Embedded Files
Mixed Methods Research: More Than “I Used Both Quantitative and Qualitative Data”
Mixed Methods Research: More Than “I Used Both Quantitative and Qualitative Data”
Mixed methods research (MMR) is often presented as a straightforward idea: use quantitative data, use qualitative data, put both in one study, and call it mixed methods. Hah!
Unfortunately, it is not that simple. Oh boy...
A study does not become mixed methods merely because it includes a questionnaire and a few interviews. But wait... you need to ask Why do I need both kinds of data, and what can I learn by integrating them that I could not learn from only one method?
That question should guide the entire study.
I bet that everyone would know, or at least have heard, these two commonly cited introductory texts.
Creswell, J. W., & Creswell, J. D. (2023). Research design: Qualitative, quantitative, and mixed methods approaches (6th ed.). SAGE.
Creswell, J. W., & Plano Clark, V. L. (2018). Designing and conducting mixed methods research (3rd ed.). SAGE.
Creswell’s work is very useful, especially for beginners, because it gives researchers accessible design labels. I know all postgrads in language education know Creswell well. However, those labels should be treated as starting points rather than fixed templates that every study must follow.
Creswell is useful, but not the whole field~
Creswell is useful, but not the whole field~
Creswell is widely cited because his work is accessible and practical. However, mixed methods research is much broader than a list of six designs.
Other scholars have developed important perspectives, for instance:
Greene focuses on the purposes of mixing methods, including triangulation, complementarity, development, initiation, and expansion.
Tashakkori and Teddlie view mixed methods as a broader methodological movement rather than a small number of fixed design categories.
Mertens develops transformative mixed methods for research involving social justice, cultural responsiveness, and equity.
This matters because MMR is evolving. Researchers should not rely only on older design diagrams or memorised labels. They should read recent methodological discussions and examine how integration is actually carried out in published studies.
Mixed Methods Research in Language Education: Please Do Not Let the Design Label Do All the Thinking
Mixed methods research is popular in language education for obvious reasons. Our field deals with complex phenomena: language learning, anxiety, motivation, participation, classroom interaction, feedback, identity, multilingualism, assessment, teacher cognition, and increasingly, AI-mediated learning.
It is understandable that researchers want more than one kind of evidence.
A questionnaire may show a broad pattern. Interviews may provide personal accounts. Classroom observations may reveal what actually happens. Writing drafts, chat logs, screen recordings, or interaction transcripts may show processes that participants themselves do not fully notice or remember.
So far, so good.
The problem begins when mixed methods research becomes reduced to a familiar methodological recipe:
Collect quantitative data.
Conduct some interviews.
Report both sets of findings.
Call the study “mixed methods.
That is not necessarily wrong. But, oh come on, it is often not enough!
The issue is not Creswell. I won't blame him, but hey, the issue is methodological autopilot.
Creswell’s designs, such as convergent, explanatory sequential, and exploratory sequential, are useful starting points. They help novice researchers understand that quantitative and qualitative strands can be arranged in different ways.
However, they should be treated as heuristics, not as permanent containers into which every study must fit.
A frequent problem in language education research is that the design label becomes more important than the research problem itself. Researchers may spend considerable effort deciding whether their study is “explanatory sequential” or “convergent,” but much less effort explaining why both datasets are genuinely necessary.
For example, a study may begin with a survey of foreign language anxiety and then follow up with interviews. That can be a sound design. But the important questions are not simply: Is this an explanatory sequential study?
The more important questions are:
What exactly do the survey results fail to explain?
Why are interviews the appropriate next source of evidence?
Will the interviews merely provide illustrative quotations, or will they challenge the interpretation of the quantitative findings?
What will become visible only when the two datasets are brought together?
Without clear answers to these questions, the study may end up as two small projects placed side by side.
“Survey first, interview later” is not always a sufficient logic!
“Survey first, interview later” is not always a sufficient logic!
One of the most common mixed methods patterns in language education is:
Administer a questionnaire.
Identify a few interesting results.
Interview participants.
Use quotations to explain the numbers.
Again, this is not automatically weak. But it becomes weak when the interview phase is treated as a decorative add-on.
For instance, imagine that a questionnaire shows that students report low willingness to communicate in English. A traditional mixed methods study may interview several students and conclude that they are afraid of making mistakes.
That is plausible... but it may also be too simple.
What if classroom observation shows that the same students speak actively in peer discussions but remain silent during teacher-led whole-class interaction? What if they communicate confidently in Thai-English mixed interaction but avoid speaking when they believe “English only” is expected? What if their silence is less about confidence and more about classroom hierarchy, task design, peer judgement, or fear of public evaluation?
At that point, the qualitative evidence is not merely “explaining” the survey result. It is changing the meaning of the result.
Perhaps the issue is not low willingness to communicate in general. Perhaps it is context-sensitive participation shaped by interactional conditions.
That is the kind of integration mixed methods research should aim for.
Language-related phenomena are rarely linear or stable!
Language-related phenomena are rarely linear or stable!
Traditional mixed methods diagrams often suggest clean sequences:
QUAN -> QUAL
QUAL -> QUAN
QUAN + QUAL
Real language learning rarely behaves so neatly.
Anxiety may rise before a presentation, disappear during peer work, return after corrective feedback, and remain low in online chat. Motivation may depend on the teacher, the task, the language being used, the classmates present, or the perceived value of the activity. Engagement may fluctuate across days, lessons, platforms, and assessment events.
For this reason, some language education questions require more iterative or adaptive designs.
A study may begin with repeated surveys, then move to learner diaries because unexpected fluctuations appear. The diary data may lead the researcher to observe particular classroom events. Those observations may then inform stimulated-recall interviews. The interviews may reveal that an apparently “individual” variable is actually tied to task structure, teacher feedback, peer positioning, or multilingual norms.
That kind of movement may look messier than a textbook diagram. But it may be methodologically more honest.
Integration should happen throughout the study, not only in the discussion section.
A major weakness in many mixed methods studies is that integration occurs only at the end.
The quantitative results are presented in one section. The qualitative themes are presented in another. Then the discussion says something like “The qualitative findings supported the quantitative results.”
That sentence is often doing far too much work.
Integration should be visible from the beginning of the study. It can occur through
developing survey items from earlier qualitative findings;
using quantitative patterns to select contrasting cases for interviews;
using interview findings to interpret surprising statistical results;
comparing questionnaire responses with classroom interaction or digital learning traces;
creating joint displays that place statistical patterns beside themes, excerpts, episodes, or learner trajectories;
revising the interpretation of one dataset in light of another.
The point is not simply to make the findings look more comprehensive. The point is to produce an interpretation that neither dataset could generate alone.
Contradiction is not failure~
Contradiction is not failure~
Older mixed methods thinking often treats convergence as the ideal outcome. Researchers seem relieved when quantitative and qualitative findings “agree.”
But disagreement can be far more interesting. Don't you think?
Suppose students report high confidence in a questionnaire, but video-recorded classroom interaction shows that they rarely initiate turns in English. Or suppose teachers report strong support for learner autonomy, but classroom observations reveal that most instructional decisions remain teacher-controlled.
These discrepancies should not be hidden or quickly dismissed as limitations. They may reveal problems in the construct, the instrument, the social desirability of self-report, or the gap between beliefs and practices.
In other words, mixed methods should not only confirm what researchers already suspect. It should also create productive methodological discomfort.
The design should follow the problem, not the research typology alone.
There is nothing wrong with using a convergent, explanatory sequential, exploratory sequential, embedded, transformative, or multiphase design. The problem arises when researchers choose a design because it is familiar, easy to label, or commonly accepted by supervisors and examiners. A stronger approach begins elsewhere; that is, you should think about...
What makes your problem complex?
What kinds of evidence are needed?
What will each form of evidence allow you to see?
How might the findings complicate one another?
What will count as a genuinely integrated conclusion?
For language education researchers, this may mean moving beyond the usual questionnaire-plus-interview model. Depending on the research problem, relevant evidence may include classroom interaction, learner corpora, drafts and revisions, platform data, AI feedback logs, multilingual exchanges, teacher feedback episodes, learner diaries, video data, and repeated measures over time.
Mixed methods research does not become stronger because it uses more tools. It becomes stronger when the methods are connected by a defensible logic.
Final thought~~
Final thought~~
Traditional mixed methods designs are not useless at all. They remain valuable for teaching, planning, and communicating methodological choices.
But they should not become methodological fossils, please.
In language education, we need designs that are responsive to complexity, multilingual realities, changing learner experiences, digital environments, unequal classroom relations, and the fact that language is not simply a score or a self-report variable.
Use design labels when they help, but do not let a diagram from a methods textbook decide what your research problem is allowed to be.
References
Bazeley, P. (2018). Integrating analyses in mixed methods research. SAGE Publications.
Creswell, J. W., & Plano Clark, V. L. (2023). Revisiting mixed methods research designs twenty years later. In C. N. Poth (Ed.), The SAGE handbook of mixed methods research design (pp. 21–36). SAGE Publications. https://doi.org/10.4135/9781529614572.n6
Fetters, M. D., Curry, L. A., & Creswell, J. W. (2013). Achieving integration in mixed methods designs—Principles and practices. Health Services Research, 48(6 Pt. 2), 2134–2156. https://doi.org/10.1111/1475-6773.12117
Fetters, M. D., & Freshwater, D. (2015). The 1 + 1 = 3 integration challenge. Journal of Mixed Methods Research, 9(2), 115–117. https://doi.org/10.1177/1558689815581222
Greene, J. C. (2008). Is mixed methods social inquiry a distinctive methodology? Journal of Mixed Methods Research, 2(1), 7–22. https://doi.org/10.1177/1558689807309969
Greene, J. C., Caracelli, V. J., & Graham, W. F. (1989). Toward a conceptual framework for mixed-method evaluation designs. Educational Evaluation and Policy Analysis, 11(3), 255–274. https://doi.org/10.2307/1163620
Guetterman, T. C., Fetters, M. D., & Creswell, J. W. (2015). Integrating quantitative and qualitative results in health science mixed methods research through joint displays. Annals of Family Medicine, 13(6), 554–561. https://doi.org/10.1370/afm.1865
Riazi, A. M. (2016). Innovative mixed-methods research: Moving beyond design technicalities to epistemological and methodological realizations. Applied Linguistics, 37(1), 33–49. https://doi.org/10.1093/applin/amv064
Riazi, A. M., & Amini Farsani, M. (2024). Mixed-methods research in applied linguistics: Charting the progress through the second decade of the twenty-first century. Language Teaching, 57(2), 143–182. https://doi.org/10.1017/S0261444823000332
Riazi, A. M., & Candlin, C. N. (2014). Mixed-methods research in language teaching and learning: Opportunities, issues and challenges. Language Teaching, 47(2), 135–173. https://doi.org/10.1017/S0261444813000505
Page updated
Google Sites
Report abuse