Cognitive interviewing helps a nursing dissertation examine how people understand and answer questionnaire items. The purpose is to investigate the response process, not to test participants or persuade them to choose the “right” answer. This guide explains how to plan neutral probes, interpret the evidence and record revisions without claiming that a small pretest validates an entire instrument.
This guide examines how respondents understand and answer particular items. Its practical output is a response-process evidence table connecting neutral probes, observed interpretation and a proposed wording revision. Begin with the how to pilot a nursing research questionnaire and record changes if you first need to establish the broader project approach. Use the detailed decisions below once that foundation is clear.
Define what cognitive interviewing will examine
Start with a specific uncertainty about the questionnaire. Perhaps respondents interpret “recent care” differently, cannot distinguish two response options or struggle to recall the requested period. A focused uncertainty helps determine what to ask and what evidence would inform a revision.
The CDC describes cognitive interviewing as a question-evaluation method that explores how respondents process survey questions (CDC, 2024). Use that purpose to keep the session distinct from a substantive interview about care experiences. You may hear important experiences, but the planned analysis here concerns how the questionnaire produces an answer.
Keep the participant separate from the item problem
Explain that the questionnaire is being examined, not the person. A participant who interprets an item differently has provided useful information about the item’s functioning in that context. Avoid language suggesting that they failed to understand something obvious.
The interviewer should not teach the intended meaning before observing the response process. If you explain the answer first, later agreement may reflect the explanation rather than the original wording. Keep any necessary clarification separate in the record so readers can see what happened before and after assistance.
Choose items for a reason
You may need to examine instructions, key concepts, response options or questions with known ambiguity. State why each selected element matters. Do not choose only the easiest items or those that participants are likely to praise.
For a long instrument, testing every item in depth may be impractical. Explain the sampling of items and the resulting limits. If permission restricts adaptation or reproduction, resolve that issue before preparing a modified form. The permissions guide covers the separate rights decision.
Prepare a brief interview plan
The plan should identify the version being tested, session purpose, administration mode and core probes. Add a process for unexpected difficulties and any agreed limits on assistance. A written plan improves consistency without requiring every conversation to follow an identical script.
Use the institution’s appropriate approval and consent arrangements. Explain whether the session will be recorded, how notes will be handled and what participation involves. Do not collect more personal or clinical information than the purpose requires. A questionnaire pretest is not a reason to copy identifiable records into a dissertation appendix.
Understand think-aloud and probing approaches
A think-aloud approach invites a person to describe what is going through their mind while answering. Probing uses follow-up questions to explore a particular part of the response process. Either approach can be helpful when suited to the participant and item.
Think-aloud work may feel unfamiliar, while probes can influence what people discuss. Explain the approach used and any support provided. Avoid presenting one approach as universally superior. The relevant question is whether it gives useful evidence about the specific interpretation uncertainty without turning the session into instruction.
Choose when to probe
Probing immediately after an item can preserve detail but interrupt the flow of completing the questionnaire. Asking afterwards may preserve a more natural sequence but depend more on recall. Consider the purpose and burden when choosing the timing.
Document the timing so that readers understand the conditions under which evidence was produced. If the interviewer switches approach because a participant is struggling, record the change. A flexible response can be reasonable, but an unreported change makes it harder to interpret why different participants produced different kinds of explanation.
Use neutral comprehension probes
A neutral probe asks what the person understood without supplying the intended answer. For an invented item about “regular contact”, you might ask, “What did regular contact mean to you here?” A leading alternative would be, “Did you understand that we meant a weekly visit?”
The leading version introduces a definition that the original item did not communicate. It may be useful later to explore a revised wording, but it does not reveal the participant’s original interpretation. Label the stage clearly and keep the response before clarification visible in the analytical record.
Explore recall and judgement separately
A person may understand the words but struggle to retrieve the relevant experience. Ask how they remembered the period or events used to answer. If the item requests a summary, explore how they selected or combined experiences without implying a preferred method.
Judgement can also involve deciding whether an experience fits a category. A participant may consider one unusual event more important than several routine contacts. That is analytically different from misunderstanding the wording. Keep these explanations distinct so that a wording revision is not proposed for a problem that actually concerns the task the questionnaire asks people to perform.
Examine the response options
After understanding and recall, the person still has to map their answer onto an available option. Ask how they chose between neighbouring categories and whether any option was difficult to use. Avoid suggesting that a middle category is a weak or undesirable answer.
Notice whether the item assumes an experience that some respondents have not had. A forced answer may conceal a need for a legitimate skip route or “not applicable” option, depending on the instrument and intended use. Do not add categories without considering permissions and scoring consequences.
Worked example: an ambiguous recall period
The following wording is invented for teaching. An item asks, “During recent visits, how often did you know what would happen next?” One participant considers the last appointment. Another thinks about all visits over the previous year. A third answers about telephone calls as well as visits.
The interviewer asks each participant what “recent visits” meant and which encounters they considered. The record shows different boundaries. It does not establish that the respondents were careless or that one person’s answer should be corrected.
The team considers a clearer period and encounter definition that fits the project question. If permission and the approved process allow a revision, the next version is tested with appropriate respondents. The finding supports a specific concern about boundaries; it does not demonstrate that all other items work well or that the revised score has established validity.
Create a response-process evidence table
| Element | Observation or account | Interpretation and next step |
|---|---|---|
| Item version | Exact approved version identifier | Prevents mixing evidence from different drafts |
| Comprehension | Participant’s explanation of a key term | Possible mismatch with intended meaning |
| Recall | Events or period used to answer | Need to clarify the requested boundary |
| Response mapping | Difficulty selecting an available option | Review the response task and permitted changes |
| Revision | Proposed wording and rationale | Retest rather than assume the concern is solved |
Keep observation separate from explanation
A pause, request for help or changed answer is an observation. The reason for it requires evidence. Do not automatically code every pause as confusion or every quick response as comprehension. Ask a neutral question where appropriate and retain uncertainty when the reason is unclear.
In the analysis, distinguish what the participant said, what the interviewer observed and what the researcher inferred. That separation makes the decision trail more credible. It also prevents a preferred revision from being justified by an interpretation that was never actually explored during the session.
Look for patterns and meaningful differences
Compare evidence across items and participants while respecting the small, purposive nature of the work. Several similar interpretations may identify a recurring issue. One unusual interpretation may still reveal an important boundary or accessibility concern.
Do not use a small pretest to estimate how common a problem is in the whole population. Counts can describe the sessions, but they do not automatically support a prevalence claim. Explain why a concern matters to the item’s intended use, rather than deciding that only the most frequent concern deserves attention.
Record revisions as decisions
A change log should show the earlier version, evidence informing the change, proposed revision and remaining uncertainty. Keep a version identifier for the questionnaire and interview plan. Otherwise, later notes may become impossible to connect to the wording that produced them.
The parent guide on piloting and recording questionnaire changes addresses the wider process. This article focuses on the evidence about answering an item. Link the two records where appropriate instead of repeating the entire pilot protocol in every section.
Retest the question that the revision creates
A revision can solve one problem while introducing another. Adding a precise period may improve recall boundaries but make the question longer or exclude a relevant encounter. Replacing a technical term may change the intended construct.
Where the approved plan permits another round, examine the specific issue introduced or left unresolved by the revision. Do not call a change successful simply because the team prefers the new wording. Record what was actually tested and avoid implying that an untested final edit has the same evidence as the version participants examined.
Report cognitive interviewing without overclaiming
Describe the purpose, participants in non-identifying terms, item selection, approach, analysis and resulting changes. Explain what evidence was available and which parts of the instrument were not examined. Avoid reporting only that “a pilot confirmed clarity”.
A precise conclusion might state that the sessions identified different interpretations of a recall period and informed a revised instruction. It should not say that the entire instrument was validated. Evidence about comprehension is valuable, but it does not automatically establish reliability, responsiveness or equivalence across groups.
Keep this method distinct from item deletion
Cognitive evidence may support retaining, revising or investigating an item. Removal is one possible decision, not the automatic outcome of a difficulty. An important concept may need clearer wording rather than deletion.
The guide on reviewing pilot survey items considers that decision alongside coverage, missingness and scoring. Keep the response-process finding visible when discussing removal. Otherwise, a nuanced interpretation problem can disappear into an unexplained statement that an item “performed poorly”.
Practise probes before the first session
Review the planned probes with your supervisor or another appropriate reviewer using fictional responses. Look for wording that embeds the intended meaning, rewards agreement or treats the participant’s answer as incorrect. A short practice can reveal that an apparently neutral follow-up actually suggests a preferred response.
For example, “Why did you misunderstand this?” assumes both a problem and its cause. “What did this phrase mean to you?” leaves room for an explanation. If a participant asks whether their answer is correct, remind them of the session’s purpose without supplying the questionnaire’s intended interpretation before it has been explored.
Plan for unexpected personal accounts
Participants may describe sensitive experiences while explaining how they answered an item. Follow the approved arrangements for handling such information and any concerns arising in the session. Do not promise an absolute confidentiality rule that conflicts with the actual participant information or local procedures.
Keep the analytical record focused on the response-process question. A detailed personal story may explain why someone selected an option, but the public dissertation may need only a less identifying summary of the relevant interpretation. Do not reproduce distinctive circumstances merely because they make the example memorable.
After each session, distinguish the notes needed for item review from information requiring a separate local response. A cognitive-interview analysis should not become an informal investigation into an identifiable care event. That boundary protects the method’s focus and supports accurate reporting of what the project was designed to examine.
Check the final summary against the evidence table
Before writing that an item was revised “for clarity”, identify the actual finding in the evidence table. Was the concern a term, recall period or response option? Naming the problem and decision makes the report more informative. It also prevents a general phrase from concealing changes that were based only on team preference.
Cognitive-interviewing checklist
- Define a specific uncertainty about answering the questionnaire.
- Identify the version, selected items and intended administration mode.
- Use the appropriate approval, consent and information-handling arrangements.
- Prepare neutral probes that do not supply an answer.
- Separate comprehension, recall, judgement and response-option issues.
- Distinguish observations from researcher explanations.
- Record revisions with evidence and version identifiers.
- Limit conclusions to the response processes actually examined.
Review the interviewer’s contribution
Include a short reflection on how the interview approach may have shaped the evidence. Consider whether probes differed between sessions, whether examples were introduced or whether participants appeared to seek reassurance. Describe material differences without assuming they invalidate every finding.
The reflection should connect to interpretation. If a concern appeared only after the interviewer supplied an example, distinguish that evidence from a spontaneously reported difficulty. If a probe was revised between rounds, identify which version produced which finding. This makes the final response-process account more transparent and helps a reader assess the strength of the proposed item change.
Use the same item identifiers in session notes and the change log so that readers can trace each decision to the wording tested.
Frequently asked questions
Is cognitive interviewing the same as a pilot survey?
No. It focuses specifically on how people understand and answer questions. A wider pilot may also examine recruitment, administration, data handling and feasibility. Explain which purpose your sessions served rather than using the terms as interchangeable labels.
How many participants do I need?
There is no universal number that makes every item adequately tested. Justify the plan in relation to the instrument, intended respondents, uncertainties and project resources. A small set of sessions can identify problems without establishing their prevalence or comprehensive validity.
Can I explain the item when someone is confused?
Necessary assistance may be appropriate, but record what happened before and after it. An explanation changes the conditions under which the answer is produced. Do not treat agreement after teaching the intended meaning as evidence that the original wording communicated clearly.
Does a single misunderstanding justify deleting an item?
Not automatically. Examine the interpretation, relevance and possible revision, alongside permissions and scoring consequences. One account can reveal an important issue, but the decision should follow a documented rationale rather than a simple frequency rule.
Can I say the questionnaire is validated afterwards?
Describe the actual evidence instead. Cognitive interviews can inform understanding of the response process, but they do not establish every measurement property. State which items, respondents and versions were examined and what further interpretation questions remain.
Develop a clearer pretesting argument
Our nursing dissertation methods support can help you review an anonymised probe plan or change log. Ask about the specific item problem you need to explain. Educational feedback should strengthen your own methodological reasoning; participant recruitment, approval and the final analysis remain within your institution’s arrangements.
Reference
Centers for Disease Control and Prevention (2024) Cognitive interviewing. Available at: CDC CCQDER (Accessed: 16 October 2026).
