Cambrige Analytica may have behaved inappropriately, even illegally, but don't get your hopes up that it could all have been very different.
The Trump election win was more or less decided by around 80,000 votes in three states and the UK's EU referendum by a 52%/48% split, so it's hardly surprising the losing sides want to hear any explanation that it was swung by anything other than their own arguments. Isn't that always the way? I still believe my dad voted for Margaret Thatcher because she sat on a tank.
Republican and Democratic data science analysts, who deal with mined information all day every day, say the type of information provided by Cambridge Analytica wasn't effective and was no more valuable than the basic data the Trump campaign could have gotten elsewhere.
Antonio García Martínez is a former Facebook employee who worked on Facebook's ad-targeting setup. “What they’re doing is bullshit, basically,” he said. “This has been tried before. There’s no reason to think it’s particularly powerful."
One Republican data scientist suggested, "Whether or not someone likes Taylor Swift or did a keg stand in college is far less useful to me than whether or not they voted in the Republican or Democratic primary in previous years. And until someone proves to me that OCEAN (openness, conscientiousness, extraversion, agreeableness and neuroticism) score targeting truly can be used to behaviorally persuade a voter, I’ll remain skeptical of the value."
Trump's campaign gave $5.9 million for Cambridge Analytica's services. As a comparison for contextual purposes, the data shows that Ted Cruz's presidential campaign spent $5.8 million on the exact same service.
Jessica Baldwin-Philippi is an assistant professor at Fordham University and studies digital political campaigns.
"The idea that they had a robust data campaign is just not true," she says. "They were doing very best practice stuff on the Facebook side, but they weren't doing this universally. It wasn't deep or robust.
"Time and again, studies have shown us that the most persuasive targeting metrics are not crazy-specific microtargeting data but public-record voting history data and geography."
In fact, if we microanalyse this information, it would seem we should be celebrating that a bunch of scammers milked Trump of 1% of his election campaign funds for a couple of tins of tartan paint.
So they targeted ads using data to appeal to people's fears. And yet we consistently see headlines about what "might" happen or what "could" happen after Brexit, in supposedly responsible newspapers, which play on fears.
Daniel Kahneman won a Nobel prize for economics and in his book, "Thinking Fast and Slow." he said:
"Tetlock interviewed 284 people who made their living 'commenting or offering advice on political and economic trends.' He asked them to assess the probablilities that certain events would occur in the not-too-distant future, both in areas of the world in which they specialized and in regions about which they had less knowledge. Would Gorbachev be ousted in a coup? Would the United States go to war in the Persian Gulf? Which country would be the next big emerging market? In all, Tetlock gathered more than 80,000 predictions....Respondents were asked to rate the probablilities of three alternative outcomes in every case: the persistence of the status quo. more of something such as political freedom or economic growth, or less of that thing.
"The results were devastating. The experts performed worse than they would have had they simply assigned equal probabilities to each of the three potential outcomes. In other words, people who spend their time, and earn their living, studying a particular topic produce poorer results than dart-throwing monkeys, who would have distributed their choices evenly over the options."
Predictions of what might happen belong in the Fortean Times, not the Sunday Times.
He also said research showed that people who think they're not easily swayed and are too clever for that are still more likely to believe something simply if it's in bold. Not you all, obviously, because you're too clever.
The Trump election win was more or less decided by around 80,000 votes in three states and the UK's EU referendum by a 52%/48% split, so it's hardly surprising the losing sides want to hear any explanation that it was swung by anything other than their own arguments. Isn't that always the way? I still believe my dad voted for Margaret Thatcher because she sat on a tank.
Republican and Democratic data science analysts, who deal with mined information all day every day, say the type of information provided by Cambridge Analytica wasn't effective and was no more valuable than the basic data the Trump campaign could have gotten elsewhere.
Antonio García Martínez is a former Facebook employee who worked on Facebook's ad-targeting setup. “What they’re doing is bullshit, basically,” he said. “This has been tried before. There’s no reason to think it’s particularly powerful."
One Republican data scientist suggested, "Whether or not someone likes Taylor Swift or did a keg stand in college is far less useful to me than whether or not they voted in the Republican or Democratic primary in previous years. And until someone proves to me that OCEAN (openness, conscientiousness, extraversion, agreeableness and neuroticism) score targeting truly can be used to behaviorally persuade a voter, I’ll remain skeptical of the value."
Trump's campaign gave $5.9 million for Cambridge Analytica's services. As a comparison for contextual purposes, the data shows that Ted Cruz's presidential campaign spent $5.8 million on the exact same service.
Jessica Baldwin-Philippi is an assistant professor at Fordham University and studies digital political campaigns.
"The idea that they had a robust data campaign is just not true," she says. "They were doing very best practice stuff on the Facebook side, but they weren't doing this universally. It wasn't deep or robust.
"Time and again, studies have shown us that the most persuasive targeting metrics are not crazy-specific microtargeting data but public-record voting history data and geography."
In fact, if we microanalyse this information, it would seem we should be celebrating that a bunch of scammers milked Trump of 1% of his election campaign funds for a couple of tins of tartan paint.
So they targeted ads using data to appeal to people's fears. And yet we consistently see headlines about what "might" happen or what "could" happen after Brexit, in supposedly responsible newspapers, which play on fears.
Daniel Kahneman won a Nobel prize for economics and in his book, "Thinking Fast and Slow." he said:
"Tetlock interviewed 284 people who made their living 'commenting or offering advice on political and economic trends.' He asked them to assess the probablilities that certain events would occur in the not-too-distant future, both in areas of the world in which they specialized and in regions about which they had less knowledge. Would Gorbachev be ousted in a coup? Would the United States go to war in the Persian Gulf? Which country would be the next big emerging market? In all, Tetlock gathered more than 80,000 predictions....Respondents were asked to rate the probablilities of three alternative outcomes in every case: the persistence of the status quo. more of something such as political freedom or economic growth, or less of that thing.
"The results were devastating. The experts performed worse than they would have had they simply assigned equal probabilities to each of the three potential outcomes. In other words, people who spend their time, and earn their living, studying a particular topic produce poorer results than dart-throwing monkeys, who would have distributed their choices evenly over the options."
Predictions of what might happen belong in the Fortean Times, not the Sunday Times.
He also said research showed that people who think they're not easily swayed and are too clever for that are still more likely to believe something simply if it's in bold. Not you all, obviously, because you're too clever.
