You may have seen an article on The Conversation kicking around on social media this morning which makes a fairly concerning claim: big data analysis suggests that the ‘No’ vote is going to narrowly win the same-sex-marriage postal vote.
The study, from Griffith University academics David Tuffley and Bela Stantic, examined 458,565 tweets from Australian users and used heuristic analysis to determine what side of the debate the users were on and how much they cared about it.
The authors of the study claim that their method is “uncannily accurate” and predicted the outcome of the 2016 US election, which – as you might recall – bucked the trends of the polling at that time.
The general gist of the study is as follows: most tweets about same-sex marriage during the postal vote were from Yes voters, but they only came from a minority of over-active users. Over-55s were underrepresented – because they obviously don’t use Twitter as much – and when Tuffley and Stantic correlated their sentiment with the ABS’s voting data, they came up with the final figure.
Here’s what they concluded:
So it is likely to be a close-run result, much closer than the earlier polls suggested, and leaning in the direction of No.
One of the problems with predicting poll outcomes is that people are often reluctant to say out loud what they really think about issues. What people say online can often be more accurate than what they say to each other in this age of political correctness.
In the lead-up to the recent US presidential election, the polls pointed to a Hillary Clinton win because many people were publicly saying “No” to Trump when asked by pollsters. But in the privacy of the booth, people quietly voted according to what they actually thought.
There’s a strong point made here: Yes voters shouldn’t necessarily be led astray by what they see posted within their own filter bubbles, as I wrote back when the campaign kicked off. Complacency is never good, and it’s easy to fall into the trap of extrapolating a national mood from your own progressive friends and family.
But it’s also worth taking stock of exactly the kind of methodology used by the study. Yes, this kind of heuristic tweet analysis did provide a more accurate result in the US election than traditional polling, which dropped the ball in a big way.
But – and I say this as a very regular user – comparatively few people really use Twitter in Australia. Well, some deeply maladjusted people such as myself do, but its hard to imagine it as an accurate or consistent sample of a broad spectrum of the electorate.
This isn’t to say that we shouldn’t note this study and be realistic about what the final result will be when it drops on November 15. But a study gleaned entirely from Aussie Twitter should probably be taken with a grain of salt for now.
Stick with each other, mates. We’ll have the result soon.