9. Mixed Methods: Integrating Different Forms of Evidence
9.1 Why combine methods?
Mixed-methods research intentionally combines qualitative and quantitative evidence within one study or programme of inquiry. It is useful when one type of evidence cannot adequately answer the question. Surveys may estimate how common a problem is; interviews may explain how it is experienced. Integration is the defining feature: two separate studies placed in the same report do not become mixed methods unless their design, analysis, or interpretation is meaningfully connected (Creswell & Creswell, 2023).
9.2 Common mixed-methods designs
An explanatory sequential design begins with quantitative results and follows with qualitative inquiry to explain patterns. An exploratory sequential design begins qualitatively and uses findings to develop a measure or larger survey. A convergent design collects qualitative and quantitative evidence in a similar phase and compares or integrates results. Embedded designs place one method within a larger design, such as interviews alongside a programme evaluation. The design label should reflect the actual sequence and purpose.
9.3 Integration and divergent findings
Integration may occur through connecting samples, building one phase from another, merging datasets, or embedding one method in the other. A joint display can compare quantitative results with qualitative themes. If survey respondents report high satisfaction but interviews describe serious access barriers, the difference should be investigated rather than averaged away. Divergence may reveal subgroup differences, question wording problems, social desirability, or different dimensions of the experience.
9.4 Cost, time, and complexity
Mixed methods often require more time, expertise, and coordination than a single-method design. The researcher needs a clear rationale for each strand, an integration plan, and sufficient competence to analyse both forms of data. Where the question can be answered credibly with one method, adding another may create unnecessary burden. Participants should not be asked to provide extensive data simply because the researcher wants a more elaborate project.
9.5 Example: digital learning and unequal access
A study might first survey learners about device access, connectivity, participation, and perceived learning support. The results may show lower participation among learners with unreliable internet access. Follow-up interviews could explore shared-device arrangements, electricity interruptions, data costs, study space, and family responsibilities. The integrated conclusion would combine the distribution of reported barriers with contextual explanations, while avoiding claims that the survey alone proves the causes of learning outcomes.