Is it time to turn social media off? After the flurry of stories on Cambridge Analytica, Zuckerberg and Congress, personal data and political advertising, the question is whether Facebook – or social media in general – has any genuinely useful social purpose. The answer, according to psychologists and social scientists, is affirmative.
For Pete Burnap, reader in data science and cyber analytics at Cardiff University, social media – in the guise of Twitter – offers a possible way to predict riots and unrest. Burnap’s team mapped data with hashtags referencing riots at the height of the street unrest in the summer of 2011.
“We developed a new way of clustering tweets using a 30 minute time window,” he explains. “We could link clusters to places, activity, links, pictures - all sorts of stuff. Using open data from the police as to when intel was received and forces deployed and the evidence suggests we could spot a riot developing earlier than the police – as well as outbreaks of violence that are not always reported to the police. If people witnessing and tweeting we can identify clusters.”
Burnap admits that he relied on Twitter for his research as it makes its data public. For private groups and Whatsapp information was hard to gather. Does this mean he’s advocating extended online surveillance? “That’s not only ethically problematic, it’s also counterproductive,” he explains. “You can’t target individuals using this – it’s useful for active event monitoring at, say, a football match. The next step is working out what other sources of open information we could use to augment the police as they struggle with cuts taking boots off the ground.”
Amy Slater at the University of the West of England Centre for Appearance Research worked on an Australian study into teenage eating disorders and discovered that as young girls increased their time spent and number of Facebook friends, body image concerns increased significantly. Slater’s conclusion? “The number of Facebook friends each girl had at the outset of the longitudinal study predicted the observed increase in drive for thinness two years later,” she explained. With current eating disorder theory suggesting that internalisation and body surveillance occurs early in a girl’s developmental, it might be possible to use Facebook to identify girls at risk of eating disorders.
Social scientist Sandra Matz, assistant professor of at Columbia Business School, was part of the Cambridge Psychometric Institute team who researched links between personality tests and Facebook likes and inspired Aleksandr Kogan to develop the controversial Cambridge Analytica research. She’s working with a bank in central America trying to help people to save money – “delayed gratification is especially problematic for people with low incomes,” she explains. “When we target Facebook ads by personality we can highlight different saving goals – extroverts would save money for very different things to introverts.”
She’s wary of attacks on social media as a research tool “because of the transient nature of psychological states, identifying such states in real time is harder than identifying psychological traits,” Matz explains. “Psychological states have traditionally been tied to questionnaire measures such as the PANAS scale for positive and negative affect. Using these to measure and act in real time on varying psychological states is impractical. But the bad guys are doing it, so we might as well be in there ourselves.”
What’s increasingly clear is that simple targeting isn’t the solution to anything much, according to Jonathan Sebire, chief strategy officer at ethical data science company Signify. He’s been tracking political influencers online in a bid to discover why figures like Jacob Rees Mogg have become popular standard bearers for the Vote Leave movement.
“Around the time stories began to appear about how the European Court of Justice would retain oversight in the UK during the Brexit transition, we found a network of people who were creating content using Jacob Rees Mogg as a lightning rod around Brexit betrayal,” he explains. “Essentially there were private groups on Facebook or Whatsapp building support and amplifying messages before suddenly they popped up, 100,000 people engaged and the stories started trending, then journalists reported on it and it entered a feedback loop.”
Remain-based or politically centrist material tends to remain in the public sphere – like Twitter – he explains. The extreme right and - to a lesser extent - the extreme-left are better at weaponising content. “If we want to understand how pressure groups and stories explode from influencers to the mass market we need more social media based research, not less. There is going to be a period of turbulence for individuals and social platforms but we think they have potential for understanding society.”
“In the past, social scientists and psychologists wanting to run experiments – like the infamous Milgram experiments on whether someone would knowingly kill a stranger is told to do so by an authority figure - have tended to work with undergraduates,” explains Curtis Jessop, research director at NatCen - Britain's leading independent social research institute. “Why undergraduates? Because they’re the only people they have to hand – but it’s safe to say they don’t really represent the population. With social media, we suddenly have a research tool that offers real potential.”