Placeholder — pending review
Research Methodology
This page is a placeholder describing the approach taken across studies on this platform.
Episode-based questions
Where possible we ask about a specific, recent episode rather than general opinion. In the BVI Banking Access Study, respondents are asked about their most recent branch visit: what they came to do, what they had tried beforehand, and what happened. Recalled episodes are more concrete than attitudes, though they remain subject to memory error.
Neutral framing
Questions are written to describe circumstances rather than to assign cause. We do not assume that compliance requirements, paperwork, or technology are responsible for in-person visits, and we do not characterise institutions as efficient or inefficient. Answer options are drafted to cover the plausible range, including preference for in-person service.
Questionnaire versioning
Each questionnaire is stored as a numbered version. Every response records the version it was answered under, and every answer is stored against a stable question identifier rather than its wording, so datasets remain comparable if wording changes later.
Sampling and limits
Participation is open and self-selected. A response cannot be taken to represent one unique resident of the British Virgin Islands: the same person could respond more than once, and respondents are unlikely to be a representative sample of bank customers. Results describe the responses received, not the population.
Data quality
Raw responses are preserved. Responses judged to be tests, duplicates, or otherwise questionable may be flagged and excluded from analysis with a recorded reason; they are not deleted. Reported figures distinguish between all responses received and the subset used for analysis.