uncertain, confidential advice is the appropriate next step.

Boundary: The matrix supports reflection. It neither establishes misconduct nor replaces the integrity policy of the relevant organisation.

Sources

  1. ALLEA: European Code of Conduct for Research Integrity, 2023.
  2. DFG Code of Conduct for Safeguarding Good Research Practice.
  3. COPE Core Practices.