Using data we collected over a year of daily time points, our paper tries to contribute 4 things that may be of use to those evaluating how to measure social media and public opinion:
First, by exploring the popularity of the Yes and No campaigns’ Facebook and Twitter accounts, we examine the validity of power-law theory (resourceful agents follow and participate early, while the bulk of people arrive late to the game) in predicting online behaviour.
Second, by using longitudinal cluster analysis we also attempt to discern between the two campaigns to see if one side was more effective than the other in how it garnered support.
Third, we examine the extent to which key events impacted upon online behaviours across the two campaigns.
Fourth, we explore the relationship between online behaviour and public opinion to see if online behaviour by unrepresentative publics is a precursor to offline behaviour by representative publics.
