USC Annenberg Online Journalism ReviewUSC

To Fix or Not to Fix: Online Corrections Policies Vary Widely


More newspapers are choosing to correct errors in their online archives and Web sites by editing stories rather than simply attaching corrections. But should archived content be tinkered with?

According to the list of valedictorians from Washington D.C.-area high schools published June 22 in The Washington Post, Andrew Carl won top academic honors at the prestigious St. Alban's School this year. A correction the next day set the record straight, noting that Andrew's last name was actually Karlin. But in the paper's electronic archives, "Andrew Carl" now and forevermore will bask in the glow of undeserved honors snatched from the real winner by someone's careless error. The hapless Karlin has been relegated to a corrections box tucked away on the right-hand margin of the computer screen.

Most papers used to follow the same policy as the Post regarding corrections in online archives -- retaining all stories exactly as they appeared in the final print edition of that day's newspaper, errors and all, and appending any corrections in a separate box. Many still do. But a growing number of papers have begun to adjust their policies. Online publishing technology makes fixing errors a snap, and they have decided it's silly to continue refraining from taking advantage of that. But tinkering with the content of newspaper archives is surprisingly controversial. Indeed, experience has shown it's not as simple a matter as you might think.

The case for fixing errors

One of the most vocal advocates of cleansing errors from the electronic record is Michael Jesse, library director at The Indianapolis Star. He used his recent stint as chair of the news division of the Special Libraries Association to chide his tradition-bound colleagues into adjusting their thinking on the issue.

"When we first started having electronic news archives, librarians were thinking of that as the official record of what was published. So they would append corrections just as they did with clippings. Once we had electronic archives and were able to do more, we generally felt that we weren't supposed to," says Jesse, attributing the hesitancy to a "deep-seated feeling among librarians around the country that it's just fundamentally wrong to make a change in a story that we published in print."

As Jesse sees it, however, microfilm preserves for the historical record a photographic reproduction of exactly what appeared in the newspaper on any give day. But an electronic archive can and should be more, the latest and best effort to disseminate the truth. With that goal in mind, he resolved to open up the original text in the archives and make any fixes called for in published corrections. "We will also do it in some situations when we haven't run a published correction," for example when a reporter finds a misspelled name in an archived story. "We're not going to publish a correction about an error that we made a year ago but it's still wrong so we will fix it."

At the top of any such altered story, librarians at the Star attach a boilerplate statement noting that the original published version contained an error that has been corrected for the archive. "Sometimes we will be more specific and explain the change. We also refer people to the published correction, which is in the archives as a separate document," Jesse says. But in every case, "we will make the smallest change necessary to correct the error. So it's not a license to rewrite the story."

Jesse acknowledges, "Sometimes it is awkward because you have to almost rewrite the paragraph to make it read correctly." But in such cases, librarians "consult with the writer or editor to determine exactly what words were wrong that we would need to change in some way to make it read correctly."

In some cases errors permeate a story to such an extent that it's impossible to fix, in which case Jesse wouldn't hesitate to remove it from the public archives, perhaps replacing it with an abstract of the story and an explanation of why it was stricken. "The logic is that if you didn't know something was wrong and published it, that's bad enough. But now you know it's wrong and you've even published a correction saying it was wrong. Yet by policy you insist on essentially republishing that every day." That's inexcusable, in Jesse's view. "What if it has libelous information? Are you going to intentionally retain libelous information just because that's what you published? It seems to me that from a legal perspective, you're on much safer ground acknowledging that you made a mistake and fixing it."

Jesse made his case for fixing errors in electronic archives in a column for News Library News in the Fall 2002, when he was chair of the SLA's news division. But he has won few converts so far. "A handful of papers have changed, but I don't think there are many," he says.

The most prominent is the Chicago Tribune. The paper's librarians began fixing errors in the text of stories destined for the electronic archives several years ago for logistical reasons, explains Debra Bade, editor of news research and archives at the Tribune. The's database at the time didn't have a way to attach corrections to stories. "So we were looking at the very likely possibility that an article that had a correction would go out on the Web site with an error but with no correction." So librarians began making the corrections in the text, attaching a note explaining what was changed "to make it clear that there was a correction," she says.

While the policy was initially adopted for practical reasons, Tribune staffers have come around to the view that there are also sound intellectual and legal rationales for the new approach, Bade says. "Certainly if we caught an error on the first press run, we'd correct it. There are continual revisions as a newspaper is produced," she explains. "We're now seeing the text that goes online as yet another edition and we want that to be correct also."

The extra workload for the library staff is manageable, Bade adds. Simple fixes of misspellings and typos don't require any input from the editorial staff. "If it's something that's more involved, we work with the editor of the appropriate section and work out how it should be worded." That's old hat for librarians who are often the ones who spot errors and end up negotiating with reporters and editors over the text of published corrections, Bade says.

The new policy has gratified the Tribune's lawyers, she adds. They didn't initiate the discussion that led to the decision to begin fixing errors in the archives, "but our lawyers participated and urged us to adopt this approach."

The case for leaving errors in place

Lawyers for papers that retain erroneous stories in their electronic archives are quick to challenge the notion that the law requires permanent erasure of old errors.

"If you're going to keep an old story up there, the correction should be displayed as prominently as the original story," says Kevin Goldberg, a media lawyer with Cohn and Marks in Washington D.C. But keeping even a libelous story online, with a conspicuous correction attached, doesn't constitute a republication of the libel. "You're not asserting it as fact any more. What you are trying to show is that in fact it is not correct," he explains.

Karlene Goller, an inhouse counsel for the Los Angeles Times, adds that retaining the erroneous story in full with a prominently displayed correction is the surest way to set the record straight. "If you go into the body of the story and fix it, people won't know that its been fixed and they won't go back and find the correct information. They'll just assume they read the story and will blow right past it," Goller says.

On that advice, the Los Angeles Times remains among the majority of papers that make no substantive changes to archived stories, whether they are accurate or not. But the paper has budged a bit from the traditional approach of treating the electronic archives as a repository for the verbatim contents of the printed paper, with no changes or additions of any sort.

Dorothy Ingebretsen, director of the Times' editorial library, explains the evolution in thinking. "The past practice was always simply to attach the 'for the record' or correction that ran with the story. The feeling was that text as published was sacrosanct. You couldn't even correct the misspelling of a name that didn't have a published correction," she says. "Then, as databases became so heavily used, we realized that misspelled names couldn't be searched, so that became a real issue.

"So we adopted a policy that the library could go in and make changes, such as fixing typos and spelling errors, to facilitate retrieval," she says. Librarians will also tinker with stories in the databases in other instances, for example when the paper in April opted to change the spelling of the name of the radical Iraqi cleric from Muqtader to Muqtada Sadr. "We added an unpublished note to old stories mentioning the new spelling, so if you search you'll find them all," adds Suzanne Oatey, manager of archiving for the paper. "But we have not been changing the text of stories at all."

On rare occasions, the library staff will make an unannounced fix of typos, but only for the purpose of assuring that the story won't be missed in a key-word search. Those occasions appear to be rare indeed. The library staff wouldn't even go so far as to add a missing "n" in the first name of Sen. Dianne Feinstein in a recent story. Once you adopt a policy of fixing "minor" errors, "It really becomes a fine line," explains Oatey. "Some corrections would be easy to make but some corrections are vague and not so easy to make, such as how legislation was characterized." To fix those, "librarians would have to rewrite stories," says Oatey. "So we find it best to leave the story the way it is."

The New York Times' policy permits a bit more tinkering. In its electronic archives, the New York paper recently changed the name of Dr. James Cleeland to Cleeman and the location of a proposed soccer stadium in Cleveland from "north of the city" to "south of the city," without leaving any indication that the printed stories were erroneous.

But a more substantive, and also more serious, error in a July 12 story was left unfixed in the archives, even though the insertion of three words would have done the trick. In the erroneous story, the Times reported that the Venezuelan President Hugo Chavez had been "charged with electoral fraud." As a subsequent correction noted, Chavez has been charged with fraud "by opposition politicians" but not, as the published story implied, in a court of law. The error remains in the archives. But a bold-faced "Editor's Note Appended" tag appears at the top, and the text of the published correction appears at the bottom of the story.

Toby Usnik, director of public relations for the Times, said the paper is guided by "common sense" in correcting errors. "For misspellings of names, places, and other fixes of this type," the keepers of the archives, in consultation with the editors of the print edition, will fix the errors in the online text as soon as the errors are spotted, whether a published correction has appeared or not. But factual errors "that require something as intricate as an editors' note to explain are not something we even try to correct in the text," he added. "We just append the note and make readers aware that we have done so."

Which approach works best?

Jesse professes to be astonished that any paper, not to mention one as esteemed as The New York Times, would knowingly leave erroneous stories in its database. He would fix all correctable errors and expunge from publicly accessible archives all stories that are beyond repair, replacing them with a retraction notice and maybe even an abstract of the erroneous story, "so that you're clearly fessing up."

In the case of stories such as the fabrications of former New York Times reporter Jayson Blair, because of their importance in the history of American journalism, Jesse might create a separate database where they could be displayed in full, clearly labeled as the frauds that they are. "But I don't think they belong in The New York Times' active database. People go to that expecting truth. If the editors at the Times know they're a bunch of lies, I can't fathom why they keep them there."

To be sure, Jesse's solution is much easier explicated than implemented. If the Jayson Blair stories should be expunged, what about all the stories in The New York Times and many other papers that lent way too much credence to now discredited reports about massive stockpiles of weapons of mass destruction in Iraq? Should those, too, be flushed out of the nation's newspaper archives?

A story in the archives of the Chicago Tribune, the most prominent adopter of the policy of rewriting erroneous stories, illustrates how difficult that task can be in practice. The story which required a subsequent correction was a major team-reported investigative piece about a regulatory board that was said to be "riddled with potential conflicts of interest."  The evidence for that claim included a finding that "campaign contributions are a common thread" among appointees to the board. But in fact, as the correction later conceded, one of the appointees had never been an arbitrator for a union that contributed to the governor's campaign.

In the archives, the text of the published correction appears prominently at the top of the story but the error remained in place more than a week after the correction was posted. A fix arguably might require not only deleting the false claim but also at least slightly toning down the sweeping claims that the board is enslaved by political interests. In a case like that, simply posting the story, error and all, with a prominent correction at the top, begins to look like the best alternative. Duly informed, readers can decide for themselves how much weight to take away from the thesis of the piece.

But Bade insists that even in a case like that, a rewrite is in order. There are no exceptions to Tribune's rule that errors will be corrected in the online text. The failure to correct the reference to the appointee was inadvertent, she says.

If a policy that calls for rewriting erroneous stories is tricky to implement, the policy of leaving errors in place certainly has perils of its own, as a perusal of erroneous stories in The Washington Post illustrates. The story about the snubbed valedictorian was a minor glitch in the larger scheme of things. A story about a bizarre visit to Capitol Hill by Rev. Sun Myung Moon contained a more substantive error. The story that ran in the paper reported that Sen. Mark Dayton persuaded his hometown paper to write about the incident. In fact, as a correction published the next day noted, the paper wrote about the incident on its own volition, not as a result of calls from the senator or his staff.

The correction appears in the Post's archives alongside the erroneous story. But it is lost in a clutter of menus well below the "fold" on the right side of the screen, with no label at the top of the story to indicate that it has been corrected. And anyone who pulls the text of the story out of the archives by cutting and pasting it into another file won't pick up the correction box.

That could easily facilitate the perpetuation of the error, which has already been spread far and wide via The Washington Post's wire service. The erroneous story has appeared without any correction in publications ranging from the Omaha World-Herald to the Drudge Report.

Jesse insists the Post could have minimized that risk by revising the story in its archives the day after it appeared in print and treating that as the "final edition" of the piece. More online news staffs are already coming around to that way of thinking.

Indeed, at The New York Times, according to Usnick, the online and print editorial staffs now operate on parallel tracks regarding corrections and updates to evolving stories. The online editors run what they call the "continuous news desk," which is "kind of an in-house re-write desk that feeds the Web site," Usnik explains. "As we know new information, we add it. As information changes, we update it. If we misspell a name we spell it right and update the story again." On the Times' Web site, that process may continue even after the final print edition has gone to bed.

To be sure, for the archives, Usnik says the Times still sticks with the final print version of stories, with minor errors fixed and more substantive corrections attached. Jesse is convinced it's only a matter of time before papers come to regard the final online version as the last, best iteration that deserves a permanent home in the archives.

A simple semantic shift regarding what constitutes the "final edition" of a newspaper "should appease those who think it's wrong to change stories in the archive," he says. "Nobody minds the fact that we're correcting errors in the final print edition of the paper. Now, we have another final version of the newspaper, an online-only edition. Once you accept that this final iteration of the story was 'published' in some manner, I think whatever qualms a lot of people have had will go away."

Related Links
Michael Jesse's column for News Library News
Special Libraries Association

Michael Jesse, library director, The Indianapolis Star