How to write corrections

Here's the second installment in an occasional series on easy ways to handle the basics -- meaning, from the desk's point of view, how to tell whether routine journalism is doing its job at peak efficiency and whether we need to step in and help. Today's topic is corrections.

Corrections do two things:
1) Acknowledge that we did something wrong.
2) Fix what we did wrong.

Like the vast bulk of this craft, corrections need to be unambiguous, concise and easy to follow. The best way to maximize all these goals and values is this:

Say what we did wrong (and when and where -- enough detail to find it, but not enough to be misleading). Then give the correct information.

We've fallen into a couple of habits lately that don't work as well, so let's look at a couple of examples.

"A Sunday story about efforts to establish a life-sciences incubator at MU should have said that the facility would be built on 2.5 acres along Providence Road."

"An article Monday contained incorrect information about support for a permanent farmers market."

The first one is Tribune style -- fine for the Tribune, but there's no reason for us to adopt unless it works. The second one looks like a new habit (I can't find it in the files before May), so it should be easy to break. Try the two formulas with this sample data set and you'll see why they're ineffective:

"Carlton Fisk hit a home run in the 12th inning to win the sixth game of the 1975 World Series against the Reds. An article on page 1B Monday contained incorrect information."

Here, the reader has no idea what we got wrong: The player, the inning, the game, the year, the opponent? Whatever the mistake is, we're making people work to figure it out.

The Tribune way is even worse:
"An article on page 1B Monday should have said Carlton Fisk hit a home run in the 12th inning to win the sixth game of the 1975 World Series against the Reds."

Now we not only don't know the nature of the purported error, we don't know whether there's been a real error at all. Does "should have said" mean we got one of the facts wrong, or that some pinhead wrote an essay about great World Series moments and inexcusably left this one out -- an issue of judgment (and damn poor judgment, to be sure) but hardly one of fact?

Here's the shortest and most efficient route:
"An article on page 1B Monday misidentified the player whose home run won the sixth game of the 1975 World Series. He is Carlton Fisk."

It's usually better to announce the mistake before giving the correct information. That way, people don't wonder why you're dragging Mr. Fisk and the sixth game into the conversation. And in most cases, it cuts down on the potential for confusion if you don't repeat the incorrect information. There are exceptions to both guidelines, but they're exactly that: Exceptions. If you think you see one, start a conversation.

As in the discussion below about cops ledes, I'm not suggesting my approach isn't formulaic. I am suggesting it's a tighter, more effective formula than the approaches we've been using.