Regret the Error

NY Times corrections database goes live

The good folks at Check Your Facts recently published an item stating that the roll out of the New York Times internal corrections database is complete. The paper is now entering all of its corrections into a central database, much like how the Boston Globe, Rocky Mountain News and a few other US papers have been doing. Reports CYF:

Check Your Facts has learned that the New York Times may finally be making good on one of the recommendations laid out in a 2005 report titled “Preserving Our Readers’ Trust.”
Fresh from a "training class," a well-placed source at the Times says a computerized tracking system for corrections is being implemented.
Correspondent Adam Klasfeld learned of the new system through correspondence with a New York Times editor (Check Your Facts is withholding the editor's name in order to maintain open communications on the matter). The NYT editor said that no tracking system as described in the report currently exists but that a “training class” was held for a new system. “We are just implementing it,” wrote the editor. “We are beginning with corrections on October 1.”

We reported on the roll out of the system in December of last year after speaking with Greg Brock, the paper's corrections editor:

A final note from the Times: Brock says the paper does have a corrections database that is being used by some departments. Each department can see its own corrections tally, and Brock has access to the total data. He says they are working to roll it out within all departments. This database was one of the recommendations of the Siegal Committee. It's good to see the paper following up on this project. We hope it's fully operational ASAP.

In an email response this week, Brock said the system was fully operational as of September 17. "So it is in place, though we are continuing to train editors who will use it only occasionally because they work on sections or special issues of magazines that publish just a few times a year," he wrote. "The main news desks are already using it."

We asked if the paper plans to release its tallies the way some other paper do (see these links: 1, 2, 3, 4), and Brock said the issue hadn't yet been raised, but that he would look into it. As of now, he is distributing reports to the relevant editors on a monthly, quarterly, and annual basis in order to show them the most common types of errors in their sections. It's good to hear the data is being put to use to help fuel preventative initiatives.

Naming the Virginia Tech shooter

An editor's note from the Washington Post:

The Washington Post today begins referring to the gunman who killed 32 people Monday at Virginia Tech as Seung Hui Cho, with his surname Cho at the end, after the family said it preferred to give the name in that order.
In earlier articles this week, The Post gave the name as Cho Seung Hui on first reference and as Cho thereafter.
Cho was an immigrant from South Korea, where family surnames are written first, followed by two-syllable given names. The given names are sometimes hyphenated. Many East Asian immigrants shift the order of the names as part of their adaptation to life in the United States.
The Post's policy is to defer, when possible, to the individual's or the family's preference on whether the surname is written first or last. In this instance, Cho himself appears to have written his name both ways, and the family had not stated a preference. The Post therefore had followed the practice of Virginia Tech and Virginia State Police, which have consistently put the family name first in references to Cho.
Yesterday, however, the Associated Press reported that the Cho family, which issued a statement regarding the tragedy, said its preference was to give the name as Seung Hui Cho. The Post thus will now place the surname at the end. The Post's style is not to hyphenate the given names.

UPDATED: The soy milk files

On Friday morning BoingBoing published a note from a contributor to Wikipedia's Reference Desk detailing what appeared to be a case of scrubbing (fixing an error but not noting it with a correction) on the New York Times' website. From the letter to BoingBoing:

On 3/27, someone submitted a question to the Wikipedia Reference Desks (where I volunteer) asking about what seemed like a silly claim in a NYT health article published that same day.
The article, which summarized a recent panel study on the health benefits of beverages, claimed that it was illegal to fortify soy milk with Vitamin D, and, because soy milk did not contain calcium, that the soy milk was not recommended as a substitute for cow's milk.
The NYT reported that claim as true, and used it to close their article.
Over the next 24 hours, the ref desk volunteers (including myself) followied the info back to its source, dicovered that the error was due to the original study's citation of a 1971 article on this point (which seems like pretty bad science, given how much nutritional laws have changed in that time)...and further tracked down plenty of evidence on both the public online documents of the Federal Register and on our own shelves which showed this claim to be absolutely false.
Wednesday night, I sent a letter to the NYT.
Sometime yesterday, the entire last section of the NYT article, which (according to the NYT website) was NYT's most emailed article for the last two days, was changed to the following sentence:
"Fortified soy milk is a good alternative for individuals who prefer not to consume cow milk," the panel said.
No letter or correction has been issued, however, calling attention to this error...

You can read the Wikipedia post about this here. The story in question, which has now been officially corrected, is here. We contacted Greg Brock, the senior editor at the Times who oversees corrections, for an explanation. He explained the issue in a letter that was also sent to the Wikipedia volunteer (UPDATE: his response is below):

...Since I oversee corrections at The Times, I thought I should give you an explanation, given some of the serious points you raised.
This error was pointed out to us by at least a dozen or more readers. We immediately started trying to reach Ms. Brody, who was traveling abroad.
Our policy is that we never publish a correction without first talking to the reporter or the editor who was responsible for the error. If that takes a few days, then we have to live with that. We only explain delays in corrections after they are more than 30 days old. You will see such explanations in our corrections column on Page A2.
As far as correcting/updating articles, we do indeed update and correct articles on the Web site. We have been doing that for as long as I can remember. I would consider it unethical if we did NOT correct an error in the Web article if we indeed were aware of the error. We make that correction as soon as we confirm the error – as we did on Friday. We do not wait to correct the Web version until a correction has appeared in print.
We do not, however, make changes in articles in The Times's in-house archives for a number of reasons, including legal ones. We have to have at least one copy of the article on file exactly as it appeared in print. Once we publish a correction in the print paper, that correction is then attached to the article and a note is put atop the article stating that the correction is appended...
And as has been pointed out to you already, we did publish a correction on Saturday morning on A2. It wasn't until yesterday that we had completed the research (we often use many sources and then even confirm that information with even more experts in the field). We may seek information from Wikipedia, but we don't rely on just one source. Our experience has been that googling a subject is the surest way to make yet another error. Web sites are filled with incorrect information, as you know.
I hope this explains our handling of this correction. If not, feel free to write me at The Times and I will be happy to try to answer any other questions.
Also, feel free to post this note on your Web site or anywhere else...

Here's the correction, as published on March 31:

The Personal Health column in Science Times on Tuesday about healthful beverages included incorrect information from the Beverage Guidance Panel about soy milk. It can indeed be legally fortified with vitamin D.

UPDATE April 4: Joshua Farber, the Wikipedia contributor, sent us a follow up note in response to Brock's letter:

Though Times editor Brock's response is illuminating, I believe the consequences of the two very virtuous goals the Times has (to wit: making changes ASAP, but issuing corrections only after an author has been contacted) is to create a short but not insignificant period during which "scrubbing" still applies, with the resultant effects and potential confusion. 
After all, we're talking about an article that, for the two days BEFORE the change was made, was the MOST emailed article on the NYT website.   After that time, it was still in the top ten.  By the time the correction had been issued, however, it had dropped off the list.  No small stakes, that.   The potential for misinformation ending up "out there" remains high.   
I sent an email saying so to Mr. Brock, and suggesting that perhaps if the public had been informed about the reasoning behind this delay that he says they "can live with", any future perception of impropriety could be mitigated.   However, I further suggested that the appearance of impropriaty could be avoided almost completely if they would only change their policy to make sure to change the PUBLICATION DATE of the web-based article at the same time as they change any facts in that article.   As an information literacy professional, I know that citation of an article must include date accessed and also date of publication, but still, the assumption for web-based content that the MLA uses is that the listed publication date OF an article represents its most recent incarnation.   If the Times were to take on this assumption, and change theor own approach accordingly, I believe the public interest -- and the potential for misinterpretation of this policy as "scrubbing" -- would be better served. 
More of my thoughts are here.

ESPN unveils cross-platform correction policy and procedures

One of the most enjoyable correction-related experiences comes at the end of every episode of ESPN’s Pardon the Interruption, a sports talk and interview show featuring Michael Wilbon and Tony Kornheiser.
As ESPN explains, at the end of every show, “researcher Tony Reali corrects any statistical fouls Kornheiser and Wilbon made in the heat of battle.” It’s a fun feature that’s often fodder for a little competition between Kornheiser and Wilbon, and it also serves the purpose extremely well: viewers get the right information and no errors go uncorrected. It’s arguably the best example of a regular TV correction anywhere in North America. (Do it better? Let us know.)
Now ESPN has taken its commitment to corrections further. After collaboration between John Walsh, ESPN’s executive vice president and executive editor, and senior staffers in the television, radio, print and online operations, along with consultation with experts at the Poynter Institute, ESPN now has a cross-platform corrections policy and related procedures.
From now on, corrections for any mistaken reporting by ESPN can be found here. A link to the corrections page appears prominently on the drop-down “ESPN” tab on the site, and readers can easily submit a correction using a form on the page. Once submitted, it goes to a senior editorial staffer who forwards it to the appropriate person for evaluation. ESPN is also entering every request for correction into a database, and logging every correction in another. This will enable them to evaluate progress and identify trouble spots. And to help with error prevention, its online operation recently added more copy editors. The end result appears to be a clear policy with appropriate procedures and tracking capabilities.
“It’s the right thing for our fans and users, and the right thing for our journalism,” Patrick Stiegman, the vice president and executive editor of ESPN.com, told us last week in an interview. “Given the fact that our news organization has really matured and we have so many moving parts across different platforms, the time was right to formalize the [corrections] policy and procedures with the idea that espn.com would be the hub for corrections across the company.”
Stiegman says corrections will still continue to* be issued in the originating medium, emphasizing that, “we have always corrected our errors.”
According to the new policy, ESPN will issue online corrections for errors that “involve a significant factual mistake, or materially change the implication or connotation of the reporting.” Stiegman says that, for example, an error in someone’s batting average in an online article would be corrected within the story but not noted on the corrections page. This is what’s called a “scrub” and it’s important that ESPN not get carried away with its scrubbing, lest it become a way to avoid running corrections.
As the policy states, “This policy is not intended to cover inconsequential factual errors, such as minor statistical mistakes, inadvertent and immaterial misidentifications, minor inaccuracies in a developing story or font errors that don't impair the viewers' understanding of a story.”
Bob Steele, the Nelson Poynter Scholar for journalism values and a senior faculty member for ethics at the Poynter Institute, reviewed the policy with three Poynter staff members. “All of us felt the new policy was substantial and meaningful,” he said. “We described it as: ‘well-thought out, thorough, professional, responsible, comprehensive...’ We suggested the ESPN guidelines on corrections ‘will set a new standard for broadcast and online transparency’ and the guidelines could be a ‘model’ for other news organizations.”
As for the increased copy editing resources in the online operation, Stiegman says, “the amount of content being produced was out pacing the copy editing resources behind it.”  ESPN also hired an ombudsman 20 months ago.
Overall, it looks like ESPN has taken the issue of corrections seriously and applied the necessary effort and resources to come up with its cross-platform policy and procedures. While the issue of scrubbing is not ideal, it could be acceptable if only applied to the smallest of typos.
ESPN’s work should also serve as notice to all the “hard news” organizations out there who haven’t deemed it necessary to create an online corrections page. That means broadcasters like CNN, CBS, ABC, Fox etc., as well as newspapers like USA Today, which recently totally redesigned its website and neglected to add a corrections page. Perhaps this will spur them into action.
Good night, Canada!

*Correction March 5: The word "to" was originally missing from this sentence. The word "procedure" in the headline also contained an extra "e" in the middle.

UPDATED: Shafer's ode to the fact checking reader

Jack Shafer shines the spotlight on a devoted Slate reader he calls the magazine's de facto fact checking department. He writes about RM "Auros" Harman:

Auros reckons that he's e-mailed Slate at least 17 requests for corrections. I met him for the first time two days ago when he e-mailed my "Press Box" alias complaining about innumeracy in a Slate piece. When I brought Auros' name up with Slate staffers June Thomas and David Plotz, both immediately recognized him—Thomas because she's the Corrections editor and Plotz because Auros e-mails him frequently to comment on his Bible blogging. After quick consultation with them, I realized that Slate does have a fact-check department, it works for nothing, and its name is Auros.

Nice to see a devoted reader getting his due. And there are plenty of* editors and ombudsmen out there who could offer the same measure of thanks to their faithful checkers. Some readers, like some editors and fact checkers, just have a knack for spotting errors. Some focus on one particular area, while others, like Harman, seem to have an admirable grasp of a wide variety of subjects. The difference between an appreciated checking reader and a hated one often comes down to how the person chooses to bring errors to light, and how willing the editorial staff is to accept a correction. "Auros calls Slate editors collegial, but I'd give him most of the credit for the magazine's graciousness: His polymathic challenges are direct and respectful," writes Shafer. "What more could a publication want?" Exactly.
This is a nice example of a publication appreciating the diligence, dedication and, most importantly, contribution of a valuable reader. As the saying goes, we are all fact checkers now. Why not offer a bit of thanks to those wonderful fact checking/proofreading readers? 
Finally, ask and ye shall receive. Shafer ends the column with this: "Before we rang off, Auros jokingly offered to vet this article for errors prior to publication. I declined, as I can't wait for him to give it the vigorous and public beating it deserves so we can post the results in the Corrections column."
Voila:

Corrections, March 1, 2007: The original version of this article misstated when RM "Auros" Harman first encountered Slate. It was in 1999 during a Microsoft internship. He become a regular reader in 2000. (Click here to return to the corrected sentence.) The article originally misspelled Quds as Qud. (Click here to return to the corrected sentence.) The article misstated that Harman's nickname evolved from his sign-on name. (Click here to return to the corrected sentence.)

What can we say, the guy is good, even if this was some low hanging fruit. He deserves all the credit he can get.

UPDATE March 5: Amy Gahran over at Poynter also loves what Shafer's column represents, and she makes a good case for why this kind of recognition and community building is good for journalism:

...Online, community is golden. Embracing conversation is at least as important as producing a quality publication. This means listening respectfully to community members, letting them have a direct voice, and praising their contributions that enhance the overall quality and value of your online venue.

I know, none of us likes having our errors corrected or our conclusions disputed, especially in public. However, it helps to take the broader view. That's why Shafer's column is smart in several ways: It turns a potential flaw or humiliation into a source of strength and pride; it rewards community and loyalty; it encourages constructive public discourse; and -- business-side folks take note -- it enhances the Slate.com brand.

*Correction, March 2: And now we'd like to thank one of our best proofreading readers for seeing that the word "of" was originally missing from this sentence. Thanks Aldous!

About those models who died due to anorexia...

The Australian Broadcasting Corporation's Media Watch program has an interesting look at how its media latched onto an incorrect tale about starving models. It's a very interesting case study. Give it a view or read here.

Inside the Guardian's corrections format

We're not sure if this was distributed on purpose or in error, but the Guardian's corrections guide recently ended up in Nexis. It's an interesting look behind the scenes of how one newspaper has standardized its correction/clarification/apology format.

The following correction/apology text must be added to the top of articles for ALL corrections, even to what you may think are minor changes such as misspellings of names:
For Guardian corrections:
The following correction was printed in the Guardian's Corrections and clarifications column, day month date year (Friday October 20 2006)Insert correction text here, using 'In the article below ...' or '... as stated in the article below' instead of the headline and date of the article.
Guardian apologies:
The following apology was printed in the Guardian's Corrections and clarifications column, day month date year Insert apology here, with 'In the article below ...' or '... as stated in the article below' instead of the headline and date of the article.
For Observer corrections:
The following correction was printed in the Observer's For the record column, Sunday October 30 2005Insert correction text here, with 'In the article below ...' or '... as stated in the article below' instead of the headline and date of the article.
Observer apologies:
The following apology was printed in the Observer's For the record column, day month date year Apology here, with 'In the article below ...' or '... as stated in the article below' instead of the headline and date of the article.

IMPORTANT: Check the correction text you've added clearly explains the mistake made or reason for the apology.
How to do corrections
1) Print out the Corrections and clarifications column for the relevant day. Check the column against the print version and make sure all the corrections have been uploaded correctly. The column should end with the contact details for the readers' editor.
2) You need to write on this print-out what you've done to each story. If you've changed the article because it was a minor factual thing (see 5) then write 'changed in text, paragraph X'. If you've just added the correction text at the top of the article, write 'Corr at top'. This will make it much easier for whoever checks your corrections.
3) Find stories by going to Eek and searching by headline through the 'Search Archive' link on the left-hand sidebar.
4) When you find the right story, go into it using the link that says 'MBT edit article' If you cannot find it using this tool, it should be easy enough to find by going into Editor's Index and changing the date to the date the article was published. It's a lengthier process, but sometimes more dependable.
5) Edit the story as normal. If the correction is something very simple, like a misspelled word or wrong date, then change it, ALWAYS add the correction to the top of the article, set the story ready to live and refresh the page.
6) If the correction is more complicated, then you must add the correction to the top of the story, using the text above. Do not change the article, but if in doubt, ask someone. Corrections are important things with legal ramifications, and it's better to talk to someone else about what to do rather than land us in enormous amounts of trouble by doing the wrong thing.
7) IMPORTANT: All apologies must have the full apology added to the top of the article - even if it looks to you like it could be easily fixed in the text. This is a legal obligation, so if the word 'apologise' is anywhere in the correction, add the full text to the top of the article with the code outlined at the beginning of these instructions.
8) Whenever you live a story, copy and paste the url of the relaunched story into a Word document or email. When you've finished all the corrections, email [email protected] with the urls.
9) Put the annotated print-out in the folder on the Production desk with your name on it.
10) Get someone else in Production to check your corrections.

Eh voila. The readers' editor is now happy and your butt is well and truly covered.

Reuters makes another adjustment to its correction policy

Following its unveiling of a new corrections style on June 5, Reuters has made another adjustment to how it tags corrections on the wire. For background on the new style, we recommend readng our interview with Reuters' Paul Holmes. And here is the original Reuters notice about its new policy. Here's the latest:

From September 1, Reuters will adopt a new policy to help customers distinguish between errors that change the meaning or significance of a news item and minor mistakes that do not alter the meaning of the story.We will continue to rectify all errors openly and as rapidly as possible.
Substantive errors, such as those with inaccurate numbers, facts or quotes, will carry the headline tag CORRECTED.
Mistakes that do not change the meaning of a story, such as a typographical error or a proper name misspelled once among several references, will carry the headline tag REFILE.
All items, regardless of whether they are tagged CORRECTED or REFILE, will explain clearly what information has been corrected and where it appears in the story.
Only senior editors will be authorized to issue a CORRECTED story or a REFILE and we will monitor these to ensure consistency of approach.
Please note that subscribers to Reuters media services will continue to see the word CORRECTED between brackets following the slug of a story, though the headline tag may say either CORRECTED or REFILE. This is to ensure that items which may be republished by subscribers can be rectified swiftly at the subscriber end.
The change follows requests from some clients for an approach that will allow them to distinguish quickly and more easily between serious mistakes and minor errors. The new policy was presented to a global sample of customers earlier this year and the response was very positive.

A look at the role of the ombudsman

Boston Phoenix media columnist (and former Boston Globe ombudsman) Mark Jurkowitz wrote his latest column about ombudsman. Some excerpts from his interesting column:

...Part internal-affairs cop, part complaint department, American news ombudsmen are truly a unique breed. They work in what has often been considered one of journalism’s most thankless jobs: getting an earful from angry (and sometimes crazed) readers and getting the cold shoulder from angry (and sometimes crazed) colleagues whom they dared to criticize, usually gently, in their columns. In addition, ombudsmen frequently toil in the shadow of public suspicion since they are paid by the same news outlets they are charged with independently evaluating.
...But despite all that, we’re beginning to see — if not quite a golden age — at least a flowering of the ombudsmanship movement. In this post–Jayson Blair era, when news outlets are recognizing the greater need for transparency, the ranks of ombudsmen are expanding both nationally and globally. The big psychological breakthrough came when the Times, which had fiercely resisted the concept for years, finally agreed to hire Okrent for the job, in the wake of the Blair fiasco in 2003. Recently, such television outlets as CBS and PBS have either hired ombudsmen or begun employing more ombudsman-ly practices.
Equally important, ombudsmen are generating more attention within their own industry. Recently, the new Times public editor Byron Calame and the Washington Post ombudsman Deborah Howell have been the focal point of significant controversies or criticism — a phenomenon that is actually a healthy sign in a business that thrives on feuds and furor.
...According to material posted on the ONO Web site courtesy of former ombudsman Arthur Nauman, the US ombudsman movement really gathered steam with the growth of the “anti-press mood” in the 1960s. Noted journalist Ben Bagdikian wrote a 1967 piece in Esquire suggesting that ombudsmen might help papers better communicate with their readers, and Times staffer A.H. Raskin authored a piece in his paper calling for a “Department of Internal Criticism.
That same year, the Louisville Courier-Journal became the first paper in the country to have an ombudsman, and shortly thereafter the Washington Post employed the first ombudsman empowered to comment on how the publication was doing. In 1981, the Post’s Bill Green penned the most-famous ombudsman column in history with his 14,000 word postmortem on the notorious Janet Cooke Pulitzer Prize–winning story about an (invented) eight-year-old heroin addict. The Boston Globe’s Tom Winship was also one of the early editors to name an ombudsman, former editorial-page editor Charlie Whipple, who was selected for the job in the ’70s.
In a nearly decade-old ONO survey, most of the ombudsmen who responded said they write columns evaluating their paper’s performance, and a large majority also said they spoke to community groups. A much smaller number, about 25 percent, said they had actually hosted reader forums or sat in on the daily news meetings at their publications. (Gina Lubrano, who has spent 14 years doing the job at the San Diego Union-Tribune, estimates that in that period of time, the ranks of US ombudsmen in ONO have increased from about 33 to about 40, and the number of foreign ombudsmen has grown to about 23. But she doesn’t see a windfall. “There’s a lot of turmoil going on,” says Lubrano. “Like everything else, it’s a matter of money.”)...

  

WSJ introduces corrections search

The Wall Street Journal has added a search feature to its online corrections page. Now you can search though the paper's corrections from the previous three months for free. We know, you're positively giddy with delight. But as far as we know the WSJ is the only publication that offers a corrections search, and it's always nice when a news organization puts some effort into a corrections-specific service. Though we do have one gripe: It appears that the WSJ has removed the links to corrections from the last few days, which it used to offer. Why not have both? UPDATE: Looks like it was a temporary glitch. The WSJ now offers the search and the links. Excellent.
The search box: