Who is winning the social media election?

And how to avoid the ‘echo chamber’.


The big problem with using Twitter and the like to predict the outcome of an election is that those using it aren’t representative of the entire electorate.


What are the stats?

To start with, 80% of 25–34 year olds may be using social networks but only 55% of them voted at the 2010 election. Compare that to the 76% of 65+ year olds who did vote, only 13% of them are actually on Twitter and Facebook.

Even out of that batch, those who are actually tweeting about the UK elections are a small subset of the general population and unlikely to be representative of us as a whole. Perhaps not the best predictor for an election outcome?

Nonetheless, pouring over social media data does unearth the odd gem in the build up to an election. One recent collaboration between The University of Sussex and cross-party think tank Demos, explored the role of Twitter in the British General Election.

Over 9 weeks they tracked 5 million tweets, looking at who retweeted them and linking them each together with a line. Like drawing out an enormously complex connect-the-dots in 3D, they effectively mapped out the political ‘twittersphere’ of the UK. And in doing so captured the different voices and how they are interacting.

Labour are the most dominant party on Twitter by far

It revealed a predictably fiery, political environment with some notable points:

  • Labour are the most dominant party on Twitter by far.
  • The SNP dominate the scottish political twittersphere.
  • Twitter accounts that supported Green, either retweet or are retweeted by Labour and the SNP, far more frequently than by the Conservatives.

Surprisingly, the map reveals that largely conversations take place within party lines and do not cross pollinate.

A report on BBC Newsnight, describes this as an example of the ‘echo chamber’ effect: you follow those on twitter who you agree with. It suggests we insulate ourselves from the views of those that differ from our own, reinforcing our own worldview and potentially closing our minds to new ideas and experiences.

One of the top Labour tweets from the election so far, a tweet from MP Amina Lone has been retweeted more than 1600 times.

But despite it’s popularity, according to Newsnight’s David Grossman it “didn’t make it far beyond the Labour galaxy, it was mostly consumed by the already converted”.

The ‘twittersphere’ showing the distance @Amina_Lone’s tweet travelled.

My echo chamber

So what of my own ‘echo chamber’? Well a quick scan of those I follow shows a predictable spread of friends, colleagues, brands, publications, the odd celeb, MP and political or social commentator. It’s fair to say, I’ve created my own twitter galaxy where I follow those who neatly reflect and reinforce my views… with one or two exceptions.

In an simple attempt to broaden my view (and consequently the guys in my office) of the political social-scape my team and I spent sometime this week creating a mini election hack.

The objective was simple:

Increase our exposure to political voices and conversations on social media.

We decided to create a big screen social wall to be displayed in our office in the final weeks leading up to the election. It pulls in the latest tweets from the leaders of the top parties (including those we don’t follow) and then surrounds them with popular and recent reactions to their posts.

Reaction to @Ed_Milliband’s latest tweet.

For the media’s take on what the politicians are saying, we selected and automatically pulled in tweets from a range of journalists and political commentators of various political persuasions.

Pulling in a tweet from the Guardian’s @IanBirrell
The display shows tweets from all the major parties and what the reactions were.

Finally, we sprinkled it with some statistical data (hey, why not?) including the number of retweets and favourites to give us an impression of how far a message is spreading or how well it’s being received.

What have we discovered?

Well let’s be clear, our election hack has given us a view of the political landscape through the prism of twitter via a sub-set of a sub-set of the population. So it’s certainly not representative. It has however exposed us to more voices and active participants in the election conversation that we wouldn’t normally see.

We built our display using social media aggregator Climb.