Mixed Methods Research: When and How to Combine Qualitative and Quantitative Approaches

Reviewed by the EditVerse Editorial Team — manuscript editors, peer-reviewed journal specialists, and research methodologists with 1,100+ manuscript completions.

Last updated: July 2026 • Reviewed by the EditVerse Editorial Team • Reading time: 14 minutes

In 2020, a systematic review funded by the U.S. National Institute on Drug Abuse examined 19 mixed-methods public health evaluations and concluded that the approach was uniquely capable of bridging the gap between statistical outcome data and the lived experience that drives those outcomes [PMID 30633710]. That gap — between what the numbers show and why the numbers look that way — is precisely what mixed methods research was designed to close.

Mixed methods research combines qualitative and quantitative methods in a single study or coordinated set of studies. A 2022 review in Kidney International confirmed the field’s growth, showing that mixed-methods use in nephrology doubled over the preceding decade as researchers recognised that quantitative endpoints alone could not explain treatment non-adherence or care disparities [PMID 35227687].

What Is Mixed Methods Research?

Mixed methods research integrates at least one qualitative and one quantitative method within the same research project to address a shared research question. The integration — not mere coexistence — is the defining feature. Three principles characterise well-executed mixed methods work. First, the qualitative and quantitative components must speak to each other at some point during collection, analysis, or interpretation. Second, the rationale for mixing must be methodological. Third, the integration strategy must match the chosen design type.

Evolution of the Approach

Systematic frameworks emerged through the 1990s, and by 2007 the field had its own journal (Journal of Mixed Methods Research). A 2024 summary in the Journal of Continuing Education in Nursing noted the approach has since expanded into clinical nursing practice, with researchers using it to plan interventions, validate instruments, and evaluate programme effectiveness [PMID 38567919].

The Three Core Designs

How to Choose the Right Design

Three questions determine which design applies. First, timing: do both datasets need to be collected simultaneously, or is one contingent on the other? Second, priority: which strand carries the primary argument? Third, integration point: at which stage do the two strands connect? Palinkas, Mendon, and Hamilton (2020) mapped integration across four stages — study design, data collection, analysis, and interpretation — and found that most errors arise when researchers fail to plan the integration point before data collection begins [PMID 30633710].

Convergent Parallel Design

Convergent parallel design collects qualitative and quantitative data simultaneously, analyses each strand independently, and merges results at the interpretation stage. A 2021 study in the Journal of Health Organization and Management [PMID 34255441] used this design to assess patient safety culture, combining staff safety climate surveys with concurrent focus groups. When merged, the two datasets revealed shift-timing conflicts invisible to the quantitative instrument alone.

Use convergent parallel when you have resources for parallel collection, when variables are already established, and when your purpose is confirmation or triangulation. The standard reporting tool is the joint display — a table pairing qualitative themes with quantitative counterparts so reviewers can assess convergence directly.

Explanatory Sequential Design

Explanatory sequential design runs in two phases. The first collects and analyses quantitative data. The second uses those results to determine who to sample and what to explore qualitatively. This design is common in health disparities research, where a survey identifies a statistically significant outcome difference between groups but cannot explain the mechanism.

After Phase 1 analysis, researchers identify specific results that need qualitative follow-up — an unexpected coefficient, a subgroup outlier, or a scale score that defies theory. The Phase 2 sampling frame is then built around those results using purposive sampling. The final integration step connects qualitative themes back to the quantitative results.

Exploratory Sequential Design

Exploratory sequential design runs qualitative inquiry first, because the researcher either lacks a validated instrument or lacks a framework to determine which variables to measure. The quantitative phase then tests or extends what the qualitative phase discovered.

The connection point — the building decision — must be documented explicitly: how exactly do the qualitative findings inform the quantitative design? Reviewers applying the Mixed Methods Appraisal Tool (MMAT) look for this documentation as a quality criterion [PMID 30905698].

Worked Example: Explanatory Sequential in Practice

Phase 1: Quantitative Survey

A cross-sectional survey of 1,329 African American adults with type 2 diabetes measured self-management behaviours, self-efficacy, social support, and glycaemic control (HbA1c). Logistic regression found two significant predictors of suboptimal HbA1c: low self-efficacy (OR 2.3, 95% CI 1.4–3.8) and low perceived social support (OR 1.9, 95% CI 1.2–3.0). The model explained only 18% of HbA1c variance — significant but practically limited.

Decision for Phase 2: Forty participants selected using maximum variation sampling — 20 with low self-efficacy and poor HbA1c (expected pattern) and 20 with low self-efficacy but normal HbA1c (unexpected pattern).

Phase 2: Qualitative Interviews

Semi-structured interviews (avg 52 min) identified three themes explaining good glycaemic control despite low self-efficacy: informal peer accountability networks, practical knowledge transfer from family members, and religious community engagement providing both emotional support and medication cost assistance.

Integration: Joint Display

Data Collection Strategies

Quantitative strands draw on surveys, experiments, secondary data (registries, administrative records), and structured observation. Qualitative strands use in-depth interviews, focus groups, ethnographic observation, and document analysis. A critical coordination decision is whether to use the same participants for both strands (same-sample approach) or different pools. Bailey et al. (2022) note that nephrology studies often use same-sample convergent designs because patient populations are small, but this risks priming qualitative responses via prior survey participation [PMID 35227687].

Analysis and Integration

Each strand is analysed using methods appropriate to its data type. The integration itself operates at three levels: data transformation (converting one type to the other — quantitizing qualitative themes was identified by Palinkas et al. as a major recent innovation [PMID 30633710]); data consolidation (merging both datasets into a joint display matrix); and multilevel modelling (entering qualitative context data as a Level 2 predictor in hierarchical regression).

Research Quality and Validity

Quality applies separately to each strand and to the integration. The quantitative strand requires standard psychometric checks; the qualitative strand requires trustworthiness criteria (credibility via member checking, dependability via audit trail, confirmability via reflexivity statements). The integration requires documentation of which conclusions are supported by one strand only versus both, and how contradictions were handled. The Mixed Methods Appraisal Tool (MMAT) provides a validated checklist for all three levels [PMID 30905698].

Common Pitfalls

Nominal mixing: collecting both data types but never integrating them. Papers with separate qualitative and quantitative sections that never connect are not mixed methods studies. Design mismatch: choosing a design that does not follow from the research question. Inadequate integration reporting: describing each strand in full detail but providing only one paragraph on how they were merged. Resource underestimation: a well-executed convergent parallel study requires two complete data collection and analysis cycles running simultaneously.

Software Tools

MAXQDA and Dedoose include explicit mixed methods features — joint displays, quantitative content analysis, and cross-tabulation of codes by numeric variables. For the quantitative strand, R’s open-source ecosystem (tidyverse, lme4, ggplot2) allows custom joint display visualisations within the same environment as the statistical analysis. REDCap can coordinate both quantitative instruments and qualitative field note templates within a single project structure for multi-site studies.

Best Practices Summary

Related Reading

For quality assessment within mixed methods systematic reviews: Quality Assessment in Mixed-Methods Systematic Reviews: Integrated Approaches.

For framing your methodology chapter when choosing between approaches: Quantitative vs Qualitative: Thesis Statement Battle.

For methodology chapter structure: Writing Effective Methodology Chapters.

References

1. Palinkas LA, Mendon SJ, Hamilton AB. Innovations in mixed methods evaluations. Annu Rev Public Health. 2020;40:423–442. PMID 30633710
2. Bailey PK, Hole BD, Plumb LA, Caskey FJ. Mixed-methods research in nephrology. Kidney Int. 2022;101(5):895–905. PMID 35227687
3. Kawar LN, et al. Quantitative, qualitative, mixed methods, and triangulation research simplified. J Contin Educ Nurs. 2024;55(7):338–344. PMID 38567919
4. Hong QN, et al. Improving the content validity of the MMAT: a modified e-Delphi study. J Clin Epidemiol. 2019;111:49–59. PMID 30905698
5. Lorenzini E, et al. Safety culture in healthcare: mixed method study. J Health Organ Manag. 2021;35(6):780–794. PMID 34255441
6. Creswell JW, Plano Clark VL. Designing and Conducting Mixed Methods Research. 3rd ed. SAGE; 2018.