By Dr. Ken Broda Bahm:
Have you heard of a company called Cambridge Analytica? Well, chances are, they’ve not only heard about you, they know quite a bit about you. Until recent months, the company had been a little-known data analysis outfit based in London. But based on a number of reports, including a very informative recent piece in Motherboard, the group has been quietly but profoundly rewriting the book on mass persuasion. Supported by a somewhat secretive investment network, and appearing to work only on behalf of conservative causes, the group specializes in big data collection and micro-targeting of voters. Their exact methods and clients are not always subject to confirmation, but based on media reports, the company is believed to have worked for the successful “Brexit,” campaign for Britain to leave the European Union, and definitely played a key role in Donald Trump’s successful presidential campaign.
Of course, campaigns and other persuaders have always used whatever is available. But the existence of “big data” — shorthand for systems to collect and apply information about all our transactions and communications — is a unique change to the landscape. And it’s only a matter of time before these new tools start to change the playing field for legal persuasion. There is already at least one company (Jury Mapping) that specializes in litigation uses of big data. While these early efforts might be, as I said in a past post, more of a shotgun than a scalpel, we can expect lawyers and consultants to explore those uses. In the wake of the recent election, however, I think that the role of big data and social media profiling carries some broad lessons for legal persuaders. In this post, I’ll take a look at four observations that have parallels in legal persuasion.
One, Social Media is History’s Most Target-Rich Environment
It has always been possible to investigate people, but now that many have chosen to advertise their “likes” and affiliations to the public, it is definitely a lot easier. And those likes can tell a lot. One psychological researcher profiled in the Motherboard piece (Michal Kosinski), for example, proved that he could accurately learn personal characteristics — skin color, sexual orientation, political leaning — with about 90 percent accuracy just by knowing 68 of a given individual’s Facebook likes. With 70 likes, he knew more than that person’s friends know, 150 would tell more than the person’s parents know, and 300 likes would tell him more than the person’s husband, wife, or partner know. It isn’t just a matter of social media to collect, it is also a matter of using it to test. Cambridge Analytica, for example, claims that on the day of the final presidential debate, their team tested 175,000 different variations of online pro-Trump advertisements that day to see which messages were resonating with what slices of the population at that point in the campaign.
Legal persuaders aren’t (and probably shouldn’t be) at that level yet. But social media searching has, just within the last year or so, become a very routine step in understanding a jury panel as well as a particular juror.
Two, Psychometrics is Better than Demographics
The term is less familiar, but “psychographics” refers to the measurement of psychological traits in comparison to an overall population. The main focus is what is called the “big five” that defines an individual’s personality: openness, conscientiousness, extroversion, agreeableness, and neuroticism. Of course, those are all factors that are well defined and have been measured extensively. The problem is that for a long time, one could only obtain the individual measurements by asking people to fill out a rather detailed and personal questionnaires. What changed that, of course, is Facebook. Now, people will voluntarily fill out a quiz labeled something like, “Find Out Which ‘Game of Thrones’ Character You Are,” just so they can share the results with their friends and invite them to take the quiz as well. And hiding in most, if not all, of those quizzes are questions that yield the big five personality traits. And given that the tests also ask participants for permission to see the user’s likes and friend lists, the sponsor of the test — reportedly, it is often none other than Cambridge Analytica — is learning quite a bit about you that has nothing to do with Game of Thrones characters.
Of course that sounds insidious, and it might well be, but the lesson that it carries is one that applies to all persuaders: psychological factors matter more than demographics. Cambridge Analytica’s CEO Alexander Nix said in a presentation last September, that previous campaigning has been based on, “A really ridiculous idea. The idea that all women should receive the same message because of their gender—or all African Americans because of their race.” I would argue that litigators trying to select or appeal to juries based on the same demographics are just as ridiculous.
Three, Micro-targeting is Better than Macro-targeting
It shouldn’t be a surprise that companies like Cambridge Analytica are collecting and applying data on people. But what might be surprising is just how much data. According to Mr. Nix, “We have profiled the personality of every adult in the United States of America—220 million people.” Let that sink in a bit: everyone. And the level of detail on each person is reported to be pretty staggering as well — an average of 5000 data points per person. Cambridge Analytica has done that by using data brokers, using Facebook, and using the sort of quizzes described above. As Mr. Nix demonstrated in a presentation in September using data points on a map, we can look at just Iowa, then narrow it to just voters, then further to just Republicans, then just gun owners, and finally just those who are strongly motivated by fear. With that laser targeting, the right message can be delivered to the social media feeds of just the right audience. Of course that is the opposite of the blanket shotgun approach that is used in modern advertising. As Mr. Nix explained in the same presentation, “My children will certainly never, ever understand this concept of mass communication.”
This focus on micro-targeting rather than macro-targeting applies more generally to persuasion as well. Litigators aren’t trying to persuade an abstract “reasonable person,” they are trying to persuade a particular and knowable judge, or a jury — and not just a “jury” but those particular members of the jury who are likely to be the higher-risk jurors or the swing jurors.
Four, Both the ‘Push’ and the ‘Pull’ Matter
We often simplify persuasion to just the act of giving people reasons to support the position we advocate. But persuasion is at least two sided: There are “pull factors” that encourage people toward our views, and “push factors” that discourage people away from an adversary’s views. These are often called “alpha” and “omega” strategies. These twin factors of push and pull are on full display during a campaign, with positive advertisements aimed at attracting voters to your candidates and negative advertisements aimed at repelling voters from your opposing candidate. The micro-targeting enabled by psychographics and social media raised this to another level. For example, the past campaign featured so-called “dark posts,” that appeared only for people with unique profiles. For example, video advertisements were aimed at African Americans featuring clips of Hillary Clinton referring to some black men as “predators,” and residents of Miami’s Little Haiti district received news stories about the failure of the Clinton Foundation following the earthquake in Haiti. These were not meant to sway anyone to Trump, they were meant to cause people who would otherwise be Clinton voters to stay home.
While the vote-suppression angle might sound unsavory, the broader point of addressing both the factors that persuasively play up your side, as well as the factors that persuasively play down their side, matter in trial advocacy as well. Most often, it translates to the advice to give jurors both, but balance your positive case with your negative case, and generally put the former first in order to build credibility. In other words, make sure your own house is in order before you start trashing theirs.
As I have written recently, this past year might be remembered as the one where statisticians got it wrong, but those were the statisticians trying to predict the results. The statisticians using psychographics and micro-targeting to influence the results might have turned out to have been much more effective.
(Thanks to Kathy Kellerman for suggesting the article)
______
Other Posts on Data Analytics:
______
Image credit: 123rf.com, used under license