Mixed Methods Research: Definition, Designs and Examples
A study can contain interviews and a survey and still fail to be genuinely mixed. The defining feature of mixed methods research is not the mere presence of qualitative and quantitative data. It is the deliberate integration of the two strands so that the combined interpretation answers the research problem more fully than either strand could on its own.
That distinction matters for students because mixed methods projects are easy to over-design. Collecting two kinds of data doubles some practical demands, but the value comes only when each strand has a clear purpose and the connection between them is planned from the beginning. This guide explains the main designs, integration strategies, sampling and analysis logic, joint displays, common mistakes, and worked examples.
What Is Mixed Methods Research?
Mixed methods research is an approach that collects and analyzes both quantitative and qualitative data and intentionally brings them together within one study or coordinated program of inquiry. Quantitative evidence can show patterns, frequencies, differences, relationships, or effects. Qualitative evidence can explain experience, meaning, context, process, or mechanism. Mixing them allows the researcher to ask what the numbers show and also what helps explain, extend, challenge, or contextualize those numbers.
NIH Office of Behavioral and Social Sciences Research (OBSSR) describes mixed methods as an approach that uses rigorous quantitative and qualitative procedures and systematically integrates different forms of evidence to address research questions that benefit from real-life contextual understanding. The integration is therefore methodological, not cosmetic.
| Core rule
If the qualitative and quantitative parts could be removed from each other without changing the study’s main conclusion, the project may be multi-method rather than truly mixed methods. A strong mixed design shows exactly where one strand informs, explains, extends, tests, or is merged with the other. |
Mixed Methods vs Multi-Method Research
The terms are sometimes used loosely, but they are not always interchangeable. A multi-method study uses more than one method. Those methods may all be qualitative, all quantitative, or drawn from both traditions. Mixed methods normally implies an explicit qualitative-quantitative combination plus integration.
| Approach | What it includes | What makes it distinctive |
| Single-method study | One principal method and one analytic tradition. | Depth and coherence within one approach. |
| Multi-method study | Two or more methods, which may belong to the same methodological tradition. | Multiple forms of evidence, but integration across qualitative and quantitative traditions is not required. |
| Mixed methods study | At least one meaningful qualitative strand and one quantitative strand. | A planned point of integration that produces a combined interpretation or meta-inference. |
For separate foundations, use the completed EssayEco guides to qualitative research and quantitative research. The broader research design guide explains where mixed methods sits alongside those two approaches.
When Is a Mixed Methods Approach Appropriate?
Mixed methods is most useful when the research problem contains more than one kind of uncertainty. A quantitative phase may identify a pattern but leave the mechanism unclear. A qualitative phase may uncover important concepts but leave their prevalence or distribution unknown. A program evaluation may need outcome measures and stakeholder explanations at the same time.
| Research need | Why mixing helps | Example |
| Explain an unexpected numerical result | Qualitative follow-up can explore why the pattern occurred. | A survey finds unexpectedly low use of a student support service; interviews explore barriers and perceptions. |
| Develop a measure or intervention from lived experience | Qualitative findings can generate concepts, language, or items for later quantitative testing. | Interviews identify dimensions of caregiver burden that are used to build and pilot a questionnaire. |
| Compare measured outcomes with participant experience | Parallel strands can reveal agreement, contradiction, or complementary insight. | Test scores improve after tutoring, while focus groups show which features students believe drove the change. |
| Evaluate implementation as well as effectiveness | Numbers estimate reach or outcomes; qualitative evidence explains context, acceptability, and implementation. | A health program tracks attendance and outcomes while interviewing staff and participants about delivery barriers. |
| Study a complex system at different levels | Different data types can capture structures, behavior, outcomes, and meaning. | Administrative data show turnover patterns while interviews explore workplace culture and decision making. |
Do not choose mixed methods because it appears more comprehensive by default. If one well-designed approach can answer the question, adding a second strand may create unnecessary cost, participant burden, analysis work, and ethical complexity.
The Three Core Mixed Methods Designs
Most student projects can be understood through three basic designs: convergent, explanatory sequential, and exploratory sequential. More advanced projects may embed one strand within another or combine designs across multiple stages. The design name should follow the actual logic of the project, not be added after data collection.
| Design | Sequence | Main integration logic | Best fit |
| Convergent | QUAL + QUAN collected in roughly the same phase. | Analyze separately, then merge or compare findings. | You need complementary perspectives on the same problem at about the same time. |
| Explanatory sequential | QUAN -> QUAL. | Use quantitative findings to select, shape, or focus qualitative follow-up. | You need to explain a numerical result, subgroup pattern, outlier, or unexpected outcome. |
| Exploratory sequential | QUAL -> QUAN. | Use qualitative findings to build a later quantitative measure, instrument, categories, or test. | Concepts are not yet well defined, or context-specific language should inform measurement. |
| Embedded | One strand sits inside a larger primary design. | The secondary strand answers a supporting question within the main study. | An experiment or evaluation needs process, implementation, or participant-experience data. |
| Multiphase / multistage | Several linked phases over time. | Different designs and methods connect across a larger research program. | Large evaluations, implementation programs, or iterative development projects. |
Fetters, Curry, and Creswell describe exploratory sequential, explanatory sequential, and convergent designs as three basic mixed methods designs and explain how integration can occur through connecting, building, merging, and embedding.
1. Convergent Design
In a convergent design, qualitative and quantitative data are collected during the same broad stage, often from the same population or closely related samples. Each strand is analyzed using its own appropriate procedures. The findings are then compared or merged to identify convergence, complementarity, expansion, or disagreement.
A convergent design is attractive when timing is limited because both strands can proceed in parallel. The challenge is integration: the researcher must decide in advance which constructs or questions are comparable across the two datasets and what to do when they do not agree.
2. Explanatory Sequential Design
An explanatory sequential study starts with quantitative data. The researcher analyzes the numeric results first, then uses those results to design a qualitative follow-up. Participants may be selected because they represent important subgroups, extreme cases, unexpected results, or contrasting outcomes. Interview questions can then probe why those patterns occurred.
The second phase should be visibly connected to the first. If the interviews are generic and could have been written before seeing the quantitative results, the explanatory logic is weak.
3. Exploratory Sequential Design
An exploratory sequential study starts with qualitative exploration. Themes, categories, participant language, or a conceptual model from the first phase are used to build the quantitative phase. The researcher may develop survey items, define variables, create response categories, adapt an intervention, or identify constructs that can later be measured across a larger sample.
This design is especially useful when existing instruments do not fit the population or context well. It also requires careful documentation of how qualitative findings were transformed into the later quantitative component.
Where Does Integration Actually Happen?
Integration can occur at more than one point. Planning integration early is one of the clearest differences between a coherent mixed methods study and two disconnected mini-studies.
| Integration strategy | What happens | Simple example |
| Connecting | Results or participants from one strand determine sampling for the next strand. | Survey scores identify high- and low-engagement students for follow-up interviews. |
| Building | Findings from one strand shape instruments, questions, variables, or procedures in the next strand. | Interview themes become survey items. |
| Merging | The two sets of results are directly compared or combined during interpretation. | Attendance rates are compared with interview themes about access barriers. |
| Embedding | A secondary dataset is nested inside a larger primary design. | Qualitative interviews are embedded within a randomized trial to study acceptability and implementation. |
| Joint display | Qualitative and quantitative findings are placed together in a table or figure to generate integrated interpretation. | A table aligns survey percentages, interview themes, and a final integrative conclusion for each construct. |
NIH/PMC integration guidance emphasizes that integration can occur at the design, methods, analysis, interpretation, and reporting levels. A mixed methods proposal should therefore identify not only what data will be collected but also when and how the strands will meet.
Priority and Timing in Mixed Methods
Not every mixed methods study gives equal weight to both strands. One component may be primary and the other supportive. Priority should follow the research purpose. If the main question is about measured outcomes and interviews are included to explain implementation, the quantitative strand may carry greater priority. If the study is developing a culturally specific instrument from participant experience, the qualitative phase may initially carry more conceptual weight.
Some textbooks use notation such as QUAN, QUAL, quan, and qual to show priority, with uppercase letters indicating greater emphasis and arrows showing sequence. Students may use that notation if their course teaches it, but the written explanation matters more: state which strand is primary, whether phases are sequential or concurrent, and where integration occurs.
Sampling in Mixed Methods Research
Mixed methods projects often use different sampling logics in the two strands. The quantitative component may need a larger, structured sample to estimate patterns or compare groups. The qualitative component may use purposive sampling to obtain information-rich cases. The challenge is not to force both components into one sampling philosophy, but to explain how the samples relate to the overall design.
| Sampling relationship | How it works | Example |
| Identical / nested | The qualitative sample is drawn from the quantitative sample or vice versa. | Twenty interviewees are purposively selected from 300 survey respondents. |
| Parallel | Separate samples come from the same broader population or setting. | A staff survey and separate manager interviews examine the same implementation problem. |
| Multilevel | Different strands sample different levels of a system. | Student surveys are combined with lecturer interviews and institutional records. |
| Sequentially connected | Results from phase one determine who or what should be sampled in phase two. | Outliers in a quantitative dataset are selected for follow-up qualitative interviews. |
The planned EssayEco guide on sampling methods in research will cover probability, non-probability, purposive, criterion, stratified, cluster, and other sampling strategies in detail.
Data Collection in Mixed Methods Studies
The methods themselves are familiar: surveys, tests, administrative datasets, experiments, interviews, focus groups, observations, documents, diaries, or digital traces. What changes in mixed methods is the relationship between them. Each method should have a reason to exist and a connection to the other strand.
| Quantitative component | Possible qualitative partner | Integration purpose |
| Survey | Interviews or focus groups | Explain patterns, explore subgroup differences, or interpret unexpected answers. |
| Experiment / intervention outcome measures | Interviews, observation, implementation logs | Understand acceptability, fidelity, mechanism, context, or participant experience. |
| Administrative or clinical records | Interviews or document analysis | Connect measured outcomes with organizational processes or stakeholder explanations. |
| Quantitative content coding | Open-ended responses or qualitative document analysis | Compare prevalence of coded features with deeper interpretation of meaning and context. |
The next planned cluster on research methods: surveys, interviews, focus groups and observation will compare these individual techniques more directly.
How to Conduct a Mixed Methods Study Step by Step
Step 1: Start With a Research Problem That Needs More Than One Kind of Evidence
Write the problem and question before choosing the design. Ask what one method would leave unresolved. If the answer is nothing important, mixed methods may not be justified. A strong rationale names the specific limitation of a single approach rather than claiming vaguely that two methods are better than one.
Step 2: Define the Qualitative and Quantitative Questions
Mixed studies often include an overarching mixed question plus strand-specific questions. For example, a quantitative question may ask whether engagement differs by study mode; a qualitative question may ask how students explain barriers to engagement; the mixed question may ask how the experiences help explain the measured differences.
Use the completed guide on how to write a research question to keep each question clear, focused, researchable, and aligned with the evidence you can actually collect.
Step 3: Justify Why Integration Is Necessary
State what the combined evidence is expected to accomplish: explanation, development, comparison, expansion, triangulation, implementation insight, or theory building. This rationale should be tied to the research problem, not to a preference for methodological variety.
Step 4: Choose the Design and Sequence
Decide whether the project is convergent, explanatory sequential, exploratory sequential, embedded, or genuinely multiphase. Record which strand comes first, whether they overlap, and whether one has priority. The design should make the integration point obvious.
Step 5: Plan Sampling for Each Strand
Choose the sample for each component using the logic appropriate to that method, then explain how the samples connect. If phase-two interviews depend on phase-one survey results, define the selection rule before collecting the second phase.
Step 6: Collect Each Dataset Rigorously
Mixed methods does not excuse weak component methods. The quantitative strand still needs defensible measurement, data quality, and appropriate sampling. The qualitative strand still needs thoughtful interviewing or observation, reflexive documentation, and a systematic analytic process.
| Two-method fallacy
A weak survey plus weak interviews do not become a strong study simply because they are combined. Each strand must meet the quality expectations of its own methodological tradition before integration can add value. |
Step 7: Analyze the Strands Using Appropriate Procedures
Quantitative data may be summarized with descriptive statistics and, when appropriate, inferential analyses. Qualitative data may be coded and analyzed thematically, narratively, through content analysis, or with another design-consistent approach. Initial strand-specific analysis is often necessary before meaningful integration.
Step 8: Carry Out the Planned Integration
Use the design logic you selected earlier. Connect participants, build one phase from the other, merge findings, embed one dataset within another, or create a joint display. Do not postpone the decision about integration until the discussion section.
Step 9: Examine Agreement, Complementarity, and Disagreement
Integrated findings can converge, complement one another, expand the picture, or conflict. Disagreement is not automatically a failure. It may reveal subgroup differences, measurement problems, timing effects, social desirability, different constructs, or a genuinely complex phenomenon. The researcher should investigate the mismatch rather than hide it.
Step 10: Develop Meta-Inferences
A meta-inference is the conclusion drawn from considering the strands together. It should be more informative than placing a quantitative conclusion next to a qualitative conclusion. Ask what becomes visible only after the datasets are connected.
Step 11: Report the Integration Transparently
Explain the rationale, design, priority, timing, sampling, data collection, strand-specific analyses, integration strategy, and combined interpretation. Readers should be able to trace how the qualitative and quantitative components influenced each other.
Step 12: Evaluate Whether Mixing Actually Added Value
In the final revision, ask a difficult question: what did the mixed design reveal that one strong method would not? If the answer is unclear, strengthen the integrated analysis rather than merely adding more description of each strand.
Joint Displays: A Practical Tool for Integration
A joint display is a table, matrix, figure, or other visual structure that places qualitative and quantitative findings in relation to one another. It is useful both as an analytic tool and as a reporting device because it forces the researcher to specify how the strands connect.
Guetterman, Fetters, and Creswell show how joint displays can support convergent, explanatory sequential, exploratory sequential, and intervention designs. Their examples include side-by-side comparisons, statistics-by-themes displays, and instrument-development displays.
| Construct | Quantitative finding | Qualitative finding | Integrated interpretation |
| Access | Only 42% of eligible students used tutoring. | Students described schedule conflicts, commuting time, and uncertainty about eligibility. | Low use appears related not only to interest but also to practical access and unclear communication. |
| Satisfaction | Users averaged 4.3/5 satisfaction. | Students valued one-to-one feedback but disliked limited appointment availability. | High overall satisfaction coexists with a capacity problem that may restrict future reach. |
| Outcome | Users gained 6 points more on average than non-users after adjustment. | Students described clearer study routines and accountability after repeated sessions. | The outcome difference is consistent with a plausible behavioral pathway, but the qualitative data do not by themselves establish causality. |
The point of the display is not decoration. It should support a new inference, identify convergence or conflict, or make the mechanism of integration visible.
How to Analyze Mixed Methods Data
Mixed methods analysis has two levels. First, each dataset is analyzed appropriately. Second, the researcher performs an integration analysis: comparing, linking, or transforming findings to answer the mixed question.
| Integration analysis | What the researcher does | Possible output |
| Side-by-side comparison | Organizes quantitative results next to qualitative themes addressing the same construct. | Narrative comparison or joint display. |
| Data transformation | Converts qualitative codes to counts/variables or translates quantitative patterns into qualitative categories for further comparison. | Merged dataset, frequency table, or interpretive categories. |
| Case-based integration | Combines multiple data types for the same participants or cases. | Integrated case profiles or cross-case matrix. |
| Follow-up explanation | Uses qualitative data to interpret selected quantitative findings. | Explanatory themes linked to specific numeric results. |
| Instrument development | Maps qualitative concepts to proposed items or scales and then tests them quantitatively. | Item-development matrix and pilot results. |
Recent mixed methods analysis guidance describes integration analysis as moving beyond separate qualitative and quantitative analyses to locate the point of mixing, use tools such as joint displays, and draw integrated conclusions.
Rigor and Quality in Mixed Methods Research
Quality has at least three layers: the quantitative component must be credible on quantitative terms, the qualitative component must be credible on qualitative terms, and the integration must itself be coherent and transparent. A flaw in one layer can weaken the whole design.
| Quality question | What to check |
| Rationale | Is there a clear reason why one method alone is insufficient? |
| Design coherence | Do sequence, priority, questions, methods, and integration fit together? |
| Quantitative rigor | Are variables, measurement, sampling, data quality, and analysis defensible? |
| Qualitative rigor | Are sampling, data collection, coding/analysis, reflexivity, and evidence trails credible? |
| Integration rigor | Is there a specific integration procedure rather than two separate results sections? |
| Interpretive fit | Does the final meta-inference represent agreement, complementarity, or disagreement honestly? |
| Transparency | Can a reader trace how one strand informed or interacted with the other? |
NIH OBSSR best-practices guidance emphasizes rigor in both components, a clear rationale for mixing, appropriate design, and explicit integration. Those principles apply beyond health research even though the NIH resource was developed for health-science investigators.
Reporting Mixed Methods Research
A mixed methods report should make the design visible early. Name the design, explain why it fits the question, show the order and priority of the strands, describe each method sufficiently, and identify the exact point or points of integration. Results can be organized by phase, by research question, by theme/construct, or through an integrated structure, depending on the design.
The EQUATOR Network GRAMMS resource highlights six foundational reporting concerns: justify the mixed approach; describe purpose, priority, and sequence; report each method; explain where and how integration occurred; consider limitations created by combining methods; and state the insights gained through mixing.
As of 2026, an updated GRAMMS 2.0 guideline is under development, so students and researchers should also follow current journal, discipline, supervisor, or institutional reporting requirements rather than treating the original checklist as the only standard.
Advantages of Mixed Methods Research
- It can connect numerical patterns with participant meaning, process, and context.
- It can explain unexpected or heterogeneous quantitative results.
- It can develop context-sensitive measures, interventions, or categories from qualitative evidence.
- It can compare different forms of evidence and reveal convergence or contradiction.
- It is useful for complex program evaluation, implementation, education, health, organizational, and policy questions.
- It can produce more actionable explanations by linking outcomes to mechanisms and stakeholder perspectives.
Limitations and Challenges
- It requires competence in two methodological traditions plus integration, not just two data-collection techniques.
- Projects can become too large for the available time, budget, participant access, or student skill level.
- Sequential designs can extend timelines because phase two cannot begin until phase one is analyzed sufficiently.
- Conflicting findings can be difficult to interpret and may expose weaknesses in measurement or sampling.
- Teams may privilege one strand so heavily that the other becomes tokenistic.
- Poor integration can leave the project as two parallel studies with no added mixed-methods value.
- Additional datasets can increase consent, privacy, data-management, and participant-burden considerations.
Mixed Methods Research Examples
| Field / problem | Possible design | How the strands integrate |
| Education: Why do online students with similar grades show different persistence rates? | Explanatory sequential | Analyze persistence and performance data, then interview selected persisters and non-persisters to explain contrasting pathways. |
| Nursing: How effective and acceptable is a new discharge-education program? | Convergent or embedded | Compare readmission/knowledge outcomes with interviews about understanding, burden, and usability. |
| Business: What drives adoption of a new employee platform? | Explanatory sequential | Survey adoption predictors, then interview high- and low-adoption groups to explain organizational and workflow barriers. |
| Psychology: How should a context-specific scale of academic belonging be developed? | Exploratory sequential | Use interviews to identify constructs and language, build survey items, then pilot and assess the instrument quantitatively. |
| Public health: Why does an intervention work better in some clinics than others? | Embedded / multilevel | Combine outcome data with staff interviews, observations, and implementation records to study context and fidelity. |
| Policy: How does a new regulation affect service use and stakeholder experience? | Convergent | Analyze administrative trends while interviewing providers and service users; merge findings by policy mechanism. |
Common Mixed Methods Mistakes
- Calling a study mixed methods merely because it contains a few open-ended survey questions.
- Writing two unrelated research questions with no overarching rationale for integration.
- Choosing a design label after data collection rather than planning the sequence and integration in advance.
- Collecting too much data for the available time and then integrating only superficially.
- Using qualitative interviews only as illustrative quotes after the quantitative analysis, without systematic analysis.
- Forcing both strands to use the same sample size or sampling logic when their purposes differ.
- Treating agreement between datasets as the only successful outcome and ignoring meaningful contradiction.
- Using triangulation as a synonym for all mixed methods rather than specifying the actual integration procedure.
- Reporting qualitative and quantitative results separately but never drawing a combined meta-inference.
Mixed Methods Research Checklist
| Check | Question to ask |
| Problem | Does the research problem genuinely require both qualitative and quantitative evidence? |
| Rationale | Can I state exactly what mixing is expected to explain, develop, compare, or extend? |
| Questions | Do the strand-specific questions and the overarching mixed question align? |
| Design | Is the study convergent, explanatory sequential, exploratory sequential, embedded, or multiphase for a defensible reason? |
| Priority and timing | Have I stated which strand comes first, whether they overlap, and whether one has greater emphasis? |
| Sampling | Is each sample appropriate to its strand, and is the relationship between samples clear? |
| Rigor | Can each component stand up to the quality expectations of its own method? |
| Integration | Have I identified where and how the strands connect, build, merge, embed, or appear in a joint display? |
| Interpretation | Does the final conclusion explain agreement, complementarity, and disagreement honestly? |
| Added value | Can I say what the mixed design revealed that one method alone would probably have missed? |
Final Takeaway
The value of mixed methods research does not come from having more data. It comes from designing a meaningful relationship between different kinds of evidence. A strong mixed study begins with a problem that genuinely needs both numerical and contextual understanding, chooses a design that makes the sequence and priority clear, conducts each strand rigorously, and integrates the findings through a planned procedure.
When integration is explicit, the researcher can move beyond two separate conclusions and develop a stronger combined explanation. When integration is vague, the project becomes larger without necessarily becoming better. The safest rule is simple: use mixed methods only when the connection between qualitative and quantitative evidence is essential to the answer.
Frequently Asked Questions
Is mixed methods research just qualitative plus quantitative research?
No. Both strands must be present, but the defining feature is intentional integration. Two datasets analyzed and reported independently do not automatically create a mixed methods study.
Which mixed methods design is easiest for students?
There is no universally easiest design. A small convergent study can be manageable when both datasets can be collected at the same time, while an explanatory sequential design can be conceptually straightforward because the quantitative result clearly guides the interview follow-up. The best design is the smallest design that answers the actual question within the available time and access.
Can mixed methods research use the same participants for both strands?
Yes. Some projects collect both forms of data from the same participants, while others use a qualitative subsample, parallel samples, or samples from different levels of a system. The relationship between samples should follow the design and be explained explicitly.
Does one strand have to be more important?
No. Some studies give both strands similar priority, while others treat one as primary and the second as supportive. What matters is that the priority reflects the research purpose and is transparent in the design and interpretation.
What is a joint display in mixed methods research?
A joint display is a table or figure that brings qualitative and quantitative findings together so the researcher can compare them and develop integrated conclusions. It can be used during analysis as well as in the final report.
What if the qualitative and quantitative findings disagree?
Investigate the disagreement rather than removing it. Different samples, timing, constructs, measurement limitations, context, social desirability, or genuine complexity may produce divergence. A well-reasoned explanation of disagreement can be one of the most valuable results of a mixed study.
Do all research papers need mixed methods?
No. Many strong research projects are purely qualitative or purely quantitative. Mixed methods is justified only when integrating both forms of evidence materially improves the answer to the research problem.
