Mixed methods research is a approach that combines quantitative data (numbers, measurements, statistics) with qualitative data (interviews, observations, open-ended responses) within a single study or program of inquiry. Rather than choosing one or the other, researchers deliberately collect and analyze both types of data, then integrate the findings to build a more complete picture of the question they’re investigating. The defining feature that separates mixed methods from simply running two separate studies is integration: the quantitative and qualitative components are intentionally connected at the design, methods, or interpretation stage.
Why Combine Two Types of Research
Quantitative research is powerful at measuring how much, how many, and how often. It can identify patterns across large groups and test whether a relationship between variables is statistically meaningful. But numbers alone rarely explain why something happens or what the experience feels like for the people involved. Qualitative research fills that gap through interviews, focus groups, and observation, capturing context and meaning that a survey cannot.
Mixed methods lets researchers use the strengths of each approach to offset the weaknesses of the other. A survey might reveal that 40% of patients stop taking a medication within six months, but interviews with those patients can explain the specific barriers they faced. Studied separately, those two pieces of information are useful. Studied together, with one informing the other, they produce findings that are richer and more actionable. This capacity to study a question from multiple perspectives, while keeping the patient or participant experience central, is one of the method’s most frequently cited advantages.
The Three Core Designs
Most mixed methods studies follow one of three basic structures. The difference between them comes down to timing: whether the quantitative and qualitative phases happen one after the other or at the same time, and which phase drives the study.
Explanatory Sequential
This design starts with quantitative data collection and analysis, then follows up with qualitative data. The quantitative phase is typically prioritized, and the qualitative phase exists to explain or expand on the numerical findings. A researcher might distribute a large-scale survey, identify unexpected results or interesting subgroups, and then conduct interviews specifically with participants from those subgroups to understand what the numbers mean in practice. The quantitative results can also be used to purposefully select participants for the qualitative phase rather than sampling randomly.
Exploratory Sequential
This design flips the order. It begins with qualitative data collection, like interviews or focus groups, to explore a concept that isn’t well understood. The themes and ideas that emerge from that exploration are then used to build a quantitative instrument, often a survey or questionnaire, which is tested on a larger sample. Researchers frequently choose this design when they need to develop a new measurement tool. The qualitative phase generates the content and language for the instrument, and the quantitative phase validates it. This makes it especially useful in areas where existing tools don’t capture what matters most to the population being studied.
Convergent
In a convergent design, quantitative and qualitative data are collected at roughly the same time, independently of each other. Neither phase depends on the results of the other. Once both datasets are analyzed separately, the researcher merges the findings and looks for areas where the two types of evidence agree, complement each other, or contradict each other. This design is particularly effective when the goal is triangulation: using multiple sources of evidence to confirm or challenge a conclusion.
Beyond these three basic structures, researchers also use advanced frameworks that embed mixed methods within larger study architectures, including multistage designs, intervention studies, case studies, and participatory research.
How Integration Actually Works
Integration is what makes mixed methods research more than two separate studies stapled together. It happens at three levels: the overall study design, the methods used during data collection and analysis, and the interpretation of results.
At the interpretation stage, several concrete techniques help researchers bring their quantitative and qualitative findings into conversation. One common approach is a convergence coding matrix, where findings from each component are displayed side by side on the same page. Researchers then systematically assess where there is agreement, partial agreement, silence (one dataset has nothing to say about a finding from the other), or dissonance between the two sets of results.
Another technique, called “following a thread,” starts with a key theme or finding from one dataset and traces it across the other, examining how it appears (or doesn’t) in different forms of evidence. A third approach focuses on individual cases rather than variables. All the data collected on a single participant, both their survey responses and their interview transcript, are examined together. These case-level comparisons can then be displayed in a matrix where rows represent individuals and columns display different types of data, allowing researchers to spot surprises and contradictions at the individual level before looking for patterns across the full sample.
Sampling in Mixed Methods Studies
Sampling decisions in mixed methods research are more complex than in single-method studies because the researcher has to decide not only who participates, but how the quantitative and qualitative samples relate to each other. Two key dimensions shape this decision: timing and the relationship between samples.
Timing can be concurrent (both samples drawn at roughly the same point, independently) or sequential (the second sample depends on results from the first). The relationship between the two samples can take several forms. In an identical relationship, the same people participate in both phases. In a parallel relationship, different people are drawn from the same population. In a nested relationship, the qualitative participants are a subset of the quantitative sample, selected because they scored in a particular way or belong to a subgroup of interest. A multilevel relationship draws samples from different levels of an organization or system, such as surveying students but interviewing their teachers.
Evaluating Quality and Rigor
Judging the quality of a mixed methods study requires looking beyond the standards for quantitative and qualitative research individually. A widely used framework focuses on two dimensions. The first is inference quality, which covers both the soundness of the design and the rigor of interpretation. Interpretive rigor means that the conclusions drawn actually follow from the data, that they’re consistent with existing theory, and that the integration of quantitative and qualitative findings is effective rather than superficial.
The second dimension is legitimation: whether the researcher has demonstrated that their inferences are credible, reliable, and confirmable. The Mixed Methods Appraisal Tool (MMAT) is one formal checklist used to evaluate these studies. It assesses five areas: whether the choice of mixed methods is justified, how well the quantitative and qualitative components are integrated, how the integrated results are interpreted, whether divergences between the two types of findings are addressed, and whether each individual method meets its own quality standards.
Strengths and Challenges
The primary strength of mixed methods is its ability to produce a more complete understanding of complex questions than either approach alone. When quantitative and qualitative findings converge, confidence in the results increases. When they diverge, those contradictions often point to the most interesting and important aspects of the phenomenon being studied. Mixed methods also keeps participant voices in the picture, ensuring that statistical patterns are grounded in real human experience.
The challenges are practical. Mixed methods research is demanding in terms of skill. A researcher (or research team) needs genuine competence in both quantitative and qualitative methodologies, plus the additional expertise to integrate them meaningfully. This isn’t a trivial requirement. Poorly executed integration, where the two strands are analyzed separately and combined only in a discussion section, is one of the most common criticisms of published mixed methods studies. The approach also requires more resources and time than a single-method study, since two full rounds of data collection and analysis must be planned, funded, and carried out.
Software for Mixed Methods Analysis
Several software tools support the integration of quantitative and qualitative data within a single platform. The most commonly used include MAXQDA, NVivo, Dedoose, Atlas.ti, and QDA Miner. These programs were originally built for qualitative analysis but have added features specifically designed for mixed methods work, such as the ability to link coded text segments to numerical variables, create joint displays, and run queries that cross both data types. MAXQDA, for example, offers an interactive quote matrix that lets researchers visualize qualitative data alongside quantitative categories. Dedoose is cloud-based and often favored by teams who need to collaborate remotely. The choice between them typically comes down to budget, team size, and which integration features matter most for a particular study design.