How Boston Police led the way on social media

Memories of the horror surrounding last year’s Boston Marathon bombing were recently revived when the US Attorney General Eric Holder announced that federal prosecutors would seek the death penalty for Dzhokhar Tsarnaev, the survivor of the two Chechen brothers accused of the atrocity.

The first anniversary of the bombing, in which three people were killed and an estimated 264 injured, fell on 15 April. The almost simultaneous ignition of the two bombs near the finish line of the Boston Marathon also unleashed a torrent on social media.

It took two minutes and 52 seconds after the explosion for tweets containing the word stems ‘explos*’ or ‘explod*’ to be sent. Most came from within the vicinity of the finish line as eyewitnesses reacted to the shock. Runners, supporters and local residents were among the first to relay the news and images. It took slightly more than three minutes for traditional news organisations, such as Boston University’s radio station WCVB, Associated Press and CNN to react.

Such was the flow of up-to-date information on social media that The Boston Globe temporarily converted its homepage to a live blog that pulled in tweets from Boston authorities, news outlets and ordinary citizens.

It has been estimated that more than 20 million tweets about the atrocity were sent in the days and weeks after the Boston Marathon. And according to Washington-based think tank Pew Center of Research, all this was lapped up by a thirsty public audience.

Four out of five Americans followed the story on television, but Pew estimates that 26 per cent tracked the news on social networking sites. This number rises to 56 per cent for 18-29 year olds.

When asked why they followed on social networks, the most common responses were that the news was ‘just there’ or that they were already online and unable to avoid the flow of information.

One in eight of those surveyed cited the ease and convenience of following news on social networking sites, while 11 per cent described it as fast, up-to-date information.

But such was the volume of tweets that it was perhaps unsurprising that disinformation started to feed through. Some of this was deliberate. Two Twitter feeds, Boston Bombing Truth and Free Dzhokhar Tsarnaev, were openly aimed at challenging the official version of events and defending the surviving suspect.

Other errors were inadvertent. When Reddit started a ‘Find the Boston Bombers’ thread, it incorrectly named as a suspect Sunil Tripathi, a student from Brown University who had gone missing a month before the marathon.

Erik Martin, Reddit’s general manager, adds: ‘We apologised privately to the family of the missing college student. We have apologised publicly for the pain they have had to endure.’ Tripathi’s sister described Reddit as ‘ugly and disgusting’.

Alexis Madrigal, senior editor at The Atlantic magazine, condemned Reddit for what he called ‘vigilantism’, adding: ‘It’s only the illusion that what we do online is not as significant as what we do offline that allows this to go on. Imagine if people were standing around in Boston pointing fingers at people in photographs and accusing them of terrorism.’

The mainstream media also erred. Associated Press, one of the world’s most respected sources, admitted it blundered. ‘We took a shellacking, a deserved one, for reporting that a suspect was in custody when, as the hours passed, that information began to look wobbly,’ explains executive editor Kathleen Carroll.

Citizen journalism also brings other challenges. Some witnesses posted horrifying photographs of bomb victims that traditional media outlets would have censored.

On the other hand, an amateur photographer produced the first high-resolution picture of Tsarnaev within seconds of the blast. The runner, who had just crossed the finishing line, used his iPhone to take a better image than that produced by the authorities.

At the centre of this social media firestorm was the Boston Police Department, who were simultaneously trying to manage the crowds, find the bombers and combat the blizzard of good and bad information swirling around the Internet.

Ed Davis, Boston’s Police Commissioner at the time, said: ‘Misinformation among our colleagues in the media was running rampant.’

But, looking back on it now, the Boston Police Department insists that social media disinformation was no obstacle.

Officer Neva Coakley, who is part of the department’s four-person media relations team, explains: ‘The public’s output did not hinder us, though we did have to correct disinformation. People want to be the first to get information out, but in the end they realised they had to follow us.’

As soon as the explosions ripped through Boylston Street, the media team realised they had to take the initiative with communications, using it both to connect with the public with routine information such as escape routes, and at the same time controlling the waves of speculation regarding possible further bombs and the identity of the bombers.

The official @bostonpolice Twitter account was used to send all the department’s tweets. It was overseen by Cheryl Fiandaca, a former television journalist who is now Boston Police’s public information bureau chief, who operated the feed as a 24 hour digital hub for information about the investigation. Fiandaca and her team were briefed by police commanders between three and five times a day during the crisis. Their first tweet to confirm that an explosion had occurred at the finish line came within an hour of the detonations.

The day after the bombings, other approved members of Boston Police started to tweet information including the hashtag #tweetfromthebeat. For example, deputy superintendent John Daley posted information on road closures and answered questions from the public on a range of topics, such as where flowers could be laid, how to submit videos and pictures as evidence and whether volunteers were needed to answer the phones.

Boston Police reconfigured its GroupTweet settings to ensure each tweet from an approved contributor containing the necessary hashtag was retweeted by the official account, reinforcing the message.

When CNN reported on 17 April that an arrest had been made, it was swiftly followed by other major news outlets, leading to more than 5,000 retweets. Boston Police swiftly tweeted a clarification that no such action had been taken, which was retweeted almost 11,000 times. ‘We ran a regular blog on our website, bpdnews.com, and put links to that on Twitter and Facebook,’ adds Coakley. While Twitter was used for rapid updates, Facebook was used to publish images of the suspects, which were shared more than 6,500 times; to provide car number plate information to support a ‘Be on the lookout’ alert; and to show maps of the cordoned-off area and updates on public transport.

Later Boston Police used its Facebook page to create memorials to the victims and send messages of condolence and support to their families and the survivors. From the moment the bombs went off until the story began to fade from view, the number of people following @bostonpolice soared from 55,000 to 300,000. Similarly, the number of ‘likes’ on the department’s official Facebook page jumped from 15,000 to 85,000.

Significantly, those numbers have held up today – at around 269,000 followers and 92,000 ‘likes’ – suggesting that the police are regarded as a trusted source of news and advice. To put these in context, the Twitter account of the Boston Globe, which has been awarded 22 Pulitzer Prizes since 1966, has 280,000 followers while that of the Boston Herald, which has eight under its belt, has just under 42,000 followers.

At their peak, 240 news organisations had people in Boston. The police did not have emails for all of them, but that didn’t matter: they followed @bostonpolice.

Coakley recalls: ‘We were working three days non-stop and we were managing 24-hour shifts. We told tweeters not to follow officers during the manhunt, because they could endanger lives, and they stopped.’ A tweet scolding those tweeting information from police scanners was retweeted more than 20,000 times.

Coakley had been home just half an hour after getting back from yet another exhausting shift, when the news broke that the second suspect had been captured. That was retweeted 142,000 times.

Boston Police have always been progressive on social media and using different platforms to communicate with the public. For example, its ‘text a tip’ hotline launched in 2007. It allows residents to send information anonymously and has been credited with helping to solve several crimes. The hotline received 333 tips about the Boston Marathon bombing.

Even before the attack, Boston Police’s Twitter feed, which included initiatives such as #MostWantedMonday, linking to photos of suspects, had more followers than any other American police force. A similar initiative ‘Boston’s Most Wanted’ was launched on Pinterest.

But over the past year, the force has moved on apace with social media. Its blog, Facebook and Twitter presences have been updated, and plans are afoot to launch a phone app. Experienced officers will be encouraged to tweet from the beat. There will be guest posts, surveys and contests on Facebook, alongside videos showing how to fill out a stolen-vehicle report form or apply for a firearms licence. Boston Police will also include cross-promotions on its social media platforms to advertise events such as Massachusetts Autism Awareness Day.

Coakley adds: ‘We have learned a lot of lessons. At the baseball World Series last October, we used social media to communicate to the crowd which streets or bars were closed, or whether a particular intersection was getting too congested. Whenever we did that, we could see on our TV screens the crowd responding as they picked up the tweets.’

While few companies or government organisations regularly face the problem of crowd control, there is no doubt that the public increasingly expect to find out the latest information about an event on their tablets or smartphones.

They are not waiting to be contacted, but are simply entering the appropriate keyword into the search box.

And those organisations that are not immediately on top of the unfolding story will very quickly look out of touch - and out of control.


Why misinformation spreads

An analysis of three false rumours that spread on Twitter in the aftermath of the Boston Marathon bombings has found that the social media crowd has the potential to self correct. Rumours, false flags and digital vigilantes: misinformation on Twitter after the 2013 Boston Marathon Bombing, an academic report by researchers at the University of Washington, considered 10.6 million tweets sent by 4.7 million authors between 15 April to 22 April 2013.

The analysis limited the number of tweets to 50 per second, and focused on those with the terms ‘Boston’, ‘bomb’, ‘explosion’, ‘marathon’ and ‘blast’. It then considered tweets containing specific hashtags to understand their context.

The most blatantly false rumour on Twitter claimed that an eight year old girl had been killed while running in the race. Its roots can be traced to a tweet from @ NBCNews announcing that an eight year old spectator had been killed. Within 45 minutes, a Twitter user ascribed a female gender to the victim and suggested she had been a participant. Four minutes later, another user added a fake picture.

The researchers found 92,785 tweets related to the rumour. Of these, 90,688 were false and just 2,046 were corrections. This is a misinformation-to-correction ratio of more than 44:1

‘Significantly, peak correction did occur roughly within the same hour interval as peak misinformation, suggesting reactionary community response,’ said the researchers, who found misinformation ‘to be more persistent, continuing to propagate at low volumes after corrections have faded away’.

The second rumour followed the release of images of dark, exploded backpacks by the FBI. Users of social media platforms like 4chan and Reddit began to collect and analyse images of individuals wearing backpacks at the scene. Within a short time, users suggested that an emblem on the hat of a suspicious individual suggested an affiliation to US military special operations. This conclusion was tweeted about 4,525 times but only 212 tweets were corrections. The rumour even returned after the identification, shoot out and capture of the Tsarnaev brothers.

When a former classmate of Sunil Tripathi, a 22 year old Brown University student who had been missing for a month before the attack, posted a tweet noting his resemblance to one of the suspects, it prompted a Reddit thread that focused on the connection. Shortly afterwards two tweets claimed Tripathi, whose body was found on 23 April, had been named a suspect by Boston Police along with Mike Mulugeta.

Within minutes, these tweets were retweeted by traditional media outlets. The misinformation was tweeted 22,819 times. The confusion was ended by the FBI releasing the Tsarnaev brothers’ names, although the research found that corrections had already started on Twitter before the official announcement.

For about six hours after the first tweet, tweets linking Tripathi to the bombing were sent at a rate of between one and three per minute. But then this swiftly spiked from 40 tweets in ten minutes to 4,690.

The researchers, who are trying to develop methods to automatically identify misinformation, concluded: ‘We find evidence of crowd correction for each rumour. Though misinformation and corrections seem to rise and fall in tandem, they exhibit different magnitudes and a lag between the onset of misinformation and the correction.’