Connect ratings with explanations, design useful follow-up questions, and interpret agreement and contradiction together.
A mixed-method survey combines quantitative questions with qualitative questions and integrates their findings to answer a shared question. A satisfaction rating and a follow-up asking why the person chose it are a simple starting point. The value comes from how the two answers are interpreted together.
A form containing a comment box is not automatically a rigorous mixed-methods study. The sampling, collection and analysis need to explain how the qualitative evidence contributes. NIH’s guidance emphasizes planning the integration of methods rather than treating the components as unrelated exercises. See the NIH mixed-methods resource.
| Design | How it works | Example |
|---|---|---|
| Rating with an immediate explanation | Ask an open follow-up next to a relevant closed question. | Rate the ease of submitting a report, then describe the hardest step. |
| Targeted branching | Ask a relevant follow-up based on an answer. | Ask people reporting a delay what happened before the delay. |
| Survey followed by interviews | Use quantitative findings to select questions or cases for deeper exploration. | Interview customers whose ratings improved and those whose ratings declined. |
| Qualitative exploration before the survey | Use early conversations to inform later questions and response options. | Interview partners about reporting barriers before constructing the questionnaire. |
The last two are broader sequences of research rather than simply question types in one form. Explain that distinction in the plan.
Use it when a numeric result alone will not explain the next decision. It can help distinguish whether a low rating relates to access, communication, process or expectations, or explore why similar scores conceal different experiences.
It is less useful to add open questions that no one will analyze. Keep the number of questions manageable and explain their purpose. Short written answers may not provide the depth needed for a complex or sensitive issue; an appropriate interview can complement them.
| Closed question | Open follow-up | Reason |
|---|---|---|
| How easy was it to complete the application? | Which step most influenced your answer? | Locate the experience behind the rating. |
| Did you use the skill in the past four weeks? | What enabled or prevented you from using it? | Distinguish capability from opportunity. |
| Was the issue resolved? | What, if anything, remains unresolved? | Check whether closure means resolution to the respondent. |
| How clear were the reporting requirements? | Which requirement needed clarification? | Identify a specific definition or instruction to improve. |
Avoid wording that assumes an improvement or failure. Ask “What influenced your answer?” rather than “Why was the new process better?” Keep the reference period and scale visible.
Suppose 50 learners answer a follow-up survey and 30 report using a skill in the last four weeks. Open responses are reviewed for reasons that enabled or prevented use. The figures below are illustrative, and each person is assigned one primary explanation only for this simplified demonstration.
| Reported application | Primary explanation | Respondents |
|---|---|---|
| Used the skill | Opportunity to practice and useful job guidance | 22 |
| Used the skill | Other or unclear explanation | 8 |
| Did not use the skill | No relevant task or opportunity | 12 |
| Did not use the skill | Needed further practice or support | 8 |
The result suggests two different follow-up actions: create opportunities to apply learning and offer additional support where needed. It does not prove the explanations caused the behavior. Review the original answers before deciding whether these categories adequately represent them.
If people can receive multiple codes, label counts as coded mentions or respondents with each theme and do not expect the percentages to add to 100%.
A published methodological paper describes integration at the design, methods and interpretation/reporting levels. Its central relevance here is that integration is a planned analytical step, not merely placing two charts beside one another. Read the integration paper.
Keep the disagreement visible. A person may give a high overall score while describing one serious problem, or a low score despite praising a particular interaction. The questions may address different aspects of the experience.
Check wording, timing and the unit of analysis before concluding that an answer is inconsistent. Do not overwrite the rating to make it match an AI sentiment label. If follow-up is appropriate and permitted, ask for clarification.
| Term | Meaning in this context |
|---|---|
| Mixed-methods research | Integrates quantitative and qualitative components to answer a research question. |
| Mixed-method survey | Uses and integrates quantitative and qualitative survey evidence, sometimes with follow-up methods. |
| Mixed-mode survey | Collects responses through more than one delivery mode, such as web and telephone. |
| Multimethod study | Uses multiple methods; definitions vary, and the methods may be of the same broad type. |
A web-and-phone survey can be mixed-mode while collecting only closed responses. A single online questionnaire can collect both numerical and qualitative material.
Retain the question version, reporting period and relevant group with both the score and explanation. Use a stable identifier only where the design needs record linkage and the permission model supports it.
AI can help classify incoming comments or prepare a synthesis across periods. Preserve the original response, classification criteria and review history. This makes it easier to investigate a trend without treating every survey as a disconnected file.
Use these existing Academy guides for the practical next step. They are suggested companion readings; follow each guide’s course navigation for the full sequence.
Explore the Loop methodology for the ongoing cycle of collection, analysis and improvement.
It creates an opportunity for qualitative evidence, but a mixed-methods claim needs an explanation of how that evidence is collected, analyzed and integrated.
It presents quantitative and qualitative findings together so the reader can examine their relationship, including agreement and contradiction.
Yes. Keeping the answers linked can support respondent-level interpretation, subject to the study design and privacy requirements.
Theme counts describe the coded data. They do not automatically estimate population prevalence, especially when open responses are optional or sampling is not representative.
It can assist organization and interpretation, but reviewers need to verify source fidelity, definitions, contradictions and the conclusions.
