Mixed-methods studies and integration
Why this level?
This hierarchy places a design by the kind of question it answers: intervention, diagnostic accuracy, prognosis, aetiology and screening intervention. This design answers a different kind of question, so no level on this scale describes it. That is a statement about the scale, not about the quality of the evidence such a study can produce. NHMRC 2009 sourceMixed methods combines strands that answer different kinds of question, so no single rung of a ladder built for effect estimates describes it. A quantitative strand that estimates an effect takes the placement of its own design, and the qualitative strand is appraised with qualitative criteria.
What question does this help answer?
Why did a quantitative result occur, how should a measure or intervention be built for this population, or do numerical and experiential findings agree, and what does it mean if they do not?
When to use it
Use it when the question genuinely has parts that need different kinds of data, such as an effect estimate and an account of how an intervention was received, or when qualitative work is needed to build a questionnaire that will then be used quantitatively.
When not to use it
Do not add interviews to a trial, or a survey to an interview study, simply because both are available. If each component would be reported without reference to the other, that is two studies, and each needs its own adequate design.
Three core designs
- Convergent: both strands are collected in roughly the same period and brought together to see where they agree, extend each other or conflict.
- Explanatory sequential: quantitative results come first and qualitative work explains them, often sampling participants according to their quantitative results.
- Exploratory sequential: qualitative work comes first and shapes a quantitative instrument, intervention or sampling frame, which is then tested.
- More complex designs, such as qualitative work embedded in a trial or a multi-phase evaluation, combine these.
What the three designs look like
Where integration happens
- At design: the reason for combining methods is stated, and the timing and priority of each strand follow from it.
- At methods: one strand determines the sampling or data collection of the other, for example choosing interviewees from survey respondents with high and low scores.
- At analysis and interpretation: results are brought together, for example in a joint display that sets quantitative results beside the related qualitative themes, and disagreement between them is examined rather than set aside.
Key requirements
- Each strand meets the standards of its own method: purposive sampling and a named analytic approach for the qualitative strand, an adequate design and sample for the quantitative strand.
- Integration is planned in advance and described, not left to the discussion section.
- Conflicting findings are reported and explored; they are often the most informative result.
- Sampling for each strand is described separately, including how participants in one relate to the other.
Clinical example
A trial of home-based rehabilitation finds a smaller effect than expected. Participants with high and low adherence are then interviewed, and a joint display sets adherence and outcome data beside themes about travel, pain and family support. The integration suggests why the programme worked for some participants and not others. This example is illustrative.
What to watch out for
Running a survey and some interviews in the same project is not mixed methods unless the two are integrated at a stated point. Integration means one strand shapes the sampling, data collection, analysis or interpretation of the other; findings reported side by side in separate sections leave the reader to do the integration.
How to report it
GRAMMS is the guideline for this design, with COREQ or SRQR applied to the qualitative strand. GRAMMS asks for the justification of the mixed-methods design, its purpose and the sequence and relative priority of the two strands, where integration actually happened, and what insight came from mixing the methods. A reporting guideline is not a design manual: use it to make the study transparent, not as a substitute for careful design. Check the target journal's own requirements as well, which this page cannot know.
Before using this in your workspace
Record the reason for mixing methods, the core design, the timing and priority of each strand, the planned integration points, how each strand is sampled and analysed, and how discordant findings will be handled.
- Achieving integration in mixed methods designs-principles and practices (opens in a new tab) · Health Serv Res · doi 10.1111/1475-6773.12117
- Integrating Quantitative and Qualitative Results in Health Science Mixed Methods Research Through Joint Displays (opens in a new tab) · Ann Fam Med · doi 10.1370/afm.1865
- Three techniques for integrating data in mixed methods studies (opens in a new tab) · BMJ · doi 10.1136/bmj.c4587
- The quality of mixed methods studies in health services research (opens in a new tab) · J Health Serv Res Policy · doi 10.1258/jhsrp.2007.007074
- Consolidated criteria for reporting qualitative research (COREQ): a 32-item checklist for interviews and focus groups (opens in a new tab) · Int J Qual Health Care · doi 10.1093/intqhc/mzm042
- Standards for reporting qualitative research: a synthesis of recommendations (opens in a new tab) · Acad Med · doi 10.1097/ACM.0000000000000388