How to Use Mixed Methods Research
Mixed methods research combines numbers and narratives in one study. Done well, it answers questions neither approach could answer alone; done badly, it is two small studies stapled together.
Many students choose mixed methods because it seems to offer the best of both worlds. Examiners, however, look for a clear reason to combine methods and for genuine integration of the results. A dissertation that runs a survey and some interviews, then reports them in separate chapters without connecting them, is not really mixed methods research. This guide explains when the approach is appropriate, the main designs, and how to integrate findings so the whole is greater than its parts.
Be clear about why you are mixing methods
Mixed methods should be driven by your research question, not by a wish to appear thorough. Good reasons include using qualitative findings to explain surprising quantitative results, using qualitative exploration to develop a survey instrument, comparing what people report in numbers with what they describe in interviews, or measuring an outcome while also understanding the process behind it. State your rationale explicitly; it is the first thing examiners look for.
Stuck on your assignment? Our expert writers can help.
Understand the main designs
Methodology texts, most notably the work of Creswell and Plano Clark, describe three core designs that most student projects follow.
| Design | Sequence | Typical purpose |
|---|---|---|
| Convergent | Quantitative and qualitative collected at the same time | Compare or corroborate two kinds of evidence |
| Explanatory sequential | Quantitative first, then qualitative | Explain or elaborate on numerical results |
| Exploratory sequential | Qualitative first, then quantitative | Explore a topic, then build or test a measure |
Choose the design whose logic matches your question, and name it in your methodology.
Decide on priority and timing
In some studies, both strands carry equal weight; in others, one is dominant and the other supports it. Timing — whether strands run concurrently or one after the other — affects your schedule as much as your analysis. Sequential designs take longer because the second strand depends on the first, which matters for a dissertation with a fixed deadline. Be realistic about what you can complete.
Address the philosophical basis
Quantitative and qualitative research traditionally rest on different assumptions about knowledge. Many mixed methods researchers adopt pragmatism, which prioritises what works to answer the research question over strict allegiance to one paradigm. Briefly explain your position and how it justifies combining methods. You do not need a long philosophical discussion, but ignoring the issue entirely can attract criticism.
Plan sampling for each strand
Each strand needs its own sampling strategy. Your quantitative strand may need a larger, more representative sample; your qualitative strand a smaller, purposive one. In sequential designs, qualitative participants are often selected from the quantitative sample, for example people with extreme or typical scores. Explain how the samples relate to each other and why.
Collect and analyse each strand properly
Mixed methods does not lower the standard for either strand. Your survey still needs a validated instrument and appropriate statistics; your interviews still need a rigorous analytic approach such as thematic analysis. Describe each strand’s methods with the same care you would in a single-method study. A weak strand undermines the whole design.
Integrate the findings
Integration is what makes the study mixed methods rather than multi-method. It can happen at several points: design (one strand shapes the other), data collection (sampling or instruments linked), analysis (data transformed or compared) and interpretation (findings brought together). Common techniques include joint displays — tables placing quantitative and qualitative results side by side — and explicitly discussing where the strands agree, complement or contradict each other.
Joint displays
A table with a row for each key topic, showing the statistical result, illustrative qualitative findings and a column interpreting how they fit together, is one of the clearest ways to show integration.
Handle disagreement between strands
Sometimes the strands point in different directions: survey results suggest satisfaction is high, while interviews reveal frustration. This is not a failure. Discrepancies are often the most interesting findings. Consider explanations such as differences in samples, what each method can capture, or social desirability in survey responses, and discuss them openly.
Structure the write-up
Mixed methods dissertations can report results by strand and then integrate them in a dedicated section, or integrate throughout by topic. Sequential designs often report strands in the order they were conducted. Whatever you choose, make the integration visible: a separate integration section or joint display should never be missing. Follow your supervisor’s guidance on chapter structure.
Evaluate quality
Assess quality for each strand using criteria appropriate to it — validity and reliability for quantitative work, credibility and transferability for qualitative work — and also for the mixed methods study as a whole. Ask whether the design suited the question, whether integration was meaningful and whether conclusions draw on both strands. Acknowledge limitations such as unequal samples or time constraints.
Plan the timeline realistically
Mixed methods projects take longer than single-method ones because you design, collect and analyse two kinds of data and then integrate them. Map each stage on a timeline with your supervisor before you start: ethics approval, instrument development, the first strand, analysis, the second strand and integration. Build in buffer time between sequential strands, because the second cannot begin until the first is analysed. If the timeline looks unworkable, a smaller second strand is usually better than rushing both.
Common mistakes to avoid
- No clear rationale for combining methods.
- Reporting two strands separately with no integration.
- A rigorous quantitative strand and a token qualitative strand, or vice versa.
- Not naming or justifying the design.
- Underestimating the time sequential designs require.
- Ignoring discrepancies between strands.
