The next session at Web Science 2016 is on information dissemination and engagement. It begins with a paper by Miriam Fernandez, whose focus is on promoting behavioural changes to combat climate change. Over the past years, there have been multiple social media campaigns that promote more environmentally responsible behaviours; what can these campaigns learn from theories of behaviour change, and how can these theories be translated into computational methods?
The focus here is on Earth Hour 15, and on the COP21 summit in Paris in 2015. The project draws on the 5 Doors Theory, which describes five stages of behavioural change: awareness of desirability, provision of an enabling context, sense of can do, generation of buzz, and invitation to act. The assumption here is that users at different stages of behaviour will engage differently, and that they can be assigned to these stages through an automated computational analysis of the posts they contribute to social media platforms.
Such automated analysis was performed using the GATE Natural Language Processing (NLP) framework. The project identified some 17-20,000 users discussing the two events on Twitter, and gathered the past 3,200 posts for each of them; these were filtered for climate change-related discussion, and any relevant posts were then processed using NLP.
Recommendations emerging from this are that most users are at the desirability stage, which means that campaigns should focus more on providing messages with concrete suggestions on how people should change their behaviours. Few users are at the invitation stage, so there is a need to identify community leaders and involve them more closely. Efforts should also be dedicated to engaging in discussions and providing more direct feedback, in order to move users to the more advanced stages.
