How we design, run and publish a study
Every publication travels the same route from question to public record — with methods recorded up front, results replicated, and data and funding disclosed on release.
Research is only as trustworthy as the process behind it. We treat method as something to be written down, reviewed and published — not assumed. This page sets out how a study moves from an initial question to a citable publication, and how we protect its quality at each stage.
The same discipline applies whether a study is a flagship report for the public record or a commissioned market assessment. Sourcing, review and disclosure do not vary with who is reading the result.
Six stages from scope to publication
Scoping & research design
Each study begins with a written research question, a pre-registered method (survey, comparative case study, econometric model or mixed methods) and an explicit statement of what the evidence can and cannot support.
Data collection
Data are drawn from public statistical agencies, primary fieldwork, licensed datasets and, where relevant, structured interviews. Sample frames, response rates and known limitations are recorded as they are gathered.
Analysis
Analysts document working files, code and assumptions. A second researcher who was not involved in the original design replicates key results before a draft is circulated.
Internal and external review
Every draft is reviewed by an internal editor for methods and framing; flagship reports and datasets also go to at least one external adviser from outside the research team.
Publication & disclosure
On release, each report states its data sources, methods and funding, and is assigned a permanent internal permalink and a copy-ready citation.
Correction & update
Errors identified after publication are corrected in a dated erratum attached to the original record; substantial revisions are issued as a new version with a changelog.
Replication is not a formality. A result that a second analyst cannot reproduce from the same data does not leave the building.
What we disclose as standard
So that findings can be checked and reproduced, each publication carries a consistent set of disclosures:
- the data sources, sample frame and, where relevant, response rates;
- the method and the assumptions behind any model, with sensitivity checks;
- named authors and, where applicable, external advisers;
- a funding line stating who paid for the underlying work;
- a permanent internal permalink, a series number and a copy-ready citation under a CC BY 4.0 licence.
Read more about how independence is protected and how conflicts are declared on our independence and funding page.
Methodology, in brief
- How do you decide what to research?
- Each study begins with a written research question and a statement of what the evidence could show. Topics are chosen by the research team on the basis of their significance and how far independent, comparative analysis can add to what is already known — not by the party funding or commissioning the work.
- What methods do you use?
- We use the method suited to the question: sample surveys, comparative case studies, econometric modelling, market sizing and mixed methods. The method, sample frame and known limitations are recorded before analysis begins and stated in the published report.
- Where does your data come from?
- Data are drawn from public statistical agencies, primary fieldwork, licensed datasets and structured interviews. We record sample frames, response rates and limitations as data are gathered, and we cite every source in the published record.
- How is a study reviewed before publication?
- A second researcher who was not involved in the original design replicates key results. An internal editor reviews methods and framing, and flagship reports and datasets are also read by at least one external adviser from outside the research team.
- How do you handle uncertainty and limitations?
- We report confidence intervals, sensitivity checks and the assumptions behind any model, and we state plainly what a study cannot support. A careful finding with its limits stated is preferred to a confident one that the evidence does not carry.
- What happens if an error is found after publication?
- Errors identified after publication are corrected in a dated erratum attached to the original record. Substantial revisions are issued as a new version with a changelog, so the history of a finding remains visible.
See the approach in the published record
Every publication states the methods and data behind it.