Start with a decision
Research should name the decision it can inform and the uncertainty it addresses. Define the audience, scope and intended use before choosing a method.
Identify the kind of publication
Label editorial guidance, original empirical studies, evidence reviews and illustrative examples clearly. A practical guide is not an original study. A proposed study is not a completed result.
Separate claims from interpretation
Distinguish direct observations, participant reports, external sources and researcher interpretation. Keep examples and modelled scenarios visibly separate from measured outcomes. Do not invent participants, results, qualifications, customer logos or testimonials.
Make methods inspectable
For original work, publish the collection dates, recruitment approach, eligibility criteria, sample, analysis methods and limitations. Provide anonymised supporting material when it can be shared safely. Describe the scope of any independent review accurately.
Respect participants
Explain participation, recording and quote use clearly. Collect only what is needed and take reasonable steps to protect identifying or confidential material. Commercial follow-up should be separate from participation in research.
Disclose uncertainty and interests
Describe recruitment bias, measurement weaknesses and limits on generalisation. Disclose sponsors and relevant commercial relationships. The current guides were prepared with AI assistance; citations and methods should be assessed on their merits.
Maintain versions
Show publication dates and a version number on guides. Substantive changes should identify what changed. New original research should include a corrections route and an update record before release.