Skip to content
PlagiatScanner.de
Research integrity · self-audit

Good research practice: a practical 20-point checklist

Good research practice means making decisions traceable, crediting sources and contributions accurately, handling data with care and correcting mistakes openly. This checklist turns those principles into evidence you can review before submitting an essay, dissertation or thesis.

Page ID INT-01Sources checked 4 September 2026

What this checklist can tell you

This is a process check, not a certificate and not a substitute for your university’s regulations, ethics requirements or subject-specific methods. Mark an item complete only when you can point to evidence: a saved search strategy, a preserved raw-data file, a methodological explanation, a version history or a documented conversation with your supervisor. Confidence that you “worked carefully” is useful, but it is not an audit trail.

The German Research Foundation’s Code of Conduct treats research integrity as a connected system of standards, responsibilities, documentation and quality assurance. The European ALLEA code expresses four fundamental principles: reliability, honesty, respect and accountability. A student project does not need an elaborate compliance system. It does need a proportionate record that allows a knowledgeable reader to understand how the work developed and where its limitations lie.

The 20-point self-audit

Planning and method

  1. Research question: I can state the aim, object and boundaries of the study in a few precise sentences.
  2. Method choice: I recorded why the method answers the question and which realistic alternatives I rejected.
  3. Requirements: I checked assessment rules, departmental guidance and supervisor agreements separately, and resolved discrepancies.
  4. Changes: Departures from the initial plan have a date and reason in my research log.

Searching and sources

  1. Search trail: Databases, queries, filters and search dates are clear enough to repeat the core search.
  2. Selection: Inclusion and exclusion follow relevant criteria rather than whether a source supports my preferred conclusion.
  3. Primary support: Strong factual claims lead, where feasible, to original studies, legislation, datasets or official guidance.
  4. Contrary evidence: Serious competing findings and interpretations are represented fairly.

Data and analysis

  1. Original state: Raw materials remain unchanged and are stored separately from processed versions.
  2. Cleaning: Exclusions, recoding and corrections are logged against a stated rule and reason.
  3. Analysis: Relevant calculations, software, parameters and versions are recorded.
  4. Uncertainty: Missing data, limitations and plausible alternative explanations appear in the interpretation.

Writing and contributions

  1. Intellectual ownership: Notes distinguish my reasoning from quotations, paraphrases and ideas encountered elsewhere.
  2. Traceability: Material derived from a source can be followed from the prose back to the exact reference.
  3. Paraphrase: I reconstruct meaning accurately in my own argument instead of merely swapping words.
  4. Assistance: Human, digital and AI-assisted contributions are disclosed in line with the rules that govern my work.

Submission and correction

  1. Credit: Named contributors reflect actual work; acknowledgements and authorship are not confused.
  2. Rights: Personal data, consent, licences and permission to reproduce material have been checked.
  3. Cross-check: Citations, bibliography, figures, tables and appendices have been reconciled systematically.
  4. Error route: I know whom to contact if I discover a material mistake and retain the evidence needed to explain it.

Connecting the audit to integrity codes

Practical orientation, not a complete restatement of either code
AreaChecksRelevant code themes
Planning and method1–4Methods, standards, quality assurance and research framework
Searching and sources5–8Honesty, reliability and responsible treatment of prior research
Data and analysis9–12Documentation, preservation and reproducible reasoning
Writing and contributions13–16Attribution, authorship and accountability
Submission and correction17–20Rights, publication choices, corrections and conflict procedures

This mapping is intentionally broad. The DFG code addresses institutions and the research system as well as individuals. Universities translate its principles into local statutes, assessment rules and support structures. Use this audit alongside those documents, and never infer that a general web checklist overrides a binding local requirement.

Turn ticks into an evidence trail

Create one modest evidence folder for each area. Planning evidence might include your proposal, a decision log and approved changes. Search evidence can include exported result lists, exact queries and screening notes. For data, retain the lawful original state, a codebook, processing records and a readable explanatory file. Writing evidence may include reference-library exports, draft history and a contribution note. Retain only material you are entitled to keep, and apply access controls where data are sensitive.

Useful records are dated, named intelligibly and linked to the project. A filename such as “final_new2” conceals its history. “2026-08-18_interview-coding_v03_after-category-review” states when and why the version exists. Proportionate documentation explains consequential decisions; it need not capture every click. For confidential research, minimisation and secure handling are part of integrity rather than exceptions to transparency.

Warning signs worth investigating

Pause if you are reconstructing all records immediately before submission, if every retained source confirms one position, if exclusion rules changed after the results became visible, or if passages can no longer be connected to their note or source. None of these signs by itself proves misconduct. Each identifies a place where reconstruction, supervisor advice or explicit limitation may be necessary. An honest account of uncertainty is more defensible than a smooth retrospective story that hides how the project actually changed.

Boundary: This checklist reviews working practices, not personal character. It does not provide a formal finding or legal advice. Specific concerns should follow the confidential procedures of the relevant university or research organisation.

A focused 30-minute final pass

Review all twenty items using three labels: green for “evidence present”, amber for “explained but not evidenced” and red for “unresolved”. Address red items affecting results, attribution, privacy or independent work first. An amber item may be closed by a clear methods sentence, a well-named file or a documented question to the appropriate person. State any genuine remaining limitation in the thesis. Used this way, the checklist becomes a last quality-control exercise rather than a ceremonial set of ticks.

Sources and further reading

  1. German Research Foundation: Guidelines for Safeguarding Good Research Practice, official English code portal, accessed 4 September 2026.
  2. ALLEA: The European Code of Conduct for Research Integrity, Revised Edition 2023, official publication, accessed 4 September 2026.
  3. German Research Ombudsman, independent information and confidential advice, accessed 4 September 2026.