Independent statistical consultancy
Evidence that holds up when somebody checks it
We work with companies, with charities and non-governmental organisations, and with academic research groups, on questions that have to be settled from data.
kinds of client: commercial, third sector, and academic research groups.
families of method, from programme evaluation to study design.
engagement ends with the code and a handover brief, not only a report.
Analysis an organisation can act on, and a reader can check
Most organisations already hold the data that answers the question they are asking about themselves. What is usually missing is a specification written before the analysis, and a report that says what the data can and cannot support.
We supply both, and hand back everything that produced the result.
Who we work with
Tailored to your needs
Companies
Commercial work is scoped to what the business asks for and stays confidential from the first conversation. The question comes from inside the organisation rather than from a research agenda, and the date is usually one somebody else has already fixed.
Charities and NGOs
The analysis is the same; the audience for it is different. A charity answers to funders, to trustees and to the bodies that commission its services, and each of those reads a different document. All three come out of one piece of work.
Academic research groups
Research groups bring us in before a bid and after an award: the statistical and power sections of a proposal, then the analysis lead on a work package across a consortium, carried through to peer-reviewed publication.
How the practice works
From a first conversation to handover
A first conversation
What the organisation does, what data exists, and the date the answer is needed. Where a question cannot be answered with the data available, we say so before the work starts.
A written specification
What will be measured, how, and what the analysis will do — agreed and dated before any of it runs. An analysis plan written afterwards is a description of what was found.
The analysis
Run against the specification, with the assumptions behind each choice recorded as they are made rather than reconstructed at the end.
Handover
The reports, the code and the brief that lets somebody else pick the work up. Whether an engagement ends in an internal report or a published paper is settled at the start.
Discuss a piece of work
The most useful first message says what you do, what data you already hold, and what date the work has to meet.