Australia Elections To Test Data-Driven Democracy

As Australia goes into election mode to decide on a new Government on May 21, the use of data and targeted social media will reach new heights of sophistication as the political parties campaign for votes.

The May 21 poll will be a digital election where the candidates will use techniques similar to tailored and personal advertising to target the right people and get their message across.

Australia might not witness the same level of foreign interference as the 2016 U.S. elections, where accounts connected to Russia spent big on targeted Facebook advertising. But the use of so-called microtargeting by many domestic stakeholders is expected to reach new heights.

Political parties in Australia enjoy a controversial exemption from aspects of privacy law. Due to this loophole, one party recently sent out hundreds of thousands of texts to voters. And the parties are not necessarily required to disclose how they came by personal information, which they then use to tailor their messages.

In recent months, Facebook ad spending by the major parties — particularly the governing Liberal coalition — has increased markedly, many of them personally targeting the Leader of the Opposition Labor Party.

Whether the Liberals are getting value for money is debatable. Around 40 of their ads in the last month have had more than 20,000 views, with approximately 70% being negative “attack ads” about their opponents.

Voters are microtargets

The other feature of the social media campaign is microtargeting.

If someone lives in a safe seat where the result is in little doubt, they may not see much in their social media feed. But if their online footprint reveals they are in one of the hot electorates, which could seal the fate of the election, then they should be prepared to be bombarded.

In regional New South Wales, for example, the Opposition Labor Party — leading in the opinion polls — desperately needs to win the seat of Macquarie. According to the Transparency Report from Google, the Labor Party has pushed out different YouTube ads to people living in other parts of the electorate.

If you live in the Hawkesbury region, on the river and its floodplain, you will receive an ad on your feed talking up Labor’s plans for the State Emergency Service. But if you live in the Blue Mountains – in the same electorate – your feed will promote a Labor ad talking up plans to fund fire prevention and control, a hot button local issue after the 2019 bushfires.

If their online footprint reveals they are in one of the hot electorates which could seal the fate of the election, then they should be prepared to be bombarded

Across South Australia, the other side of politics uses similar techniques based on age demographics. Older females in the seat of Barker are reportedly more likely to see ads about Liberal agricultural policies, while younger male voters will be exposed to ads on road funding.

Another feature of the upcoming election is the anticipated strength of an independent movement of candidates agitating for action on climate change. This group, calling themselves Climate 200, is focused on voters between the ages of 18 and 34 who earlier research has identified as being more engaged on climate issues.

Other lobby groups are also pushing their agenda during the election. Research by data journalists at the Australian Broadcasting Corporation found that the Pharmacy Guild is sponsoring ads about the cost of medicines but is explicitly targeting younger people about the cost of both medicine and rent.

Addressing opinion poll bias

As with elections elsewhere, the inaccuracy of opinion polls has been a significant talking point. This was a major issue in the 2019 election, won by the Liberal coalition despite almost all opinion polls pointing to a resounding Labor victory.

In the years since, the polling industry has been working hard on its models to understand what went wrong. It concluded that polls are almost always biased towards Labor because the most engaged and educator voters favor the party and are more likely to respond to political surveys.

The polls are created from polling well-established panels of voters and not random samples. So, it has taken the polling industry some time to understand that these panels are not representative of the wider population.

As a result, the major pollsters now reweight their results to account for this bias. Although they have declined to reveal how much, it is believed to be around 2.5%.

With a tight vote expected on May 21, the polls' microtargeting and reweighting are assuming increasing importance in the most data-driven election in Australian political history.

Lachlan Colquhoun is the Australia and New Zealand correspondent for CDOTrends and the NextGenConnectivity editor. He remains fascinated with how businesses reinvent themselves through digital technology to solve existing issues and change their entire business models. You can reach him at [email protected].

Image credit: iStockphoto/twinsterphoto