Mixed Methods Research: Integrating Quantitative and Qualitative Evidence

Establish why integration is needed

Mixed methods research combines quantitative and qualitative approaches within a coherent design. The important question is not whether the project contains numbers and quotations, but what the combination allows the researcher to understand. Explain the added value before committing to the additional workload.

A quantitative component might identify a pattern, while a qualitative component investigates how that pattern develops. Alternatively, qualitative inquiry might help define concepts before a wider measurement exercise. NIH's mixed-methods resources emphasise thoughtful design and the contribution of integration. ^[1] Using two methods without a reasoned connection can increase cost without improving the answer.

Choose the sequence and priority

A sequential design collects one form of evidence before another. For example, survey findings may guide interview recruitment, or interviews may inform questionnaire development. A concurrent design collects components during a similar period and brings them together during interpretation or another planned stage.

Decide whether the components have equal priority or whether one supports the other. Explain the timing, dependency and decision rules. If the second stage depends on first-stage findings, allow time for analysis before recruitment continues. Avoid writing a timeline in which interviews begin before the survey has produced the information needed to select participants.

Design each component rigorously

The quantitative component needs suitable measures, sampling and analysis. The qualitative component needs a coherent question, recruitment rationale and analytic approach. Mixed methods does not compensate for weak execution in either component. Ensure that the team has the required expertise and that responsibilities are explicit.

Think carefully about the relationship between samples. The same people may contribute both forms of evidence, or different samples may address different aspects of the question. Record how participants are linked, how identities are protected and whether consent covers the intended combination of records. A merged dataset can reveal more about a person than either source alone.

Plan concrete points of integration

Integration can occur when one component informs sampling, instrument design, analysis or interpretation of another. Write these points into the protocol. A joint display can place numerical patterns alongside qualitative explanations, relevant context and the interpretation that follows from considering them together.

Do not use quotations merely to decorate a table of results. Ask whether qualitative material explains, challenges or extends a numerical pattern. Equally, do not treat a survey result as a verdict that automatically overrides a detailed account. The meaning of disagreement depends on differences in questions, measures, timing and participants.

Illustrative example: an online course redesign

Suppose course records show that learners frequently stop at a particular module. The records identify where participation drops but cannot establish why. The team interviews learners who completed, paused and withdrew, with appropriate permissions and recruitment that avoids pressure.

Interviews suggest that some learners struggle with unclear instructions while others cannot access a required application. The integrated analysis links these experiences to device access and timing in the records. A redesign can then address distinct barriers. The study should still avoid claiming that either barrier caused every withdrawal, especially when interviews involve a selected group and records have limited contextual detail.

Examine convergence and disagreement

When components agree, check whether they genuinely provide independent support or simply repeat the same underlying measure. When they disagree, examine definitions, recall periods, sample composition and the influence of the data-collection setting. Disagreement may expose a meaningful difference between reported confidence and demonstrated performance, for example.

Document how the combined interpretation was developed. Identify conclusions supported by both components, those supported mainly by one and those that remain unresolved. Reporting resources such as SRQR can support transparent description of the qualitative component, while the quantitative component needs guidance appropriate to its design. ^[2] The integrated design also requires its own explanation.

Review feasibility and reporting

• State the integration question and why it matters.
• Show timing and dependencies between components.
• Justify sampling and participant linkage.
• Allocate expertise, budget and analysis time.
• Plan how contradictory findings will be explored.
• Explain each component's limitations and their combined effect.
• Present a conclusion that emerges from integration.

A mixed-methods project can offer a richer account, but complexity is not a substitute for coherence. If resources cannot support both components properly, a well-designed single-method study may be more useful. The final judgement should be based on what the design contributes to the question, not on the number of methods listed.