This is the first in a series of articles I’m going to write about Cambridge Analytica and political messaging in general. Read part two, on Irish politics, here; but first, a general introduction.
This week saw a powerful reminder of the importance of top quality journalism, with the revelations by the Guardian, New York Times, and Channel 4 News that British company Cambridge Analytica used Facebook profile data, without users’ knowledge, to influence the outcome of the last US Presidential election and possibly the Brexit referendum.
It’s an enormously complex story, of course, but the essential elements break down, in most of the coverage, to two points:
- The means of obtaining the data from Facebook
- How the data was used to target and influence voters
The first point raises obvious concerns over data protection and personal information online. A Facebook personality quiz asked for access not only to the profile of any user taking the quiz, but also the profiles of all that user’s friends, without any participation or consent from those friends.
The second point is not so straightforward. The data was used to target particular voters, by noting correlations between specific likes and interests and some political opinions, and aiming online advertising at demographics with those likes and interests. That’s also pretty straightforward, and is really just refining the methods of online advertising to reach the right people. It’s not new: in campaigns I’ve run, or been involved with, we used online advertising targeted at particular keywords in Google searches, for example.
Access to the Facebook database would have allowed better targeting. Rather than just having ads showing when users search for keywords relating to candidates or issues, the ads could be targeted at all online activity by people in the selected demographics. So the advertising would be targeted more specifically to people who might be sympathetic, and those people would see the ads more often – much more efficient return on investment than simple keyword search targeting. Thousands of variations of ads can be run, microtargeted at specific groups of people. A simple example might be that users whose favourite colour is blue might see ads with a blue background, while those who prefer green will see a green version.
That much is simple, and I don’t particularly have a problem with it, beyond where the data comes from. The complexity comes in the content of the advertising.
There has been much discussion of the fact that the ads used by the Trump and Brexit campaigns were generally emotive, rather than factually-based. At the heart of this are two overlapping issues. On the one hand, emotive advertising; and on the other hand, so-called fake news. Both campaigns blurred the lines.
Some journalists and many users online have expressed surprise at the first concept, the use of emotions in political campaigns. But this isn’t new: it’s long been known in political circles that tapping into emotions is much more powerful than purely fact-based messaging. There’s a rather excellent book on the subject by American psychologist Drew Westen called The Political Brain, which I urge you to read if you’re interested in the topic. Amongst the American examples he cites are ads for presidential candidates going back decades.
This is probably the single most powerful political advertisement ever, from Lyndon Johnson’s 1964 campaign against Barry Goldwater.
There’s not a single piece of factual information in that ad, except the date of the election. It is laden with emotionally charged imagery and sound. Little girl, aw cute. Daisy petals – most people have done something like that as a kid, so nostalgic. Miscounting – awwww, just adorable. And then the massive contrast to a countdown, with a sombre ominous voice, ending in a nuclear explosion. “The stakes are too high to stay home” – playing on Cold War fears; remember, this was only a few years after the Cuban Missile Crisis.
Not a single fact, nothing to say why Johnson was more likely to avoid catastrophe. But the association was enough. Who would you have voted for after watching that? You probably hate Goldwater already, and you know nothing about him.
Johnson, ahem, blew Goldwater away in the election, 486-62 in the electoral college. It was a wipeout. This ad is why.
Trump’s team actually leaned on some elements on Reagan’s 1984 campaign. “Make America Great Again” is not a million miles from this:
There’s some facts in this, but the power comes from the emotional impact. Those facts were carefully selected for their emotional impact. First time buying a home. Getting married. Morning. Dawn. Hope. It’s not the facts that have power here, they just provide the foundation for the emotion. Reagan crushed Mondale 525-13 in the electoral college.
More recently, Obama relied heavily on emotional messaging. Remember this?
Obama spoke of hope and change – but was a bit lighter on the subject of what he was hoping for, and what he wanted to change to.
Sound familiar? The Brexit campaign did exactly the same. They associated negative emotions with the EU, and positive emotions with a future outside the EU. “Take Back Control”. “We Want Our Country Back”. Voting day would be their “Independence Day”. Nothing about how any of this would work, or what the future outside the EU might look like, just that it would be better.
The same for Trump. “Make America Great Again” – but not so much detail on how.
Both Trump and Brexit campaigns relied not just on associating positive emotions with who and what they wanted people to vote for; they associated negative emotions with what they wanted people to vote against. “Crooked Hillary” as a national security risk, with her insecure emails and Benghazi. The EU as corrupt and undemocratic. Both played on fears of immigration: “Take Back Control” and “Build The Wall”.
Where they seriously crossed the line was in stoking people’s fear, and outright lying. Using emotions to emphasise your political message and policies is generally fine; generating that emotion in the first place through lying and deceit is not.
The Brexit campaign lied about immigration, about the levels and the impact, and that’s why the “Take Back Control” message worked. It resonated with fears of terrorism and job losses they’d already created through manipulation and deceit. Trump’s border wall with Mexico worked for exactly the same reason. Never mind that the vast majority of Muslims in Britain and America are just as horrified by Al Qaeda and ISIS as everyone else, and most immigrants are benefiting the economy and not taking anyone’s job or claiming massive benefits.
The real question that should be asked is not why emotional messaging works, but why does fake news work.
I put the blame on three groups, who in my mind, have blurred the lines to such an extent that it’s now hard to tell the difference between truth and fiction.
First, politicians; and not just Trump and the Brexit campaign. How many times have we seen pre-election promises broken? How often have we seen politicians twist facts to suit their case? We’ve reached a point where we simply don’t know what to believe when politicians speak.
Second, the media, particularly tabloids. The Daily Mail is an excellent case in point here, constantly full of contradictions and hypocrisy – but it’s the difference in coverage of issues between different media sources that’s the real issue. Read the Guardian and you’ll get one view; read the Mail or the Sun and the news is presented in a completely different light. Who are we supposed to believe? The likes of the Mail attract people in with stories about celebrities, generating familiarity and interest, then hit them with a right-wing agenda. It’s not particularly likely that the Guardian is then read for balance; but even if it was, the contradictions would generate confusion. So, most people default to believing what their usual sources tell them: the concept of a trusted messenger. It doesn’t matter what the message is; what matters is where it comes from.
Third – and this one might be controversial – religion. Through schools and churches, most people are brought up as children with the concept of faith – belief without evidence. In my mind, this is the single most dangerous idea in history. That we should be expected to believe in God and saints and heaven, or variations thereof depending on which religion, without any evidence whatsoever – we have been brainwashed in our youth to believe the trusted messenger without question. In religion, that messenger is the Bible or Qu’ran etc, as well as the priests, ministers, rabbis, imams and other clergy. Is it any surprise, then, that we generally don’t seek evidence from our trusted messengers in politics and other areas of life?
For example – most of Trump’s core vote is the American evangelical Christian community – despite the fact that he is almost completely anathema to their core religious beliefs. Why? Because he said the right things to become their trusted messenger, and they unquestioningly believed what he said.
In fact, once the trusted messenger has spoken, contradictory messages don’t tend to change minds – they provoke a defensive response. The alternative message isn’t believed, and the recipient gets angry. The original message is actually reinforced by contradictory new information, bizarre as that may seem.
A final factor is that in modern life, most people don’t have time to fact-check. Life is stressful for many people, especially with the recession over the past decade. People are worried about rent or mortgage payments, jobs, bills, childcare, healthcare, and much more. This is why so many people, for example, are not paying the lowest available rates for things like bank charges, or electricity bills. We don’t have time to sort through all the options – there is so much choice out there, there’s almost too much choice. We want to know that when we put the kettle on, it will boil and we can have a cup of tea; as long as that works, we’ll live with it. We don’t have time to fact-check everything we see in the news.
So, that’s why it worked for Trump and Brexit. Emotional messaging based on fears stoked by manipulation from trusted sources which people had been conditioned to believe, and didn’t have the time or the inclination to fact-check.
Cambridge Analytica may have facilitated both campaigns, but they can’t claim full responsibility. Trump had already won the Republican nomination and the Brexit referendum had already been called before they got involved. More than that, the seeds of both had already long been sown.