Qualitative Methodologies and Research Designs

Introduction

Qualitative research investigates how people interpret experience, construct meaning, interact with institutions, and respond to social and organizational conditions. In health care management, qualitative inquiry is particularly valuable because many important outcomes—trust, dignity, safety culture, professional identity, service accessibility, organizational change, and patient experience—cannot be understood adequately through numerical indicators alone.

A qualitative methodology is not simply a collection of interview or observation techniques. It is a coherent relationship among a research question, philosophical assumptions, design, sampling strategy, data-generation methods, analytic procedures, and criteria for quality. The central methodological task is therefore alignment: the design must fit what the researcher is trying to know.

For example, a manager who asks, “What is it like for patients to live with treatment-resistant chronic pain?” is asking about lived experience and may select phenomenology. A researcher asking, “How do nurses develop strategies for managing medication interruptions?” may use grounded theory to generate an explanatory process model. A study of communication norms in an emergency department may require ethnography, whereas an in-depth investigation of one hospital’s failed digital transformation may be a case study.

Qualitative designs overlap in practice, but they are not interchangeable. Each has a distinctive intellectual purpose, preferred form of evidence, analytic emphasis, and relationship with participants.

Learning Objectives

By the end of this chapter, the reader should be able to:

1. Explain the philosophical and methodological foundations of qualitative inquiry.
2. Distinguish phenomenology, grounded theory, ethnography, case study, narrative inquiry, qualitative description, participatory action research, and critical inquiry.
3. Match qualitative designs to health-care management questions and contexts.
4. Evaluate sampling, data-generation, analysis, reflexivity, rigor, and ethics across designs.
5. Develop a defensible design rationale for a postgraduate health research proposal.

1. Foundations of Qualitative Design

1.1 Ontology, epistemology, and methodology

Ontology concerns assumptions about reality. A qualitative researcher may view reality as singular and observable, multiple and socially constructed, historically situated, or shaped by unequal power relations. Epistemology concerns how knowledge can be produced. Interpretivist epistemologies emphasize understanding meanings from participants’ perspectives; constructionist approaches examine how meanings are produced through language, interaction, and institutions; critical approaches ask how power and ideology shape what is accepted as knowledge.

Methodology is the logic that connects these assumptions to research practice. Methods are the practical techniques used to generate and analyze data, such as interviews, focus groups, observation, document analysis, and visual methods. The same method can be used within different methodologies. An interview does not automatically make a study phenomenological, grounded-theory based, or narrative.

1.2 The logic of qualitative inquiry

Qualitative inquiry is commonly characterized by:

• attention to context rather than isolated variables;
• emphasis on participants’ meanings and interpretations;
• flexible, iterative data collection and analysis;
• purposive or theoretically informed sampling;
• the researcher as an instrument of interpretation;
• rich textual, visual, or audio data; and
• findings expressed through concepts, themes, narratives, typologies, or theories.

Qualitative research is not necessarily small-scale, subjective in an unexamined sense, or incapable of systematic analysis. Its rigor depends on transparency, reflexivity, coherence, ethical practice, and a convincing connection between evidence and interpretation.

1.3 Design alignment

A useful design decision begins with the question’s dominant verb:

This table is a starting point, not a mechanical rule. Complex health-care problems may require a consciously stated combination of perspectives, but researchers should avoid combining designs merely to appear comprehensive.

2. Phenomenology: Understanding Lived Experience

Phenomenology studies how people experience and make sense of a phenomenon. Its focus is not the external measurement of an event but the meaning and structure of experience as lived by individuals. Health-care topics suitable for phenomenology include living with a stoma, receiving a serious diagnosis, experiencing moral distress, becoming a first-time caregiver, or undergoing professional burnout.

Classical phenomenology is associated with Edmund Husserl and the description of experience. Interpretive or hermeneutic phenomenology, influenced by Martin Heidegger and later scholars, emphasizes interpretation and the researcher’s situated understanding. In applied health research, researchers should state clearly whether they aim primarily to describe essential features of experience or to interpret how experience is situated in social, historical, and professional worlds.

Design features

Participants are selected because they have directly experienced the phenomenon. Interviews are often open-ended and encourage concrete descriptions: “Can you describe the moment you realized the treatment was not working?” Researchers may also use diaries, reflective writing, or observation. Analysis commonly involves identifying significant statements, meaning units, experiential themes, and a synthesized account of the phenomenon.

Bracketing, sometimes called epoché, refers to making preconceptions visible and attempting to prevent them from silently determining interpretation. In interpretive phenomenology, complete suspension of pre-understanding is often regarded as impossible; reflexive engagement is therefore more realistic than claiming total neutrality.

Strengths and limitations

Phenomenology can reveal dimensions of care that routine satisfaction surveys overlook, such as vulnerability, altered identity, uncertainty, or loss of control. Its limitations include the difficulty of defining a shared phenomenon, the risk of reducing complex accounts to generic themes, and the danger of using the label without a genuinely phenomenological analysis.

3. Grounded Theory: Developing Explanatory Theory

Grounded theory is designed to generate an explanatory account of social processes from systematically collected and analyzed data. It is appropriate when existing theories do not adequately explain how people act, adapt, negotiate, or make decisions in a health-care setting.

A grounded theory study might ask: “How do multidisciplinary teams develop shared responsibility for delayed discharge?” The intended result is not merely a list of barriers but a conceptual explanation of stages, conditions, strategies, and consequences.

Core logic

Data collection and analysis proceed iteratively. Early interviews inform later sampling and questioning. Researchers code data, compare incidents, develop categories, identify relationships, and seek cases that refine or challenge emerging explanations. Constant comparison means examining similarities and differences across interviews, participants, events, and categories.

Different grounded-theory traditions vary. Glaserian approaches emphasize emergence and theoretical sensitivity; Strauss and Corbin provide more structured coding procedures; constructivist grounded theory, associated with Charmaz, recognizes that the emerging theory is co-constructed by researcher and participants. A proposal should identify the chosen tradition rather than treating all grounded theory procedures as identical.

Theoretical sampling and saturation

Theoretical sampling selects additional participants or data sources because they can develop an emerging category. For example, if “informal workarounds” becomes central in a study of electronic prescribing, the researcher might sample junior clinicians, pharmacists, and units with different levels of system integration.

Saturation is not a universal numerical target. It refers to an analytic judgment that additional data are no longer adding meaningful properties, relationships, or variation to the developing categories. Researchers should document how this judgment was reached.

Strengths and limitations

Grounded theory is useful for management because it can explain implementation processes, coordination, adaptation, and organizational responses. It is demanding: premature closure, superficial coding, and confusing thematic description with theory are common problems. A grounded theory should show relationships among concepts, not simply present many categories.

4. Ethnography: Studying Culture and Practice

Ethnography examines shared meanings, practices, norms, language, rituals, and forms of membership within a social group. In health care, the relevant “culture” may be a clinical department, ambulance service, professional community, patient advocacy group, or integrated-care network.

Ethnography traditionally involves prolonged engagement and participant observation. Researchers seek to understand what people do, what they say they do, and the often-unspoken assumptions that connect the two. Field notes, informal conversations, interviews, documents, artifacts, and observations are integrated into a contextual account.

Emic and etic perspectives

The emic perspective represents participants’ meanings from within the setting. The etic perspective is the analytic interpretation brought by the researcher. Strong ethnography moves between both perspectives without treating either as automatically complete.

An ethnographer in an emergency department might observe handovers, informal allocation of tasks, interactions with relatives, and moments when formal protocols are bypassed. The purpose is not to label staff as compliant or non-compliant, but to understand how local norms and pressures shape practice.

Focused and institutional ethnography

Classic ethnography may require extended residence in a setting. Focused ethnography concentrates on a specific problem and may use shorter but intensive fieldwork. Institutional ethnography begins with people’s everyday experiences and traces how texts, policies, classifications, and administrative systems organize those experiences.

Strengths and limitations

Ethnography reveals tacit routines and contradictions between policy and practice. It is especially valuable for patient safety, infection prevention, interprofessional collaboration, and organizational culture. Access, observer effects, confidentiality in small teams, and the time required for fieldwork are significant challenges.

5. Case Study: Investigating Bounded Systems

Case study research explores a contemporary phenomenon in its real-life context, especially when the boundaries between phenomenon and context are not clear. A case may be a hospital, program, ward, policy implementation, merger, patient-safety incident, or digital health project. The case must be bounded by time, place, organization, population, or activity.

Case studies may be single or multiple, and exploratory, descriptive, or explanatory. A hospital’s implementation of virtual outpatient clinics could be studied as a single case using interviews with managers and clinicians, patient records, meeting minutes, implementation documents, and service data. Multiple cases might compare implementation in three hospitals.

Case logic and evidence

A case is not simply a topic. “Patient safety” is a topic; “the implementation of a medication reconciliation program in Hospital A between January and December 2025” is a bounded case. The researcher should establish the unit of analysis and explain why the case is strategically informative.

Triangulation may involve comparing different data sources, investigators, theories, or methods. Triangulation does not mean that all sources must agree. Contradictions can reveal competing perspectives or changes over time.

Strengths and limitations

Case study is well suited to complex management questions involving context, implementation, and organizational relationships. Its findings can be transferable through analytical reasoning rather than statistical generalization. Weak case studies lack clear boundaries, rely on one convenient interview, or use “case study” as a label for any qualitative project.

6. Narrative Inquiry: Experience as Story

Narrative inquiry studies how people organize experience through stories. Narratives have plots, characters, turning points, temporal sequences, identities, and audiences. A narrative study may explore how a patient reconstructs life after a stroke, how a chief executive describes organizational crisis, or how migrant health workers narrate professional belonging.

Narrative researchers attend both to what is said and how it is said. They may analyze chronology, metaphors, silence, genre, positioning, and the relationship between personal stories and dominant cultural narratives. A participant’s account of “surviving” a restructuring is not merely a report of events; it is also an identity claim and a way of assigning responsibility and meaning.

Narrative inquiry can use life-history interviews, longitudinal conversations, diaries, letters, blogs, or institutional narratives. Ethical care is essential because stories can contain identifiable and emotionally sensitive material. Researchers should avoid fragmenting a participant’s account so extensively that its narrative structure disappears.

7. Qualitative Description: Staying Close to the Data

Qualitative description aims to provide a comprehensive, low-inference account of an event, service, experience, or group in everyday language. It is often appropriate for service evaluation, needs assessment, implementation research, and questions such as: “How do patients describe barriers to attending cardiac rehabilitation?”

This design does not mean “no theory” or “easy research.” It requires purposeful sampling, clear data-generation, careful coding, and disciplined reporting. Findings are commonly organized into descriptive categories and themes supported by representative quotations.

Qualitative description is especially useful when the intended audience includes clinicians, service managers, policymakers, and community organizations that need a clear account for action. Its limitation is that it may not explain deep structures, cultural systems, lived experience, or causal processes as fully as other designs. Researchers should not make stronger interpretive claims than the design supports.

8. Participatory Action Research: Knowledge Through Collaboration and Change

Participatory action research (PAR) treats participants as co-researchers rather than merely sources of data. It links inquiry with action and commonly proceeds through cycles of planning, action, observation, reflection, and revised planning. PAR is suitable for improving culturally safe services, addressing health inequities, redesigning patient pathways, or strengthening community control over health programs.

A PAR project may involve patients, community representatives, clinicians, and managers jointly defining the problem, collecting evidence, implementing a change, and evaluating its effects. The research question can evolve as participants’ priorities become clearer.

Participation has degrees. Consultation is not the same as shared decision-making, and inviting participants to review findings does not automatically make a project participatory. Researchers should specify who controls the research questions, data, interpretation, dissemination, and decisions about action.

PAR offers practical relevance, empowerment, and locally grounded knowledge. Challenges include unequal power among partners, competing timelines, role conflicts, tokenistic participation, and the possibility that action pressures compromise careful inquiry. Governance agreements should address ownership, authorship, confidentiality, compensation, and the right to withdraw.

9. Critical Inquiry: Knowledge, Power, and Transformation

Critical inquiry examines how power relations, institutional structures, language, history, and ideology shape health and health-care practice. It asks not only “What is happening?” but also “Who benefits, who is marginalized, whose knowledge is privileged, and what alternatives are possible?”

Critical studies may draw on critical theory, feminist theory, critical race theory, intersectionality, postcolonial theory, disability studies, or Foucauldian analysis. A study of triage categories, for example, might investigate how apparently neutral classifications distribute credibility, urgency, and access across gendered, racialized, or classed groups.

Critical discourse analysis examines how language constructs problems and subjects. Institutional analysis traces how policies, forms, metrics, and professional categories organize everyday work. Intersectional analysis examines how multiple social positions interact rather than treating categories such as gender or ethnicity in isolation.

Critical inquiry can expose structural injustice and support institutional change. Researchers must nevertheless avoid assuming that participants share the researcher’s political interpretation. Reflexivity, careful evidence, and respectful engagement remain essential.

10. Comparing Designs in Health-Care Contexts

11. Sampling, Data Generation, and Analysis

Sampling

Purposive sampling seeks information-rich participants. Maximum-variation sampling includes diverse perspectives; criterion sampling includes everyone meeting defined conditions; typical-case sampling examines ordinary settings; critical-case sampling examines cases of particular strategic importance; and snowball sampling can assist access to hidden populations. Grounded theory may use theoretical sampling as categories develop.

Sample adequacy depends on the design, question complexity, heterogeneity of the population, quality of data, and analytic purpose. Researchers should justify adequacy through information power, depth, variation, and analytic sufficiency rather than relying on an arbitrary universal sample size.

Data generation

Interviews may be structured, semi-structured, or unstructured. Focus groups reveal interaction and collective norms but may silence minority views. Observation captures practice and context, while documents reveal institutional classifications and official accounts. Digital data, patient portals, online communities, and electronic records may be useful but raise substantial privacy and consent issues.

Analysis

A general analytic sequence may include familiarization, initial coding, category development, comparison, interpretation, and synthesis. However, analysis is design-specific. Phenomenology seeks experiential meaning; grounded theory builds relationships among categories; ethnography interprets cultural patterns; narrative inquiry preserves plot and temporality; and critical inquiry examines power in discourse and institutions.

Computer-assisted qualitative data analysis software can support organization, retrieval, memoing, and audit trails. It does not interpret meaning, establish credibility, or replace researcher judgment. Generative AI may assist with transcription cleaning, preliminary text organization, or question development only under strict governance. Sensitive health data should not be uploaded to systems without approved privacy, security, and data-processing arrangements. AI-generated coding must be treated as provisional and checked against the original data.

12. Quality, Reflexivity, and Ethics

Quality criteria vary across paradigms, but commonly include credibility, dependability, confirmability, transferability, reflexivity, and authenticity. Credibility may be strengthened through prolonged engagement, triangulation, peer debriefing, negative-case analysis, member reflection, and close attention to disconfirming evidence. Transferability depends on thick description that allows readers to judge relevance to another context.

Reflexivity requires researchers to examine how their professional role, institutional position, identity, assumptions, relationships, and emotional responses influence the study. A senior manager interviewing subordinates may obtain data shaped by fear or loyalty. A clinician studying patients may overlook institutional assumptions that feel normal. Reflexive journals, team discussions, positionality statements, and transparent reporting can make these influences visible.

Ethical issues include informed consent, privacy, confidentiality, distress, dual relationships, coercion, data ownership, incidental findings, and the consequences of publication. In small clinical teams, removing names may not prevent deductive identification. Researchers must consider whether quotations, job titles, rare conditions, or distinctive events make participants recognizable.

Ethics is not completed when an approval letter is obtained. Qualitative projects evolve, relationships change, and new risks emerge during fieldwork. Consent should be treated as an ongoing process, particularly in observation, longitudinal work, PAR, and digital research.

13. An Applied Case: Choosing a Design for Hospital-at-Home Implementation

Background

This illustrative case concerns a regional hospital introducing hospital-at-home services for older adults with acute medical conditions. The program promises improved patient convenience and reduced inpatient demand, but early implementation is uneven. Some clinicians support it, while others worry about clinical risk, workload, and inequitable access.

Problem

The management team initially proposes the question: “Why is hospital-at-home implementation inconsistent across clinical teams?” This question includes organizational processes, professional meanings, patient experience, and power relations. No single design automatically answers every aspect.

Design reasoning

A case study could examine the program as a bounded implementation case using policy documents, interviews, observations, and service records. If the central aim were to develop a theory of how teams negotiate risk and responsibility, grounded theory would be more appropriate. If the focus were the lived experience of older adults receiving care at home, phenomenology could be selected. If the project aimed to redesign the service with patients and staff, PAR would be appropriate. A coherent program of research could contain several linked studies, but each should retain a clear question and design.

Suppose the management team selects a qualitative case study. The case is bounded as the hospital-at-home program during its first twelve months. Purposive sampling includes physicians, nurses, allied health professionals, patients, family caregivers, administrators, and community providers. Data include interviews, observations of referral and handover meetings, implementation documents, and patient information materials.

Analysis and outcome

Analysis identifies four interacting themes: uncertainty about clinical accountability, reliance on informal coordination, unequal suitability assessments, and the importance of family capacity. Comparison across professional groups reveals that “patient choice” is sometimes used to describe genuine preference and sometimes to justify limited service capacity. The final case account shows that inconsistent implementation is not simply staff resistance; it results from unclear accountability, fragmented information systems, and assumptions about capable families.

The management response includes a clarified escalation protocol, shared referral criteria, protected coordination time, and a patient-facing explanation of alternatives. A subsequent PAR cycle could involve patients and staff in revising the referral process.

Learning points

1. The wording of the question should precede selection of the design.
2. A case requires explicit boundaries.
3. Multiple perspectives do not automatically constitute triangulation; their relationships must be analyzed.
4. Findings can support managerial action without claiming statistical generalization.
5. A future study may use another methodology if the research purpose changes.

14. Selecting and Defending a Design

A postgraduate proposal should justify design selection through a chain of reasoning:

1. State the practical and scholarly problem.
2. Identify the kind of knowledge sought: experience, process, culture, bounded explanation, story, description, action, or critique.
3. State the philosophical orientation and relevant theoretical commitments.
4. Define the unit of analysis and setting.
5. Explain participant selection and anticipated variation.
6. Align data-generation methods with the design.
7. Describe design-appropriate analysis.
8. Anticipate researcher positionality and ethical risks.
9. Explain how rigor will be established.
10. Specify what the study will and will not claim.

A design is defensible when its assumptions, procedures, and claims are coherent. A phenomenological study should not promise a causal model; a qualitative description should not claim to uncover an entire culture; and a PAR project should not describe participation that is merely consultation.

Contemporary Developments

Contemporary qualitative health research increasingly addresses complex systems, virtual care, algorithmic decision-making, climate-related health risks, migration, and structural inequity. Multimodal and digital ethnographies examine platforms, online communities, and hybrid clinical spaces. Co-design and community-led research challenge conventional divisions between researcher and researched. Indigenous and decolonizing methodologies emphasize sovereignty, relational accountability, collective benefit, and the limits of extractive research.

These developments do not make foundational design distinctions obsolete. Rather, they make methodological clarity more important. Researchers must explain how digital traces were obtained, how algorithms shape the setting, how communities govern data, and how institutional power affects participation.

Chapter Summary

Qualitative methodologies provide different ways of understanding health-care experience, practice, organization, and inequality. Phenomenology investigates lived experience; grounded theory develops explanations of social processes; ethnography studies culture and practice; case study examines bounded systems in context; narrative inquiry interprets experience through stories; qualitative description offers a clear, low-inference account; participatory action research links collaborative knowledge with change; and critical inquiry examines power, discourse, and structural injustice.

The appropriate design depends on the research question, philosophical assumptions, unit of analysis, context, participants, and intended contribution. Strong qualitative research aligns design, sampling, data generation, analysis, ethics, reflexivity, and claims. In health-care management, the value of qualitative inquiry lies not only in describing problems but also in explaining implementation, revealing hidden practices, improving services, centering patient and community knowledge, and challenging inequitable arrangements.

Key Terms

Methodology: The coherent logic linking philosophical assumptions, research design, methods, analysis, and claims.

Phenomenology: Inquiry into the meaning and structure of lived experience.

Grounded theory: An approach for developing explanatory concepts and theory from systematic, iterative analysis.

Ethnography: Study of culture, shared practices, meanings, and social life in context.

Case study: In-depth investigation of a bounded phenomenon in its real-life setting.

Narrative inquiry: Study of experience as organized and communicated through stories.

Qualitative description: A comprehensive, low-inference account of events, experiences, or services.

Participatory action research: Collaborative inquiry linked to collective action and change.

Critical inquiry: Research examining power, inequality, ideology, and possibilities for transformation.

Reflexivity: Critical examination of how the researcher influences and is influenced by the research.

Purposive sampling: Deliberate selection of information-rich participants or cases.

Theoretical sampling: Selection guided by emerging concepts in grounded theory.

Saturation: An analytic judgment that further data are no longer adding meaningful development to categories or themes.

Triangulation: Comparison of perspectives, sources, methods, investigators, or theories.

Credibility: The plausibility and trustworthiness of an interpretation.

Transferability: The extent to which readers can judge relevance to other contexts.

Bracketing: Making preconceptions explicit and managing their influence on interpretation.

Emic perspective: Understanding a setting from participants’ internal meanings.

Etic perspective: The researcher’s external analytic interpretation.

Positionality: The researcher’s social, professional, and epistemological location in relation to the study.

Review Questions

Short Answer Questions

1. Why is alignment between a research question and qualitative design important?
2. Distinguish methodology from method.
3. What is theoretical sampling, and in which design is it especially associated?
4. Give two examples of ethical risks in qualitative health research.
5. What distinguishes qualitative description from phenomenology?

Long Answer Questions

1. Compare phenomenology and grounded theory in terms of purpose, sampling, analysis, and findings.
2. Explain how ethnography could be used to study patient safety culture in a hospital.
3. Discuss the strengths and limitations of case study research for health-care management.
4. Explain the role of participation, power, and action in participatory action research.
5. Develop and defend a qualitative design for studying the implementation of a new electronic health-record system.

Analytical / Application Questions

1. A researcher wants to understand how families experience the transition from hospital to home after a child’s complex surgery. Which design would you recommend, and why? What data-generation and analytic strategies would be appropriate?
2. A hospital reports that a sepsis protocol is followed in policy documents but inconsistently applied in practice. Compare how an ethnography, case study, and critical inquiry would frame and investigate this problem.
3. Community members want to reduce inequitable access to primary care and to lead the research process. Design a participatory action research project, specifying partners, cycles of action and reflection, ethical safeguards, and likely outcomes.

Multiple Choice Questions with Answers

Question 1

Which qualitative design is most directly concerned with the meaning and structure of lived experience?

A. Phenomenology

B. Grounded theory

C. Ethnography

D. Case study

Correct Answer: A. Phenomenology

Question 2

A researcher iteratively samples participants to refine emerging categories about how nurses manage interruptions. Which design is being used?

A. Narrative inquiry

B. Grounded theory

C. Qualitative description

D. Phenomenology

Correct Answer: B. Grounded theory

Question 3

What is the defining feature of a case in case study research?

A. It must involve quantitative measurements

B. It must represent a national population

C. It is bounded by time, place, organization, population, or activity

D. It must be selected randomly

Correct Answer: C. It is bounded by time, place, organization, population, or activity

Question 4

Which design most strongly emphasizes shared practices, norms, and meanings within a social group?

A. Ethnography

B. Phenomenology

C. Narrative inquiry

D. Qualitative description

Correct Answer: A. Ethnography

Question 5

A study examines how patients construct identity through stories of recovery over time. Which design is most appropriate?

A. Case study

B. Narrative inquiry

C. Grounded theory

D. Critical inquiry

Correct Answer: B. Narrative inquiry

Question 6

Which statement best describes participatory action research?

A. Participants are passive sources of data

B. It separates research from practical change

C. It combines collaborative inquiry with action and reflection

D. It requires random sampling

Correct Answer: C. It combines collaborative inquiry with action and reflection

Question 7

Critical inquiry primarily asks which question?

A. What is the average treatment effect?

B. How do power and inequality shape the issue?

C. How many participants are needed for statistical significance?

D. What is the reliability coefficient?

Correct Answer: B. How do power and inequality shape the issue?

Question 8

Which approach is generally most suitable when a service manager needs a clear, low-inference account of patients’ reported barriers?

A. Qualitative description

B. Phenomenology

C. Ethnography

D. Grounded theory

Correct Answer: A. Qualitative description

Question 9

What does reflexivity require from a qualitative researcher?

A. Removing all interpretation from the study

B. Ignoring the researcher’s professional identity

C. Examining how the researcher’s position influences the research

D. Using only numerical data

Correct Answer: C. Examining how the researcher’s position influences the research

Question 10

Which practice is most consistent with responsible use of AI in qualitative health research?

A. Uploading identifiable patient transcripts to any free tool

B. Treating AI-generated codes as final findings

C. Using AI only under approved governance and checking outputs against original data

D. Allowing AI to decide participant consent

Correct Answer: C. Using AI only under approved governance and checking outputs against original data