Services
Qualitative and participatory research
We find out why. Interviews, group discussions and community sessions explain what the numbers can’t: why people stay away, how decisions get made, and what a programme felt like to the people it was for.
Questions this answers
- Why are people not using the service?
- How are decisions made in the household, the agency or the community?
- What did the programme mean for the people it was meant for?
The kinds of qualitative work we run
Most studies combine two or three of these.
Key informant and in-depth interviews
One-on-one conversations with the people closest to a decision or an experience, where a group setting would flatten what they actually think.
Focus group discussions
Structured conversations with a small group, usually five to eight people, guided by a moderator trained not to steer the outcome. Held in a setting participants find comfortable, not wherever is easiest for us.
Case studies
A close look at one state, facility, school or community, to show how something plays out in a real place rather than in the abstract.
Life histories
A person’s own account of their experience over time, used where the sequence of events matters as much as the outcome.
Community mapping, scorecards and participatory appraisals
Tools that let a community show us the picture, rather than us describing it to them: a map they draw, a service they score, a priority they rank.
Norms research, and research led with the people concerned
Studies of social norms and household decision-making, and work designed and sometimes co-led with organisations of the people the study is about, such as a disability organisation or a women’s group.
How we run a group or an interview
The method changes what you can trust the finding to mean. We state ours in every proposal.
A trained moderator, not a form-reader.
Someone skilled enough to follow where the conversation goes without steering it toward an answer we expected.
More than one group per audience.
A single focus group can be dominated by one or two voices. We run at least two groups for any audience we need to draw a conclusion about, and compare them before we report a pattern.
A setting people can speak freely in.
Private where the topic needs privacy, accessible for people with disabilities, and never a setting that puts anyone at risk at home or in their community afterwards.
A referral pathway, agreed before fieldwork.
If someone raises a protection or health concern, our team follows an agreed route rather than deciding in the moment. We map the health and protection services in each study area and agree the route with them before fieldwork, and fieldwork does not start until it is in place.
How you know a finding is trustworthy
Qualitative research has its own version of a margin of error. Researchers call it trustworthiness, and it rests on four questions, set out by Lincoln and Guba in 1985 and still the usual reference point.
Credibility
Is this really what people think, or what the researcher expected to hear? We compare across interviews, sources and researchers (triangulation), take findings back to participants where we can to see if they recognise them (member checks), and ask a colleague outside the study to challenge our reading (peer debriefing).
Transferability
Does this finding travel beyond the exact people we spoke to? We describe the setting and the participants in enough detail (thick description) that you can judge where else it is likely to hold.
Dependability
Would another careful researcher, following the same steps, reach a similar reading? We keep a dated record of how the study was run and how codes and themes were built, so the whole process can be audited.
Confirmability
Is this the participants’ view, or the researcher’s? Every theme traces back to the interview records and quotes behind it (an audit trail), so you can check that the finding came from the people we spoke to, not from us.
From our own work
Real work of ours that used this kind of research, not an invented example.
Adebayo Adeleke LLC
Real exampleIndigenous Supply Chain Management System
84 in-depth interviews and 315 structured surveys. Seven states and the FCT: Lagos, Kano, Borno, Akwa Ibom, Benue, Edo, Anambra and Abuja.
The same cross-border energy infrastructure operator
Real exampleWhat traditional rulers and community leaders said, in their own words
12 traditional rulers and 12 focus group discussions. Communities along the right-of-way in Lagos and Ogun States.
Reading a qualitative finding
IllustrativeTheme: mothers avoid the clinic at night, even when it is open
““I don’t walk that road after dark, even if my child has a fever. It’s not the clinic I don’t trust, it’s the road.” — mother, in-depth interview, example location”
This finding was said by 9 of 14 mothers interviewed across two example communities, and it matched what community health workers described independently in a separate interview — that is triangulation. It is reported as a finding only because more than one source and more than one respondent said it.
Questions worth asking a qualitative research provider
There is no single industry checklist for buying qualitative research, the way ESOMAR publishes one for online survey samples. We built our own five questions from the trustworthiness criteria above, and we answer all five in every proposal.
How many people did you actually speak to before you stopped?
We state the number of interviews or groups per audience, and whether we reached the point where new interviews were not producing new themes.
Who checked the coding, besides the person who did it?
A second analyst reviews how a theme was built before it is reported as a finding.
How are quotes anonymised?
Names, places and identifying details are removed or altered before a quote leaves our team, and we agree with you in advance what level of detail is safe to publish.
How were participants chosen, not just who was easiest to reach?
The selection method is stated in the proposal: who was eligible, how they were approached, and who declined.
What happens if someone becomes distressed during an interview?
The interviewer pauses and follows the referral pathway agreed for that study area before fieldwork began, so no one is deciding what to do in the moment.
What you receive
Findings organised by theme, with anonymised evidence and what they mean for design.
The findings
Organised by theme, not transcribed conversation by conversation.
The evidence
Anonymised quotes and examples that support each theme, so you can see the finding is grounded, not asserted.
What it means for design
What the finding implies for the decision you are making, not left for you to work out.
Questions people ask about qualitative research specifically
How many interviews is enough?
Enough is not a fixed number. We keep interviewing an audience until new interviews stop producing new themes, a point researchers call saturation, and we state in the report how many interviews that took.
Can you name who said what?
Only with the person’s explicit, informed consent, agreed in writing before publication. By default, every quote is anonymised.
How does consent work for someone who cannot read?
The consent information is read aloud, in the person’s own language, and we record their agreement rather than relying on a signature alone.
Why is this more expensive per person than a survey?
Forty in-depth interviews need a trained moderator, a transcriber and two analysts, and each interview is transcribed, coded and checked line by line. A survey team can collect and process thousands of short questionnaires in the same time. The cost reflects the depth, not the headcount.
Have a question a number can’t answer?
Tell us what you need to understand, and who it involves. We will say whether qualitative research is the right tool, and roughly how many conversations it would take.