Studying Institutions in Practice: Qualitative and Mixed Methods
Notes integrating Sara Stevano’s seminar, ‘Qualitative and Mixed Methods for Economic Research’ (29 October 2026), NCRM lectures on ethnography, and selected readings.
Tao Louie Xu
Development research examines outcomes such as investment and employment, alongside the social processes that produce them. Qualitative enquiry can support us to better establish how actors understand their circumstances, how decisions are made, and how institutions operate in practice. Combining methods is most useful when different forms of evidence contribute to a shared explanation.
1. Qualitative Enquiry: Meaning, Process, and Reflexivity
Qualitative research studies meanings, practices, and relationships in their social and historical settings. Design can develop through fieldwork, moving between empirical material and theoretical concepts. Three principles are particularly useful.
- Question the categories. Interviews and observation can reveal distinctions that standard indicators overlook. Stevano’s example of food consumption in Mozambique illustrates how participants’ understandings can improve the construction of survey questions.
- Practise reflexivity. Researcher positionality, trust and power relations determine access, accounts, and interpretation. Rao (2022) proposes cognitive empathy, narrative evidence, attention to process, and participant involvement as ways to make economic enquiry more reflexive.
- Sample for explanation. Interview selection follows the research question and the emerging reflection. Elmholdt, Gill and Nielsen (2026) distinguish opening, focusing and closing sampling moves. The aim is conceptual depth and analytical or theoretical generalisation, with claims calibrated to the evidence.
2. Institutions, Everyday Practice
Interviews, fieldnotes, observation and documents can reveal gaps between formal policy and actual routines. The NCRM ethnography lectures offer two complementary approaches:
- Ethnographic techniques attend to everyday activities, participants’ interpretations and the researcher’s relationship with the field. Short research encounters, narrative interviews and participant or non-participant observation can be combined according to access and the research question.
- Institutional Ethnography, developed by Dorothy E. Smith, begins from people’s everyday work and traces its connections to ruling relations: translocal forms of institutional coordination and power. Policies, contracts, forms and classifications are studied as texts that organise action when people use them.
Orly Benjamin’s account of contracted care work illustrates the second approach. Employment conditions experienced by workers were shaped by procurement contracts and administrative judgements about costs, skills and staffing. The broader methodological insight is to trace how decisions and classifications travel through organisations and shape what actors can do. Policy documents should also be situated in their institutional and historical contexts.
3. From Description to Causal Explanation
The readings connect these empirical practices to the philosophical foundations of research. Sumner and Tribe (2008) distinguish ontology (what exists) from epistemology (how it can be known). Critical Realism offers a useful explanatory orientation: observable experiences, events and underlying causal structures represent distinct aspects of social reality (Easton, 2010).
- Retroduction asks which structures, powers and mechanisms could account for an observed outcome, given the specific context. Explanations develop iteratively through evidence and comparison.
- Process tracing examines a proposed causal sequence within a case. Ricks and Liu (2018) recommend specifying hypotheses, constructing timelines and causal links, identifying alternative paths, and seeking evidence that discriminates between rival explanations.
- Interpretation remains central. Coding organises material; analysis establishes relationships, examines discrepancies and revises explanations. Software and AI can assist with transcription and coding while substantive judgements remain grounded in the sources.
4. Mixed Methods as an Integrated Research Design
Methods can be combined sequentially or concurrently, with equal or different analytical weight. Their contribution depends on how evidence is linked at the stages of conceptualisation, sampling, analysis and inference.
For research on industrial policy and state capacity, a coherent design might proceed from policy texts and administrative classifications to interviews and observation of implementation, then to quantitative patterns in firms or sectors. Process tracing can investigate how particular decisions emerged; quantitative evidence can establish their distribution and wider associations. Each stage can revise the questions and categories used in the next.
The organising principle is to follow the research problem across levels of evidence: from actors’ meanings and everyday practices, through institutional rules and decision sequences, to the economic outcomes they help explain. Critical Realism provides a framework for investigating causal mechanisms; pragmatic method selection keeps the empirical design responsive to the research question.