What is the difference between a correction, a clarification and a retraction?

They describe three escalating levels of error, and how an outlet uses them is one of the better signals of whether it takes accuracy seriously.

A clarification addresses something that was accurate but capable of being misread, or that lacked context. Nothing was wrong exactly; the impression created was. Outlets often prefer clarifications because they concede less.

A correction addresses a factual error — a wrong figure, name, date or claim. The correct information is stated, and good practice is to say what was originally published so the reader can see what changed.

A retraction withdraws the piece or a central claim entirely, because the story's foundation was wrong. This is rare, serious, and often follows a complaint, a regulator, or legal action.

What good practice looks like:

The correction is attached to the article, not only published on a separate corrections page nobody visits.

It says what was wrong, rather than silently editing the text. Quiet post-publication edits — sometimes called stealth editing — are widely criticised, because the record changes with no acknowledgement.

It is dated and visible, not appended in small text at the foot.

Prominence is proportionate. A front-page error corrected in a footnote is not really corrected. Regulators in some jurisdictions can require due prominence.

Why corrections deserve to be read positively. An outlet that publishes corrections is one that checks and admits. An outlet with no corrections is far more likely to be one that does not look than one that never errs.

Where to complain: in the UK, most publishers belong to IPSO or IMPRESS, and the BBC and Ofcom-regulated broadcasters have their own processes. Most countries have an equivalent, and going to the outlet first is normally required.

Social posts are rarely corrected, so the wrong version often outlives the fix.