Apple introduced its latest update, iOS 17, promising several upgrades to many features which include built-in autocorrect functionality. At the Worldwide Develop Conference, it…
Seven top tips for verifying tweets
Twitter is like the coffee machine in the newsroom. Get a shot of caffeine, chat with colleagues, exchange useful information, but also get the latest gossip and rumours. The difference is that you probably won’t publish what you hear at the coffee machine. Twitter works differently: gossip, rumours, and other unverified information are published and sometimes cause a hoax. A tweet about the suspension of the famous CNN talk show host Piers Morgan over the phone hacking scandal is a recent example of a Twitter hoax. Another one from the beginning of 2011 saw a single tweet spreading the rumour that Nelson Mandela had died.
Think before you (re)tweet would generally be good advice. Journalists should not, however, only rely on good advice in a situation where Twitter has become an important news source. Of course, “check the information” is the golden rule for journalists. How do you do that, because social media and the internet work differently from traditional sources? This question tends to evoke much debate among journalists. Craig Silverman wrote a story about the different solutions in the Columbia Journalism Review. Most of the journalists working with Twitter have found ways to check or verify tweets, which can be summed up in a list of best practices.
1. Check the account
This is the first rule of journalism: Is the source credible? On Twitter, you should check the account of the tweep. Obvious items that can be easily checked are: When was the account created? When was it last updated? What is the ratio of their followers to the number they’re following? Last but not least: is there a picture? One should be suspicious when dealing with accounts that are not brand new, regularly updated, and with suspicious followers. In the Piers Morgan case, the tweet originated from a fake account (@danwooden) meant to parody Dan Wootton, a former show biz editor for News of the World.
2. Check the person
Looking at the tweets from the Dan Wootton parody account, it is easy to see that it is an isolated account with no interaction at all. With other accounts, you can do a more detailed search and try to find out how important the account holder is, and whether they play a central role (hub) in the Twitter network. Klout score can give a fairly reliable estimate of a person’s influence. Next: there is always the possibility of a web search: check a person’s name on the web and see what comes out. Identify, a Firefox extension, searches more specifically, and can give you a social profile of the person. In the case of the Mandela death hoax, a simple check on the originator of the tweet would have been enough to denounce its credibility.
3. Follow the lead
In February 2009 an airplane crashed on the runway of Schiphol airport in Amsterdam. Because the runway is close to the highway, the accident was witnessed by one of the commuters on the highway. He sent a tweet and within seconds the message was spread around the world. Journalists were trying to contact the mobile phone of the tweep who sent the message. They not only wanted a live report of the accident but also a check on the information. Others were watching and also sent follow-ups. This is another possible avenue for checking the facts of a story. In the case of the crash, the location was also important, some tweets contained maps of the place of the accident.
4. Corroborate the story
When a journalist hears that Piers Morgan is suspended, the first thing they would do is ask CNN or Piers Morgan himself. Of course Piers could be difficult to contact, but CNN is easier. In the Mandela case, simply contacting the official institutions — the ANC, or the Nelson Mandela Foundation — would have falsified the message. Of course official institutions are slow, but they have no interest in letting false information do their devastating work.
Checking the web could be tricky as well, the design could be eye candy and the latest information was updated. But how do you know that this is an official web page? martinlutherking.org, for instance, looks official at first glance but a search in the allwhois.com, the database of registers of domains, shows that it is owned by the Stormfront, an American white nationalist movement. Another check on the credibility of web sites is the importance or centrality in the network of web pages.
Several services are available for checking a site’s page rank. Alternatively, check delicio.us, a Firefox plugin for sharing bookmarks, to see if the site is regularly bookmarked. Technorati has a nice set of tools and figures to check a blog
If nothing helps, send out a tweet to your followers asking for help. Ask if there is someone who could substantiate the claim, true or not. An interesting example is the work of Andy Carvin. When several news organizations posted stories claiming they had evidence of Israeli munitions being used in Libya, “my Twitter followers and I investigated the claims and ultimately debunked them”, says Carvin. Here’s the story.
6. Damage control
At the end of the day there will also be a decision to make: publish or not? There are two things to consider here. First, how urgent is it? And second, what is the damage in case of falsehood? This is difficult to measure, it varies from to case to case; and one has to try to balance the good over the bad. If you think it is not urgent, your competitor could do otherwise. Falsehood damages your credibility, but if true, you win the race with a scoop.
Verifying images is a special element in the overall process of corroborating a story. In the world of Photoshop and the Gimp, two software programs for editing images, I’ve stopped believing what I see in the picture. I will never forget seeing a designer take a picture of the Reichstag — the German Parliament — and put it on fire with a few keystrokes including people running away. Or how a photo editor chose an old picture (by mistake) to illustrate a recent flood.
Every picture taken with a digital camera has information about how and where it was taken, called Exif data. Geo information of the picture could, for example, be combined with the local weather report. A reverse lookup in Tinyeye could help to find others who used the picture.
To establish truth from falsehood is difficult on Twitter. Asking for a lie detector or filter, which controls the information sent, would be a limitation of the freedom of speech. I would argue that the damage by control is greater than damage by false tweets. Because false tweets can be detected. Skilled journalists should be the first in line to successfully combat falsehood.