A topic in the Open Knowledge Graph — a free, open map of 15,290 topics and the order to learn them in.
Synthesizes qualitative and quantitative methods for comprehensive understanding of social phenomena. Covers sequential, concurrent, and explanatory designs; integration at design, data, analysis, and interpretation stages; and philosophical foundations of mixing methods. Addresses methodological pluralism.
Design a mixed methods study with explicit integration points, analyze qualitative and quantitative data separately then jointly, write integrated findings, reflect on how methods informed each other.
You already know what each tradition can do on its own: ethnography produces thick, contextual accounts of how people experience and make sense of their worlds; linear regression identifies patterns and estimates causal effects across populations. The core insight of mixed methods research is that these are not competing approaches to the same question but complementary tools that illuminate different facets of the same phenomenon. A regression might show that attending a job training program increases employment by 15 percentage points — but it cannot explain *why*, or why the effect is stronger for some subgroups than others. Ethnographic fieldwork can answer those questions but cannot tell you how representative its findings are. A well-designed mixed methods study does both.
The most important design decisions concern *sequence* and *purpose*. In a sequential explanatory design, you run the quantitative component first, then use qualitative methods to explain surprising or anomalous findings. This is the most intuitive sequence: the numbers raise a question, the fieldwork answers it. In a sequential exploratory design, qualitative work comes first to generate hypotheses or develop instruments that are then tested on a larger sample — useful when existing theory is poorly suited to the population you're studying. In a concurrent design, both strands are collected simultaneously and integrated at the analysis stage; this demands more methodological infrastructure but avoids the risk that early findings unduly constrain the later strand.
Integration — not merely collection of both types of data — is what makes a study genuinely mixed methods. Integration can happen at multiple points: at the design stage (qualitative sampling strategies informed by quantitative distributions), at the data stage (embedding a survey within an ethnographic field site), at the analysis stage (using qualitative themes to interpret regression coefficients), or at the interpretation stage (building a unified explanation that could not emerge from either strand alone). The benchmark is joint display: presenting findings from both strands together so that convergences, divergences, and complementarities are visible. When quantitative and qualitative findings point the same direction (convergent validity), confidence in the conclusion increases. When they diverge, that divergence itself becomes a finding worth explaining.
A common misconception is that mixing methods automatically triangulates away error or bias. This conflates triangulation — using multiple methods to verify a finding — with mixed methods more broadly. In practice, qual and quant methods can be biased in different but correlated ways (both may, for example, oversample accessible populations). The more honest framing is that mixed methods achieves complementarity: each strand answers questions the other cannot, and together they produce a richer account of the phenomenon than either could alone. The philosophical foundation is methodological pluralism — the position that no single method owns a monopoly on social truth, and that the choice of method should be driven by the research question rather than by disciplinary tradition or paradigmatic loyalty.
Topics in reflective domains aren't scored by quiz answers. Read, reflect, and mark when you've thought it through.