Junk News Bubbles: Modelling the Rise and Fall of Attention in Online Arenas

Despite the ever growing availability of traces produced by digital media, so far no large-scale empirical research has been devoted to online attention cycles, with few remarkable exceptions. To encourage such research, we propose a toy model inspired by one of the most influential accounts of attention dynamics: the “public arenas model” introduced in 1988 by Stephen Hilgartner and Charles Bosk (H&B) [4]. Despite its clarity and insightfulness, H&B’s framework has never been mathematically formalized because of its complexity and lack of formal description. We streamline H&B’s model focusing on the rise and fall of attention matters and proposing a ready-to-test version that we hope will encourage further investigation with empirical data. The main consideration of H&B work that we included in our formalization can be summarized as follows:
1. a population of “matters of attention”(or “social problems”, as in H&B original formulation) defined self-referentially as the entities that compete to capture public attention in a media arena;
2. a fixed attention capacity (or “carrying capacity”, in H&B terms) by each debate arena, justified by the limited attention of media audiences;
3. a competition mechanism that primes trendiness. Each attention matter see its visibility promoted or demoted according to its previous variation multiplied by a parameter that amplifies or dampens this dynamics.
The simulations suggest that trendiness boost the steepness in the rise-and-fall of attention matters and affects both dimensions of the media cycle: the life time of an attention matter, and the diversity of the public debate. On the one hand, the stronger is trendiness boost, the shorter is the lifecycle of attention matters, allowing a higher number of matters to enter and exit the arena. On the other hand, a higher trendiness boost increases the maximum visibility reached by attention matters and, most importantly, amplifies the difference between successful and unsuccessful attention matters, creating a situation in which most of the available attention is captured by a minority of over-visible items. We claim that debate arenas that reward to trendiness excessively end up displaying a syncopated rhythm of attention that is at the same time increasingly dispersed and increasingly concentrated. We call such phenomenons junk news bubbles: speculative weaves that destroy rather than create richness in public debate.