How to: verify content from social media

The mass of information now available and being shared online offers a fantastic arena for journalists to engage with online communities and pick up on breaking news at the same time. This means journalists are also having to sharpen their verification and fact-checking skills in a digital environment.

This how-to features advice from a panel of experts on the key considerations, questions and tools journalists should have in mind when carrying out verification of content that surfaces via social media, be it a news tip, an image, a piece of audio or video.

The process covers three main stages: monitoring of social networks and the online community before news breaks, checking the content when it comes into play and subsequently reporting that content once verified. The comprehensive advice outlined in this how-to guide offers practical steps, specific questions and cross-checks journalists can make at each stage, as well as online tools to support them.

Our panel of experts on this topic includes:

• Claire Wardle, director of development and integration, Storyful
• Fergus Bell, senior producer and UGC lead, Associated Press
• Paul Bradshaw, course leader of online journalism MA at Birmingham City University, visiting professor at City University London and publisher of Online Journalism Blog and HelpMeInvestigate
• Craig Silverman, founder, Regret the Error, now published on Poynter

1) MONITORING

• Be across platforms

Storyful's Claire Wardle outlined the different social media platforms on which content can be found, from Twitter and Facebook, where she recommends 'liking pages' of certain organisations to gain access to a stream of related information, to YouTube, SoundCloud and AudioBoo for multimedia content.

"People think, 'how do you do this?'. There is a level of automation but a lot of it is knowing good people and knowing the people who are uploading the best content.

"That takes you half way through the process of verification, because you already know who this person is, you trust them, you've spoken to them. So when they upload the latest video you obviously have to do the same checks but you're halfway there."

She also said Bambuser, which enables users to stream live video has "really come to the forefront, particularly around Syria". She added that having built a network of users on the platform you can get "incredible content from people who are just using it to stream from where they are".

"The biggest thing is how do you stop your eyes from glazing over? It comes back to old fashioned journalism and knowing who the right people are beforehand on a number of different stories. The next phase is to know those people who at the moment we don't know."

• Spot and understand trends

Fergus Bell of AP added that a key part of the monitoring process is to identify trends which can be followed up.

"I use Twitter and Hootsuite, which helps me to interpret Twitter in a way that's most useful to me. I can be monitoring what's going on on social media on a specific story just by having a page up on my screen and I can take that away and use that on a mobile device. People might use Tweetdeck because they find that more user friendly, it is much of a muchness, it just depends on what your preference is.

"I like to use Trendsmap, which helps me plug into what people are talking about in the real world, so when I'm looking at something I get some context about what I am seeing, e.g. why is this word trending in Paris and also in Newcastle? It helps me understand where I should be looking and have I identified a story that's worth us chasing.

He also emphasises the importance of effective searching on these platforms, such as Twitter advanced search, YouTube filtering, Hootsuite search or Facebook search. The platform he uses depends on the story in hand.

"I don't use each tool for every situation, the tool I use depends on the news event, what kind of people are sharing information. If it's a news event where there are likely to be teenagers there I'll look in one place, if it's likely that there were slightly older people or people with slightly older phone technology I will look in another place."

• Build a network of contacts before the story breaks

One way to reduce the stress when a news event breaks, is to have already built up a trusted network of sources around the story or location. As Craig Silverman says, you have to "lay the groundwork".

If there are particular geographic areas which you as a reporter are really focused on it makes a lot of sense to try and identify credible sources in these areas before the news breaksCraig Silverman, Regret the Error

"What I mean by that is if there are particular geographic areas which you as a reporter are really focused on it makes a lot of sense to try to identify credible sources in these areas before the news breaks. This goes to a larger point in terms of decision making in these moments, when the stress is there and the time pressures are there, we don't function as well as we do when those pressures aren't in place.

"When stress is at a high level and time pressures are really urgent we usually go into fight or flight mode, and that's not when we make our best decisions. When it comes to finding sources and getting that material and verifying it well, what really matters to a huge extent is what you do before the moment arises. Find those sources, reach out to them, build Twitter lists, build virtual Rolodexes, so you can go in and contact these people and see if they can help."

2) CHECKING

• Apply the Too Good To Be True test

Paul Bradshaw says the 'too good to be true' test is "the main one" for journalists to bear in mind when they are starting to verify material online.

If you're about to leap out to start writing this up, just hold yourself back for a minute and ask yourself, is this as wonderfully news worthy as you think it is, or is there something else at work here?Paul Bradshaw, journalism academic

"It's just that gut feeling. If you're about to leap out to start writing this up, just hold yourself back for a minute and ask yourself, is this as wonderfully newsworthy as you think it is, or is there something else at work here?

"Beyond that it's treating it as you'd treat any source – where is this information coming from? Does it sound like that person? does it ring true? And that's before you get into any of the technical kinds of checks, it's just that gut feeling."

Silverman added that journalists "need to be wary of the one great video, or one great photo that seems to encapsulate an event".

"News organisations have to be extra cautious because that's when the hoaxers and fraudsters tend to come out. Be aware and beware of the amazing shot in a breaking news situation".

• Verify the source (part 1) – speak to them and cross reference answers with social data

According to Wardle (and the other experts) "the gold standard we would always say is to try to talk to someone if you haven't spoken to them before".

Before you do so, there are checks journalists can make to arm themselves with questions to cross-check with the source direct, such as by looking at location history which may be available on Twitter, or data about the photo and equipment used to take it on Flickr.

"When you speak to them you can check it reflects the information you got from the picture. On any type of information you want to do as much as you possibly can, the simplest things are just cross referencing. Does it match what they've been tweeting about before around location or subject matter, on Twitter who are they following, who is following them, have they been listed by other people as somebody who knows something?"

Location can also be cross-checked using Google Street view, she added. "Then when you get them on the phone, it's about saying, 'what can you see out the window right now?' It's about using tools up until the point, then can we go to the next level. You can't phone everybody, that's why as the networks grow you get a list of people who you've spoken to or you trust or you know they are there."

• Verify the source (part 2) – look at social media history across platforms

Look for the history of that person, have they been on those platforms for some time, if so they are more likely to be that person. If they're trying to hoax you the account is likely to be more recentPaul Bradshaw, journalism academic

"The obvious things to do is to look at that name on other social media platforms," Bradshaw says. "Look for the history of that person, have they been on those platforms for some time? If so they are more likely to be that person. If they're trying to hoax you the account is likely to be more recent.

"Look to see whether the information on that account correlates – do they talk in the same way, do they have the same details? Remember there may be more than one person with the same name, even working in the same field."

He also recommended using email addresses and phone numbers to search for users, "which are specific to that person".

"On Facebook if you enter an email address in the search box that will take you direct to that person rather than using their name. It's a much better way of tracking a person on Facebook. There's also a Firefox extension called Identify which will very quickly bring up related social network accounts when you're on a particular page, and Polaris does the same on Chrome for companies."

Wardle also advises making use of the network. "When you're dealing with a network you need to evaluate the network as part of your verification process", she added.

"It's not just about that photo or video, it's about the person who provided it to you, who are they connected to and how have they been using the network to deliver [information] to you and prior to them sending this to you."

Bell added that "social history" is key to this part of the process.

Everyone that we interact with in the process of news gathering have a social history whether they realise it or notFergus Bell, Associated Press

"I have a social history which shows I am a journalist, I tweet stories the AP report, it shows I am interacting with people who are sharing information, it shows I am exchanging information with people who are interested in journalism and social media and verification. That's my social history and everyone that we interact with in the process of news gathering has a social history whether they realise it or not.

"So I am able to look at someone and say 'OK, this is where they have been in the past, these are the interactions they have had, this is who they are talking to'. If they are showing some information from this specific location, I can see they have been there for a year, or lived there for a year, or been talking to other people in the same place about that story."

But he added that at the AP "we'd never just go on information shared in a tweet".

"We'd use it to direct our coverage but we'd always seeks the AP verification before we were able to run with it."

• Use Whois tools to verify websites

It is not just users you can work to verify, you can also look into websites and online companies. Bradshaw recommends using a "number of tools called Whois tools", which he said "will tell you who registered a web address, what physical address that person/company has".

"These all paint a picture, they come together to give you an indication as to how likely it is that website is owned by the company or individual that it's supposed to be. If the registration information is private then it's likely to be an individual and not a company.

"If a server hosting a website is shared with a number of other websites then it's less likely to be a big corporation", he added, which "tend to have websites hosted on their own servers".

• Use online tools to examine images

Wardle recommends sites such as TinEye, which journalists can use to perform a reverse image search to identify different versions of an image that may exist, a process which can also now do through Google Images.

You can also use the WolframAlpha search engine, she said, to ask certain questions such as the weather in a certain place at a certain time, to verify images or video.

• Check for photoshopping or repetition in images

Another thing for journalists to look out for when verifying images in particular is "things like airbrushing and cloning", Bradshaw adds.

"Spend a bit of time looking at the image and asking yourself, ''is the person actually in that place', 'is the light coming in at the same direction on all of the people/objects in that photo', does the weather match up to what it was supposed to be like?', 'does the equipment match up, e.g. that army's guns or that police force's badge?'.

• Question edited footage

Bell says he is "personally sceptical when I see video being uploaded that's edited".

"It might be genuine but what am I not seeing? I would definitely need to have that conversation with the creator of it to know why there's an edit point in there, why they've changed the angle of the camera, why have they gone from portrait to landscape, was it the same camera used for both sequences?"

• Harness online discussion boards and experts

And of course the internet provides journalists with a host of resources when it comes to others who may have experienced the same issues, Bradshaw says.

Journalists need to not just access technical tools but also the human resources available online and speak to people about the expertise they have that they can bring to bear on pieces of informationPaul Bradshaw, journalism academic

"Any problem that comes about online, it's very likely that someone else has encountered that problem and tried to solve it, so you do find many tools and individuals online who will be much more accessible than they would ever have been before the internet, who can help you check the background of images or a piece of information.

"Sites like Snopes debunk hoaxes or urban myths. If you're coming across a rumour, it's quite often a different version of an old one."

He added that discussion boards can be useful places to find others talking about the same issue, while specialists in particular fields or geographic areas can be found online.

"Journalists need to not just access technical tools but also the human resources available online and speak to people about the expertise they have that they can bring to bear on pieces of information."

• How urgent is it – could more steps be taken to verify before you publish?

Silverman encourages journalists to take a step back when they find themselves getting caught up in the 'need for speed' when news is breaking, and ask themselves 'is this photo, or this video or piece of information, really so essential or urgent that we can't wait and investigate other avenues with it?'

Social media means information is out there so quickly and it gets passed around and passed around. There is a pressure to try and confirm that or to maybe miss out a step when you're confirming it, but we don't believe that's the way it should be doneFergus Bell, Associated Press

"I would caution journalists in that moment to sit back and ask whether it's essential you go without taking an extra step here or there. I think a lot of the time it's probably not. There are definitely times I'm sure when news organisations have to make a call, e.g. everything we've looked at tells us this is accurate even though we have a barrier to getting hold of a person. Take a distance and evaluate – do we need to put this out now, or can we take hour or two and figure out what more we can provide to confirm or deny this? You also need to evaluate why you can't talk to that person. Have you exhausted all avenues?"

Bell agrees that accuracy is vital. "Social media means information is out there so quickly and it gets passed around and passed around. There is a pressure to try to confirm that or to maybe miss out a step when you're confirming it, but we don't believe that's the way it should be done. We'd rather be a couple of seconds slower but know that we're right, than to rush something out because we're caught up in the social media wave."

• Crowdsourced verification – 'be judicious in the ways you put unconfirmed information out there'

One form of verification which has been used effectively by a number of journalists is crowdsourcing, asking the online audience to come forward with any information they have which can help to verify a claim or content.

Wardle said Storyful crowdsources privately but has not "done anything public yet". "We are thinking about how can we do more of that, in a way that makes sense and is helpful, and is valuable to the audience as well as us," she said.

Silverman added that it is a "valid method" and offers "lots of wonderful opportunities" but "it is all about what you're trying to confirm and what training you've had to do it in an ethical and responsible way". It can be "fraught with some serious potential pitfalls and danger", he warned.

"You don't just start throwing things out there and saying 'is this true?', that's not the responsible way to do it".

He spoke about having discussed the issue with Andy Carvin of National Public Radio, who is known for using crowdsourcing in an effective way.

"What you're seeing on Twitter is only one part of the process that he put together," Silverman explains. "A really significant part is that he is on Skype with lots of sources checking things out in other areas, using the phone and Skype while also doing it on Twitter.

The old school device of the phone is still incredibly essential when trying to do crowdsourced verificationCraig Silverman, Regret the Error

"The old-school device of the phone is still incredibly essential when trying to do crowdsourced verification."

And "things you don't share are just as important as the things you do," he added.

"When you decide to retweet somebody and ask whether what they're reporting is confirmed – you're making a big decision there."

He outlined "two concerns to keep in mind".

"Don't just repeat what somebody said, find a way to rebroadcast what they've said but place it in context of adding some kind of hedging to it – you want to add text to say 'this is unconfirmed, does anybody have any news?' A really core question in any verification is how did you get this information?"

"Be aware that when you're retweeting something, you need to make sure it's not something so sensational, and so unconfirmed it could actually cause problems. Be aware even if you do retweet and add 'anyone else reporting this?', realise that sometimes when people rewteet you, even if you've been responsible in adding that context and questioning, they may strip that away just because of limitations of characters."

This is "another argument to be judicious in the ways you put unconfirmed information out there", he added.


3) REPORTING

• Consider any permissions and crediting which may be necessary

Bell said that when it comes to reporting material obtained from sources on social media asking permission is essential.

"Because part of our verification process means we have to talk to the individuals, we always ask for permission to be able to run that content. Apart from in extreme or very exceptional circumstances we only run something if we have permission from the person who has copyright for that content or that video, the person that has filmed that video, the person who has taken that photograph.

"When it comes to crediting we will ask them how they want to be credited, some want to be credited with a YouTube handle or Twitter handle. Some people want to be credited with their real name. We don't guess on that, we ask them."

• Clearly communicate the level of verification this story has been given

Wardle said that an important part of the reporting process is being transparent about the level of verification which has been achieved.

"When we put out content you get the video but underneath we put about the levels of verification we've gone to and then its up to the news organisation to take that information and warnings and then decide how it fits with their own standards."

• Made a mistake or new information come to light? Issue a clear and networked correction

Twitter is forcing more news organisations to be honest and offer corrections when things go wrongClaire Wardle, Storyful

And the standard of being transparent at the time of reporting should continue when mistakes are made or material turns out to not be as it first seemed, Wardle added, with the Twitter environment "forcing more news organisations to be honest and offer corrections when things go wrong".

Silverman, who runs the Regret the Error blog in the US, has some detailed advice for news outlets on how to handle such corrections in "a networked environment".

"It's not enough to just put the correction out once in one place. If you put the error out on Facebook, on Google Plus and on Twitter then the correction needs to flow into all those places.

"But even more than that your job now is to promote that correction. Just putting out one tweet of a correction, some people might retweet that but Twitter is a stream and not everyone who saw the incorrect information is going to see that one tweet.

That's the fundamental thing news organisations need to understand about corrections in the networked world, is it's your job to promote it and to reach out and get it to spread out just as far as the incorrect informationCraig Silverman, Regret the Error

"So I strongly encourage news organisations to push the correction out to all platforms where the error was and realise that once is not enough. They need to actively work to promote that correction.

"If you can identify people who saw the incorrect information either by retweets, or by likes, or by re-shares, then it's your job to reach out to these people individually and make sure they see the correct information.

"The reason to do that is you activate the correction network," he said. News outlets need to work to contact those who retweeted the incorrect information and ask them to retweet the correction also, he added.

"You're now starting to engage the same network effect that pushed the incorrect information out there. That's the fundamental thing news organisations need to understand about corrections in the networked world, is it's your job to promote it and to reach out and get it to spread out just as far as the incorrect information.

"What really upsets the public is not when we make mistakes, that certainly gets on peoples' nerves, but they realise journalists make mistakes, what they won't tolerate is if we wilfully make mistakes or if we cover them up and don't work hard to correct them."

So to summarise, the top tips from our panel of experts on an effective verification process from start to finish are:

• Monitor across platforms (including Facebook, YouTube, Twitter, Soundcloud, AudioBoo, Bambuser)
• Spot and understand trends (using tools like Hootsuite, Tweetdeck and Trendsmap to create lists and identify trending topics)
• Build a network of contacts before the story breaks and limit the stress
• Use online tools to examine evolution of images (including TinEye, Google Images and WolframAlfra)
• Verifying sources – speak to them and cross reference answers with social data
• Verifying sources – look at social media history across platforms
• Use Whois tools to verify websites
• Check for photoshopping or repetition in images
• Apply the Too Good To Be True test
• Harness online discussion boards and experts (use sites like Snope to spot urban myths and common hoaxes early on)
• Question edited footage
• How urgent is it – could more steps be taken to verify before you publish?
• Crowdsourcing – 'be judicious' about how you send out unconfirmed information
• Consider any permissions and crediting which may be necessary
• Clearly communicate the level of verification a story has been given
• Made a mistake or new information come to light? Issue a clear and networked correction

You can also listen to the experts discuss this topic in our latest podcast here.

Free daily newsletter

If you like our news and feature articles, you can sign up to receive our free daily (Mon-Fri) email newsletter (mobile friendly).