What is evaluation ethics? A practical checklist for personal information, consent and equity

Respect participants, avoid harm and value equity, privacy, transparency and professional integrity. Check potential disadvantages, excessive emphasis on one perspective and concealment of unfavorable results.

Consult program staff, personal-information managers and, as needed, affected people or external experts rather than leaving judgment to evaluators alone. This reduces omissions. Check applicable laws and organizational rules too.

Informed consent clearly explains purpose, questions, required time, who handles information and how, and the choice not to participate. Providing a technical document alone is insufficient explanation.

Take special care with people who find refusal difficult, including minors, support recipients and subordinates. Consider separating data collectors from support providers or assessors, anonymous methods and separate explanation settings.

• Briefly explain purpose and scope of use
• State that participation and responses are voluntary
• Explain stopping midway and the scope and deadline for withdrawal of submitted data
• Check recording, photography and secondary use separately
• Provide contacts and complaint routes

Collect only what questions require. Names, contacts, detailed characteristics and open accounts can identify people when combined, so check purpose and necessity for each.

Manage collection, storage, analysis, sharing and disposal. Decide access, devices or locations and deletion timing beforehand; avoid casual spreadsheet sharing.

Replacing names with numbers may still permit identification through linkage tables, characteristics or accounts. Do not judge legal compliance from this practical guide alone; check laws and organizational rules for the information type and intended use.

1. Identify only information needed for questions
2. Check direct and indirect identifying items
3. Decide storage, readers, sharing and retention
4. Separate identifiers from analytical data
5. Recheck cases and small-group tables in reports
6. Dispose of unneeded data through agreed methods

Examine experiences of people not reached or potentially harmed alongside averages. When presenting minority cases, avoid identifying wording and check permission to publish as needed.

Do not claim effects beyond what data support. State limitations, respondent bias and non-comparable conditions, treating unfavorable results as improvement material too.

• Consider who benefits or loses from findings
• Do not erase minority voices or exceptions
• Avoid case descriptions allowing identification
• Disclose conflicts and evaluator positions
• Present evidence strength and limitations together