What is mixed methods research?
Mixed methods research intentionally combines qualitative and quantitative approaches within a study or program of inquiry. Its defining purpose is integration: using the relationship between the strands to develop an understanding that the separate results would not provide alone.
Integration is not limited to placing a person's score and comment on one database row. It can involve using one phase to shape another, connecting samples, bringing findings together or interpreting how the strands agree and differ. The research question determines the appropriate relationship.
For example, a team might survey member organizations about a reporting process, then interview selected representatives to investigate the findings. Another team might begin with interviews to improve the questions in a later survey. Both need a plan explaining how the phases inform one another.
Why use mixed methods research?
A mixed approach can help when a question spans measurement and experience, when an unfamiliar issue needs exploration before measurement, or when a numerical pattern needs a more focused investigation. It can also reveal differences that a single source would leave hidden.
It is not automatically the strongest choice. Each strand requires suitable collection and analysis, and the team must do the integration work. If one well-designed approach answers the question, adding another may create burden without useful learning.
Before committing, complete this sentence: “We need both because reading these findings together will help us understand…” If the answer is only “to make the report more persuasive,” the design needs more thought.
Three common mixed-methods designs
Exploratory sequential, explanatory sequential and convergent designs are common starting points. Fetters, Curry and Creswell discuss these designs and integration at design, methods and reporting levels in Achieving Integration in Mixed Methods Designs.
Scroll horizontally to see all columns →
| Design | Basic sequence | Example purpose | Planning question |
|---|---|---|---|
| Exploratory sequential | Qualitative exploration informs a later quantitative phase | Understand local terminology before building a survey | How will the exploration change the instrument or measures? |
| Explanatory sequential | Quantitative findings guide a later qualitative phase | Investigate a difference between groups in a survey | Which findings and perspectives need follow-up? |
| Convergent | Related strands are brought together for interpretation | Review implementation measures alongside accounts of experience | Where will the findings meet, and how will differences be examined? |
These descriptions do not prescribe one sample size, one platform or identical participants in every strand. For a fuller design walkthrough, see mixed-method design; for extended illustrations, see mixed-methods research examples.
A study example: understanding a member reporting process
This example is fictional. A network wants to improve its annual return without imposing identical local questionnaires. The central team knows that some returns arrive late, but it does not yet know whether the problem is wording, access, local record availability or something else.
Its study question is: “How do local representatives prepare the annual return, and which parts of the process should we change?”
Plan the quantitative strand
The team defines submission timeliness and the kinds of clarification requested, then reviews the records for the reporting period. It identifies the expected organizations, checks incomplete records and separates a late return from a return that has not arrived.
Suppose the records show that 30 of 40 expected organizations submitted by the deadline. That is 75% on-time coverage for this defined cycle. It describes the recorded submissions; it does not identify the cause of lateness.
Plan the qualitative follow-up
The team selects representatives with varied experiences, including timely and late submissions and different local operating contexts. Interviews explore how they assembled the return and what happened during the process. The selection is explained; the interviews are not treated as a representative poll.
The interview guide leaves room for unanticipated issues. If representatives describe difficulty interpreting a shared measure, the team examines that account rather than assuming every delay is a training problem.
Integrate the findings
The team compares the collection records with the interview interpretation at the appropriate organizational and process level. It asks whether the accounts clarify a pattern, complicate it or address a different question.
A useful result might be a more focused hypothesis: the shared measure needs clearer guidance, while local questions should remain flexible. That interpretation would still need support from the actual accounts and consideration of contrary evidence.
The next step is a bounded test of revised guidance, followed by a review of subsequent experience and collection quality. The study does not prove that guidance alone will solve late reporting.
Write a study plan before collection
Scroll horizontally to see all columns →
| Decision in the plan | What to specify |
|---|---|
| Purpose | The question and why integration is needed |
| Design | The sequence and relationship between strands |
| Sources | What each strand will collect and from whom |
| Selection | How sources or participants are chosen and what gaps may remain |
| Analysis | The quantitative and qualitative approaches, separately |
| Integration | What will be connected, compared or used to inform another phase |
| Governance | Permissions, access, record context, version history and sharing arrangements |
| Reporting | The audience, expected output, limitations and next decision |
Assign responsibility for integration. It should not become a final task left to the person building slides. Plan when the people responsible for each strand will examine the relationship between their findings.
How samples and sources can relate
The same participants may contribute to both strands, a subset may be selected for follow-up, or different sources may contribute at another level. Explain which arrangement you use and why it answers the question.
If you want to examine an individual's change, reliable and appropriate matching matters. If you are studying an organizational process, a manager's interview may provide context for administrative records without representing the experience of every person in those records.
Do not claim a direct link that the data cannot support. Equally, do not dismiss a valid integrated interpretation merely because the sources were stored separately or came from different participants.
Analyze each strand before forcing a combined story
For numerical data, inspect the measures, units, coverage, missingness and assumptions. Report the relevant estimate and uncertainty where appropriate. For qualitative material, use a suitable analytical approach and show how the interpretation follows from the accounts.
Then bring the findings together around the study question. A joint display can help: place the related findings, their relationship and the resulting interpretation in a table. The important work is the reasoning, not the format.
For a practical integration process, see mixed-methods data analysis. For a narrower ratings-and-comments exercise, see qualitative and quantitative analysis.
What to do when findings disagree
Check whether the sources refer to the same period, population and aspect of the issue. A survey rating of the overall service can coexist with a critical account of one episode. Different response patterns may also produce different pictures.
Examine whether one strand reveals a limitation in the other. Perhaps the survey omitted a topic that emerged in interviews, or the interview sample underrepresents a group visible in the records. Preserve the disagreement when it remains unresolved.
Do not choose the result that makes the cleaner story. Explain the tension and what further evidence would help. Disagreement can improve the next question without yielding an immediate conclusion.
Strengths and limitations
A well-designed mixed study can connect measurement with context, improve instruments, investigate unexpected patterns and make differences easier to understand. Those benefits depend on doing both strands well and explaining their relationship.
The challenges include more planning, varied analytical skills, coordination and the risk of collecting material that cannot be meaningfully integrated. One weak strand can limit the combined conclusion. A larger volume of evidence does not automatically resolve bias or uncertainty.
Mixed methods also does not establish causation by default. A change in a measure and a plausible interview account may suggest an explanation without ruling out alternatives. Match the strength of the claim to the design.
Plan for local flexibility and shared evidence
In federated networks, schools or distributed service teams, define the small shared core required for a cross-site question. That may include a unit, reporting period, measure definition and relevant context. Keep local questions where local needs differ.
Document definitions and changes in a data dictionary. Keep stable reference details once where appropriate, and update changing context with dates. A shared field should not conceal different eligibility or collection conditions.
Some findings may belong in separate local views. Aggregate only when the meaning and coverage are compatible, and explain the boundaries. This is a design decision, not something a dashboard can repair after the fact.
How Sopact reduces coding and reporting work
Research design remains a human responsibility. For recurring applied studies, the operational advantage is reducing repeated coding and integration work while keeping the evidence open to review.
A workflow with repeated manual work
- Define from an initial sampleRead material and agree on the codebook.
- Apply it across the datasetCode responses and check the result.
- Revise a definitionReturn to affected material and recode it.
- Reconnect the numbersReconcile coded results with ratings and context, then rebuild the view.
The Sopact workflow
- Your team owns the definitionsDecide what each code means and improve it as you learn.
- Apply coding across the eligible dataAutomate application; people review quality and exceptions.
- Reprocess after a definition changesReapply the revised definition across the configured scope instead of recoding each response by hand.
- Ask across coded text and numbersKeep the response, rating and relevant record context connected; inspect the evidence behind the result.
This compares workflow patterns, not a claim that every research tool requires manual coding or separate files. Some already automate parts of this work; compare the complete cycle.
For this codebook-based workflow, the main saving is repeated application and reconnection—not the removal of human judgment. A changed definition can be reapplied across the configured data while reviewers concentrate on quality, exceptions and interpretation. Coded text stays connected to the relevant ratings and context.
Count the recurring work in ownership cost. Include setup, coding, recoding after revisions, source reconciliation, review and reporting, plus your actual platform and processing expenses. A worked scenario of four cycles of 4,000 responses illustrates 272 fewer annual staff hours; it is an assumption-based example, not a customer benchmark. Existing automation, review needs and implementation effort can substantially change the result.
Adjust the workload assumptions and compare total effort →
A reliable assistant should calculate from the selected records and let a reviewer open the supporting evidence. Check the data scope, definition, denominator and access permissions. Reproducible arithmetic does not make every AI interpretation correct.
Watch: Why Qualitative Analysis Stays Small — And How to Scale It
See why revising a codebook creates repeat work, and how connected coding and quantitative analysis change that workload.
Use tools to support the design
Multiple tools can support a valid mixed-methods study. The operational question is how much repeated preparation and reconciliation the team must manage, and whether sources, definitions and decisions remain accessible.
Evaluate Sopact around the recurring workflow: collection, connected context, reviewed analysis and governed reporting. Test the intended source relationships, changed questions, incomplete records and review process. Verify the features needed for the study rather than assuming integration is automatic.
AI can assist with parts of preparation and draft analysis. It does not choose a defensible research design or remove the need to inspect evidence and interpretation. Keep responsibility for the study with the people conducting it.
What a mixed-methods report should contain
Explain why the study used both approaches, how each strand was conducted and how integration changed the interpretation. Include source selection, coverage, method, limitations and disagreements. Distinguish the findings from proposed actions or future hypotheses.
For broader reporting support, use How to Write an Impact Report and browse report examples.
Watch: mixed methods research
This companion introduces combining qualitative and quantitative evidence. Use the study-plan questions above to adapt the approach to your own context.
Mixed methods research: combining qualitative + quantitative
Frequently asked questions
What makes research mixed methods?
It intentionally combines qualitative and quantitative approaches and integrates them to address a study question. Simply collecting both does not explain how they contribute together.
Does integration require one participant record?
No. A participant record can support some designs, but integration can happen through sequencing, source relationships and interpretation at different levels.
Can I use separate survey and interview tools?
Yes. A valid design can use separate tools. Plan the source relationships, analysis and integration, and account for the maintenance and reconciliation involved.
What are the main mixed-methods designs?
Common starting points are exploratory sequential, explanatory sequential and convergent designs. Choose according to the study question and how the strands will inform one another.
Does mixed methods research prove why an outcome changed?
Not automatically. It can contribute to explanation, but causal claims depend on the design and evidence, not merely the presence of numbers and interviews.
What if the strands disagree?
Examine their populations, periods, questions and methods. Report unresolved differences and what they imply rather than forcing agreement.
Choose your next step
ASME
Explore how a membership organization uses quantitative and qualitative evidence. The proposed research study above is fictional.
Read the customer story →