What we do
Research & Evidence
NoviThea designs and runs research for people who have to decide something. We take one question, choose the methods that can actually answer it, and report what the evidence carries and where it runs out.
Research earns its place when it changes what someone does next
A study is worth commissioning when it gives a clearer account of what is happening, whose experience is visible in it, and what is still unresolved. These are policy, programme and organisational questions that a quick answer would get wrong.
- What is changing, and why?
- Which groups experience a problem differently?
- Which mechanisms appear to shape outcomes?
- What evidence is strong enough to support a decision?
- What can be learned before a programme or policy is taken further?
Ways of working within this
Applied research
Research designed around a live policy, programme, organisational or societal question, with attention to the decisions it needs to inform.
Qualitative research
Interviews, focus groups, participatory approaches, coding, thematic analysis and process-oriented inquiry to understand experience and context.
Quantitative research
Survey and structured-data analysis, statistical work, comparative or longitudinal analysis, and careful visual interpretation of findings.
Mixed-methods research
Bringing qualitative and quantitative evidence together to test patterns, explain mechanisms and examine where findings converge or differ.
Evidence synthesis
Literature and evidence reviews, structured synthesis, evidence mapping and comparison across sources when a question needs a wider view.
Evaluation and learning
Evaluation design, contribution and process-oriented approaches, learning questions, and evidence that can inform adaptation and decisions.
Foresight and anticipatory research
Inquiry into emerging risks, possible futures and uncertainty when preparedness matters as much as explanation of the present.
Research design and methodological support
Support with research questions, sampling, instruments, analytical strategy, protocols, documentation and reproducible practice.
The design comes after the question, not before it
An organisation that only runs surveys will find survey-shaped answers. We settle what is being decided and what would count as an answer, then choose and combine methods for that, which sometimes means telling a partner the study they asked for will not tell them what they need.
Who this is for
Public institutions, funders, international organisations, civil-society groups, research organisations, universities, businesses and consortia — anyone who needs an answer they can defend.
Evidence is not uniform, and treating it as though it were is the common error
Two studies can point the same way and deserve very different weight. More data does not settle a question that was badly posed, and a large sample does not rescue a measure that captures the wrong thing. The work is in the judgement about which findings will bear the load.
- A design chosen for the question in front of us, rather than the method the team is fastest at.
- Provenance kept with the finding: where a figure came from, who was counted, who was missed, and what was done to it before it reached the page.
- Uncertainty reported as part of the result, not removed to make a recommendation easier to read.
- Where a model or AI tool contributes, its output stays traceable and open to review by someone who can tell when it is wrong.
- Interpretation tested against people who would notice if it were mistaken, and against the possibility that we asked the wrong question.
Start with the research question
A conversation can begin with the question, the decision ahead and the evidence that is already available.