Why a Revised Figure Is Not a Wrong One

Every year, a familiar complaint circulates in newsrooms: a growth figure published three months ago has changed. Some reporters treat this as evidence the earlier number was wrong, or worse, that it was manipulated. In most cases, neither is true. Revision is a normal, expected part of how statistical agencies work, and understanding why makes the difference between a fair account and an unfair one.

Preliminary estimates are built from partial data. A quarterly output figure released six weeks after the period ends is compiled from returns that are still arriving. As more returns come in, the estimate is updated to reflect a fuller picture. This is not a correction of an error so much as a natural refinement, the same way a weather forecast narrows as the day approaches. Agencies publish preliminary figures because waiting for complete data would mean waiting months, and a delayed number is often less useful than an approximate one released on time.

Reading revisions without alarm

The practical skill is knowing which figures are provisional and by how much they typically move. Most statistical offices publish a revision history alongside their releases, showing how far earlier estimates for the same period have shifted over successive updates. A reporter who checks this history before writing "growth slowed sharply" is far less likely to publish a claim that dissolves under the next revision.

A single quarter's preliminary estimate should be treated as a first sketch, not a finished portrait.

It also helps to distinguish between a routine revision and a definitional change. Occasionally an agency updates its methodology entirely — a new base year, a redefined sector boundary, a different survey frame. These changes can produce jumps that look dramatic but reflect a change in what is being measured, not a change in the underlying reality. Comparing figures across such a boundary without noting the change misleads readers, even unintentionally.

Writing the caveat into the sentence

The clearest fix is often the simplest: state the status of the figure in the same sentence that reports it. "A preliminary estimate suggests" reads differently, and more honestly, than a bare number presented as settled fact. Readers rarely object to appropriate caution; what erodes trust is a number stated with false certainty and then quietly abandoned a quarter later.

None of this argues for avoiding preliminary figures altogether — they are often the most current information available, and timeliness matters. The goal is simply to carry their limits into the text, so that when a revision does arrive, it reads as expected rather than as a retraction.