Qualitative Research: Data Collection, Analysis, and Quality
❦
Qualitative research explores meaning, process, and context. It is suited to questions beginning "how" and "why", to new phenomena, and to understanding the views of participants (Creswell & Poth, 2018; Flick, 2018; Merriam & Tisdell, 2016).
Designs
| Design | Focus | Typical business use |
|---|---|---|
| Case study | A bounded case in its real context, with multiple sources of evidence (Yin, 2018; Eisenhardt, 1989) | A firm's turnaround, strategy, or digital transformation |
| Phenomenology | The essence of a lived experience | Experiences of redundancy or entrepreneurship |
| Grounded theory | Theory generated from data (Glaser & Strauss, 1967; Charmaz, 2014); the Gioia methodology is widely used in management (Gioia et al., 2013) | Processes such as succession in family firms |
| Ethnography | Culture and practices through immersion | Workplace culture |
| Narrative inquiry | Stories and life histories | Career narratives |
| Content analysis | Systematic categorization of texts (Krippendorff, 2018) | Annual reports, social media |
| Action research | Solving problems with participants | Improving processes |
Sampling
Use purposive sampling (choosing information-rich cases), maximum variation, snowball sampling, or theoretical sampling (in grounded theory). Plan for saturation: the point at which new data add little new (Guest et al., 2006). The number depends on the aim, the diversity, and the quality of the data (Malterud et al., 2016). Many student studies use about 8 to 20 interviews, but the justification matters more than the number.
Data Collection Methods
| Method | Notes |
|---|---|
| Semi-structured interviews | Prepare an interview guide with open questions and probes; begin with easy questions; record with consent; take notes (Kvale & Brinkmann, 2015) |
| Focus groups | About six to ten participants; a moderator; explore group views and interaction (Krueger & Casey, 2015) |
| Observation | Structured or unstructured; field notes |
| Documents and records | Policies, reports, minutes, websites |
| Open-ended survey questions and diaries | Reach more people; limited depth |
| Visual and digital data | Photographs, social media posts |
Practicalities: choose a quiet place; build rapport; explain confidentiality; avoid leading questions; listen more than you speak; use participants' preferred language, and plan for translation and transcription (a Sinhala or Tamil interview may need to be transcribed in the original language and translated, with a second person checking accuracy).
Analysis
Thematic analysis is the most widely used approach (Braun & Clarke, 2006, 2021). The six phases are: familiarization; coding; generating initial themes; developing and reviewing themes; defining and naming themes; and writing up. Nowell et al. (2017) show how to demonstrate trustworthiness at each phase.
From data to themes: an example
| Data extract | Code | Category | Theme |
|---|---|---|---|
| "I leave home at 6 and return after 9 at night." | Long hours | Workload | Work demands erode personal time |
| "My manager does not allow leave even for family events." | Leave denied | Managerial control | Work demands erode personal time |
| "The branch manager listens to us." | Supportive manager | Supervisor support | Support sustains commitment |
Manual coding: print transcripts or use Word comments or a spreadsheet; highlight and label segments; build a codebook with definitions; group codes into categories and themes on index cards or a table.
Software: in NVivo, import transcripts (Import), create nodes (codes), code by selecting text and dragging it to a node, run queries (word frequency, text search, matrix coding), and use case classifications for participant attributes. ATLAS.ti and MAXQDA work similarly. Software organizes data but does not analyze it; the interpretation is yours.
Content analysis: define the unit of analysis and the coding scheme; code reliably (use two coders and measure agreement such as Cohen's kappa where appropriate); count frequencies and interpret (Krippendorff, 2018).
Grounded theory: open, axial, and selective coding with constant comparison and memo writing (Charmaz, 2014).
Trustworthiness
Lincoln and Guba (1985) proposed criteria parallel to the quantitative ones:
| Criterion | Meaning | Techniques |
|---|---|---|
| Credibility | Findings are believable | Prolonged engagement, triangulation, member checking, peer debriefing |
| Transferability | Findings may apply in other contexts | Thick description of context and participants |
| Dependability | The process is consistent and documented | Audit trail, detailed methods |
| Confirmability | Findings are shaped by the data, not by bias | Reflexivity, audit trail |
Add reflexivity: reflect on how your background and position influence the research. Report your procedures according to guidelines such as COREQ or SRQR where relevant.
Reporting Qualitative Findings
Organize by themes; define each theme; use short, anonymized quotations (with participant codes such as P7) as evidence; explain what each quote shows; present a theme table and, where useful, a thematic map. Avoid listing quotes without interpretation.
Common Mistakes
- Treating interviews as if they were questionnaires.
- Reporting topics rather than themes.
- Presenting results as statistical generalizations.
- Not documenting how themes were developed.
CHAPTER 13