News Column

Why Is It So Difficult To Separate Good Tweets From Bad?

Page 2 of 1

Rapper Chris Brown was declared dead on Feb. 15. A week earlier, so was actor Keanu Reeves.

Fidel Castro, Bill Cosby, Jon Bon Jovi, Tiger Woods and North Korea's new leader Kim Jong Un have all faced the same demise -- on Twitter.

None of it was true.

"To the people behind the foolishness, I'm not sure you see how upsetting this is," an exasperated Cosby said on his own Twitter account in August 2010 after he was knocked off for the fourth time on the social networking service.

It's easy to blame the Internet and social media -- Twitter, Facebook, Sina Weibo or even Pinterest -- for unreliable reports, but more important to figure out how to suss out the false ones. In the case of Whitney Houston, tweets about her Feb. 11 death proved true.

Experts say that in addition to the development of new technology to ferret out what's accurate on social media, a community-based, citizen-policed system might be the answer to bad information.

Consider the online encyclopedia site, Wikipedia, where contributors often flag entries that are suspicious or unreliable.

"How did Wikipedia, with an unstructured base, become a depository for reliable information?" said Ari Lightman, professor of digital media and marketing in the Heinz College at Carnegie Mellon University.

"It came about because people aggregate around specific sources, and that creates a mechanism that helps determine how best to identify the clear signal amid all the noise."

Lately, "the noise" might be anything from a blog that erroneously reports the death of Penn State's Joe Paterno, to the weekly tweets and retweets of yet another celebrity death. For mainstream news sources, it's a daily challenge to determine which reports are scoops and which are false.

In the case of Paterno's prematurely declared death on Jan. 21, picked up the news from Onward State, an online news organization serving Penn State students, faculty, staff, alumni and State College, and published it to its website. The report was quickly denied -- via Twitter -- by Paterno's two sons, but not before it had made the rounds via social media.

"Accountability is prized in social media," said Anthony DeRosa, social media editor for Reuters. "Once you've lost trust it is hard to earn it back. I rely less on CBS because of that one-time lack of attribution."

CBS later posted an apology on its website.

At the Pittsburgh Post-Gazette, the report came as the paper was preparing its Sunday edition, where a story on Paterno's grave illness was planned. The paper began to update its obituary on Paterno while checking on the report of his death.

Seeing the tweets and retweets "kind of put the pressure up," said PG Sunday editor Lillian Thomas, but the paper determined it would not run anything without verification by its beat reporter or from credible sources such as the Associated Press.

"AP had been really on top of the story, and they weren't reporting it," she said, adding the Post-Gazette's Penn State beat writer also could not confirm Paterno's death. "When we see [social media reports] we can't go with it, but you can't ignore it, either."

Paterno was pronounced dead at 9:25 a.m. the next day at Mount Nittany Medical Center of complications from lung cancer.

"The trick is to find the source of it," Jacob Sanders, Post-Gazette breaking news editor, said about the tweeted information. "Context is important, and as far as I'm concerned it [posting online or tweeting information] cannot move faster than your reporting."

For a brief time last summer, hackers used @FoxNewsPolitics on Twitter to declare President Obama shot dead. But the tweet was so outlandish and easily discredited, the "news" quickly died.

Not so Feb. 11, when @AjaDiorNavy tweeted "omgg, my aunt tiffany who work for whitney houston just found whitney houston dead in the tub. such ashame & sad. : -- (

According to Topsy Labs, which measures the strength of social media news, the information generated 2.5 million tweets and retweets in the first hour. In this case, the news was real.

Reuters' DeRosa said that when it comes to reporting what began in social media, life and death is just that. "I am more careful about this type of tweet than any other," he said. "You don't want to mis-report someone's death so I will hold off until I get information from a legitimate source and can then trace it back to another legit source, where both got it independently."

Other examples of Twitter sounding the first alarms of big news include the man who took a photo from a ferry of the USAirways "Miracle on the Hudson" flight, and IT guy Sohaib Athar accidentally live tweeting about the first stirrings of the raid on Osama bin Laden's compound in Pakistan.

In fact, the Twitterverse was alive with credible news relating to the raid and bin Laden's death well before the official White House announcement.

"People are starting to wake up and smell the coffee, to see that citizen journalists can bring something to the table," CMU's Lightman said.

He said a great example was the G-20 summit in London, where the BBC and the Guardian newspaper equipped staffers with social media tools. When reports of violence broke out, they were able to report back quickly to their editors to send more reporters.

"I think that's where things are heading, toward a lot more collaboration. This is not a passing fad," Lightman said.

But what to make of outright erroneous reports? Anyone can kill off a celebrity. There are a number of companies, including Lithium and Radian6, that monitor social media and will provide clients with tools to understand data across a variety of platforms.

A joint project from Rutgers University and Microsoft created "Seriously Rapid Source Review" (SRSR), which helps journalists determine what is trustworthy by determining where social media input is coming from, how it is gathered and filtering out what is likely unreliable information.

SRSR acknowledges that a lot of social media content is provided by non-journalists, but that eyewitness reports such as the one about Ms. Houston can be credible if supported by other factors.

Services such as SRSR can be helpful in this era of 24/7 news coverage where there's as much pressure of "getting it first" as "getting it right." Citizen journalism exploded with the rise of Twitter, where for some, retweeting is an obsession.

"This gets into behavior patterns: Why do people retweet stuff?" Lightman said. "It might be because we believe in spreading awareness, and some of the content might not be of a factual nature but just something they think their followers should know about."

Journalists and everyone else using social media will continue to have to struggle to sort the factual from the false. "But when you get down to it," Lightman said, "the utility and value of Twitter surpasses all that other crap."

Story Tools Facebook Linkedin Twitter RSS Feed Email Alerts & Newsletters