• by Erik Sass , Staff Writer @eriksass1, June 24, 2016

The explosion of social media over the last decade has given rise to a whole cottage industry of what might be termed opinion polling by proxy, in which pundits attempt to predict the outcomes of important events by raking over social media analytics. The reasons for doing so are easy enough to understand: social media has been so widely adopted it may appear ubiquitous, and users freely express their feelings about every subject under the sun, all with convenient hashtags which make them easier to track. Above all, it’s a lot cheaper than hiring pollsters to conduct independent surveys.

The only problem is that it doesn’t really work, as was neatly demonstrated in the case of #Brexit, Britain’s stunning vote to leave the European Union on Thursday. Indeed, social media analysts tracking sentiment about the referendum sound just as surprised as the rest of us – and now some of them are trying to unpack the reasons why social media sentiment didn’t prove a reliable predictor of the actual outcome.

There’s no question that the overall trend on social media seemed to favor “Remain”: according to one social media analytics and monitoring tool, Talkwalker, which had #Remain leading #Leave hashtags by 58.4% to 41.6% in the final 24 hours before the vote. Considering the final outcome was 52% to 48% in favor of Leave, there was clearly a disconnect between social media and reality.

One obvious shortcoming of social media as a predictive tool is that despite its appearing ubiquity, not everyone is actually using it. As marketers well know, young people are more likely to use social media than older folks – and in Britain they were also more likely to favor Remain, creating an appearance of stronger support than actually existed. What’s more, old people (while lagging on social media adoption) are also more likely to actually vote. Did some young people who talked a good game on social media fail to follow through with a visit to the polling station? There’s no way to know for sure, but it certainly looks that way.

There’s also the fact that most of the big personalities on social media – including a whole raft of celebrities, politicians, business people, pundits and so on – favored Remain, and propagated this message across their large number of online followers. But just because people are idly talking about what celebrities are talking about doesn’t mean they share these opinions themselves, or embrace them strongly enough to make the trip to the voting booth.

The outsized role of celebrities and other public figures in social media, and the disparity between their sentiments and the final outcome, also illustrates the basic anti-institutional thrust of the whole Brexit movement. Above all the vote reflected the deep alienation of a large section of the British population, mostly older, with less education and lower incomes, who have arrived at a position of radical skepticism and distrust of politicians, experts, and everyone else associated with the “establishment.” The social media debate over Brexit was dominated by establishment types, but for pro-Brexit voters this was a meaningless truism – yet another sign that media coverage was “fixed.”

It’s worth noting that the overall picture is complicated, as not every social media platform showed a pro-Remain majority. One analysis of the EU referendum debate on Instagram, tracking almost 16,000 users, found that 32.6% supported Brexit, compared to just 12.3% in favor of Remain (obviously with a large proportion of ambiguous or undecided participants). The same study found that Brexit supporters were much more active and organized in spreading their message on Instagram, with posts from Leave supporters receiving 26% more likes and 20% more comments.