To read the story purchase the article from SAGE website or check a preprint version. You can also check this TEDx talk in French, where Casilli is presenting the study. You even download the computer simulation (technical description of the model, Netlogo and Java applet versions).
“Social Media Censorship in Times of Political Unrest – A Social Simulation Experiment with the UK Riots” by Antonio A. Casilli and Paola Tubaro, published July 02 2012 in Bulletin of Sociological Methodology
London, (July 02, 2012). Is social media censorship a means to quell a modern uprising? Some politicians and law enforcers during the political turbulence of 2011 thought so but recent research suggests that uncensored citizens experience less violence and longer periods of peace between outbursts than communities subject to censorship. These new findings appear in the Bulletin of Sociological Methodology, published by SAGE.
A consensus is forming around Internet censorship in the wake of last year’s uprisings, extending from the Arab Spring to the UK, according to Antonio Casilli, associate professor in digital humanities at Telecom ParisTech, France and Paola Tubaro, senior lecturer in economic sociology at the University of Greenwich, UK. The authors used sophisticated computer modeling to find out if the assumptions that actors’ use of media – such as Twitter – fueled mob action through greater awareness were true. Ambiguously, current narratives among the European political establishment suggest social media can be either the tools of liberation (in developing countries) or threats to values of peace and freedom (in Western countries).
The researchers used state-of-the-art agent-based modelling as a starting point. Political conflict is often described as cumulative, involving ‘escalating’ conflict and sometimes ending with regime change. However, in reality, periods of relative stability punctuated with violent outbursts are more typical. Existing models include a variable called ‘vision,’ an individual agent’s ability to scan his/her neighbourhood for signs of police officers and/or active protesters. Higher vision means greater awareness of one’s surroundings and a larger range of possible actions.
In Casilli and Tubaro’s computer simulation, censorship narrows down vision. It interrupts the flow of communication and decreases the ability of individuals to appreciate their environment. In this sense, censorship blinds social actors to their own context.
The researchers found that all possible scenarios led to initial outbursts of violence but how the situation evolved was significantly influenced by government social media censorship. In a total censorship scenario, similar to the Egyptian riots, violence levels remained at a maximum. Stronger censorship led to an increase in the average level of endemic violence over time.
According to the model, the “no censorship” situation at first appears bleak, with incessant, high-level violent outbursts that seem larger than in other scenarios. However, looking at average violence levels over time, the uncensored scenario still has the least aggression. Although agents protest, sometimes violently, they are able to return to relative calm for longer periods in-between. The decision to maintain peace is the choice of agents themselves, rather than due to police repression.
This research offers an interesting methodological bridge that shows how rules operating at the micro or individual level can account for collective dynamics. This is particularly interesting at a time when research is trending into two camps, either using micro-motives (such as personality, culture, and morals) or macro-indicators (such as poverty and social stratification) as explanatory factors.
“In the absence of robust indicators as to the rebelliousness of a given society, the choice of not restricting social communication turns out to be a judicious one for avoiding the surrender of democratic values and freedom of expression for an illusory sense of security,” say Casilli and Tubaro.