Mixed-Method Study Selection is the discipline of picking the research method whose strength matches the specific decision under uncertainty, then combining methods when neither method alone would close the gap. Qualitative methods reveal why behaviour happens but cannot say how often. Quantitative methods reveal how often but cannot say why. Most product decisions live at the intersection where both questions must be answered, and the mixed-method study is the operational pattern that lets a researcher cover the gap.
Creswell and Plano Clark (2017) defined mixed-method research as the deliberate integration of qualitative and quantitative methods inside one study, with explicit reasoning about how the methods complement each other. Their core insight is that the integration must be planned, not retrofitted. A team that runs a survey and then a follow-up interview round without planning the integration produces two studies, not one mixed-method study.
Rohrer (2022) at Nielsen Norman Group operationalised the selection problem with a three-dimensional framework: behavioural versus attitudinal, qualitative versus quantitative, and context of product use. The framework places 20 UX research methods on the axes and helps practitioners pick the one whose dimensions match the decision. The Rohrer framework is the most-cited industry reference for method selection in 2026.
The principle: Name the decision under uncertainty. Pick the method whose strength addresses it. Add a second method only when the first cannot close the gap.
Mixed-method research has a long methodological tradition in social science before it became a UX practice. The UX adoption pattern in the 2010s and 2020s adapted the methodology to product-decision contexts where the questions are smaller and the timelines tighter than in academic mixed-method work.
Creswell and Plano Clark (2017) wrote the canonical methodological reference. Their typology of mixed-method designs (convergent, explanatory sequential, exploratory sequential, embedded) is taught in graduate research-methods courses and grounds the UX adaptations that follow. The explanatory sequential design (quant first, then qual to explain the quant) is the most common pattern UX teams use; the exploratory sequential design (qual first, then quant to validate the qual at scale) is the second-most-common.
Nielsen (2000) wrote the foundational small-n usability article that anchors qualitative method selection in UX. His claim that 5 users surface roughly 85% of usability problems has held up across two decades of replication and remains the industry default for moderated usability studies. The Nielsen finding is specifically about formative usability evaluation; it does not generalise to attitudinal or preference research, where samples must be larger to support inference.
Sauro and Lewis (2016) in Quantifying the User Experience extended the quantitative side. Their book establishes when a sample of 5 is sufficient, when 30 is required, and when 100+ is the minimum (typically for survey-based attitudinal research with subgroup analysis). The Sauro and Lewis framework pairs directly with Nielsen's: small-n qualitative for formative work, larger-n quantitative for inference and prioritisation.
Rohrer (2022) at Nielsen Norman Group produced the most-cited industry method-selection framework. His three-dimensional model (behavioural/attitudinal, qualitative/quantitative, context of use) helps practitioners select among 20 methods including A/B testing, intercept surveys, diary studies, card sorts, and eye-tracking. The Rohrer framework is what most product teams reference when scoping a study in 2026.
The combined finding across these sources is consistent: teams that pick the method whose strength matches the decision and pair methods when the gap requires it report substantially higher decision confidence than teams running a single method by default. The mixed-method study is not always required; the discipline of selection always is.
For UX Researchers: Method selection is your highest-leverage upstream decision. A well-selected single method is more useful than a poorly selected mixed-method study. The discipline of asking "which method's strength matches this decision" before scoping protects you from the most common UX research failure: running the method you are most comfortable with rather than the method that fits the problem.
For Product Managers: When you commission research, the method-selection conversation is where you most influence the eventual usefulness of the findings. Ask your researcher to name the decision under uncertainty before they name the method; the order matters.
For Designers Running Their Own Studies: Mixed-method discipline keeps your studies scoped. A common failure mode is running an unmoderated usability test and then trying to interpret the results as both "what users do" and "what users want" simultaneously. Mixed-method discipline forces you to pick the method for each sub-question explicitly.
For Engineering Leadership: Engineering decisions about performance, instrumentation, and platform features benefit from method-aware research too. Behavioural log analysis tells you what users do; it does not tell you why they abandoned a checkout step. The mix matters for engineering decision quality as much as for design quality.
Mixed-method study selection works at any scale, from one-day micro-studies to multi-week formal research programs. The decision framework stays the same.
Name the decision under uncertainty before naming the method. Write down the specific decision the study must inform and the timeline for the decision. The decision constrains the method choice; the method should not constrain the decision framing. "We need to choose between two navigation patterns by Friday" yields a different method choice than "we want to understand navigation behaviour generally."
Pick the method whose strength addresses the decision risk. Qualitative methods (interviews, moderated usability, diary studies) reveal why behaviour happens, what users feel, and where the mental model breaks. Quantitative methods (surveys, analytics, A/B testing, intercept surveys) reveal how often behaviour happens, at what magnitude, and across which segments. If the decision risk is a why question, qualitative is the right starting method; if it is a how-often question, quantitative is.
Add a second method only when the first cannot close the gap. A 5-user moderated usability test can answer most formative design questions on its own. A survey can answer most prioritisation questions on its own. The mixed-method pattern is for decisions where the gap genuinely requires both: "users abandon checkout at step 3, but we do not know why." Run analytics to identify the step, then qual interviews to explain it.
Sequence the methods deliberately. Explanatory sequential (quant first, then qual to explain) works when you have the quant signal already and need to understand it. Exploratory sequential (qual first, then quant to validate at scale) works when you have a hypothesis from interviews and want to confirm it across a broader sample. The choice of sequence is part of method selection, not an afterthought.
Match sample size to method, not to academic comfort. Nielsen's 5-user finding applies to formative usability evaluation. It does NOT apply to attitudinal preference research, where samples of 30-100+ are typically required. Sauro and Lewis (2016) provide the sample-size framework for quantitative work. Matching n to method is what makes the findings actually generalise.
Cap the scope at the decision. A mixed-method study that runs longer than the decision window is theatre. If the decision locks Friday, the study must conclude by Thursday. Method selection is downstream of the timeline; pick methods that fit the timeline.