Skip to content
EMPOBase
en

How to use qualitative data in MEAL reports

By the EMPOBase team · · 4 min read

Short answer

Qualitative data is information in words, such as interviews, group discussions and observations. It explains why results happened and how people experienced them. To use it well, plan questions in advance, record sources, code the answers into themes, check them against numbers, and present short quotes with context. Never let one story stand for everyone.

Key points

  • Numbers show how much; qualitative data shows why and how.
  • Plan the questions, the sample and the consent before you collect.
  • Group answers into themes, and count how common each theme is.
  • Use short, anonymous quotes with context, and say how many people held the view.

What is qualitative data?

Qualitative data is information that is not numbers. It comes from interviews, group discussions, open questions on a survey, observation, complaints and feedback, and diaries or photos. It tells you what people think, feel and do, and why.

Quantitative data tells you that 600 people attended a session. Qualitative data tells you why some women did not come back. Reports are stronger with both.

When does qualitative data help most?

  • Explaining a result. A number rose or fell, and you want to know why.
  • Checking relevance. Are activities what people actually need?
  • Finding unintended effects. Both good and bad.
  • Hearing from groups that surveys miss. For example, people who cannot read or travel.
  • Designing indicators. Early interviews help you word questions and choose measures.

How do you collect it well?

  1. Define the question. What do you need to understand, and for what decision?
  2. Choose the method. Individual interviews suit sensitive topics. Group discussions suit shared views. Observation suits behavior.
  3. Choose who to hear from. Include different groups by sex, age, disability and place. Avoid only asking people who are easy to reach.
  4. Plan consent and safety. Explain the purpose, say that taking part is voluntary and protect identity.
  5. Use a guide. A short list of open questions keeps interviews consistent.
  6. Record accurately. Take notes or recordings with permission, and write them up soon after.
  7. Record the source. Note date, place, group and method, but not names where this is unsafe.

How do you analyse it?

You do not need special software for a small data set.

  1. Read everything. Look for what comes up repeatedly.
  2. Code the text. Give short labels to each point, such as “distance to water point” or “fear of reporting”.
  3. Group codes into themes. A few clear themes are better than a long list.
  4. Count and compare. Note how many people or groups raised each theme, and whether it differs by group.
  5. Check against numbers. Do the themes explain what the indicators show? Look for conflicts.
  6. Pick quotes. Choose short ones that illustrate a theme, not the most dramatic one.

Our article on accountability to affected populations shows how feedback fits in.

How do you present it in a report?

  • Say how you collected it. Method, number of people and groups.
  • Lead with the theme, then the quote. “Many women said the session time clashed with farm work: ‘We can only come after midday.’” The quote supports a point; it is not the point.
  • Show how common it is. “Raised in six of eight groups” is better than “some people said”.
  • Keep quotes anonymous. Remove names and details that could identify someone.
  • Link to indicators. Place themes beside the numbers they explain.
  • State limits. Who you did not hear from, and what that may mean.
  • Say what you will do. End each finding with an action.

What mistakes should you avoid?

  • Using one vivid story as proof of broad change.
  • Asking leading questions.
  • Hearing only from group leaders.
  • Quoting without consent.
  • Reporting themes without saying how many people held them.
  • Collecting stories and never analysing them.

What does a short qualitative section look like?

A short section in a donor report can follow this pattern:

  1. Question. What did we want to understand?
  2. Method. Eight group discussions, two per site, with men and women in separate groups.
  3. Main themes. Three themes, each with a short quote and a note on how many groups raised it.
  4. Link to numbers. How the themes help explain the indicator results.
  5. Limits. Who was not heard, and what that means for the findings.
  6. Action. What the team will change.

Keeping to a pattern makes sections quick to write and easy to compare between reports. For a deeper look at how reviewers use such evidence, see how to check field data quality. Keep the section to one page where you can, and move longer transcripts to an annex.

Where EMPOBase fits

EMPOBase holds Complaints & Feedback (CFM) records, narrative report types such as situation, activity, monitoring and assessment reports, and Risk & Learning records, which can all serve as qualitative evidence. A submitted narrative report does not automatically establish an approved indicator result, so link themes to numbers yourself. EMPOBase does not code or analyse interviews for you. Keep analysis files in your own tools, and take care with sensitive records.

Questions people ask

Is qualitative data less reliable than numbers?

Not by nature. It answers different questions. It is reliable when you collect it carefully, record how you did it and compare it with other evidence.

How many interviews are enough?

It depends on the question and the groups you need to hear from. Stop adding interviews when new ones stop giving new themes, and make sure each key group is included.

Can we use stories from beneficiaries in donor reports?

Yes, with informed consent, and in a way that does not identify or endanger anyone. When in doubt, remove names and details.