Journalists and news organizations need to do a better job of avoiding involvement in the spread of lies and unconfirmed rumors.
Accuracy and credibility are the heart of good journalism, and Craig Silverman‘s study Lies, Damned Lies and Viral Content documents widespread disregard for both in the spreading of digital reports by pro.
I won’t attempt to summarize the report here, though I will use some favorite quotes from it at the end of this post. I hope you will read the full report (it’s 164 pages) and consider what it says about you and your news organization.
What I want to focus on here are some suggestions for news organizations and individual journalists, some of which repeat Craig’s own suggestions and some of which are my suggestions, inspired by his report:
Confirming and debunking rumors
To start, I don’t think chasing rumors is necessarily the highest form of journalism, though admittedly, great journalistic investigation starts with a tip that’s indistinguishable from a rumor. But in general, I would encourage a journalistic approach that seeks to find and publish new information rather than chasing rumors.
That said, some rumors are newsworthy and can’t be ignored, and some news organizations have at least part of their operation devoted to aggregation and curation of material published first elsewhere.
Newsrooms should discuss and set standards for whether and how they debunk rumors. Early in my career, many an editor dismissed the suggestion that we should ever publicly address rumors we couldn’t confirm. I remember more than one editor snarling something to the effect that “we don’t publish rumors, we publish facts.”
Of course, we often researched rumors and published them if we confirmed them as facts. That’s still a good approach. But we generally didn’t debunk the rumors we found out were false. Perhaps that was the right approach back when rumors circulated by word of mouth. But, with rumors being published today in social media, and sometimes amplified by other media sources, I think we should attempt to shoot down the false rumors we bother to check out.
I don’t offer my suggestions as a finished set of standards for you to follow, but as a discussion-starter. The discussion might lead to some better standards that I would gladly adopt in place of mine. But here are my suggestions for an approach newsrooms and individual journalists should take to verifying or disproving rumors:
- Don’t publish rumors until you have checked them out (with one exception I’ll note in detail in the crowdsourcing section below).
- Don’t check out a rumor unless you would consider it newsworthy if it turns out to be true. In other words, I don’t think news organizations need to chase every rumor that makes it into social media or the competition. But if it’s newsworthy, you should check it out before publishing.
- If a rumor would be newsworthy if it’s true, and you check it out, it’s probably worth debunking, especially if it’s circulated widely in social media or has been published in other professional media.
- One exception to the point above: If a rumor would be harmful to someone’s reputation and hasn’t circulated widely or been published in other professional media, you should consider whether debunking the rumor would actually give it some level of credibility, and more circulation, and be more harmful to the person’s reputation. As Craig says in the report: “Repeating the rumor bomb detonates it—regardless of context.” You might ask the person how much they’ve been damaged by the rumor and whether they would prefer for you to disprove it or let it die quietly without further attention.
- Especially for a local news organization, I don’t think you need to feel responsible for correcting every error that makes it into social media or even all the errors that competing media make. But an occasional story such as the Washington Post’s What was fake on the Internet this week is both fun content and helps set the record straight (and shows a commitment to accuracy).
- You should be clear and specific about who erred (when you know) and about how you verified or debunked a rumor. Do it in a factual, non-accusatory way. You’ll make your share of errors, too, and did-not-did-too pissing matches don’t serve anyone well.
Crowdsourcing rumor verification (and debunking)
Your community can help you get to the bottom of a story. Andy Carvin of Reported.ly (and formerly of NPR) is exceptionally good at this. I’ve used his example of quickly shooting down a rumor that Israelis were arming Libyan rebels as an excellent example of crowdsourcing, and his debunking of the Gay Girl in Damascus blogger’s fake disappearance was another. But he used a technique I would use sparingly, and with great care: He says what the rumor is before he’s learned whether it’s true.
— Andy Carvin (@acarvin) March 12, 2011
That link only leads now to the Al Manara Press Facebook page and the image is not there any more. But I’ve posted it here and it plausibly looked like Israel’s Star of David on the side of the shell, along with a crescent, appearing to symbolize some sort of alliance between the Israelis and a Muslim group. In posting the rumor (or speculation), Andy raised a question about it and asked his Twitter followers for help in determining the truth of the rumor. In doing so, he risked that others might spread the rumor without spreading his questions about it. But he also quickly got to the bottom of the rumor: That star has marked illumination shells since World War I, before the creation of modern-day Israel. And the crescent is a parachute. The illumination shell drops slowly into an enemy area at night, glowing brightly to help artillery units determine where to aim the live shells. Many nations manufacture illumination shells using those symbols.
I have used crowdsourcing some myself and worked with newsrooms that have crowdsourced successfully. I think it’s a technique that helps in verification of rumors, but should be used carefully. Some tips for crowdsourcing as you seek to verify or disprove rumors:
- Don’t crowdsource verification of a rumor that could harm someone’s reputation. I did a fair amount of reporting on sexual assault in my career, and you simply can’t crowdsource a story like that by tweeting that you’re looking for people who have been abused by a particular coach, clergy member or politician whom you’re investigating. On the other hand, verifying some information is a great way to do some crowdsourcing. Once I nailed down a story about abuse by a particular priest, publication of that story would often bring forward tips or direct contact with more victims of that priest or other priests. And that was without any crowdsourcing requests. If you have a verified story, consider whether a crowdsourcing request for others who have experienced this offense to come forward so you can advance the story.
- Consider whether you can word a crowdsourcing request generally, rather than repeating a rumor. For instance, if you’ve heard a police-scanner report of shots fired at a local high school (and police-scanner reports should be treated like rumors; police and fire departments chase more false alarms than they do big news stories), you don’t have to tweet: “Can anyone confirm that shots were fired at Yourtown High School?” That could unnecessarily panic parents if it’s a false alarm. Of course, you’re going to dispatch reporters and photographers immediately to the school, but you also can tweet, while they’re en route, “Does anyone know what’s happening at Yourtown High School?” That might get you quick confirmation that something serious is happening. Or it could get you a quick response that someone was overreacting to a car that backfired.
- Address people who are discussing a rumor in social media, asking them how they know what they’re saying and letting them know that you’re trying to nail down the fact. In my case study of how the Austin American Statesman covered a 2010 breaking story, I noted how carefully Robert Quigley, then the social media editor, responded to people tweeting about rumors:
@chicklitter We’re trying to get to the bottom of it. No official explanation out there yet. We are working on it.
— Austin Statesman (@statesman) February 18, 2010
- Collaborate on how to crowdsource. Confer with your editor (or your staff) in whether and how to crowdsource confirmation of a rumor. Even if everyone seems to agree, encourage someone to make the opposing arguments, so you can consider all options.
- Don’t just turn your verification over to the crowd. You start with them, but they might be wrong, so be diligent in asking questions and assessing their answers. Ask how someone knows what they’re telling you. If they’re giving you secondhand information, ask if they can connect you with the firsthand source. If they witnessed something, ask if they took photos or videos. Ask for details and documentation. Ask who else witnessed something or who else knows something about this.
- Thank (and credit) people who provide helpful, accurate information.
When sharing links or retweeting
You don’t have time to be active in social media and thoroughly check everything before you share it, but you want to be a credible source in social media.
- If it sounds too good to be true, that should be a tip-off that it might be. The more incredible a story sounds, the more you should consider what “incredible” really means: not credible. Look those stories over carefully before you share them.
- Consider the nature of an article before you share it. If it’s a humor piece or a commentary on a situation that’s been in the news, and the underlying facts have been widely reported, you probably don’t need to worry much about posting it. But if the heart of the piece is a fact or a piece of news you haven’t heard before, you should at least consider the source and examine the verification the source presents.
- Consider how serious a mistake it would be if you retweeted or shared something that turned out to be bogus. For instance, last week’s llama-chase story would have been pretty harmless if it turned out to be bogus, so I’m not going to fault someone who shared that story without checking it out. If, on the other hand, a story might damage someone’s reputation or cause a panic, that requires extensive checking out. Many stories fall somewhere in between.
- Evaluate the source. Lots of those llama-chase links were live video feeds of bona fide media organizations covering the story. It was pretty easy to see quickly that this was really happening.
- Check to see if someone has already debunked or verified the rumor. Particularly is a rumor would be of national interest, Snopes.com, Hoax-Slayer and Doubtful News (sources Craig cited as diligent in disproving rumors) might already have addressed a rumor. Or if you’re dealing with a political statement, PolitiFact, Factcheck.org or Fact Checker might already have assessed it. If the person, organization or issue has a Wikipedia page, see whether the rumor is addressed there. Wikipedia entries can be wrong, and Wikipedia cautions against using it as a primary source, but it requires citations, and you can check the links in the citations to find helpful primary sources.
- If a reputable media source has done original reporting, I’m fine with sharing its report. I didn’t verify that Edward Snowden had revealed National Security Agency snooping to the Guardian. I couldn’t duplicate the Guardian’s reporting, and I trusted it. But if a story from a trusted media source is later corrected, you should share the correction, too. For instance, I don’t fault anyone who last week shared Jezebel’s report about Scott Walker proposing to delete language from Wisconsin state law requiring reporting of sexual assaults on campuses. Jezebel does solid reporting on gender issues and should be a reliable source on a story such as this. I didn’t share the link, but I might have “liked” someone else’s post about it. But if you shared the story, you should also note that it was updated to note that the University of Wisconsin asked for deletion of the state requirement, which was redundant with its federal reporting requirement. On the other hand, asking skeptical questions when you read media reports is a good idea, and I’d applaud anyone who read the original report and wondered whether the university would still have to meet the federal requirement.
- Many popular media organizations, whether they do original reporting or not, also follow up on stories reported elsewhere, either adding their own reporting or often just aggregating the original report. As Craig’s report documents, many professional media organizations have passed along bogus reports in this way. Try to get to the original source of a story you share, so you can make a better assessment of it (and link to the source that deserves the credit, if it’s true).
- Watch out for news sources that deliberately publish hoax stories. Of course, some people fall for the satire in The Onion or the New Yorker’s Borowitz Report, and I guess you can’t do anything about that, except maybe share such a link with a humorous comment and hope people can figure it out. The Daily Currant isn’t as funny, but describes itself as satirical (and I’ve laughed at some of its pieces, though it’s not as funny as Borowitz or The Onion). Craig counts the Currant in a group of fake news sites that just seek to fool people (and generate traffic) by churning out fake news. Others he lists are National Report, Civic Tribune, World News Daily Report and WIT Science. (I didn’t find satire disclaimers on any of those sites. In fact, they generally had official-looking about-us pages designed to make them look like legitimate sources of news and information.) You should not share information from these organizations. And generally, if you’re not familiar with a news site, you should be doubly careful about sharing anything from it.
- Check a person’s social profile. If a friend or social contact you know and trust shares something (especially if it’s a personal statement, not just passing along a link you can examine to assess that source’s credibility), you have a basis for trusting what that person says. But if you see something reported by a source you don’t know, take a look at that person’s profile and assess what you can. For instance, I follow @badbanana because he’s funny, but a quick look at his tweets would tell you not to take anything he tweets seriously. He’s always joking, but you might not see that from an individual tweet. If a profile is fairly new, that might be a tip-off that it’s a fake. If the person offers a link to a website, blog or LinkedIn profile, that might give you more basis for assessing credibility.
- Share with some skepticism. If you can’t verify a report (or don’t have time, but think it’s important to share), post it with a disclaimer, such as “if this is true …” or ask your social circle: “I haven’t heard this before. Anyone know if it’s true?”
Correct quickly and thoroughly
When you learn that you have circulated a bogus report, I think you have a responsibility to correct it and to do what you can to stop its spread. Here’s what I recommend:
- Delete the original post if it’s a tweet. I generally don’t recommend deleting tweets, but you can’t edit a tweet, and an inaccurate tweet stands out there on its own, subject to being read and retweeted, so I would delete a tweet that’s egregiously in error.
- If it’s a Facebook post, you can edit to correct it, so I would decide which approach is better in the circumstance: editing or deleting.
- Acknowledge your error. If you delete a tweet or Facebook post, I think you should tweet or post that you deleted the earlier post and that it was inaccurate and why. You might post a link to the accurate information, particularly if it’s something like a Snopes report that debunks what you originally posted.
- I think you should go one step further, even if it’s a big step: Call your correction to the attention of those who have retweeted or shared your original post. Both Twitter and Facebook show who shared or retweeted your posts (be sure to check before you delete the post). You can post your correction to a person’s Facebook wall (with thanks to them for sharing and an apology for your error) or you can tweet at a person. Frankly, I think you should call out your correction to people who’ve liked it or favorited the tweet, too, but you should especially correct those who have passed your error along. When I’ve suggested this before, some people have said that’s too much work in cases where something has been shared dozens or even hundreds or thousands of times. I disagree. Credibility is essential, and correcting one error hundreds of times will make you more cautious next time you want to share something without checking it out. Cutting and pasting will speed the process. I also wish Facebook and Twitter would develop a correction function where you could, with one correction, notify the people who’ve interacted with an inaccurate post.
What are your tips?
I’d be happy to share other tips for verifying rumors, sharing media reports responsibly on social media. What are some steps you take to assess credibility when sharing? What are some lessons you’ve learned the hard way?
Quotes from ‘Lies, Damn Lies and Viral Content’
As I said above, I’m not going to summarize Craig’s report. I hope you read it in full. It’s worth the time for any journalist, journalism student or journalism professor (or a serious news consumer). But here are some quotes from the report that stood out to me:
Lies spread much farther than the truth, and news organizations play a powerful role in making this happen. News websites dedicate far more time and resources to propagating questionable and often false claims than they do working to verify and/or debunk viral content and online rumors. Rather than acting as a source of accurate information, online media frequently promote misinformation in an attempt to drive traffic and social engagement. …
Many news sites apply little or no basic verification to the claims they pass on. Instead, they rely on linking-out to other media reports, which themselves often only cite other media reports as well. The story’s point of origin, once traced back through the chain of links, is often something posted on social media or a thinly sourced claim from a person or entity. …
News organizations are inconsistent at best at following up on the rumors and claims they offer initial coverage. This is likely connected to the fact that they pass them on without adding reporting or value. With such little effort put into the initial rewrite of a rumor, there is little thought or incentive to follow up. The potential for traffic is also greatest when a claim or rumor is new. So journalists jump fast, and frequently, to capture traffic. Then they move on. …
The data collected using the Emergent database revealed that many news organizations pair an article about a rumor or unverified claim with a headline that declares it to be true. This is a fundamentally dishonest practice. …
Once a certain critical mass is met, repetition has a powerful effect on belief. The rumor becomes true for readers simply by virtue of its ubiquity.
Meanwhile, news organizations that maintain higher standards for the content they aggregate and publish remain silent and restrained. They don’t jump on viral content and emerging news—but, generally, nor do they make a concerted effort to debunk or correct falsehoods or questionable claims. This leads to perhaps my most important conclusion and recommendation: News organizations should move to occupy the middle ground between mindless propagation and wordless restraint.
If it’s newsworthy, such as the report of flooding at the NYSE, at least a few journalists are likely to jump on it prior to practicing verification. If one (or more) credible outlet moves the information, others are quick to pile on, setting off a classic information cascade. When a rumor or claim starts generating traffic and gets picked up by other media outlets, then it’s even more likely journalists will decide to write something. The danger, aside from journalists becoming cogs in the misinformation wheel, is that it’s incredibly difficult to make corrections go just as viral.
Present someone with information that contradicts what they know and believe, and they will most likely double down on existing beliefs. It’s called the backfire effect and it’s one of several human cognitive factors that make debunking misinformation difficult. The truth is that facts alone are not enough to combat misinformation. …
News organizations must recognize the value of being smart filters in a world of abundant, dubious, and questionable information. …
Journalism’s fundamental value proposition is that it provides information to help people live their lives and understand the world around them. This is impossible to do when we are actively promoting and propagating false information. It’s impossible when we do a poor job of communicating the uncertainty and complexity of claims that circulate on networks and elsewhere. …
The very act of pointing to a rumor or claim adds a level of credibility. …
One of the easiest ways to avoid becoming part of a chain of dubious propagation is to take a few minutes and search/read closely to see where the claim or rumor originated. Don’t point to a rumor unless you have located the original source and evidence and evaluated it. …
I also advocate developing standard language that can be used in content. For example: ‘This claim has not been independently verified by [news org] and therefore should be treated with skepticism. We published it because [insert reason].’ If you don’t feel comfortable explaining why it needs to be reported at any given moment, that’s a sign you shouldn’t publish. …
Provide a counter narrative: This is one of the most important debunking strategies. The goal is to replace the existing narrative in a person’s mind with new facts. It’s more effective than a piecemeal approach to refuting rumors. Humans are attracted to stories, not a recitation of information. … Journalists should use all the storytelling tools available to make a debunking compelling and persuasive. Don’t be a spoilsport denier—tell a great story.
American Journalism Review’s Rumors, Lies and the Internet: 7 Questions for Craig Silverman
Verification Handbook (edited by Craig Silverman, with a contribution from me)
My earlier posts on verification: