Cambridge Analytica and the science of personality

If you tune into the news at the moment, you’re likely to see a familiar name – Facebook – and a less familiar one – Cambridge Analytica. The story surrounding these two companies could be one of the biggest tech scandals ever, raising interesting questions about our personal data and personalities.

What has happened so far? Cambridge Analytica is a company that specialises in analysing huge amounts of consumer data and combines it with behavioural science to identify people who organisations can target with marketing material. According to its founder, Alexander Nix, it was set up “to address the vacuum in the US Republican political market” that became evident after Mitt Romney lost the 2012 US presidential election – Nix had noticed that the Democrats were leading the tech revolution, and so saw an opportunity.

The controversy stems from an allegation that Cambridge Analytica harvested the data of around 50 million Facebook users. To put it simply, some Cambridge academics designed a quiz that could use your public Facebook ‘Likes’ and other uploaded information to predict your personality (using the OCEAN scale, or the ‘Big Five’ personality traits). Cambridge Analytica, seeing potential in this quiz, started a subsidiary which promoted a new version of this quiz designed to harvest data. This was heavily promoted to US users through advertising and paying users to take it.

The controversy stems from an allegation that Cambridge Analytica harvested the data of around 50 million Facebook users

It is believed that around 270,000 people took a version of the quiz; however, the issue is that, in giving Facebook access to their accounts, players also granted access to the public profile information of all their friends. Was this against the rules? Technically not. At the time, Facebook’s API (a set of tools that developers can use to build Facebook features into their websites or apps) allowed this information to be harvested. Facebook apps also told users what information they could access as they requested authorisation – it was feared (correctly) that people didn’t properly read the conditions.

Whether or not these acts technically fell within the rules is not the issue, though. The acts violate users’ reasonable expectations of privacy and responsible data handling, and it could fall foul of EU data rules. Although it fell within the rules for companies to gather this data, it was not authorised to be shared with others. For its part, Cambridge Analytica claims that it never used the data, and deleted it when Facebook asked it to (former employee Chris Wylie claims it was not). Facebook has also come under fire, such that the Guardian reports that they knew about this breach for two years and did nothing to protect its users.

The acts violate users’ reasonable expectations of privacy and responsible data handling, and it could fall foul of EU data rules

Did this information affect the Brexit vote and the US election, as has frequently been claimed? The answer is probably not. The company profiles people, but that profiling does not really lend to changing behaviour on a mass scale (add in the fact that every candidate used some form of profiling and micro-targeting to persuade voters, and the idea that one company had such a massive sway becomes markedly less probable).

How does this profiling actually work? Cambridge Analytica essentially analyses which Facebook ‘Likes’ correlate with personality traits, or other ‘Likes’. Crudely, imagine you were a Democrat wanting to reach women. If the analysis showed that women who liked ‘The Democrat Party’ also liked things ‘A’, ‘B’ and ‘C’, you could target your ads at those things and be more likely to find people who are receptive to your message, but who may not necessarily already back you.

The company profiles people, but that profiling does not really lend to changing behaviour on a mass scale

However, the company is regarded by industry experts as “snake oil salesmen”, and the validity of the science of personality profiling is debated. Scientists have spent lots of time analysing personality tests and debunking them; although they are used in order to help employers choose candidates who will fit into a particular workplace, critics argue that these tests (the Myers-Briggs Type Indicator is probably the most famous) are inherently flawed. People, they say, are too complex to be broken down into a small number of personality ‘types’, and these tests are often designed such that the ‘right’ answers are obvious – people can rarely apply the objective eye required to accurately analyse their own personality. It doesn’t help that there is no objective standard of personality and research often produces vague and conflicting results.

Perhaps the most interesting theory comes from psychologist Walter Mischel. Mischel was employed by the Peace Corps as a personality consultant, and after a large research programme, he came up with the conclusion that personality doesn’t really exist – it’s entirely dependent on context (hence we are not the same person in every situation).

Personality doesn’t really exist – it’s entirely dependent on context

A warrant has now been sought to look at the databases and servers used by Cambridge Analytica, and with Mark Zuckerberg now facing calls to appear before lawmakers on both sides of the Atlantic, this story is only going to grow. More questions about online security and how our data is used will be asked – the answers to those questions could have huge ramifications for social media and its users.