What is qualitative data collection?
Qualitative data collection gathers accounts, observations and other non-numeric material to understand experiences, meanings, behavior and context. Common sources include interviews, focus groups, open-ended survey responses, field notes, documents and diaries.
The best method depends on what you need to understand. Asking people about a process is different from observing them use it. A useful study may combine methods, but each source needs a clear purpose and an analysis plan.
Which qualitative method should you use?
| Method | Good for | Example | Watch for |
|---|---|---|---|
| Individual interviews | Detailed accounts and follow-up questions. | Why customers leave after an unresolved issue. | Interviewer influence and incomplete recall. |
| Focus groups | How people discuss a shared experience. | How staff interpret a new process. | Power differences and reluctance to disagree publicly. |
| Open-ended survey questions | Short explanations alongside standardized answers. | What most influenced a satisfaction rating. | Thin responses and unequal willingness to write. |
| Observation | What happens in practice. | Where people struggle during an application. | The observer’s interpretation and effect on behavior. |
| Document review | Existing written evidence and institutional context. | How partner reports define the same indicator. | Documents were created for a purpose that may differ from yours. |
| Diaries or repeated logs | Experiences recorded close to when they occur. | How a new workflow affects a person over several weeks. | Participant burden and missing entries. |
A case study is usually a broader study design that may use several of these methods. Ethnography is an approach to understanding a social setting, often through sustained engagement and observation. Do not treat design, collection method and software as interchangeable terms.
Method, tool and question guide: what is the difference?
The method is how evidence is collected—for example, a semi-structured interview. The tool supports collection—a recorder or video-conferencing service. The question guide or observation protocol sets out what the researcher will ask or attend to.
Choosing a tool does not settle sampling, consent or analysis. A polished form can still ask leading questions. A detailed transcript can still omit the contextual information needed to interpret it.
Plan collection around the decision
Suppose a team wants to understand why people leave a workforce program. “Collect feedback” is too broad. A more useful question is whether early departures relate to scheduling, the content, support needs or a mismatch with participant goals.
- Identify the groups whose experiences could differ.
- Choose a method that lets those groups describe the relevant experience.
- Plan how to reach people who disengaged, not only current participants.
- Decide what contextual information is necessary and appropriate.
- Specify how evidence will be analyzed and who will review it.
Write down what another method would need to establish. Interviews may explain a barrier; a suitable survey or administrative dataset may be needed to estimate how widespread it is.
How to choose participants
Qualitative sampling often seeks relevant experience and useful variation rather than a statistically representative sample. Explain the criteria: service stage, location, type of experience or another factor linked to the question.
Avoid presenting convenient recruitment as if it covered every perspective. Record who was invited, who participated and which important groups remain unheard. Sample size should follow the study’s purpose, diversity, depth and analysis approach rather than a fixed rule.
For a study of early departures, people who completed the program are useful comparison cases but cannot speak for everyone who left.
Collect detailed evidence without leading people
Ask people to describe an event before asking for a general judgment. “Tell me about the last time you submitted a report” gives a starting point. “Was our process simple?” narrows the answer too early.
For observation, distinguish what you saw from what you inferred. “The user reopened the instructions three times” is an observation. “The user lacked confidence” is an interpretation that needs further evidence.
CDC’s qualitative-data guidance describes the importance of the collection setting and interviewer role. It also notes that individual interviews can provide space for experiences people may not share in a group. Read the field guidance.
An example: study departures with complementary sources
| Source | Question it can help answer | Context to retain |
|---|---|---|
| Exit interview | What sequence led to the decision to leave? | Date, participant stage and interview guide version. |
| Open response at a check-in | What concern was visible before departure? | Question wording and timing relative to the event. |
| Attendance record | When did participation change? | Session dates and definitions of attendance. |
| Staff observation | What barriers were seen during delivery? | Observer, date and distinction between observation and interpretation. |
| Program documents | What support was available or promised? | Document version and applicable period. |
These sources may disagree. That is a finding to investigate. A staff note may describe low engagement while a participant describes transport problems. Preserve both accounts and examine the context rather than forcing them into a single label.
Prepare data for analysis as it arrives
Retain the original source and create an analysis copy with appropriate access controls. Attach the collection date, source type, question or guide version and relevant group or record identifier. Avoid collecting identifiers that the study does not need.
Start reviewing early enough to identify unclear questions or missing perspectives. If you change the guide or codebook, keep the version history so later comparisons remain interpretable.
AI can help organize text, suggest codes and locate passages. Keep the analysis tied to the source, review uncertain classifications and look for material the model omitted. Faster processing does not remove the need for judgment.
What makes qualitative evidence trustworthy?
- A clear link between the research question, sampling and collection method.
- Enough context to interpret what was said or observed.
- An explicit analysis process, with definitions and examples where codes are used.
- Attention to contradictory cases and the researcher’s assumptions.
- A traceable account of how findings were developed from sources.
- Honest reporting of access, participation and other limitations.
Qualitative quality is not determined by a single inter-coder score. Some approaches use structured agreement checks; others emphasize reflexive interpretation. Explain the approach you used and apply it coherently.
Ethics and participant burden
Explain why evidence is being collected, how it will be used and who can access it. Take particular care when participants depend on the organization for employment or services: apparent agreement may not mean they feel free to decline.
Do not promise anonymity if a person could be identified from the record or quotation. Limit sensitive detail, establish a retention approach and obtain appropriate permission for public quotations. Choose a collection plan that respects people’s time as well as the organization’s reporting needs.
Continue learning in the Academy
Use these existing Academy guides for the practical next step. They are suggested companion readings; follow each guide’s course navigation for the full sequence.
Use the practical guides to connect the model to a collection plan, review new evidence and decide what to improve.
How Sopact reduces coding and reporting work
Plan how the collected material will be used after arrival. In a codebook-based workflow, repeated application and revision can become a larger burden than the initial collection.
A workflow with repeated manual work
- Define from an initial sampleRead material and agree on the codebook.
- Apply it across the datasetCode responses and check the result.
- Revise a definitionReturn to affected material and recode it.
- Reconnect the numbersReconcile coded results with ratings and context, then rebuild the view.
The Sopact workflow
- Your team owns the definitionsDecide what each code means and improve it as you learn.
- Apply coding across the eligible dataAutomate application; people review quality and exceptions.
- Reprocess after a definition changesReapply the revised definition across the configured scope instead of recoding each response by hand.
- Ask across coded text and numbersKeep the response, rating and relevant record context connected; inspect the evidence behind the result.
This compares workflow patterns, not a claim that every research tool requires manual coding or separate files. Some already automate parts of this work; compare the complete cycle.
For this codebook-based workflow, the main saving is repeated application and reconnection—not the removal of human judgment. A changed definition can be reapplied across the configured data while reviewers concentrate on quality, exceptions and interpretation. Coded text stays connected to the relevant ratings and context.
Count the recurring work in ownership cost. Include setup, coding, recoding after revisions, source reconciliation, review and reporting, plus your actual platform and processing expenses. A worked scenario of four cycles of 4,000 responses illustrates 272 fewer annual staff hours; it is an assumption-based example, not a customer benchmark. Existing automation, review needs and implementation effort can substantially change the result.
Adjust the workload assumptions and compare total effort →
A reliable assistant should calculate from the selected records and let a reviewer open the supporting evidence. Check the data scope, definition, denominator and access permissions. Reproducible arithmetic does not make every AI interpretation correct.
Watch: Why Qualitative Analysis Stays Small — And How to Scale It
See why revising a codebook creates repeat work, and how connected coding and quantitative analysis change that workload.
Frequently asked questions
What are the main qualitative collection methods?
Common methods include interviews, focus groups, open-ended questionnaires, observation, document review and diaries. They can be combined within a suitable study design.
Can an online survey collect qualitative data?
Yes. Open-ended responses can provide qualitative evidence, but short responses may need follow-up or complementary methods for depth.
Should the codebook be written before collection?
Some studies start with predefined concepts; others develop codes from the data or use a combination. Document the approach and revisions.
How many interviews are enough?
There is no universal number. Consider the question, participant diversity, depth of evidence and whether important perspectives remain missing.
Is qualitative evidence less rigorous than quantitative data?
No. Rigor depends on appropriate design, collection, analysis and transparent reporting. The methods answer different questions and have different limitations.
