A reaches, on average, a small fraction of the audience that saw the error.
The is structural rather than a failure of effort by anybody.
An error travels because it is new, surprising and consistent with something a reader already suspects.
A has none of those properties and is in addition an by the publisher.
Publishers have responded by making s more prominent, which helps less than expected.
determines who sees the and not what they do with it.
A reader who encounters a to a claim she never saw acquires the claim and the doubt together.
Repeating a false statement in order to deny it leaves the statement more familiar than it was.
is one of the strongest of whether something is later judged true.
The recommended practice is therefore to lead with the correct statement and mention the error once.
'The bridge opens in March' works better than 'it is not true that the bridge opens in June'.
This is well evidenced, inexpensive and contrary to the of everybody who writes s.
The is to name the error first because that is what the is about.
A second finding is more uncomfortable for publishers and concerns who is corrected.
Errors about powerful institutions are corrected at several times the rate of errors about individuals.
The difference is not editorial policy but who complains and how they can threaten.
A newspaper that corrected in proportion to harm rather than to pressure would look very different.
No newspaper has tried it, and the reason is not that anyone thinks the present arrangement is right.