Diagnosis before design
A social programme is almost never defeated by a lack of goodwill. It is defeated earlier, at the point where somebody decided what the problem was. A budget line is written, a target population is named in a sentence, activities are chosen because they are the activities everyone chooses — and from then on the programme is busy. It runs workshops, it fills attendance sheets, it produces a final report. Whether it moved anything is a question nobody can answer, because nothing was measured before it began.
We exist to interrupt that sequence. Episteme works in two movements that are really one: we investigate, and then we intervene on the basis of what the investigation says — including when it says something inconvenient.
What investigation means here
It means going to the territory. Household and user surveys with a sampling frame we can defend, semi-structured interviews with the people who receive a service and the people who deliver it, focus groups where disagreement is allowed to surface, documentary and administrative data read against each other, and mapping that puts need and provision on the same sheet of paper. Qualitative material is coded, not quoted selectively. Quantitative material is reported with its limits attached. Where the evidence is thin, we write that it is thin.
What intervention means here
It means a design that the team who will actually run it can run: an explicit theory of change, activities tied to the mechanisms the diagnosis identified, indicators that someone can collect on a Tuesday afternoon without a research degree, and a baseline recorded before the first session so that evaluation is possible later. We stay through implementation, because designs meet reality and have to be adjusted in the open rather than abandoned in silence.
Why institutions ask us
Public bodies and NGOs are accountable to boards, funders, auditors and the communities themselves, often simultaneously. That accountability is easier to carry when the reasoning is written down: why this population, why this mechanism, what we expected, what we observed, what we would change. Our deliverables are built to be read by a technical team and defended in front of a council or a donor committee.
We do not promise outcomes; honest research cannot. We commit to method, to transparency about uncertainty, and to ethical fieldwork — informed consent, protection of identifying data, and returning the findings to the communities that produced them.
