We know they used an app to analyse users' personality traits, then correlated these traits with, on one hand, patterns of liked pages, and on the other hand, behavioural predictions, and specifically predictions of susceptibility to different ad campaigns.<p>All of this has been discussed and criticised from different points of view; what I seem to have missed is discussion on whether this is just a fraud or it actually works. Specifically:<p>* Does an app allow to determine the personality traits of its users? Won't users take it too lightly, being it an app rather than a white coat wearing serious professional asking questions?<p>* Do tests for personality traits actually make usable and accurate predictions on people's behaviour?<p>* Do correlations between like patterns and personality test results hold? In other words, could you really predict the result of the personality test of people at large having their like patterns?<p>* Is there evidence that people with different personality traits reacts differently to stimuli ad campaigns?<p>* Did people's voting behaviour actually change? How strong was the effect?<p>All these questions may be framed differently, that is: did Cambridge Analytica actually try to disprove the null hypothesis? How did they do it, or how could it be conceivably done? Did they succeed, or did the null hypothesis win?<p>N.B. I'm not asking: do personality tests work in general? I'm asking: did this stunt actually work? How would you tell a genuine CA style proposal from a "let's win the elections using tarots, talismans and acupuncture?
Here is the basic science on personality traits:<p><a href="https://en.wikipedia.org/wiki/Big_Five_personality_traits" rel="nofollow">https://en.wikipedia.org/wiki/Big_Five_personality_traits</a><p>Now if by "did it work?" you mean "did it change the result of the election?" I think that some people are going to believe that it did no matter what because the election was as close as it was. That is, to change the course of that election you would only need to change maybe 50,000 votes in the right places. This election therefore is a "ground zero" for counterfactuals -- one can easily blame Comey's last-minute announcement or many other factors and people probably will.<p>Possibly the people at CA will deny that it was effective if they are facing trouble in court, but I am sure they would tell you it was very effective if they were selling their services to you!<p>I think it is also consistent with the way Bannon thinks which is certainly a few steps ahead of most people on the left and the right in terms of psychology. For instance, the "new right" and the "politically correct left" are always at arms with each other, but Bannon believes that really they have the same psychology. He realized early on that BELL HOOKS was on to something: if a black women can put together a custom marginalized identity, somebody else can feel like they are "down" because they are a conservative on a liberal campus, got bullied as a kid, misunderstood as an anime fan, has alopecia, etc.<p>Bannon does realize some things, such as the fact that white and black people in the U.S. could possibly join up in opposition to immigration (look at the way race is pictured in the movie Falling Down) but what he doesn't realize is that it takes one stupid racist comment and there is a black person who will never vote for you for as long as they will live.