Despite the backlash against Cambridge Analytica, psychological targeting is not going to disappear from the political stage anytime soon.
As a researcher who has been involved in the development of psychological profiling and targeting from the very start, I am often approached by companies trying to enter the market of psychological profiling. Judging from the number of requests I receive, there will soon be more capacity for psychological targeting, not less. That means that the debate over psychological targeting has to continue — and we might find that psychological targeting does not have to necessarily be a bad thing.
On July 17, the UK’s Electoral Commission sanctioned the pro-Brexit Vote Leave campaign for violating its spending limit. The campaign had funneled £500,000 through another campaign group to pay AggregateIQ, a firm linked with Cambridge Analytica — a political consultancy which has allegedly deployed psychological microtargeting on behalf of the Trump campaign.
I am somewhat puzzled by the media coverage around the alleged role of psychological targeting in the 2016 US presidential election and in the Brexit referendum. On the one hand, psychological targeting is described as the new wonder weapon that single-handedly delivered the Leave vote and Donald Trump. On the other hand, it is described as a sham that under no circumstances could have contributed to either outcome.
Let's take a step back here and acknowledge that the world is rarely as black and white as those opposing views. Is there reason to believe that psychological targeting is effective? Yes. Was it the main driver behind Brexit and Trump? No.
Political events such as the US presidential election or the EU referendum cannot be chalked up to a single cause: a vast number of factors, motives and circumstances is always at play. That does not mean, however, that clever campaigning focused on people’s fundamental motivations, fears, and preferences could not contribute to a certain result. Even less so does it mean that psychological targeting as an approach to persuasion is ineffective.
In fact, academic research, including my own, shows how tailoring persuasive messages to people’s psychological profile makes them more effective in influencing not only people’s attitudes and beliefs but also their behaviour. Consider this: do you communicate in the same way when you speak with your three-year-old, your best friend, or your boss? The idea sounds almost ridiculous. We are so used to tailoring our language in face-to-face interactions, that we often forget that the tailoring is there. We intuitively know that it is going to be more effective.
If we look at the presidential election or the Brexit vote, all it took to swing the outcome was a few extra votes from those who remained undecided until the end. This means that psychological targeting could have indeed been effective enough to have an impact. Even if the extent to which Cambridge Analytica itself used psychological targeting remains unclear, what is important moving forward is answering the following questions: how effective is psychological targeting in comparison to other data-driven approaches such as the ones used, for instance, by Barack Obama in 2008 and 2012? And if psychological targeting is found to provide a competitive advantage over other approaches, should we ban it from the democratic process?
The answer to the first question is empirical. We don’t know enough to claim that psychological targeting is certainly more effective than other approaches, but there is reason to believe that it could be. That is because psychological targeting offers advice on how to communicate with people: it is less about finding the perfect and most receptive audience for a specific message or candidate, and more about providing a human understanding of one’s audience that makes it possible to cater persuasive messages to the audience’s specific needs and motivations. From the few research findings we have to date, it seems that this assumption holds— but it is too early, and we do not have enough evidence, to know for sure.
The answer to the second question is a normative one that can only be collectively answered by society at large. We might decide that we do not want psychological targeting anywhere near our democratic process, because the risks of undetected voter manipulation are too large. If psychological targeting is indeed effective, the story of Cambridge Analytica certainly suggests that it could pose a threat to our democracy. I am not opposed to this view, but let me play devil’s advocate for a second and introduce a different scenario. Imagine that not Trump but Clinton had used psychological targeting in her campaign, and won. Not to discourage people from voting, but to try and understand what they truly care about, and to communicate with them in a personal way that encouraged them to be part of the political process.
I cannot know for sure, but my best guess is that she would have been celebrated. Just like Obama was celebrated for his clever use of data in his campaigns. Her usage of ‘advanced tools’ such as psychological targeting might have become a badge of honour.
Call me naïve, but I truly believe that the same way psychological targeting can pose a threat to democracy, it can also provide a solution for an even bigger threat: the fact that most people are fed up with politics right now— whether in the US, in the UK, or elsewhere in Europe. Politicians are often accused of not caring about their constituents enough and of not speaking their language. That is exactly what psychological targeting is about.
Instead of condemning the technology altogether, we might decide to initiate a public debate around how we could regulate psychological targeting in a way that makes it harder to abuse (for instance, by creating more transparency around political advertising), but that allows us to leverage the power it could bring to the very idea of democracy itself.
If we do so, let’s not solely focus on Facebook the way that we have done so far. Psychological profiles can be built from pretty much any digital footprint out there, whether that is your Twitter profile, your email history, your credit card summary, or the sensor data collected by your smartphone. It is not just Facebook, it is all of our digital lives. We all benefit from the technologies associated with these data, and it is the responsibility of all of us to make sure we use it in our best interest and the best interest of the societies we live in.
Sandra Matz is a computational social scientist and an assistant professor of management at Columbia Business School in New York.
This article was originally published by WIRED UK
