Freshness and reliability

47 published measures · 46 indexed sources · review date November 2, 2026.

Latest available and common year answer different questions

The atlas and country profiles now show the latest captured observation for each economy, with its year or survey window. Latest values may describe different periods and are not automatically like-for-like. No missing year is filled with a neighbouring observation.

The experimental ranking retains its eight inputs at one endpoint. Complete cases: 2023: 170; 2024: 0; 2025: 0. Life expectancy reaches 2024 in the captured source; PM2.5 stops at 2023, while governance and tax data include 2025. Updating every input label to 2025 would misrepresent older evidence.

Specialised explorer datasets

The atlas register below covers 47 measures. Additional tools use 304 OECD household cases, 290 Eurostat regions, and 1731 financial-strain observations. Inspect their separate source and validation audit. These are not silently added to the five-domain index.

How we assess evidence

A means provenance and dimensions were rechecked in the current batch. B means an earlier captured series is retained but was not re-audited in this batch. C is a research-only or blocked source, excluded from publication as a metric. These are audit-status grades, not accuracy percentages. Official statistics, models and surveys measure different constructs; no universal provider league table would establish their accuracy.

We inspect provenance, definition fit, population and period, dimensions, coverage, uncertainty, duplicate keys and numeric validity. Sampling error, nonresponse, model assumptions, revisions and cross-country comparability remain separate questions. A recent model estimate is not necessarily more accurate than an older survey. Revision error and independent benchmarking have not yet been quantified.

Download the metric register (JSON) · Ingestion checks and raw-file hashes · Full discovery catalogue, including pending and blocked sources