Data Revision

A data revision is a scheduled update to an already-published statistic, made once more of the underlying source data has arrived. Surveys receive late responses, tax and administrative records catch up, seasonal factors are recalculated, and the agency reissues the period with a better estimate. The revision is part of how publication is designed, not the correction of a mistake. Two things follow for anyone reading a release. First, the headline of the day may arrive alongside revisions to earlier periods that change the trend the new figure sits in — a rise against a previous month revised downwards is a different statement from a rise against the number originally reported. Second, a historical chart drawn today does not show what was known at the time, because it carries the revised values. Some agencies publish real-time or vintage datasets precisely so the figure as first published can still be retrieved.