Draft a Cognitive Interview Guide for Survey Items

Evidence and applicability: use only supplied evidence for historical facts, source values, completed checks and approvals. Cite source labels and row, item or section references. Separate observed facts, assumptions, proposed procedures and unresolved questions. Preserve conflicting claims with provenance; absent documentation does not prove a practice never occurred. Do not invent data, permissions, findings, validation or sign-off. A draft plan grants no operational, ethical or legal clearance.

Use only material supplied in this conversation. Ask for the exact draft survey items with stable item IDs, shared instructions, response categories and routing; intended construct and use; population and relevant language variety/context; collection mode and accessibility needs; session resources; and applicable consent, safeguarding, recording and data-handling limits. A survey title or topic alone is insufficient for item-specific probes. Request safe instrument text and nonidentifying context, not real respondents' personal answers or contact information. If items or construct definitions are absent, identify missing inputs and provide only a session structure, not invented survey questions or an item-specific guide.

Draft an item-linked cognitive interview guide to investigate how intended respondents understand and answer the supplied survey items. Keep this distinct from a full pilot logistics plan or a substantive qualitative interview about the program.

1. Establish scope from the supplied items, construct, language and mode. Retain exact item wording, item IDs, instructions, response options, reference periods and version. Identify priority uncertainties as hypotheses to investigate, not demonstrated defects. Link each proposed probe to the intended construct and a specific item or instruction. Do not assume a concept is understood, a translation is equivalent, or respondents have experienced the topic. Flag missing local-language competence or cultural context for qualified review rather than claiming local fit.
2. Draft a brief session introduction explaining that the questions are being tested, not the participant, and that honest reactions including uncertainty are useful. Explain voluntary participation, ability to skip a question, decline a probe, pause or stop, and the actual purpose, expected time, data uses and confidentiality limits only where supplied. Do not promise that services or benefits are unaffected unless that assurance is verified. Provide unresolved consent details as a completion list outside the script rather than inventing contacts, compensation, guarantees or approvals. Recording requires separate, informed permission and an authorized procedure; participation does not imply recording or quotation permission. State where any locally required assent, guardian permission or safeguarding review must be resolved before use. Do not pressure participation through program staff or authority relationships.
3. Offer a proportionate administration approach, using neutral retrospective probing after the participant answers or an optional think-aloud method where appropriate to language, accessibility and burden. Explain the approach in plain language and allow the participant to decline it. Administer the original item in its intended mode first; record the version, exact probe used and any deviations. Do not teach the intended interpretation before eliciting understanding, correct the response, praise a preferred answer, or ask every possible probe mechanically. Use a short neutral practice exercise only if appropriate, clearly separate from evidence about target items.
4. Produce an item-by-item probe table with exact source item reference, intended construct, uncertainty to investigate, neutral primary probe, optional follow-up, cognitive stage and interviewer observation. Tailor comprehension, retrieval/recall, judgment and response-mapping probes to the item. Candidate forms include: 'In your own words, what is this question asking?'; 'What did that term mean to you here?'; 'How did you work out your answer?'; 'What time period did you have in mind?'; 'Was anything difficult to remember or decide?'; and 'How did you choose that response option?' Use only relevant forms, avoiding assumptions that the participant misunderstood or had a particular experience. Ask about reasons for difficulty only after difficulty is indicated. Do not embed the intended meaning, an expected answer, a presumed service experience or a leading example in a probe. Avoid yes/no comprehension checks as the only evidence of understanding. For sensitive topics, ask about the question-answering process without demanding disclosures or repetition of distressing experiences.
5. Give interviewer process notes: use neutral acknowledgments and pauses; respect refusals; distinguish the original answer, participant explanation, interviewer observation and interpretation. Do not infer confusion from silence or refusal alone. Record what happened without diagnosing motives, literacy or ability. If a participant requests an explanation, follow the approved protocol, record the explanation and its timing, and distinguish any subsequent response from the unassisted response. Adapt pace and mode without silently changing the construct; log adaptations and any unresolved accessibility issue.
6. Provide a blank observation and revision log, without sample findings. Fields: session code with no direct identifier, language/mode, item/version, actual probe, safe evidence reference or permitted minimal excerpt, observation, interpretation/hypothesis, alternative explanation, construct implication, proposed revision, decision status, authorized reviewer if supplied, and retest need. Do not populate participant quotes, difficulties, frequencies, completed sessions or improvement claims unless actual safe session evidence is supplied. Keep raw notes restricted under the applicable rules and share only necessary de-identified findings; small samples and distinctive quotations can remain identifiable.
7. Propose a post-session debrief and evidence-to-revision process: separate isolated observations from recurring patterns within the sessions actually completed, compare relevant language/context groups cautiously, preserve contrary evidence, and adjudicate changes against the intended construct. Material changes require further testing in the relevant version/mode before claims about use. Sample scope should be justified by item risks, diversity of relevant experiences, safety and resources, with no universal count or claim of statistical representativeness. Cognitive sessions can reveal response-process concerns; they do not alone establish reliability, validity, measurement equivalence or field readiness. Provide unresolved decisions and next testing needs for the responsible team, not final clearance.

Output: a scoped session introduction, item-linked neutral probe guide, interviewer process notes, blank observation/revision log, and conditional debrief/retest plan. Mark unresolved script details clearly outside the participant-facing text. This is a proposed guide, not evidence that sessions occurred or a claim that the instrument has been validated.