Mandate-method-data-currency grid
Understand institutional remit and viewpoint
Identify the unit responsible for the publication. A ministry, regulator, statistics office, audit body and commissioned research institute perform different tasks. The report may monitor compliance, evaluate a programme, describe a population or recommend policy. That purpose shapes selection and presentation without making the document inherently reliable or unreliable. Look for a clear commission, responsible authors and declared limits.
Distinguish author from publisher. An external team may have conducted the work while a department hosts the final report. Record who designed the study, who supplied data and whether the publication includes a disclaimer. Cite a corporate author only when the document presents it that way; otherwise preserve named or institutional authorship rather than flattening responsibility.
Start with definitions and methods
Read the methods appendix and table notes before extracting a headline figure. Check the target population, sampling or administrative coverage, data collection, non-response, weighting, missing values and revisions. An evaluation also needs a credible comparison, stated outcomes and attention to alternative explanations. Administrative records usually measure recorded activity. A rise may reflect a real change, new eligibility, altered reporting or enforcement capacity.
Official reports do not always follow a journal article structure. Search for transparent decisions rather than particular headings. If a necessary method detail is missing, narrow the claim and document the gap. “The report recorded” is often supportable where “the policy caused” is not. A large sample cannot repair a badly defined measure.
Keep source data separate from official interpretation
A report analyses evidence; its underlying dataset may be a separate object. If you reproduce a reported table, cite the report. If you download the data, apply filters and calculate a new result, cite the dataset as well and label the calculation as your own. Save release number, download date, variables and transformation steps. This avoids attributing a later analysis to the public body.
Check units, price basis, seasonal adjustment, provisional status, rounding and suppression. Small symbols and footnotes can change the meaning of an official table. Prefer machine-readable tables over values estimated from a chart. When HTML, PDF and spreadsheet releases coexist, note which file actually contains the number.
Separate observation period, release and revision
The date displayed on a landing page may differ from the report date and from the period covered by the data. Record all three where they matter. A report released in 2026 using observations to 2023 does not automatically describe conditions in 2026. Search the series or publication collection for later releases, corrected PDFs and revised historical values.
Provisional statistics may be updated as more records arrive. Use the release that fits your research cut-off and label its status. Do not silently replace an older figure if your project analyses what decision-makers knew at that earlier time. For a current-state claim, however, relying on superseded provisional data requires an explicit reason.
Turn the grid into a proportionate claim
Draft a narrow evidence sentence: name the reporting body, measured phenomenon, reference period, data basis and material limitation. Then add your interpretation. Phrases such as “according to the agency” clarify provenance but do not absolve you from checking the method. On contested issues, compare independent research or another oversight body instead of presenting one institutional account as the whole literature.
The reference should include the responsible author or body, exact title, series or report number, version, year and persistent URL. Add an access date for a changeable page. Keep the completed four-part grid with your notes. It explains why this report supports this proposition and makes later updates much easier to audit.
Worked example: interpreting an official rate
Suppose a report compares registered incidents across regions. Before calling the result a prevalence difference, identify the denominator: residents, reports received, cases investigated or proceedings completed. Check whether regions apply the same recording rules and whether missing values were estimated. A higher rate could reflect better reporting or more intensive enforcement rather than more underlying events. Cite the observed administrative measure first; a claim about real-world prevalence requires additional evidence.