Winning friends, influencing people, and measuring it — 2019 election style

Australia had an election this weekend. The result — reelection of the Coalition Government — ran counter to many predictions. In fact, one betting agency even paid punters on the result two days before the election — a multi-million dollar mistake, given that they got the result wrong! Understandably, many people have asked how the polls got it so wrong, so I was interested to see this article on the topic in The Sydney Morning Herald.

The article talks mainly about the role of social media in the election. The idea of using social media to predict events or analyze an election isn’t new, but the Australian angle is interesting.

Given that Australian opinion polls, which had always been based on landline phone numbers mapped to the electoral roll, seem to be struggling in a post-landline world, it’s worth thinking about better ways to gauge sentiment. Social media gives us one option, although there are plenty of caveats, including filtering the real from the fake, and the fact that it’s only really possible to analyze public social media posts.

It’s also interesting to read the comments in the article on the different social media methods that the two major contenders in the election used, with an observation that the Liberal Party of Australia (the larger party in the Coalition) had been much more effective in engaging users with its online content that the Opposition party, the Australian Labor Party.