EU referendum: The pollster party is over – big data drawn from social media like Twitter is where we’ll learn if the UK is voting Brexit or Bremain
The EU Referendum is looming and at this stage, it’s still anyone’s guess which way it will go.
David Cameron’s ‘In’ campaign is running at full steam but Boris Johnson’s surprise announcement to back the ‘Out’ vote may yet derail his plan.
Already, there are dozens of polls taking place desperately trying to judge which way people are swaying. However, when it comes to the pollsters’ view of the forthcoming referendum, how accurate will they actually be? Frankly, if we’ve learnt anything from the UK election, they don’t hold much weight.
They use an archaic methodology which may have worked decades ago when the public was still communicating via letters and telephones, but things need to evolve if they’re going to keep up.
In a society engulfed in social media, public sentiment is not only easy to read but very reliable too. This is because people are more likely to voice their honest opinions on social media than anywhere else, rendering results more accurate. And why is this? It is because monitoring social media removes social desirability bias.
Social desirability bias is a social science research term that describes the tendency of survey respondents to answer questions in a manner that will be viewed favourably by others – and this is where the pollsters are falling down. Ultimately, they can’t compensate for the Hackney hipster who says they voted Green when in fact they’re a closet Conservative.
And when the pollsters fall down, the big data engines stand up. People on social are not being polled (at least they don’t know they are) they’re being 100 per cent honest, and that is where true and accurate sentiment can be monitored.
So, how can we learn from this when it comes to the EU referendum?
Weeks ahead of the UK election our big data engine, which was tracking conversation on social media, spotted that the Conservatives were a clear leader in the polls.
It also flagged that Nicola Sturgeon’s popularity was increasing by the second as her own televised debate went on. This was because it was based on analysis of tens of thousands of social media conversations, not a poll.
For those still not convinced about the influence of social media, the European Central Bank (ECB) released a paper last year on using online conversations to predict the investment markets. It concluded that Twitter bullishness has a “statistically and economically significant predictive value in respect of share prices in the United States, the United Kingdom and Canada".
Yes, this is referring more specifically to share price value, but if the ECB is ready to listen to Twitter when it comes to financial information, shouldn’t we all listen to it?
Of course, pollsters offer a very useful service, and it’s not a case of doing away with them altogether, but they do need to embrace technology and listen to the real audience – and that audience is online.
What we do know is that, now even more so following Johnson revealing where his vote lies (considering the influence he has), this referendum will be called on social media.