Last week Arthur Brisbane, the new public editor of the New York Times, posted an illuminating exchange between a reader of the paper and one of its top editors.

The reader asked: What’s with the way stories change all the time on the website? “How does the newspaper of record handle this? I read something, and now poof, it’s gone without a trace.”

Jim Roberts, the paper’s associate managing editor, responded: “We are constantly refining what we publish online.” He added that the paper often”uses the final printed version as the final archived version that stays on the Web.” But not always! There are “many exceptions.”

The headline over the column reads “Revising the Newspaper of Record.” But what the exchange reveals is that, right now, there is no record of the newspaper of record. The Times is revising its copy online all the time. No doubt the vast majority of these “refinements” are trivial or uncontroversial. But some of them are likely substantial. (Here was one right on the edge that was filed at MediaBugs.) If I understand Times policy correctly, when a change fixes an outright error, it is supposed to be marked with a correction notice. But there’s no record of these changes, so the Times could be cutting corners here and we’d never know.

When I raise this issue I sometimes hear back some variation on “What’s the big deal? Wire services change their copy all the time. Newspapers have always revised stories from edition to edition. How’s the Web different?”

I’ll tell you how: When newspapers change a story from the early to the late edition, the early edition is still out there for people to read and compare. When you change a Web page, the older version disappears, unless you take active steps to save it.

That, of course, is precisely what the Times — along with every other news outlet that’s committed to accountability — ought to do. Whenever a published story is changed, the paper should make the previous versions available to its readers. (I’ve outlined this idea here, written about it more here, and there’s now a WordPress plugin to demonstrate it in action.)

Let the world see the changes. This is all published, public material; there’s nothing to hide. With this one change to its publishing practices online, the Times can make good on the promise of the old “paper of record” moniker and become a website of record — while giving itself real freedom to keep improving the stories it has already posted.

Here are some potential objections, and my responses:

(1) What about actual errors that you’ve corrected? Unless they’re libelous and there’s some legal need to take them down, you can leave the errors visible behind a “revised version link” — while clearly marking them as errors that have been corrected. This is the most foolproof way of keeping your correction process transparent and trustworthy. When material is removed for legal reasons, a note can indicate that.

(2) Why provide so much excess material to readers? Aren’t we all drowning in too much information already? You can hide the previous versions behind links, the way Wikipedia (or our WordPress plugin) does. Most readers will ignore them — except every now and then when they notice that something’s changed and they want to see why. The Web has more than enough “space” for this data; it’s all just files on disk drives or data in databases.

(3) What you ask for makes sense, but it’s a ton of work to make that sort of change on a website as complex as a major newspaper’s! Right. Sure. I don’t expect this to happen tomorrow. But it’s worth beginning to plan now. I’m firmly convinced that this is an essential “best practice” for trustworthy news publishing online. It will happen, eventually. Why not get the ball rolling?

UPDATE: Mahendra Palsule pointed me to his account of a situation last month where the Wall Street Journal’s modifications to live stories made it look as if it might have scrubbed a controversial quote from a story (though it hadn’t done that at all). In comments there, Zach Seward of the Journal’s online team mentions that the paper is discussing the revisions-display idea. Maybe a little healthy competition will get this practice adopted!

Craig Silverman had a fascinating column last week about changes that Slate has made in its corrections policy in the wake of an embarrassing dustup with Politico. Here’s Craig’s pithy summary of this bizarre Escher-esque episode (which I also wrote about at the time):

In July, Slate published an article that provided evidence that Politico was routinely scrubbing errors out of its stories without adding a correction or similar notice for readers. Slate’s story, of course, was noticed by many journalists and many of us, myself included, weighed in with criticism and reaction. Then the tables began to turn: Politico pointed out several mistakes in Slate’s reporting and, in the end, Slate admitted that its correction policy had a provision that allowed for it to scrub factual errors from stories without the addition of a correction. (Provided that the error was spotted within twenty-four hours of publication, and it was spotted by someone at Slate.) Pot, meet kettle.

Slate is not the only publication with a post-publication window for tinkering with stories. At MediaBugs, we keep encountering this issue — as with this bug filed about a New York Times story earlier this summer. If you read that discussion, you’ll see a Times editor admitting the paper made substantive changes (though not technically the correction of an outright error) to an article after it was posted on the Web. Does the Times have a window, like Slate’s, in which it considers it okay to alter published articles online? How about headlines? Both Slate and the Times have always been extra meticulous in thinking through their approach to corrections, yet this post-publication change policy is a problematic area for both.

As a result of the Slate-Politico fracas, Slate has now decided to close its 24-hour window. This seems to be the upright, ethical response. It’s plainly well-intentioned and carefully thought out. But I think it’s the wrong direction to go: it’s a step backwards. It continues the tradition of allowing the limitations of print-think to govern our online behavior. The Web lets us improve our stories after we publish them. Why should we tie our hands?

We do so, of course, out of a sense of accountability. We don’t want to seem to be pulling fast ones on our readers. We don’t want it to appear that we’re “scrubbing” the record, as Silverman puts it.

There is a simple solution to this problem: Throw open the windows — and keep them open! Let yourself make changes; just don’t hide them. The same technology that lets us keep tinkering with words we’ve published also lets us store and display every alteration we make.

Instead of agonizing over the duration of the period in which we allow ourselves to change an already-published story, we should just be transparent about all the post-publication changes we make. Let’s give readers the “history” of our articles the way Wikipedia shows all the changes to each of its pages. As James Bridle puts it: Everything should have a history button!

This achieves two useful goals at once: It frees your hand to keep improving your articles even after you’ve published them, which is one of the great advantages Web publishing offers over print; and it makes sure that your readers will never think you’re hiding anything from them. Most readers will never need or want to see the revisions you’ve made; they just want to read the current version of the story. But in the rare case that someone raises a question or a dispute about a change, the record will be public.

If and when there are major substantive errors that you need to correct, you can still go the whole distance and add a traditional correction notice to any page. But if you’re fixing small errors or little details that don’t warrant that, you wouldn’t have to bother.

The technical hurdles to such an approach are minimal; most modern content management systems store all the revisions to each story in a database anyway. There’s a small amount of design work in figuring out how to expose the revisions to your readers. But it’s not a hard problem. I’m already showing all my revisions here on this blog — just look at the bottom of the full-page version of this, or any, post. (And if you run WordPress, you can install the WordPress Post Revision Display plugin too if you like!)

I’m firmly convinced that over the next few years this practice will become as commonplace in online publishing as “print this page” buttons and comments. So my challenge to Slate, the Times, and every other online publisher is: why not become a leader in this realm? Stop closing your “windows.” Instead, open them up and show your changes.

One of the things we’re trying to accomplish with MediaBugs is to encourage a change in newsroom culture. Journalists are still often reluctant to admit error, or even discuss the possibility of a mistake, for fear that it undermines their authority. But today a growing number of them understand that accuracy is best served, and authority best preserved, by being more open about the correction process.

That is the attitude we’ve encountered at most of the Bay Area news institutions where we’ve demoed MediaBugs. Unfortunately, it’s not what we found at Bloomberg when we tried to obtain a response on behalf of blogger Josh Nelson, who’d filed an error report at MediaBugs about a Bloomberg story.

Nelson raised a specific and credible criticism about the headline and lead on a Bloomberg report based on a national poll. Bloomberg’s coverage, Nelson argued, didn’t accurately reflect the actual question that its pollsters had asked about the Obama administration’s ban on deepwater oil drilling in the Gulf. (The story and headline said that “Most Americans oppose President Barack Obama’s ban” on such drilling, but the poll asked about a general ban on all Gulf drilling, while Obama has placed a temporary hold on deepwater drilling.) Bloomberg, as we described recently, circled the wagons in response.

The news service, of course, has every right to “stand by its story.” But since Nelson has raised a reasonable question, Bloomberg’s public deserves a reasonable response. It would be useful for its readers — and its colleagues at publications like the San Francisco Chronicle, which reprinted the story — to hear from the editors why they disagree with Nelson. Apparently they believe their copy accurately reflects the poll they took, but they have yet to offer a substantive case explaining why.

Institutional behavior of this kind always leaves me scratching my head. A comment posted on our previous post on the Bloomberg bug over at the PBS MediaShift Idea Lab proposed an intriguing theory: A former Bloomberg journalist suggested that the company’s personnel policies came down so hard on employees who made errors that they were reluctant to admit them at all.

These standards, which are meant to make people super-careful before publishing a story, actually serve as a perverse incentive and cause people at all levels of the newsroom to resist correcting stories after they are published if there is any way to justify leaving the story as is.

This was, we thought, worth a follow-up, and so we contacted the commenter. He turned out to be Steven Bodzin, who’d worked as a reporter in San Francisco and Venezuela for Bloomberg for four years before leaving the company in March. My colleague Mark Follman spoke at length with Bodzin last week.

Bodzin said he “rarely saw complaints from the public get ignored.” He told us that Bloomberg’s culture is actually “hypersensitive” to public response but especially focused on issues raised by sources or by customers who subscribe to its terminal service (Bloomberg’s business was built on selling real-time market data to the financial industry over its own network — only later did it begin distributing news and information on public networks).

Bodzin described his own “prolific” first year as a Bloomberg correspondent, during which five of his stories were cited as exemplary in the company’s weekly internal reviews. He also had an unusually high number of corrections that year — which he attributed to the intense pace of the job — and got the message from his superiors that “you really have to bring that down.” He says that made him more careful. But he observed that the stigma that Bloomberg attached to corrections also encouraged a sort of silence in the newsroom in the face of potential problems.

Certainly there were situations where you realize something is wrong but you’re gonna say “I didn’t see that” or just forget about it.
At Bloomberg that’s considered a really serious offense … but at the same time, if you or nobody else mentions it … no harm no foul. I think it happens. One time a colleague of mine, who’d already had one correction that day, saw one and said to me: “I am Olympically burying this error.”

We asked Bodzin about the specific issue Josh Nelson raised about the drilling-ban poll.

They see this case as a question of interpretation, a judgment call — this is their own poll, a lot of reporters and editors are involved, so they [would all] get a correction. So they aren’t going to want to do it.

What we’re looking at here isn’t some revelation of blatantly irresponsible behavior but a subtler insight into the complex interplay of motivation inside a big organization. Bloomberg is hardly the only company where such a dynamic may be at work. What’s important is that the people who lead such institutions understand the need to change the dynamic — to rebalance the incentives inside their newsrooms.

Unfortunately, this incident suggests that Bloomberg’s culture today clings to the wagon-circling habit. As so much of the rest of the journalism field moves toward more open models, it remains an old-fashioned black-hole newsroom, happy to pump stories out to the world but unwilling to engage with that world when outsiders toss concerns back in. Bodzin explained, “Staffers aren’t supposed to talk to press at all — you’re supposed to send reporters to the PR department.”

And that’s exactly what we found when we tried to get comment from Bloomberg about the issues Bodzin raised. When we asked senior editors at Bloomberg to discuss their own policies and newsroom culture, they shunted us over to Ty Trippet, director of public relations for Bloomberg News, who wrote back:

Our policy is simple: If any Bloomberg News journalist is found to be hiding a mistake and is not transparent about it, their employment with Bloomberg is terminated.

So Bloomberg looks at a nuanced psychological question of newsroom behavior and responds with an “Apocalypse-Now”-style “terminate with extreme prejudice.” Doesn’t exactly give you confidence about the company’s ability to foster a culture of openness around the correction process.

Earlier this week Bloomberg announced the hire of Clark Hoyt — the Knight Ridder veteran who for the last three years served as the New York Times’ public editor. In that ombudsman-style role he served as a channel for public concerns about just the sort of issues we are raising here about Bloomberg.

Though Hoyt’s new management job at Bloomberg’s Washington bureau isn’t a public-editor role, it does put him squarely in the chain of command for stories like the oil-drilling poll. So maybe he’ll look into this, and also more generally at how Bloomberg handles public response to questions of accuracy. Right now, the company’s stance is one that hurts its reputation.

News organizations’ default response to criticism is to circle the wagons.

“We stand by our story!” is a stirring thing to say, and sometimes it’s even the right thing. But in the web world of 2010, where everyone has a public platform, ignoring critics can also squander a news outlet’s credibility and alienate its audience.

The basic premise of MediaBugs — which I laid out in this video — is that news organizations can begin winning back the public trust they have lost by engaging civilly, in public, with people who criticize them about specific errors. Whoever is right in the end, and whether the newsroom decides to run a correction or not, the editors are better off explaining their thinking than slamming the door on dialogue.

For an example of precisely the wrong way of handling legitimate questions about coverage, consider the controversy over a recent Bloomberg opinion poll.
[Read more…]

Last month, the hardworking gang at Politico got into a dustup with critics after an editor made a change in an already-posted story. The story was about the Rolling Stone/General McChrystal affair; the change removed a phrase that described how beat reporting works; the phrase had drawn considerable attention, and so did its disappearance.

I’m not going to add to the volume of commentary on that affair. I’m interested here in the larger issue of the mutability of online content, and how responsible news organizations deal with it.

A story posted at Slate yesterday sheds considerable light on this issue, in the course of making a few stumbles of its own. (The story includes quotes from a recent post I wrote about best practices in online corrections.) It’s remarkable that, after 15 years of Web publishing experience, we haven’t gotten better at handling changes to news published online. Before this post is done, I will offer a straightforward, concrete proposal for doing so.

[Read more…]