How we research
A transparent workflow for definitions, evidence, calculation and uncertainty.
1. Frame the question
We reduce broad prompts to a claim that evidence could support or challenge. Terms, geography, period and unit are recorded before research begins.
The question record also notes why the issue matters and what the page will not attempt to answer.
2. Build a source register
Primary documents lead: official statistics, issuer filings, index methodologies and contract specifications. Each entry records publisher, URL, date, unit and revision status.
Secondary commentary can aid discovery, but it cannot replace the document closest to the underlying claim.
3. Reproduce the measure
Where feasible, we recreate arithmetic from stated inputs. A result that cannot be reproduced is labelled as reported rather than independently confirmed.
Inputs, formula, rounding and units are shown so a reader can inspect rather than simply trust the result.
4. Test competing readings
We ask which alternative mechanism could produce a similar pattern. This guards against treating correlation, timing or a persuasive narrative as causation.
Contradictory evidence is reported when material; it is not removed merely to create a cleaner narrative.
5. Add Irish context carefully
We distinguish Irish, euro-area and global evidence. Central Bank of Ireland material may be linked as an official resource; that never implies authorisation or endorsement.
Domestic relevance is explained without implying that every Irish reader has the same exposure or tax position.
6. State the boundary
Every substantive page names data, model, market and reader limitations. No publication can determine personal suitability or capacity for loss.
Limitations are written beside the claim they qualify, not hidden in a generic footer.
7. Review and correct
Pages show a review date. Material corrections should identify what changed and why; stylistic edits need not alter the review record.
Review timing reflects change risk, while corrections preserve the original error and its effect.
8. Keep commercial pressure out
No deposit prompts, urgency devices, transaction links, performance claims, product rankings or lead-pressure sequences shape editorial conclusions.
Funding, contributor interests and commercial changes require specific disclosure under the publishing standards.