This is the first comprehensive statement on best practices that we know of. Craig Silverman in particular has argued for years that news organizations need to be more transparent about how they handle corrections.

How does a reader know if information in a story was changed? Is there a difference between reporting a typo or an error of fact? Should all stories, no matter how old — be updated? The guiding principle of the document is transparency — that we don’t simply “scrub” content and hope that no one has noticed.

These are some of the questions we tried to address. Two key recommendations are:

When we verify factual errors in digital content, we should amend the copy to make it correct. In all but the most insignificant errors, we should also append a clearly visible note to the article to tell readers that the material was changed/edited/corrected from a previously published version and provide explicit details about what was corrected. For example: An earlier version of this article misstated the overnight price of a litre of gas as $2.40.

and

We have the ability – and responsibility – to correct digital content as soon as we verify something is wrong and no matter how long ago it was published. There is no time limit on making things right.

The report follows a document that addressed a similar issue — “unpublishing” content.

We’re hoping this document becomes a foundation for news organizations to establish and build their own policies for how they publish, update and correct content.

It concerns the strange condition that afflicted Los Angeles CBS reporter Serene Branson. Here it is:

The problem is, Branson may not have had a stroke at all. A Globe and Mail story later Monday suggested that Branson had not been hospitalized — in fact she was examined by paramedics and allowed to go home.

Sure, I used the weasel-word “may.” But I had speculated on someone’s health situation, which is never a good thing. However, that was just one of the things wrong with this tweet. As this post by Patrick Smith rightly points out, the Telegraph story I linked to was horribly sourced. No sources, no facts, no byline. No one should be linking to stories like this. What’s worse? I ignored guidelines for retweeting that I, myself, had authored for the Canadian Association of Journalists last year. Yikes.

How could this have happened?

I remember reacting emotionally to the ridicule being heaped on Branson online in my Twitter stream, with many tweets punctuated with the #fail hashtag, which I dislike. The judgments seemed mean-spirited. Branson has been nominated for two Emmy awards — she’s no amateur and it’s highly unlikely she just forgot what she was going to say or had a tough time spitting out her words. The situation looked like an obvious health issue. Here it is:

I did a quick search online to learn more and found the Telegraph article suggesting she had been hospitalized. Credible news organization, right? Alas.

Nearly two days after the event, we still don’t know what happened to Branson. My pride was assuaged somewhat today by Tara Parker-Pope’s blog post on the New York Times site. Parker-Pope quoted neurologist Daniel Labovitz regarding Branson’s alleged treatment:

I very strongly suspect this was a stroke or transient ischemic attack … Even if it wasn’t a stroke, you need to get it checked out. It’s a tremendous opportunity for her to talk about what stroke is and what T.I.A. is, and what to do. You don’t go home.

Still, I led with my heart, not my head on this one. It’s just one more reason we need a mechanism within Twitter to issue corrections. Craig Silverman has some excellent ideas about how to make this happen.

Shakeups such as this don’t happen often — especially when they are related to publication of a single story. (That said, the T-J, under editor Shawna Richer, made a number of questionable moves in recent months.)

The New Brunswick news outlet, controlled by Irving-owned Brunswick News, announced this morning “there was no credible support” for the paper’s July 8, 2009 report that Prime Minister Stephen Harper “slipped the thin wafer that Catholics call ‘the host’ into his jacket pocket,” during a funeral mass for former Governor-General Romeo LeBlanc. That’s bad enough. But the paper went on to apologize to the bylined reporters on the story for adding “inaccurate” statements “in the editing process” without their knowledge. Wow.

Still, I’m waiting to see what the paper does with the original story in its archives.

The July 8 story has no note either on the T-J’s site or in the NewsScan archive database that I referenced early on July 29. No mention of the correction or the apology. And no hypertext link between the two. Granted, it’s been only 18 hours since the T-J issued the statement and a notation could still be coming in a data dump. However, the paper should have made any change in conjunction with its public statement.

How should news media deal generally with stories they have determined to be substantially wrong?

Reuters released its style guide, called the Handbook of Journalism, on July 9, 2009. It outlines its internal filing procedures when “the story is fundamentally flawed” and notes that “this will alert our online colleagues to pull each version of the story from websites and to contact online customers to ask them to remove it.”

But simply removing all online evidence of an egregious error may not be the best practice. How can the audience be convinced an apology is comprehensive if the source document no longer exists? How can future researchers learn of errors in published stories? Would it not represent a step backward from the days when microfilm gave us the full published record, warts and all?

The New York Times sets a higher standard on these matters. The Times website still hosts the full stories published by Jayson Blair, the reporter the Times revealed in 2003 as having committed “frequent acts of journalistic fraud.” In addition, it annotates each page of his stories with known inaccuracies. For example, this 2002 story on the investigation into the Washington-area sniper contains a long list of corrections.

Many news organizations still lag in fully linking corrections and apologies to the originally published items. Many more succumb to the temptation to simply wipe clean the source of the embarrassment. I hope this won’t be the case here. There is an opportunity for the T-J to turn a major embarrassment into a minor source of pride.

Alexander blames the Post’s layoff of copy editors since 2005; it cut the number of full-time copy editors by almost half — from 75 to 43.

Obviously, errors of this type are bad. Sloppy presentation suggests sloppy reporting. But what’s the best way to get clean copy? Are layers of copy editors really the solution?

I’ve been playing around with a new tool for crowd-sourced editing called GooseGrade. With this free widget installed on a site, curmudgeonly readers can easily apply their virtual red pen to suggest corrections. It’s an open-source approach that empowers the audience to improve content.  Note: I’m not using it currently because it locked up my WordPress (publishing system) last week. But I liked the product and will try it again.

How great is the problem Alexander highlights? A 2007 study by Michael Bugeja and Jane Peterson from Iowa State University suggests it’s around 8% of all errors. Spelling, grammar and typos account for about 4% of errors; using the wrong word — 3.6%. (Misidentifying a person or reporting a wrong date — that’s another matter).

And who actually finds errors? It’s difficult to assess the effectiveness of copy editors. (They no doubt make a big difference.) But a 2005 study by University of Nevada academics Donica Mensing and Merlyn Oliver suggests that adding levels of oversight may not be the most effective approach. In the survey of editors at 300 small-town dailies in the United States, titled “Editors at Small Newspapers Say Error Problems Serious,” here’s who discovered the error once a story was published:

• a person mentioned in the story (50%)
• a reader (26.5%)
• a member of the newsroom staff (23%)

Further, even when published, minor errors of the kind that Alexander points out do not seem to significantly affect a media organization’s credibility. In his 2002 study “Getting it right? Not in 59 percent of stories” University of Oregon professor Scott Maier, states: “By several measures, the relationship between errors and newspaper credibility was statistically significant but weak.” He adds that errors have a slightly greater effect on how people judge the story itself, but generally only when the reporter’s mistake involves an error of judgment.

It’s interesting that many of the people who commented on the Post’s article were critical, but not about the increase in errors. They were more concerned that the Post had failed to account publicly for a controversial fundraising event that came to light last week. In a flyer, the Post advertised a “salon” at the publisher’s home that offered lobbyists access to Obama administration officials — and the Post’s editorial staff. (Publisher Katharine Weymouth did comment publicly on the issue the same day.)

As one commenter pointed out, the bigger problem is that many news organizations lack a system for quickly addressing errors — factual, subjective and ethical — pointed out by the public.

It’s this lack of accountability that seems to be an issue with media credibility — not the errors themselves. A study by the Canadian Media Research Consortium concluded in 2008 that only about half of Canadians believed the media was doing its job with full accuracy (51.6% said news organizations generally got their facts straight; 55.7% said news organizations were careful to check and verify information). But the kicker was that only 33.8% said they believed news organizations were willing to admit mistakes.

Better procedures for finding and correcting published errors may be what’s needed — not more copy editors.

July 10, 2009: Correction made on silly usage error (see Comment 4 below). Thanks, Carol!