Method

Identify the object, geography, time window and outcome being examined. A sector-wide number and a project-level record may inform the same question without measuring the same thing.

Use original records where available. Preserve the source’s date, version, units and stated limits. An announcement, estimate, operating result and legal instrument answer different questions.

Published analysis distinguishes the reported observation from our interpretation, explains the reasoning and identifies unanswered questions. A conclusion is limited to the evidence available for the stated scope and period.

Signals state a bounded judgment, supporting evidence, counterevidence and an observation that could change the view. Status applies only to the stated test.

Material corrections and evidence updates identify the affected publication, explain what changed and describe its effect on the argument. Follow the correction link on a publication to inspect its recorded changes.

Follow the originating document for its full context and terms. Citation does not transfer copyright, and a library entry is not an endorsement or partnership.