A mixed-method survey carries closed and open-ended questions on one record. A worked questionnaire example, the four formats, and how to design one.
A mixed-method survey is a questionnaire that intentionally collects and integrates quantitative and qualitative evidence to answer the same research question. Structured questions establish patterns, frequencies, ratings, or change; open-ended questions explain meaning, context, exceptions, or mechanisms. The instrument becomes mixed-method through planned integration, not merely by containing two question types.
Watch: Mixed methods research: combining qualitative + quantitative.
A mixed-method survey is narrower than a mixed-methods study. A broader study may combine surveys with interviews, observations, documents, or behavioral data in sequential, embedded, or convergent designs. For survey-led programs, Sopact's preferred architecture keeps decision-critical ratings and their explanations on the same persistent participant record, which preserves the connection without presenting one architecture as a universal methodological rule.
Key takeaways
A survey becomes mixed-method when quantitative and qualitative questions contribute to the same research question and the analysis deliberately integrates their findings. Closed questions measure prevalence, level, difference, or change. Open questions explain meaning, mechanism, context, and exceptions. Integration tests what the two forms of evidence say together.
A questionnaire can contain Likert items and comment boxes without being meaningfully mixed-method. The design becomes mixed-method when the team specifies in advance which open response explains which result, how the two will be joined, and how convergence or contradiction will affect interpretation. The integration principles described by mixed-methods researchers distinguish collecting two data types from producing one integrated inference.
Four practical patterns cover most survey-led mixed-method designs: rate then explain, select then elaborate, measure change then explain, and segment then explore. A strong instrument uses the pattern only where the qualitative answer can clarify a decision, reveal a mechanism, or challenge an apparent quantitative result.
| Pattern | Structured question | Open follow-up | Best for |
|---|---|---|---|
| Rate → explain | Rate confidence from 1 to 5 | What is the main reason for your rating? | Explaining scores |
| Select → elaborate | Which barrier affects you most? | How does that barrier affect you? | Understanding categories |
| Change → explain | Higher, unchanged, or lower? | What contributed most to the change? | Outcome tracking |
| Segment → explore | Cohort or demographic field | Describe your experience of the service | Understanding subgroup differences |
Use a mixed-method survey when a decision requires both measurement and explanation: what or how much is happening, and why or how it is happening. Useful pairings include satisfaction and its reason, outcome change and its contributors, barrier prevalence and lived experience, training confidence and examples of application, or service quality and a specific improvement suggestion.
Do not add open-ended questions merely for color. If the team will not analyze the responses, connect them to a quantitative finding, or use them in a decision, the questions add respondent burden without creating mixed-method value. A short structured survey is often better when prevalence is the only question; interviews may be better when depth matters more than comparison.
The practical rule is to pair each decision-critical result with an explanation only when the reason could alter interpretation or action. Optional prompts, skip logic, and focused wording reduce fatigue. The broader choice among explanatory, exploratory, and convergent studies belongs on mixed-methods research examples.
A joint display is a table, matrix, or visualization that deliberately places quantitative findings beside related qualitative findings so a team can examine convergence, complementarity, and divergence. A joint display is easier when both forms of evidence share a respondent identifier, but broader mixed-methods studies may integrate different samples through a governed design and explicit inference.
Sopact calls the architecture One Record, One Contact ID: the rating and the open-ended reason are captured on the same form, bound to the same participant, so the score and the story never separate. The three mixed-methods designs — explanatory, exploratory, convergent — and worked studies are on the mixed-methods research examples page, and the analysis on mixed-methods data analysis.
This page is the instrument node: the questionnaire that carries both kinds of question. The broader case for using both methods at all is on the qualitative and quantitative methods page, and the instrument architecture beneath it on survey design.
Start with the decision and research question, identify which results require explanation, write specific open follow-ups, define the join key, and decide the qualitative analysis plan before collection. Keep scales consistent across waves, avoid double-barreled ratings, do not force a narrative where a respondent has nothing to add, and never rely on one generic comment box to explain an entire questionnaire.
Sopact recommends writing a data dictionary and a starter codebook with the instrument. The dictionary fixes field meaning and allowed values; the codebook defines how open responses will be themed while allowing genuinely new themes to be reviewed. The full instrument-construction discipline is covered on survey design.
This six-question workforce instrument measures job-search confidence, identifies the main barrier, and explains change over time without asking respondents to narrate after every item. The rating-and-reason and barrier-and-elaboration pairs are decision-critical; the change question connects the current response to the prior wave on the same Contact ID.
| # | Question | Response |
|---|---|---|
| 1 | How confident are you that you can find suitable employment? | 1, not at all → 5, very confident |
| 2 | What is the main reason for your rating? | Open text |
| 3 | Which barrier affects your job search most? | Transport, childcare, skills, confidence, employer access, other |
| 4 | How does that barrier affect you? | Open text |
| 5 | Compared with three months ago, is your confidence lower, unchanged, or higher? | Five-point change scale |
| 6 | What contributed most to that change? | Open text |
Mixed-method analysis should test whether findings converge, complement one another, or diverge instead of assuming that qualitative evidence merely explains the quantitative result. Convergence means both strands support a similar interpretation. Complementarity means one adds detail or mechanism. Divergence means the findings point in different directions and require investigation.
A satisfaction average of 4.4 out of 5 beside repeated comments about poor access is not a failed study. The contradiction may reflect subgroup differences, social-desirability bias, a vague rating item, a threshold effect, or a problem affecting a smaller population. Sopact's One Record, One Contact ID lets the team inspect which respondents and segments produced each pattern rather than averaging the tension away.
| Quantitative finding | Qualitative finding | Integrated interpretation |
|---|---|---|
| Confidence rose from 2.8 to 3.9 | Interview practice is a frequent theme | Practice may help explain improvement; the design does not prove attribution |
| 18% remain at low confidence | Childcare and transport dominate their comments | Skills support alone does not address structural barriers |
| Satisfaction averages 4.4 | One site reports repeated access problems | The average conceals a site-level divergence that requires follow-up |
A mixed-method survey integrates quantitative and qualitative evidence within a questionnaire; mixed-methods research is the broader study design; a mixed-mode survey uses more than one administration channel, such as web, phone, paper, or in-person collection. Mixed-method and mixed-mode are therefore different concepts even though search results often blend them.
A mixed-methods study may connect surveys, interviews, observations, documents, or behavioral records and may use the same or different samples. Integration can occur at design, collection, analysis, or interpretation. Harvard Catalyst's mixed-methods research resources describe the broader methodological family; this page owns the questionnaire-level instrument.
One persistent participant record is not the definition of mixed methods; it is Sopact's preferred way to preserve survey-level integration across questions, segments, and waves. Read the last column as an operational architecture, not a universal research rule.
| Design choice | Two separate surveys | A mixed-method survey |
|---|---|---|
| Pairing | Ratings and open-ends in different tools | Each rating paired with its explaining question |
| Identity | Reconciled by name at the end | One Contact ID from first response |
| Open-ends | A generic comment box, unread | Themed against a codebook on arrival |
| Output | A bar chart and a pull quote | A joint display: result beside its reason |
| Timing | Merged in the final report | Joined at collection, read as it lands |
Every row collapses to one requirement: the number and the reason have to share a record. That is One Record, One Contact ID — the architecture that turns a rating and an open-end into a joint finding instead of two files someone staples together.
When the quantitative and qualitative sides are joined only at reporting time, the join is a manual merge that ages immediately. Theming the open-ends on arrival means the joint display exists continuously — the score and the reason are linked the moment both are collected. That is the premise of the Loop, Sopact's method for continuous impact intelligence: collect clean at the source, analyze the moment data arrives, improve while you can still act.
The Loop is also what makes a mixed-method finding defensible. The quantitative result and the qualitative theme both trace back to the same respondent, so a claim about why a number moved resolves to the person who said it. That standard has its own chapter in traceability and transparency.
One method, three moves that never stop
Then the cycle runs again, a little sharper each wave. Read the method: the Loop methodology →
The fastest way to feel the difference is to pair one rating with its reason on one record. Each prompt below pastes into Sopact Sense's Assistant, or reasons through with your team; the arrow above each links the Academy walkthrough that shows the expected output and the tips.
Academy walkthrough → Pair decision-critical results with reasons
For this program, design a mixed-method survey where each decision-critical quantitative result is paired with an open-ended question only when the explanation could change interpretation or action: [PASTE PROGRAM + OUTCOMES]. List the structured questions, focused open follow-ups, persistent identifier, and segments. Flag unnecessary open questions and any important result that lacks needed context. Return the instrument.
Academy walkthrough → Theme the reasons on arrival
Theme these open-ended responses against the codebook and join each to the respondent's rating: [PASTE CODEBOOK + RESPONSES with respondent_id + rating]. Return respondent_id, rating, assigned theme, and flag where the theme and the rating disagree. Then show the theme distribution within each rating band.
Academy walkthrough → Define both sides once
Turn this mixed-method instrument into a data dictionary covering both the quantitative and qualitative fields: [PASTE INSTRUMENT]. For each field give the name, definition, answer type, and allowed values or codebook reference. Flag any field two waves would treat differently. Return a table.
Academy walkthrough → Build the codebook for the open side
Draft the codebook the open-ended half of this survey will be themed against: [PASTE OPEN-ENDED QUESTIONS + FRAMEWORK]. Give 6-10 codes with definitions and include/exclude rules, and map each code to the rating it most often explains. Return the codebook.
Each walkthrough is short and practical: what to do, the prompt to run, the output to expect, and the tips that keep it reliable.
A mixed-method survey intentionally collects and integrates quantitative and qualitative evidence in a questionnaire to answer the same research question. Sopact's One Record, One Contact ID architecture keeps decision-critical ratings and explanations connected without treating one-record collection as the universal definition of mixed methods.
Start with the decision and research question, identify which quantitative results require explanation, write specific open follow-ups, define the respondent join key, and plan the qualitative analysis before collection. Sopact pairs decision-critical results with reasons while avoiding unnecessary open-question burden.
A workforce program asks participants to rate their job-search confidence and, on the same form, to explain what changed it; the rating shows how much and the open-end shows why, for the same person. A sequential design might follow a survey with interviews of specific respondents. Sopact's examples keep the score and the reason on one record so the joint display is automatic.
A mixed-method survey is the instrument — one questionnaire carrying both kinds of question; mixed-methods research is the broader design that may combine surveys, interviews, and documents in explanatory, exploratory, or convergent sequence, covered on the mixed-methods research examples page. Sopact treats the survey as the collection node feeding the wider design.
A sequential design collects one type of data, then the other, to build on it: explanatory sequential runs the survey first and interviews to explain the results, while exploratory sequential interviews first and surveys to test at scale. Both need the two waves linked by a persistent identifier. Sopact's One Contact ID keeps the sequence joined at the participant across every stage.
Compute the closed questions, theme the open-ends against a fixed codebook, then join each theme to the rating it explains for the same respondent — producing a joint display rather than a bar chart beside a pull quote. Sopact themes on arrival and keeps both on one record, so the analysis is a query; the deeper analysis workflow is on the mixed-methods data analysis page.
Quant-plus-qual surveys fail when the team collects both data types without planning how they answer the same question or how findings will be integrated. Sopact's One Record, One Contact ID makes respondent-level integration easier, while a governed joint display tests convergence, complementarity, and divergence.
Mixed-method research combines and integrates qualitative and quantitative strands. Multimethod research uses several methods that may remain within one tradition or may be analyzed separately. Sopact treats integration as the defining requirement: the methods must produce a combined inference, not merely coexist.
A mixed-method survey integrates quantitative and qualitative evidence. A mixed-mode survey administers a survey through several channels such as web, phone, paper, or in person. Sopact can support mixed-method evidence even when collection modes vary, but the two terms describe different design choices.
Disagreement is a finding to investigate, not an error to hide. Sopact's joint display identifies the respondents, sites, or segments behind the divergence so teams can test wording, bias, subgroup effects, and contextual explanations before drawing a combined conclusion.
Next: see the three designs worked out on the mixed-methods research examples page, or the case for both methods on the qualitative and quantitative methods page.
