Big Data and the Art of Political Persuasion
In a recent article, I reported the research of Robb Miller and Mathew Feinberg from Stanford and the University of Toronto. The two authors present evidence through conducting a series of experiments that political salesmanship or persuasion is most successful when it addresses values that the considered person holds rather than values that might be more widely held, or that the person is holding that is trying to persuade someone.
In their experiments, they show that conservatives are significantly more likely to support gay marriage if it is framed around values that conservatives tend to pride themselves on. Similarly, liberals are significantly more likely to support higher defence budgets if this is framed as a public works program that will create employment rather than appealing to patriotic values.
The authors argue that the success of political persuasion can be significantly increased through what they termed “moral reframing”.
In his newest book “WTF” Robert Peston discusses the enormous impact of new campaigning strategies on recent elections such as the Brexit vote in the UK. Peston discusses how the Leave campaign employed a new social media-driven strategy that targeted voters with individualised messages and considered how likely they were to vote for Leave or Remain. Peston reports how the Leave campaign obtained data on the likelihood of different voters to prefer Leave or Remain and ultimately was able to target voters with specifically tailored campaign messages.
In the case of Brexit, it seems like the campaign slogan surrounding the supposed possibility to redirect the spending of 350bn GBP to the funding of the NHS every week, should the UK be able to leave the EU, was most successful. This specific policy item seemed to strike the biggest chord with the voters in the UK and seemed to also capture the biggest criticisms that people had towards the EU. This leads one to assume that anti-EU sentiment in the UK was predominantly fuelled by the perceived need to employ more resources at home as certain social services are underfunded, as well as the notion that the UK was contributing considerable sums to Brussels without adequate benefits resulting in return. Peston also mentions the important detail that the campaign slogan regarding the 350bn GBP was altered for different audiences – there were different versions of this slogan.
So where is the connection between the research on the relevance of political values for political persuasion and big data used in political campaigns?
The more insight into the values, hopes, and fears political campaigners can have, the more targeted can their political messages be to maximise the persuasion power that their arguments have for that person. Hence, big data derived from social media enables political campaigns to create political messages that have the highest persuasive power for every single person by appealing to their individualised core values.
This way of convincing someone does not just have implications for targeted campaign slogans. But also, can be used in a wider or rather more comprehensive political persuasion campaign. If it is necessary for my political campaign to convince you of a certain belief or argument, and I have extensive information about your beliefs, interests, and values, I can theoretically through the power of targeted and personalised social media streams, create a political and social reality in your news feed that triggers the right buttons in your brain and refers to the corresponding values you hold, for you to be swayed by a particular argument.
This does not even require fake news!
In the wake of Donald Trump’s election campaign and ultimate success, as well as because of the now obvious meddling of Russian troll farms in Western elections, the topic of Fake News has gained status among the fears of political commentators. What is concerning about the interaction of value-based persuasion techniques and the ability to gage surprisingly accurately what different people will react most to on the basis of their virtual footprint, is the realization that fake news is not necessarily required to create distinct political realities for people online.
A very primitive way of doing this would be to bombard you with news stories of immigrant attacking people within the native population, for you to activate your sense of us-versus-them, but also your need to defend your tribe, and a sense of urgency for you to vote against a pro-immigration bill.
It might be true that there are in fact attacks of the described nature so this would not be fake news per se, but your news feed might withhold important information or underreport violence of the opposite side, and so create a perceived reality that significantly differs from what is actually going on.
This is of course not necessarily a completely new phenomenon. The research of Miller and Feinberg merely very neatly describes how and why this process works. I do not believe that the findings are counterintuitive to anyone reading about them. What is new, however, is social media’s power to shape what you regard as reality. Your political opinions have become more a function of the time you spent on various social media and the cues you receive while being subject to influencing factors on those platforms. Because of the sheer amount of time we spent on social media and the brilliance of these platforms to deduce our interests and values through our behaviours on those platforms enables political campaigns/cues to be almost perfectly tailored to your value and belief system and so have the strongest influential effect conceivable. This means that a political campaign could not only show you just certain parts of reality but also present them in a way that is perfectly suited to convincing you specifically.
These marketing techniques have been used for a long time to make us buy things that we had previously no idea we needed- through targeted ads. But now the grey political class seems to slowly catch on and arm their political campaigns with the newest insights on political persuasion and the power amplifying tools of new technologies.
What will be the result of this? Is this dangerous?
Well, it is incredibly powerful, which means it is crucial who uses these tools and to what ends. And as there is no way to control that, and no real way to determine who should have this level of power and who shouldn’t – this is absolutely dangerous!