The psychological study that gave Trump the victory

"With 70 likes, Kosinski knows more about an individual than their closest friends. With the analysis of 150 likes, he discovers more than their family members."

Donald Trump, when running for president, had a campaign strategy that was shocking to purists: he spent almost nothing on television advertising and snubbed opinion poll experts. He made it seem as though his high-impact statements were the result of his deep understanding of what he calls the “forgotten citizens.”

The German magazine Das Magazin, however, seems to have found the key to Trump's political marketing: a British company called Cambridge Analytica, specialized in harvesting and analyzing personal data on the internet and turning it into raw material for online micro-targeting—that is, individualized political advertising. Did anyone think of the dystopia created by George Orwell in the book 1984? It is that and much more.

It all started with a Polish psychology student at Cambridge University in England. During his PhD in psychometrics (the scientific attempt to measure a person's personality), Michal Kosinski and a colleague created a Facebook app that encouraged users to answer a psychological questionnaire in exchange for a personality “profile.” Millions of people responded, and based on this material, Kosinski developed a method that allows individual characteristics to be discovered simply by analyzing Facebook likes. With 70 likes, Kosinski knows more about an individual than their closest friends. With the analysis of 150 likes, he discovers more than their family members.

In 2014, Aleksandr Kogan, an assistant professor at Cambridge, approached Kosinski and proposed his participation in a highly paid project to analyze the profiles of 10 million Americans on Facebook. The company making the offer specialized in elections. Kosinski refused the offer and reported his colleague to the university administration. For him, using his method in elections was a form of manipulation. Aleksandr moved to Singapore. Kosinski went to teach at Stanford University in California.

Then, in November 2015, Kosinski heard about a company that was providing data analysis services for the campaign for the UK to leave the European Union. It was the same company that had approached him a year earlier: Cambridge Analytica. As he feared, his method was being used to manipulate voters—and, as became clear in June 2016 with the victory of the Brexit vote, successfully.

In September 2016, Cambridge Analytica gained a new spotlight at a scientific conference in New York. One of the speakers, Alexander Nix, CEO of Cambridge Analytica, explained on that occasion how they were cross-referencing personal data of American citizens to use in Donald Trump's presidential pre-campaign. And he made a shocking claim: by cross-referencing data purchased from different sources with Facebook like histories, his company managed to compile a complete psychographic profile of every single American adult—no fewer than 220 million people. His team was able to define the consumer tastes, political opinions, habits, and cultural traits of every voting-age citizen in the country.

An example of how this data was used throughout the campaign: on the day of the third presidential debate between Trump and his opponent Hillary Clinton, the Republican's digital team sent out 175,000 variations of online advertisements, each targeted to specific groups of citizens. The differences between these ads were subtle, but designed to hit their targets dead-on, aiming to appeal to them based on their greatest fears, hopes, and preferences.

Cambridge Analytica, according to what its CEO told the authors of the Das Magazin report, received 15 million dollars from Trump's committee for the service.

Is this political use of social media data ethically condemnable, or is it no different from the traditional dilemmas of political marketing? This is a debate that is just beginning.