13 Qualitative and Mixed Methods
Here’s a scenario you might find familiar. You’re reading a research article about a health intervention in rural Kenya, maybe a community health worker program aimed at improving maternal health outcomes. The numbers look good: clinic visits increased by 37%, births with skilled attendants rose from 42% to 61%, maternal mortality declined. Success, right?
But then you visit that community. The gains are real. They are also unevenly shared. The women still missing from those averages are the poorest and the most remote—some walking past the new clinic to a traditional birth attendant three villages away because the health workers, despite their training, aren’t trusted: they don’t speak the local dialect. And a rising rate of skilled attendance says nothing about the woman who reached the clinic only to die after it referred her to a distant hospital her family couldn’t afford. The numbers aren’t wrong. They’re just incomplete, technically accurate but missing something essential about what’s actually happening on the ground.
This gap between what we can count and what we need to understand drives the need for qualitative research in global health. Don’t get me wrong: numbers matter enormously. We need to know if interventions reduce mortality, if treatments improve health outcomes, if programs reach their intended beneficiaries. But when we’re working across cultures, within complex social systems, and with vulnerable populations facing intersecting challenges, numbers alone can’t capture the full story. We need approaches that can illuminate how and why things work (or don’t work), that can reveal cultural meanings and social dynamics, that can capture the lived experiences of the people our research is meant to serve.
Qualitative research focuses on understanding phenomena through in-depth exploration of meanings, experiences, and contexts, typically using methods like interviews, observations, and document analysis rather than numerical measurement.
13.1 Foundations of Qualitative Research
Before getting to specific approaches, we need to lay some groundwork: what philosophical assumptions distinguish qualitative from quantitative research, what characterizes it as a method, and when should you actually choose it? I promise to keep the philosophy brief, but these foundational concepts matter because they shape everything from how you design studies to how you analyze data to how you judge quality.
PHILOSOPHICAL UNDERPINNINGS
Quantitative research generally operates within a positivist or post-positivist paradigm. The core assumptions: there’s an objective reality out there, we can measure it (though perhaps imperfectly), and systematic observation following scientific methods can reveal truth. When we conduct randomized controlled trials or analyze survey data, we’re usually working within this paradigm.
Qualitative research typically operates within interpretive or constructivist paradigms. The core assumptions here are different: reality is socially constructed, meaning emerges through interaction and interpretation, and multiple valid perspectives exist on any phenomenon. Knowledge isn’t discovered but co-created through the research process itself.
Constructivism is a philosophical stance holding that humans construct meaning through interaction with the world, rather than discovering objective truth.
What does this mean practically? It means that when conducting qualitative research, you’re not trying to eliminate your influence as a researcher to achieve objectivity (the quantitative ideal). Instead, you’re trying to be reflexive about your influence—acknowledging how your own background, assumptions, and relationship with participants shape what you see and hear. Good qualitative research doesn’t pretend the researcher is invisible; it makes the researcher’s positionality and potential influence explicit and thinks carefully about how these might affect findings.
Before conducting qualitative research, ask yourself: What is my relationship to this community? What assumptions do I bring? What privilege or power do I hold? How might these influence what I see and hear?
Imagine you’re a Western researcher studying maternal health practices in rural Afghanistan. Your positionality—your nationality, education, gender, institutional affiliation—profoundly shapes what you can access and what people will tell you. Pretending you’re an objective observer would be naive. Good qualitative research acknowledges this and thinks strategically about it. Maybe you partner with local researchers who can build relationships you can’t. Maybe you’re explicit about your outsider status and how that limits your understanding. Maybe you employ community-based participatory approaches that share research power with community members. The point is: you think about positionality rather than pretending it doesn’t exist.
KEY CHARACTERISTICS OF QUALITATIVE RESEARCH
While qualitative research traditions vary (we’ll get to those), they share several common characteristics:
Naturalistic settings. Qualitative researchers typically study phenomena in their natural contexts rather than in controlled laboratory settings. If you’re studying how families make decisions about childhood vaccination, you might observe community meetings, visit homes, or attend clinic visits rather than bringing people to an artificial setting.
Inductive reasoning. While quantitative research typically tests pre-existing hypotheses (deductive reasoning), qualitative research usually builds understanding from the ground up (inductive reasoning). You start with observations and let patterns, themes, and theories emerge from the data rather than starting with a theory and testing it.
Emergent design. Unlike quantitative studies where you specify everything in advance (sample size, measures, analysis plan), qualitative studies often adapt as you learn. Maybe your initial interviews reveal that a factor you didn’t anticipate, say, transportation costs, emerges as crucial. Good qualitative design allows you to adjust your interview guide to explore this more deeply in subsequent interviews.
Holistic perspective. Qualitative research tries to understand phenomena as complex wholes rather than breaking them into discrete variables. When studying health worker motivation in low-resource settings, you wouldn’t just measure “salary” and “supervision” separately. You’d explore how these factors interact with community relationships, professional identity, workplace culture, and personal circumstances to create a holistic understanding of motivation.
Researcher as instrument. In quantitative research, measurement instruments (surveys, tests, diagnostic tools) collect data. In qualitative research, you are the primary instrument, working through your observations, your interview questions, your analytical choices. This makes researcher training, reflexivity, and quality processes crucial.
These characteristics flow from the interpretivist philosophical foundations: if meaning is socially constructed and context-dependent, then of course you need to study phenomena in context, reason inductively from what you observe, maintain flexibility to follow unexpected insights, and think holistically about complex interactions.
TYPES OF QUESTIONS QUALITATIVE RESEARCH ANSWERS
With these foundations in mind, the practical question becomes: when should you choose qualitative research over (or in addition to) quantitative approaches? This isn’t about one being “better” than the other. They serve different purposes and answer different kinds of questions.
Qualitative research excels when you’re asking how, why, or what does this mean questions rather than how many or how much questions. Consider these contrasting examples:
| Quantitative question | Qualitative question |
|---|---|
| What percentage of pregnant women in rural Tanzania attend at least four antenatal care visits? | What factors influence pregnant women’s decisions about whether to attend antenatal care in rural Tanzania, and how do they weigh competing demands on their time and resources? |
| Does task-shifting mental health counseling to peer health workers improve depression outcomes? | How do patients and community members perceive peer health workers delivering mental health interventions, and what characteristics make peer workers acceptable and trustworthy in different cultural contexts? |
Notice the difference? Quantitative questions seek to measure outcomes, test hypotheses, and quantify relationships. Qualitative questions seek to understand processes, explore meanings, and probe mechanisms. Both matter, but they require fundamentally different approaches.
Qualitative research becomes especially valuable when you’re dealing with:
- Cultural contexts and beliefs that shape health behaviors in ways quantitative data can’t capture
- Implementation processes where understanding how an intervention works (or fails) matters as much as whether it works
- Patient and provider experiences that reveal dimensions of healthcare quality invisible in clinical outcome data
- Complex social dynamics involving power, stigma, gender, or other dimensions that resist simple measurement
- Emerging phenomena where you don’t yet know enough to develop standardized measures or structured surveys
A WORKED EXAMPLE: STUDYING THE SAME QUESTION WITH DIFFERENT METHODS
To make this concrete, let’s walk through how quantitative and qualitative researchers have tackled the same question differently. The question: how should perinatal mental health support be delivered to women in low-resource settings, and who should deliver it?
The Quantitative Approach
Kumar and colleagues used a discrete choice experiment (DCE) with 153 pregnant adolescents (ages 14–18) recruited from two primary healthcare facilities in Nairobi’s informal settlements (Kumar et al., 2023). Rather than ask participants to rate features individually on a 5-point scale—an approach that lets people endorse everything as “important”—a DCE forces tradeoffs. Each participant completed 10 choice tasks, picking between two hypothetical depression treatment packages that varied across eight attributes: where information was delivered, whether caregivers participated, number of sessions (4 vs. 8), delivery agent (community health volunteer vs. facility nurse), training pathway, support type (peer support vs. parenting skills), service setting (adolescent-friendly vs. mixed with older mothers), and incentives.
A discrete choice experiment elicits preferences by asking participants to choose between hypothetical scenarios that bundle several attributes. Because picking one option means giving up something in the other, the modeled choices expose how much each attribute weighs against the others—tradeoffs people may not articulate when asked directly.
Because each package bundled several features, participants couldn’t endorse everything; choosing one meant giving up something in the other. The modeled choices expose which features participants actually trade off for which. Adolescents preferred facility nurses over community health volunteers, eight sessions over four, parenting-skills content over peer support, and adolescent-friendly services over clinics mixed with older mothers (Kumar et al., 2023).
The DCE quantifies relative preference weights across many participants, mimics real-world decision-making more faithfully than direct rating, and can be analyzed at the subgroup level. The findings produce directly actionable guidance: build something that looks like this configuration of features, not that one.
The Qualitative Approach
Bhushan and colleagues asked a closely related question in Lilongwe, Malawi (Bhushan et al., 2026). They conducted 20 in-depth interviews, 3 focus group discussions, and 10 social-support mapping sessions with 42 participants: adolescents living with HIV experiencing depression (ages 13–19), their caregivers, healthcare providers, and people who had previously delivered or received a similar lay-counseling intervention. Rather than ask participants to choose between bundled options, they asked open-ended questions about where, when, how, and from whom adolescents would want mental health counseling.
Across groups, participants converged on a small set of counselor qualities (trustworthy, youth-friendly, non-judgmental, able to keep confidences) and on a clear preference for healthcare providers over other community authority figures such as village chiefs or teachers. One adolescent explained the reasoning: “[At the clinic] there are people who studied the mind of a person… they subscribe to strict code of ethics so they can’t be going about to people and expose you”. The institutional structure of the clinic, more than any individual attribute, made trust possible.
But the same conversations surfaced things a forced-choice survey would not. A minority of adolescents wanted counselors who were themselves living with HIV, valuing lived experience over credentials. Preferences for session timing and frequency varied widely. And a structural tension emerged: adolescents and caregivers wanted frequent, flexible, relational counseling; counselors and implementers emphasized workload, staffing limits, and the need to maintain professional boundaries.
What Each Approach Reveals
The two approaches point in the same broad direction, but we learn something different from each one.
The DCE delivers a result like this: when asked who should deliver intervention sessions, respondents preferred facility nurses over community health volunteers (52.8%, β=6.04, p<0.001) (Kumar et al., 2023). The number is precise, the direction is clear, and the finding plugs straight into program design.
Bhushan’s adolescents also preferred clinic-based providers over community-embedded figures, but with their qualitative approach we learn why. Clinics are convenient, the staff are generally competent and trained in how to maintain confidentiality.
This is what qualitative research uniquely offers: understanding of why and how the choice is constructed, and visibility into the dissent, exceptions, and operational tensions that a quantitative summary smooths over.
WHY CHOOSE QUALITATIVE APPROACHES
Beyond the type of question you’re asking, several specific scenarios call for qualitative methods:
When you need depth over breadth. A landmark ethnographic study examining antiretroviral therapy (ART) adherence in Nigeria, Tanzania, and Uganda conducted 414 in-depth interviews and 136 field observations with patients, treatment partners, and healthcare providers (Ware et al., 2009). The researchers spent extended time with communities, observing daily life and building relationships that allowed participants to explain their experiences honestly. What emerged was a crucial insight that quantitative adherence data alone might not reveal: extraordinarily high ART adherence in these poverty-stricken settings wasn’t primarily driven by fear of disease progression. Instead, adherence reflected individuals’ recognition that staying alive through ART adherence preserved critical social relationships that enabled their economic and social survival in conditions of extreme poverty.
Ethnography involves prolonged immersion in community settings to understand cultural practices, beliefs, and social dynamics through observation and participation.
When context is everything. An interpretive ethnographic study in Northern Ghana among the Dagomba people examined why traditional medicine and biomedical systems remain separate despite policy efforts to integrate them (Kwame, 2021). The ethnographers spent three weeks conducting detailed observations of healing practices and structured interviews with traditional healers, biomedical practitioners, and community members. The findings revealed that the integration challenge wasn’t primarily about clinical efficacy or safety concerns. Instead, integration failed because of power imbalances: formally educated biomedical practitioners held institutional authority and recognition that traditional healers lacked, regardless of their knowledge or community trust. Understanding why integration efforts fail required understanding the cultural, historical, and political context. The ethnographic approach, including detailed reflexive fieldnotes documenting the researchers’ own emerging understanding, allowed them to trace how power operates through institutional structures and professional hierarchies, not just individual attitudes.
When participants’ own perspectives matter. A grounded theory study with 25 people living with HIV in Austria explored their subjective experiences of living with HIV/AIDS, coping with the disease, and managing stigma (Beichler et al., 2023). Through rigorous qualitative analysis with open, axial, and selective coding, researchers identified a “paradigm shift” in how participants understood HIV: the recognition that modern ART has transformed HIV from a terminal diagnosis to a manageable chronic condition fundamentally changed how people thought about their future, their relationships, and their identity. This paradigm shift emerged from participants’ own sense-making processes. It wasn’t a framework researchers imposed; it was a conceptualization people living with HIV developed to make sense of their experience.
Grounded theory is an inductive methodology that develops theoretical frameworks directly from empirical data rather than testing pre-existing theories.
MAKING THE CHOICE: A DECISION FRAMEWORK
So how do you actually decide whether to use qualitative, quantitative, or mixed methods for your research question? Here’s a practical framework to guide your thinking:
Start with your research question. Write it down explicitly. If your question asks “how many,” “how much,” “what percentage,” or “is there a difference between,” you probably need quantitative methods. If your question asks “how,” “why,” “what does this mean,” or “what is the process,” you probably need qualitative methods. If you’re asking both kinds of questions, you probably need mixed methods.
Consider your knowledge state. Are you exploring a phenomenon you don’t understand well enough to develop valid measures? Use qualitative methods to build understanding first. Do you have well-validated measures and want to test relationships at scale? Use quantitative methods. Do you want to both test relationships and understand mechanisms? Use mixed methods.
Think about context and culture. If cultural meanings, social context, or local power dynamics are central to understanding your phenomenon, qualitative methods become essential. If you’re measuring outcomes where cultural variation is minimal (e.g., mortality from a specific disease), quantitative methods may suffice.
Assess what counts as evidence for your stakeholders. Some audiences prioritize large sample sizes and statistical significance; others value rich description and contextual understanding. Some want both. This shouldn’t be your only consideration—you need to use methods appropriate for your question—but it’s worth thinking about how you’ll convince people your findings matter.
Recognize that “either/or” thinking is often wrong. Many research questions benefit from combining approaches. The question isn’t always “qualitative or quantitative?” but “qualitative and/or quantitative in what sequence and for what purposes?” We’ll explore this more in the mixed methods section.
13.2 Five Primary Qualitative Approaches
Now for the practical part: what are the main types of qualitative research, and when would you use each? Think of these as different tools in your methodological toolkit. You wouldn’t use a hammer for every carpentry job, and you wouldn’t use ethnography for every qualitative question. Different approaches suit different purposes.
These categories aren’t as neat and tidy as textbooks (including this one) sometimes make them seem. In practice, researchers often blend elements from multiple traditions, and you’ll see studies that combine ethnographic observation with phenomenological interviewing or case study design with grounded theory analysis. But understanding the distinctive features of each approach helps you make informed methodological choices.
Research traditions in qualitative work emerged from different disciplinary roots—anthropology (ethnography), sociology (grounded theory), philosophy (phenomenology)—each bringing distinct assumptions about knowledge and how to generate it.
ETHNOGRAPHY: UNDERSTANDING CULTURAL CONTEXTS
What it is: Ethnography involves prolonged immersion in a community or cultural setting, combining participant observation (watching what people do) with interviews (asking what people think) to understand beliefs, practices, and social dynamics from participants’ cultural perspective.
The key word here is prolonged. We’re not talking about a quick site visit or a few interviews. Classic ethnography involves months or even years of fieldwork, building relationships, learning the language (both literal and cultural), and witnessing the rhythms of daily life. You’re trying to understand the world as your participants see it, which requires time and presence.
When to use it: Choose ethnography when culture, context, and social dynamics are central to your research question. If you’re trying to understand how health beliefs and practices are embedded in cultural systems, ethnography is your friend. It’s particularly valuable when you suspect that what people say they do and what they actually do might differ, a common phenomenon in health behavior.
Example in action: Recall the ART adherence study in Nigeria, Tanzania, and Uganda we discussed earlier (Ware et al., 2009). The ethnographic approach—414 interviews plus 136 field observations over extended time periods—enabled researchers to see how economic survival strategies shaped adherence behavior. They observed patients borrowing money for transportation to clinics, witnessed treatment partners ensuring medication taking, and heard stories of “doing without” in other life areas to prioritize treatment. These observations revealed patterns that became apparent only through sustained fieldwork: patients employed strategic approaches like borrowing and begging for travel funds and making “impossible choices” to maintain adherence despite extreme poverty.
The ethnographic methodology proved crucial because the researchers employed African researchers trained in ethnographic methods who were themselves from the study countries. This enabled cultural competence and community trust that external researchers would struggle to achieve. It’s a reminder that ethnography isn’t just about methods; it’s about relationships and positionality.
Methodological considerations: Ethnographers must grapple with their own positionality constantly. Are you an insider or outsider to the community you’re studying? How does your presence change what you’re observing? How do you build trust while maintaining analytic distance? The answers to these questions shape what data you can access and how you should interpret it.
The ethnographic tradition emphasizes reflexivity: explicit, ongoing questioning of your own assumptions, biases, and influence on the research process. Good ethnographers keep detailed field notes not just about what they observe but about their own reactions, assumptions, and evolving understanding. This reflexive documentation becomes part of the analytical process.
Strengths: Rich, contextualized understanding; captures taken-for-granted cultural patterns; reveals discrepancies between what people say and do; builds deep community relationships; generates insights about social structures and power dynamics.
Limitations: Time-intensive (often months to years); requires language and cultural competence; findings may not generalize beyond the specific cultural context; can be difficult to publish in journals favoring large samples and statistical analysis; demands substantial resources for extended fieldwork; raises complex ethical questions about representation and power.
GROUNDED THEORY: BUILDING CONCEPTUAL UNDERSTANDING
What it is: Grounded theory is a systematic methodology for developing theoretical frameworks directly from empirical data through iterative cycles of data collection and analysis, using specific coding procedures (open, axial, and selective coding) to identify patterns and develop conceptual categories.
Here’s what makes grounded theory distinctive: you’re not starting with a hypothesis to test. Instead, you’re building theory from the ground up (hence “grounded”) through careful, systematic analysis of empirical data. The theory emerges from patterns you identify in the data rather than being imposed from external frameworks.
This doesn’t mean you approach the research as a blank slate. That’s impossible and probably not desirable. But it does mean you’re open to being surprised by your data, to discovering patterns you didn’t anticipate, to developing conceptual insights that challenge existing theories.
When to use it: Choose grounded theory when your goal is to develop new theoretical understanding about a process or phenomenon, especially when existing theories feel inadequate or culturally inappropriate for your context. Grounded theory is particularly valuable for understanding social processes: how things unfold over time, how people navigate systems, how decisions get made.
The analytical process: Grounded theory involves specific coding procedures that move from description to conceptualization:
- Open coding: Breaking data into discrete parts, examining them closely, and developing initial categories
- Axial coding: Relating categories to subcategories, exploring connections and relationships
- Selective coding: Integrating categories around a core category to develop a theoretical framework
This sounds abstract, so let me show you what it looks like in practice.
Example in action: A grounded theory study with 25 people living with HIV in Austria and Germany explicitly walked through the three coding steps (Beichler et al., 2023). Researchers conducted semi-structured interviews about coping, adherence, stigma, and disclosure, then analyzed transcripts using Strauss and Corbin’s procedure with MAXQDA software.
Open coding fragmented the data into discrete chunks of meaning—statements like “the diagnosis crushed me,” “I write it down every day, it’s ritualized,” and “I don’t want to be reduced to HIV”—and grouped them into initial categories such as shock, ritualized adherence, and resistance to being defined by the diagnosis.
Axial coding organized those initial categories into five higher-order ones: fast coping with diagnosis through paradigm change, psychosocial burden of HIV, ART as a need, trust as a requirement for disclosure, and stigmatization and discrimination. The team then mapped how these categories related to each other: psychosocial burden as a causal condition; relationship with health-care providers, health literacy, and shared decision-making as intervening conditions; stigma and discrimination as contextual conditions; ART as the central strategy; and trust-and-disclosure as a consequence.
Selective coding integrated everything around a single core category: “fast coping with diagnosis through a paradigm change.” The claim is that modern antiretroviral therapy has transformed HIV from a terminal diagnosis to a manageable chronic condition, and adherence, identity, and disclosure all hang off that shift. Stigma is what hasn’t caught up.
That core category wasn’t a hypothesis the researchers imported. It emerged from comparing accounts across a heterogeneous sample—men and women, MSM and heterosexual, recently diagnosed and infected in the 1980s—until the same pattern kept surfacing.
Why this matters: A decontextualized framework would catalogue barriers to adherence and barriers to disclosure as separate problems. The grounded theory analysis instead linked them through a single shift in what HIV is to the person living with it, and located the lived burden in the social response rather than the disease itself. That kind of integrative claim, built up from open coding to a core category, is what grounded theory is built to produce.
Strengths: Generates new theory grounded in data rather than imposing external frameworks; systematic analytical procedures enhance rigor; comparative approach across settings yields nuanced understanding; directly addresses the “so what?” question by building conceptual insights; produces frameworks that guide implementation in new contexts.
Limitations: Can be labor-intensive with multiple rounds of data collection and analysis; requires analytical sophistication to move beyond description to theory development; findings may feel abstract rather than immediately actionable for practitioners; the constant comparative method demands large datasets; debates about “true” grounded theory versus adapted versions create methodological confusion.
Constant comparison is central to grounded theory: you’re always comparing new data to existing categories, refining your developing theory, sometimes returning to collect more data to test emerging concepts. It’s iterative, not linear.
PHENOMENOLOGICAL RESEARCH: CAPTURING LIVED EXPERIENCE
What it is: Phenomenological research explores how people experience and make sense of specific phenomena, focusing on the subjective, first-person perspective of lived experience. The goal is to understand the essential meaning or “essence” of experiencing something.
Phenomenology asks: what is it like to experience this phenomenon? Not what caused it, not how to measure it, but what the quality of the experience itself feels like from the inside. This might sound touchy-feely, but phenomenology employs rigorous analytical methods to systematically examine lived experience.
When to use it: Choose phenomenology when you want to understand what it’s like to experience a particular health condition, treatment, or situation from the patient or caregiver perspective. Phenomenology is particularly valuable when subjective experience itself is clinically relevant: for quality of life, wellbeing, treatment acceptability, or understanding patient priorities.
Example in action: A phenomenological study with 15 intensive care unit patients in South Korea explored their lived experiences of transfer from ICU to general ward (Lee et al., 2021). Using in-depth interviews and Colaizzi’s phenomenological method—requiring researchers to repeatedly review data while suspending preexisting assumptions—researchers identified four main themes: “hope amid despair,” “gratitude for being alive,” “recovery from suffering,” and “seeking a return to normality.”
The phenomenological analysis revealed that transfer from intensive care wasn’t merely a clinical transition but a profound existential transition involving psychological recovery alongside physical healing. Participants described fear about losing intensive monitoring, grief about leaving skilled ICU nurses, and profound relief at gaining autonomy and normalcy.
The researchers analyzed these experiences through what phenomenologists call existential dimensions: the participant’s embodied self, the physical space of the ward, temporality (sequence of experiences), and relationships with other patients and caregivers.
For instance, the physical layout of the ICU—visible monitoring equipment, the nurse call button within easy reach—had shaped patients’ sense of security. Transfer to a general ward lacking these visual anchors created anxiety despite medical stability. This level of insight requires phenomenology’s systematic attention to the quality of subjective experience.
Think about the clinical implications. Understanding that ICU transfer creates existential anxiety, not just logistical change, might influence how healthcare providers prepare patients for transfer, how they design general wards, or how they structure transitional support. This knowledge emerged from phenomenological attention to lived experience, not from clinical outcome measures.
How phenomenology compares to ethnography and grounded theory: All three approaches are inductive, interpretive, and resist imposing prior theory on the data. But they aim at different objects, and that difference drives how they collect and analyze it (Table 13.1).
| Ethnography | Grounded Theory | Phenomenology | |
|---|---|---|---|
| What it produces | A description of a culture or community from the inside | An explanatory framework built up from the data | The essential structure of a lived experience |
| Primary focus | A setting or group | A social process | A subjective experience |
| Typical data | Long-term participant observation, field notes, interviews | Interviews, focus groups, documents, often iterative | A small set of in-depth interviews |
| Analytic focus | Thick description; reading practices in context | Open → axial → selective coding toward a core category | Bracketing assumptions; distilling to an essence |
What sets phenomenology apart most sharply is that last move: bracketing (sometimes called epoché). The other traditions ask researchers to be reflexive about their assumptions and write them into the analysis. Phenomenology asks them to suspend those assumptions so participants’ descriptions can be encountered fresh. Different phenomenological schools (Husserlian, Heideggerian, hermeneutic) prescribe the move differently, but the underlying commitment is the same: the experience as the participant lives it is the object, not the researcher’s prior theory of what that experience should be.
Another example examined patients’ experiences of being observed during clinical teaching, when a supervisor watches a resident physician (Rietmeijer et al., 2021). The phenomenological analysis revealed that patients experienced these situations as involving “two doctors interacting with one another and with them,” with patient comfort depending substantially on the quality of the working relationship between resident and supervisor. When tensions were apparent between the doctors, patients felt uncomfortable; when interactions seemed positive, patients relaxed. Importantly, patients’ lived experiences diverged from assumptions held by both residents and supervisors about how patients experience observation, illustrating how phenomenology can reveal gaps between stakeholder assumptions and reality.
Strengths: Deep insight into patient and caregiver experiences; reveals dimensions of experience invisible in clinical outcome data; particularly valuable for understanding quality of life and wellbeing; centers participant voice and perspective; generates understanding directly relevant to patient-centered care.
Limitations: Findings are descriptive rather than explanatory (tells you what the experience is like, not why it’s that way); may be dismissed by clinicians as “just subjective experience”; requires skilled interviewing to access deep reflection; can be challenging to translate findings into specific intervention recommendations; different phenomenological schools have conflicting methodological prescriptions; demands analytic sophistication to move beyond summarizing descriptions.
CASE STUDY ANALYSIS: UNDERSTANDING COMPLEXITY IN CONTEXT
What it is: A case study takes one or more cases—a programme, an organization, a community, an event—and examines it in its real-world context (Greenhalgh, 2025). The case itself, not a variable or a hypothesis, is the unit of analysis. The aim is to develop a deep understanding of how the case relates to the social, political, economic, and cultural context that surrounds it.
Different scholarly traditions approach case studies differently. Social science case studies tend to be prospective, qualitative, and built around the meanings participants make of their world. Public health case studies are often retrospective accounts of how a health threat emerged and was managed, leaning on routinely collected data alongside interviews and commentaries. Implementation science case studies document how an intervention was actually put into practice in a particular setting, often combining quantitative fidelity measures with formative qualitative work (Greenhalgh, 2025). What unites them is a commitment to thick description: a full, warts-and-all narrative that juxtaposes multiple perspectives, including findings that disconfirm the researcher’s preferred interpretation.
That commitment is what drives the methodological signature of case study research: multiple sources of evidence. A serious case study typically pulls together documents (plans, protocols, reports, meeting minutes), interviews, ethnographic observation, and descriptive quantitative data, then triangulates across them.
When to use it: Case study methodology is well suited to “how” and “why” questions about complex phenomena in real-world settings, especially when the boundary between phenomenon and context is hard to draw. It is a strong choice when an intervention is woven through processes, people, organizations, and policy, or when you want to understand what actually happened when something was implemented (not whether it worked on average in a trial). Cases can be single (one organization or event examined in depth) or multiple (a small set compared across contexts).
Example in action: A mixed-methods case study examined vaccination governance in northwest Syria during protracted conflict (Baatz et al., 2024). The case was bounded by geography (opposition-controlled northwest Syria) and by time (the protracted-conflict period after 2012, when the state withdrew from these areas). The team adapted Siddiqi’s health-governance framework and assembled four data sources: 14 key informant interviews with stakeholders from the Syria Immunisation Group, the Idlib Health Directorate, and partner organizations; a 15-participant validation workshop in Gaziantep; ethnographic observations from daily field immersion; and a review of reports and policy documents, including SIG annual reports.
The four sources were triangulated theme by theme. Sometimes they converged: interviews praised cold-chain reliability, observations confirmed well-maintained facilities, and SIG’s 2021 annual report documented the distribution of 1.5 million routine vaccines and 350,000 COVID-19 vaccines (Baatz et al., 2024). Sometimes they diverged in productive ways. Informants rated the strategy positively, but the documents revealed what was missing from them: no detailed analysis of vaccine losses, no linkage between outbreak data and coverage statistics, an over-reliance on paper-based systems that made comprehensive coverage analysis impossible. That gap in the formal record was itself a finding, one no single data source could have produced.
Why case study methodology mattered here: Syria’s conflict context makes it a unique case. Generic frameworks for vaccination governance, developed in stable settings, might not account for the specific challenges of operating health systems amid active conflict. The case study approach allowed researchers to understand how governance actually functioned in this specific, complex context.
Triangulation in case studies: Notice how this study used triangulation—multiple data sources examining the same phenomenon. When interview data, workshop discussions, and observations converge on the same finding, confidence increases. When they diverge, that’s interesting too. It prompts investigation into why different data sources reveal different facets of the phenomenon.
Strengths: Holistic understanding of complex phenomena; uses multiple data sources for triangulation; findings are richly contextualized; can examine rare or unique situations; generates detailed understanding of how things work in practice; allows investigation of phenomena where experimental manipulation is impossible or unethical.
Limitations: Findings may not generalize beyond the specific case (though case studies can provide “analytical generalization”—insights about processes or mechanisms that may apply elsewhere); can be criticized as lacking rigor if methods aren’t systematic; requires substantial resources for comprehensive data collection; analysis can be overwhelming with diverse data types; risk of getting lost in detail without clear analytical framework.
Analytical generalization differs from statistical generalization. You’re not claiming your case represents all similar cases statistically, but you are developing theoretical insights that may apply to other contexts.
Document review is so central to case study work that it sometimes appears as a method in its own right. Document analysis is the systematic examination of written materials—policy documents, medical records, meeting minutes, media coverage, historical archives—to understand policies, processes, or institutional practices. It is sometimes used alone, for instance to trace how a national policy evolved over a decade, but more often it is one strand within a case study or other mixed-methods design.
Documents are not neutral records. They are created by particular people, for particular audiences, with particular purposes, and that shapes both what they contain and what they leave out. Rigorous document analysis means asking who wrote a document and for whom, what perspectives are centered, what is missing, and how the document relates to formal institutional power (Kayesa et al., 2021). A methodological review of health-policy document analysis in low- and middle-income countries flagged a specific challenge for global health researchers: formal archives are often thin, and a great deal of policy work happens through informal decision-making that is never written down (Kayesa et al., 2021). Documents must therefore be read for what they hide as well as for what they reveal.
PARTICIPATORY AND COMMUNITY-BASED APPROACHES
What it is: Community-based participatory research (CBPR) involves community members as active partners throughout the research process—from defining questions through data collection to interpretation and dissemination—rather than treating communities merely as study subjects.
CBPR represents a fundamental shift in research power dynamics. Traditional research extracts knowledge from communities, with academic researchers controlling questions, methods, interpretation, and dissemination. CBPR redistributes this power, recognizing community members as knowledge holders whose priorities, insights, and interpretations are essential to research quality and relevance.
When to use it: Choose participatory approaches when working with communities that have been historically marginalized or exploited by research, when community buy-in is essential for intervention success, or when you want research to directly benefit and empower participants. CBPR is particularly valuable in global health settings where researchers from high-income countries work in low- and middle-income contexts, as it can help address power imbalances and ensure research serves community priorities.
Example in action: A participatory action research study during COVID-19 in rural Pakistan engaged community members and leaders through iterative cycles of planning, action, and reflection (Moran et al., 2023). Rather than researchers deciding what questions to ask, the community identified challenges to following public health recommendations and designed locally relevant solutions.
Through focus groups and participatory action research meetings, the research team learned that alongside medical recommendations like masking and hand hygiene, critical social and economic factors shaped preventive behavior: community belief structures around disease transmission, employment concerns requiring continued work despite pandemic risks, and family decision-making processes dominated by male family heads who might not prioritize women’s health concerns.
The participatory approach generated not merely descriptive evidence but actionable evidence, with community members themselves designing and implementing solutions reflecting their values, resources, and social structures. That’s the power of participatory research: it honors community knowledge and agency while generating evidence directly useful for local action.
What participation looks like in practice: In this Pakistan study, community members weren’t just interview subjects. They were co-researchers who identified problems, designed data collection processes, interpreted findings, and developed solutions. The iterative cycles of planning, action, and reflection allowed the community to test solutions, learn from what worked and didn’t, and refine their approach.
Think about what makes this different from traditional research. A conventional study might survey community members about COVID-19 prevention barriers, analyze responses using predetermined frameworks, and make recommendations from outside. The participatory approach positioned community members as experts on their own context who could identify barriers researchers might never have thought to ask about and design solutions that fit their specific social structures and resources.
A scoping review of participatory approaches in primary health care across LMICs reinforces the distinction: programs that engaged communities only in awareness-building or health education rarely produced durable change, while those that engaged communities in jointly identifying problems, prioritizing actions, and monitoring progress generated genuine improvements in service quality, participation, and accountability (Saif-Ur-Rahman et al., 2022).
Strengths: Centers community priorities and knowledge; builds community capacity; increases likelihood findings will be used locally; addresses power imbalances in research relationships; generates culturally appropriate solutions; creates shared ownership of research and interventions; can transform research from extractive to empowering; builds trust between researchers and communities.
Limitations: Time-intensive with extensive community engagement required; can be challenging to balance academic rigor with community priorities (academic journals may not value community-defined outcomes); may create tensions if academic and community goals diverge; requires sustained commitment and resources; demands researchers develop skills in facilitation and collaboration beyond traditional research training; can be challenging to maintain participatory principles under funding and timeline pressures.
CBPR sounds great in principle, but practitioners will tell you it’s challenging in practice. Academic researchers face pressure to publish in high-impact journals, secure grants, and complete projects on schedule. Community members may prioritize action over publication, immediate problems over research questions, and local knowledge over academic theories. Navigating these tensions requires explicit negotiation, sustained relationship-building, and willingness to compromise from all partners.
CHOOSING AMONG QUALITATIVE APPROACHES
You might be thinking, “These approaches seem to overlap. How do I actually choose?” Fair question! Here is a rough guide.
Start with your research question:
- If your question is about cultural context and social systems \(\rightarrow\) consider ethnography
- If you’re trying to build new theory about processes \(\rightarrow\) consider grounded theory
- If you want to understand lived experience of a phenomenon \(\rightarrow\) consider phenomenology
- If you’re examining a specific, complex case in context \(\rightarrow\) consider case study
- If community empowerment and action are priorities \(\rightarrow\) consider participatory approaches
- If you’re studying policy or institutional processes \(\rightarrow\) consider document analysis
Consider your resources:
- Time: Ethnography requires months/years; phenomenology or grounded theory might be completed in weeks/months
- Access: Ethnography requires deep community relationships; document analysis requires archive access
- Skills: Different approaches demand different analytical skills
- Budget: Ethnographic fieldwork is expensive; document analysis may be cheaper
Think about outputs: What will make your findings credible to your audience? Policymakers may want case studies; clinicians may value phenomenology; theorists may appreciate grounded theory.
The qualitative approaches we’ve covered represent well-established traditions, each with distinctive analytical procedures and quality criteria. Understanding these distinctions helps you make informed methodological choices and conduct research that’s rigorous by the standards of the tradition you’re working within.
BEYOND THESE TRADITIONS
The approaches above are the most common in global health, but they aren’t the whole map. Photovoice hands cameras to community members and asks them to document their own realities; the photographs then anchor group discussion and advocacy. It is often used inside a CBPR frame and has been particularly powerful for amplifying voices that rarely shape health policy.
Other approaches you’ll encounter include narrative inquiry (people’s stories as the unit of analysis), discourse analysis (how language constructs and enacts power), arts-based methods like body mapping or theatre, and rapid qualitative inquiry, a family of accelerated techniques developed for outbreaks and humanitarian settings. The point is that the toolbox is larger than what we’ve covered. When a question doesn’t fit one of the traditions above, that’s a signal to look further before forcing a fit.
13.3 Mixed Methods Research: Combining Approaches
We’ve been treating qualitative and quantitative research as separate approaches, but many research questions require both. Numbers can tell you what and how much, but you often need qualitative insight to understand why and how. Mixed methods research systematically integrates quantitative and qualitative data collection and analysis to provide more comprehensive understanding than either approach alone.
WHAT IS MIXED METHODS RESEARCH?
Mixed methods research isn’t just doing a quantitative study and a qualitative study in parallel and stapling them together in the discussion section (though I’ve seen people try). True mixed methods involves integration—bringing qualitative and quantitative strands together in ways that create insights neither strand alone could provide.
RATIONALE FOR MIXED METHODS
Why go through the extra complexity of mixed methods? There are several compelling reasons:
Complementarity: Qualitative and quantitative methods address complementary questions. A randomized trial might show an intervention improves health outcomes (quantitative) while qualitative interviews suggest that the intervention works by changing social norms around health-seeking behavior (qualitative). Together, you understand both that it works and generate ideas about how it works.
Development: Findings from one strand can shape the design of the other. You might run qualitative interviews first to identify the constructs that matter to a population, then build a survey instrument around them. Or run a quantitative scan to find outlier communities, then design a qualitative case study to understand why they’re outliers.
Expansion: Mixed methods can extend the range of what a study covers by using different methods for different sub-questions. A type-2 hybrid implementation-effectiveness trial, for example, might use a randomized design to test whether a treatment works and a parallel qualitative strand to study how clinics actually adopted it. The strands aren’t two angles on the same finding. They’re answering distinct questions that together describe the full picture.
Triangulation: Sometimes called validation, this involves using one method to check or validate findings from another. If survey data suggest a barrier to care and interview data confirm and elaborate on this barrier, your confidence in the finding increases.
Explaining unexpected findings: Quantitative analysis might reveal a surprising pattern—say, an intervention worked in some settings but not others. Qualitative follow-up can help explain why.
WHEN TO USE MIXED METHODS
Mixed methods makes sense when a single method genuinely can’t answer your question. If you need both breadth and depth—a survey-scale view of who is affected and an in-depth view of how they experience it—neither approach alone will do. The same logic applies when you care about outcomes and mechanisms: a trial can tell you that an intervention works, but only qualitative work can tell you what about it mattered to participants. Complex interventions in diverse contexts almost always demand mixed methods. The same program rarely behaves the same way across settings, and you need both quantitative variation and qualitative explanation to see why.
There is also a more pragmatic reason. Different audiences want different kinds of evidence. Funders and ministries of health often want numbers; communities and frontline implementers usually want stories that reflect their experience. A well-designed mixed-methods study can speak to both without contorting itself.
A word of caution. Mixed methods is expensive and demanding, requiring teams with both quantitative and qualitative skills and the methodological sophistication to integrate them. Don’t reach for it just because both kinds of data feel useful. Reach for it when your question genuinely requires both.
THREE BASIC MIXED METHODS DESIGNS
Mixed methods designs vary in complexity, but three basic designs cover most situations you’ll encounter. These differ in timing (which method comes first) and priority (which method is emphasized). Think of these as blueprints for how you’ll structure your study. Each blueprint serves different purposes and requires different implementation strategies.
Convergent Design: Parallel Integration
What it is: You collect quantitative and qualitative data simultaneously (or close to it), analyze each separately, then bring results together to compare and contrast findings. Think of it as running two parallel investigations of the same phenomenon, then merging them to see where they converge and where they diverge.
Imagine two parallel streams flowing side by side, then merging into a single river. That’s convergent design. Both strands maintain their own analytical integrity before converging at the integration point, where findings are compared, contrasted, and synthesized.
When to use it: Choose convergent design when you want to validate or corroborate findings across methods (triangulation), or when you want to obtain different but complementary data about the same phenomenon. It’s particularly useful when you have resources to conduct both strands simultaneously and when your research question benefits from multiple perspectives on the same issue.
Implementation steps:
Design both strands: Develop your quantitative instruments (surveys, assessments) and qualitative protocols (interview guides, observation frameworks) simultaneously, ensuring they address the same overarching research questions from different angles.
Collect data in parallel: Conduct quantitative and qualitative data collection during the same time period. You might even recruit from the same population, though not necessarily the same individuals.
Analyze independently: Complete quantitative analysis (statistical analysis) and qualitative analysis (coding, thematic analysis) separately, without letting one strand influence the other. This maintains the integrity of each approach.
Create joint displays: This is where integration happens. Develop tables, matrices, or visual displays that bring quantitative and qualitative findings side by side for direct comparison. For example, a table might show survey results about healthcare satisfaction alongside interview themes about patient experiences.
Interpret integration: Look for convergence (both strands support the same conclusion), divergence (strands suggest different conclusions), and expansion (one strand provides deeper understanding of patterns identified in the other).
Integration point refers to the specific stage in your analysis where you bring qualitative and quantitative findings together. In convergent design, this typically happens after independent analysis of both strands is complete.
Example in action: The vaccination governance study in northwest Syria used a convergent approach (Baatz et al., 2024). Researchers collected quantitative vaccination coverage data while simultaneously conducting 14 key informant interviews and observations about governance processes. The quantitative data provided evidence about program reach across multiple dimensions—effectiveness, efficiency, inclusiveness, data availability. Meanwhile, qualitative interviews explored stakeholder perspectives on these same governance dimensions. Integration happened when researchers created joint displays comparing vaccination coverage data against stakeholder reports of governance effectiveness, revealing crucial discrepancies: despite positive stakeholder perceptions of program effectiveness, quantitative records revealed lack of demographic data and over-reliance on paper-based systems. This divergence prompted deeper investigation that neither strand alone would have revealed.
Strengths: Efficient (data collection happens in parallel rather than sequential); provides validation through triangulation; relatively straightforward to implement; often appeals to stakeholders who value both numbers and stories; can compensate for weaknesses in each individual method.
Challenges: Requires substantial resources to collect both types of data simultaneously; integration can be difficult if findings diverge (but divergence can also be illuminating!); need expertise in both quantitative and qualitative methods on your team; timing must be carefully coordinated; determining how to reconcile contradictory findings requires sophisticated methodological thinking.
Explanatory Sequential Design: Numbers Then Stories
What it is: You start with quantitative data collection and analysis, then use qualitative methods to explain or elaborate on the quantitative findings. The quantitative phase identifies patterns, associations, or outcomes; the qualitative phase explores why those patterns exist or how the mechanisms work.
Sequential means you complete Phase 1 before beginning Phase 2. The qualitative phase is informed by quantitative results—you might adjust your sampling strategy, interview questions, or focus group topics based on what the quantitative data revealed.
Picture it as a detective story: quantitative data reveals what happened (the pattern or outcome), then qualitative inquiry investigates why or how it happened. The second phase is deliberately designed to explain the first.
When to use it: Choose explanatory sequential when you have quantitative results that need explanation—maybe surprising findings, significant differences between groups, outlier cases you want to understand better, or relationships that require mechanistic understanding. It’s particularly valuable when stakeholders are comfortable with quantitative data and need qualitative insight to interpret or act on the numbers.
Implementation steps:
Conduct quantitative study: Design and implement your quantitative data collection and analysis first. This might be a survey, secondary data analysis, clinical trial, or cohort study. Complete the statistical analysis before moving forward.
Identify areas requiring explanation: Review quantitative results to identify patterns requiring qualitative follow-up. This might include unexpected findings, significant group differences, outlier cases, non-significant findings where you expected effects, or associations requiring mechanistic understanding.
Design targeted qualitative phase: Develop qualitative protocols specifically designed to explain quantitative findings. Your interview guides, focus group topics, or observation frameworks should directly address questions raised by the quantitative phase. Purposively sample participants who can illuminate the patterns you’re investigating.
Conduct qualitative inquiry: Collect and analyze qualitative data. The goal is understanding why the quantitative patterns emerged or how the mechanisms operate.
Integrate through explanation: Write up findings showing how qualitative results explain, contextualize, or elaborate quantitative patterns. The narrative typically moves from quantitative findings to qualitative explanations.
Example in action: A study of perioperative care in three teaching hospitals in southern Ethiopia is a clean example of explanatory sequential design (Mulugeta et al., 2024). The quantitative phase used the WHO/Harvard Surgical Assessment Tool to walk through each hospital and measure capacity across five domains: infrastructure, workforce, service delivery, financing, and information management. The survey identified specific gaps. The surgical, obstetric, and anesthesia specialist workforce was just 0.58 per 100,000 population; surgical volume was 115 procedures per 100,000; blood products were available only 25–50% of the time; perioperative management protocols were used in just 1–25% of cases; and over 90% of patients lacked health insurance.
These numbers raised the obvious question: why? Twenty semi-structured interviews with surgeons, anesthetists, and nurses formed the qualitative second phase, designed to explain the patterns. The interviews surfaced what a walkthrough survey could never see. Protocols were not just underused; clinicians described a poor patient-safety culture, fragmented interprofessional communication, and frustration that repeated quality-improvement projects produced “no tangible change.” Blood scarcity wasn’t only a logistics failure but a downstream effect of declining donations, broken collection practices, and unresolved national-level policy issues. Three themes emerged that the survey instrument could not have produced at all: the influence of national governance, sociopolitical turmoil, and global market volatility on local perioperative care. The team integrated the two strands through a joint display table, mapping each quantitative finding alongside the qualitative themes that explained it.
Strengths: Phased approach is easier to manage than simultaneous data collection; qualitative phase can be precisely designed to address specific gaps from quantitative findings; often resonates with traditionally quantitative audiences (numbers first, then explanation); results in compelling narratives showing both “what” and “why”; efficient use of resources by focusing qualitative work on areas most needing explanation.
Challenges: Time-consuming (sequential rather than parallel); requires waiting for quantitative results before designing qualitative phase, which delays final results; initial quantitative phase may miss important phenomena that qualitative exploration could have identified earlier; requires flexibility in timeline and budget to conduct two distinct phases.
Exploratory Sequential Design: Stories Then Numbers
What it is: You start with qualitative exploration to identify important variables, develop measures, or understand a phenomenon, then follow with quantitative data collection to test patterns or generalize findings to a larger population. The qualitative phase explores new territory; the quantitative phase tests or measures what you discovered.
Think of it as exploring uncharted territory first (qualitative), then mapping it systematically (quantitative). The first phase generates hypotheses, identifies variables, or develops measures; the second phase tests them.
When to use it: Choose exploratory sequential when you’re studying something poorly understood where existing quantitative measures may be inadequate or culturally inappropriate, or when you want to develop an intervention or instrument based on qualitative insights and then test it quantitatively. It’s particularly valuable in global health research where Western-developed measures or theories may not apply in different cultural contexts.
Implementation steps:
Conduct qualitative exploration: Begin with in-depth qualitative research to understand the phenomenon. This might involve interviews, focus groups, ethnographic observation, or participatory methods. Analyze data to identify key themes, patterns, or variables.
Generate quantitative elements from qualitative findings: Use qualitative results to develop quantitative instruments, measures, hypotheses, or interventions. For example, interview themes might become survey items, or qualitative insights about barriers might inform intervention components.
Design and pilot quantitative tools: Develop survey scales, intervention protocols, or assessment tools based on qualitative insights. Pilot them with a small sample to ensure they’re understandable and appropriate.
Implement quantitative phase: Collect quantitative data from a larger sample to test patterns identified qualitatively, measure prevalence, assess associations, or evaluate intervention effectiveness.
Integrate by showing continuity: Demonstrate how quantitative elements emerged from qualitative work and what quantitative findings add. Did the quantitative phase confirm qualitative patterns? Reveal unexpected prevalence? Show effectiveness at scale?
Example in action: A study developing a perinatal depression screening tool in rural Kenya is a good example of exploratory sequential design (Green et al., 2018). The qualitative phase started with the suspicion that screening instruments designed in high-income settings, like the Edinburgh Postnatal Depression Scale (EPDS) and the Patient Health Questionnaire-9 (PHQ-9), might miss how Kenyan women actually describe perinatal distress. The team assembled 365 items from 17 existing screening tools and pulled out 171 unique cover terms. Two free-listing and card-sorting groups of pregnant women and new mothers, and six groups of community health volunteers in Bungoma County, then ranked which of those terms matched local experience and generated new local idioms of distress that no Western tool included. A panel of eleven Kenyan mental health professionals reviewed the results and signed off on a 60-item blended pool: items from the EPDS and PHQ-9, plus locally-generated items reflecting the language women had used.
Instrument development is a common use of exploratory sequential design. Qualitative interviews reveal how participants think and talk about a construct, then those insights inform survey item wording, ensuring questions are culturally appropriate and meaningful.
The quantitative phase then validated this blended tool against a gold standard. A random sample of 193 pregnant women and new mothers completed all three scales (EPDS, PHQ-9, and the new 60-item pool) and a Structured Clinical Interview for DSM-5 administered by trained Kenyan counselors. Item analysis and combinatorial optimization across more than 600,000 possible scale combinations narrowed the pool to a 9-item Perinatal Depression Screening (PDEPS) tool. The PDEPS outperformed both established scales: sensitivity and specificity of 0.90/0.90, compared to 0.70/0.72 for the EPDS and 0.70/0.73 for the PHQ-9 in this population. The qualitative phase ensured the tool was grounded in how Kenyan women actually talk about distress; the quantitative phase confirmed that grounding paid off in diagnostic accuracy.
Strengths: Grounds subsequent quantitative work in contextual understanding; ensures measures and interventions are culturally appropriate rather than imposed from external frameworks; can identify unexpected variables or phenomena missed in existing literature; generates hypotheses from lived experience rather than assumptions; particularly valuable for understudied populations or phenomena.
Challenges: Time-consuming with sequential phases requiring completion of qualitative work before beginning quantitative design; may be challenging to convince traditionally quantitative audiences (especially funders or ethics committees) to start with qualitative exploration; requires resources and expertise for both phases; moving from qualitative insights to quantitative operationalization requires thoughtful methodological translation.
Integration is Key
You might forget the official names of these mixed methods designs, but remember this: the defining feature of mixed methods isn’t merely doing both qualitative and quantitative research, it’s integrating them. Ask yourself: How will I bring these strands together? What insights emerge from integration that couldn’t emerge from separate analyses? If you can’t answer these questions, maybe you’re doing two separate studies rather than true mixed methods.
Integration happens through:
- Joint displays: Tables or figures showing quantitative and qualitative findings side by side
- Data transformation: Converting one type of data to the other (e.g., quantifying qualitative themes or “qualitizing” quantitative data)
- Narrative weaving: Writing that explicitly connects findings across strands
- Meta-inferences: Interpretations that draw on both quantitative and qualitative insights
The best mixed methods studies make integration explicit and central, not an afterthought.
13.4 Quality and Rigor in Qualitative and Mixed Methods Research
Here’s a question students often ask: if qualitative research doesn’t have statistical significance, confidence intervals, or effect sizes, how do we judge whether it’s any good? Qualitative research does have quality criteria. They are just different from the ones quantitative research uses, and they reflect a different epistemology.
QUALITATIVE QUALITY CRITERIA
Most qualitative researchers work from a four-part framework laid out by Lincoln and Guba: credibility, transferability, dependability, and confirmability. Each maps loosely onto a quantitative concept, but the underlying logic is different. Quantitative research tries to control for bias through design. In contrast, qualitative research treats bias as inevitable and demands that researchers make their decisions, assumptions, and analytical moves transparent enough that readers can judge them.
| Dimension | Quantitative Term | Qualitative Term | Key Question |
|---|---|---|---|
| Truth value | Internal validity | Credibility | Are causal claims valid? |
| Applicability | External validity | Transferability | Do findings apply elsewhere? |
| Consistency | Reliability | Dependability | Could others reach similar conclusions? |
| Neutrality | Objectivity | Confirmability | Are findings grounded in data, not bias? |
Credibility asks whether the findings believably represent participants’ perspectives. The classical tactics for building it are prolonged engagement (enough sustained immersion to see patterns and distinguish typical from atypical experience), triangulation across multiple data sources, member checking with participants, and peer debriefing with colleagues who are willing to challenge your interpretations. The Syria vaccination study, for example, triangulated key-informant interviews, a validation workshop, and ethnographic observations (Baatz et al., 2024); when all three converged on the same governance gaps, that convergence was the source of credibility.
Transferability asks whether findings travel. The key tactic is thick description: enough detail about context, participants, and processes that readers in different settings can judge applicability themselves. Purposive sampling that deliberately captures variation strengthens transferability further. The peer mental health worker study in Goa and Rawalpindi sampled across distinct contexts and found that the relative importance of peer-worker characteristics varied by setting (Singla et al., 2014), a finding that itself tells readers something about when and where the work might transfer.
Dependability asks whether another researcher following the same process could reach similar conclusions. The answer hinges on documentation. A serious qualitative project maintains an audit trail of methodological decisions, codebooks showing how categories were defined and applied, analytic memos showing how interpretation evolved, and drafts showing how the argument was sharpened. “I read the transcripts and themes emerged” does not meet this standard.
Confirmability asks whether findings are grounded in data rather than the researcher’s prior commitments. Three habits keep work confirmable. The first is reflexivity: explicitly examining your own assumptions, positionality, and potential biases, and naming them in the paper. The second is clear linkage to data: every major claim should trace to a quote, field note, or document excerpt that readers can see. The third is negative case analysis: actively seeking and addressing data that do not fit your emerging interpretation. If your analysis claims all participants experienced stigma but three interviews suggest otherwise, those exceptions are the finding, not a problem to hide.
A scoping review of qualitative research’s contributions to clinical and epidemiological understanding makes the same point: rigor emerges through systematic collection, transparent analysis, and ongoing questioning of researcher assumptions, and only when work is conducted that way can qualitative findings reliably illuminate causal mechanisms, identify important subgroups, or explain how associations operate (Meuleman et al., 2025).
MIXED METHODS QUALITY CRITERIA
Mixed methods studies have to clear the quality bar for both strands and then clear one more bar that neither strand alone has to: the quality of integration. This is where most weak mixed methods studies fail. The numbers and the stories are each fine in isolation, but they sit side by side on the page without ever truly talking to each other.
Three dimensions of integration quality are worth checking.
Design integration asks whether the choice of design (convergent, explanatory sequential, exploratory sequential) actually fits the question, and whether the timing and relative priority of the strands are justified. The Syria vaccination study used convergent design because the team needed coverage data and stakeholder perspectives in parallel to understand governance comprehensively (Baatz et al., 2024); the Ethiopia perioperative study used explanatory sequential design because they had a quantitative picture of capacity gaps and needed clinicians’ perspectives to explain them (Mulugeta et al., 2024).
Analytical integration asks whether the team actually brought the two strands into contact (usually through joint displays, narrative weaving, or data transformation) and whether they explicitly explored convergence, divergence, and expansion.
Interpretive integration asks whether the discussion leverages both strands to produce meta-inferences: claims that neither strand could have made alone. When quantitative coverage data showed reach but qualitative interviews revealed documentary gaps, the Syria team did not dismiss the contradiction; they explored it, and the gap became the finding.
A useful self-test: if you removed either the quantitative or the qualitative section from your paper, would the remaining section still tell a complete story? If yes, you probably have not integrated. True integration means both strands are necessary for the full picture.
13.5 Closing Reflection
Global health faces complex challenges: pandemics, health systems under strain, the long road to universal health coverage, persistent inequities. These do not fit neatly into single methodological boxes. Understanding whether an intervention works calls for randomized trials. Understanding how it works calls for qualitative inquiry into mechanisms. Understanding whether it will work in a new context calls for mixed methods implementation research. Understanding how to adapt it calls for participatory approaches that center community knowledge.
The most interesting global health research is increasingly comfortable moving across these approaches, not as a compromise but as a strength. As you develop your own work, resist the pressure to pick one methodological camp and stick to it. Let the question lead. Sometimes that means quantitative work; sometimes qualitative; often mixed methods; and increasingly, participatory approaches that share research power with the communities the work is meant to serve. The goal is not methodological purity. It is evidence that is rigorous, contextually grounded, and useful for improving health and advancing equity.