With Facebook omnipresent in modern life, most of us are aware that we see adverts, and that they are often tailored to us individually. There is now also new evidence that the amount of personalisation can be used to change your opinion on political questions. A study from the University of Bath tested the effectiveness of psychographic targeting, the use of psychological profiles, built from information available online, to target adverts at specific audiences.
Facebook even has a dedicated website advertising itself to political campaigns, explaining how to create an effective ad campaign and using the headline “Find your voters on Facebook”. And while an ad might not change your views in the long term, even a small change in turnout can be crucial.
Psychographic targeting means using psychological profiles to target ads at specific users. Artificial intelligence sorts through users’ data such as interests, age, location, gender, etc., and creates psychological profiles based on it. It means that the people making adverts can show you the ones that someone with your personality is more likely to respond to.
Psychographic targeting is used frequently in marketing – if a company knows who they are advertising to, it’s much easier for them to know how to sell something to their users. For example, children’s products are marketed differently to different types of parent. While demographics are concerned solely with the basic information of users (age, location, and so on), psychographics explores their interests and the reasons why they purchase – or vote – the way that they do. They make it possible to use interests, beliefs, and attitudes, as well as demographics to target adverts as precisely as possible.
In Dr. Joanna Bryson’s study at the University of Bath, researchers first had a group take a test that determined whether they had authoritarian views or not. Then, researchers created profiles based on indicators such as gender and interests to guess a user’s views. They then showed adverts tailored to these tendencies to users. It is hardly surprising that users on both sides of the spectrum found the adverts tailored to their views effective and persuasive.
This shows that just by using data readily available on Facebook – not party alignment or political views – it is possible to guess how you would vote, and most importantly, why.
Psychographic targeting has been used heavily in previous elections, so you will probably have seen ads aimed specifically at your personality type, trying to change how you will vote. For instance, during the EU referendum, Vote Leave’s campaign spent over a third of its budget on a small Canadian IT company, AggregateIQ. Similarly, in the 2015 UK general election, the Conservative party spent £1.2 million on Facebook campaigns and the Labour party £16,000.
using personal data might invade privacy – and thus break the law.
Cambridge Analytica, which specialises in using data mining and data analysis for political purposes, was involved with the Leave campaign, Ted Cruz’s campaign, and was later accused by a Time report of spreading Russian propaganda during the US presidential election. Even more recently, The Left in Germany shared an article criticising the far-right AfD – but only on the feeds of Saarland inhabitants who liked the AfD or Pegida on Facebook.
Even though social media, and Facebook in particular, are known to be influential, the Electoral Commission receives very little information on how parties use psychographic targeting. The Information Commissioner launched an investigation into campaigns using data analytics in May as using personal data might invade privacy – and thus break the law. Fines could go up to £500,000. An example of clarity in their use of “dark ads”, the Green Party in Germany has made all of its ads available on its website, explaining that since it is normal practice in in advertising that not all users can see the same adverts, they are making an effort to be more transparent.
It isn’t clear yet if adverts will change people’s views in the long term. However, the researchers say that their effect might simply be of discouraging people from voting. “We know it’s really easy to convince people not to go to the polls,” Dr. Joanna Bryson, the artificial intelligence expert leading the research, told the New Scientist. “Prime at the right time and you can have a big effect. It’s not necessarily about changing opinions.”