Epstein, Ziv. The dynamics of attention in digital ecosystems. Diss. Massachusetts Institute of Technology, 2019.

Abstract

With more than half of the population of earth now using social media, it has become a critical tool for individuals to gain information about the world around them and to connect with others, and for researchers to build a broad understanding of human behavior. However, myopic design patterns of these platforms, as well as the broader attention economy, have enabled the proliferation of misinformation and undermined collective intelligence. These situating factors motivate a guiding research question for the emerging field of platform design: how does attention operate in digital ecosystems, and how can we design platforms to mitigate misinformation and promote collective intelligence? This dissertation addresses key aspects of this question by arguing that attention operates in a two- stage process (“Try” and “Buy”) on social media. The Try phase (stage 1) involves initial exposure to content, and the Buy phase (stage 2) involves engagement with con- tent conditional on having been exposed to it in the first place. Due to the difficulties of measuring Stage 1 exposure in standard survey experiment settings, most research has focused only on psychological determinants and design interventions for stage 2, neglecting the crucial role attention plays online. To understand the attentional dynamics of stage 2, I will discuss studies on a scalable accuracy prompts that “moves the spotlight of attention” towards peoples’ existing but latent capacity for discerning truth from falsehood. To under- stand the attentional dynamics of stage 1, I will discuss how social influence can impact the content users are exposed on social media environments, via a large scale field experiment with AI-generated hybrid animals (“GANimals”). To directly measure attentional expo- sure to content in addition to engagement, I introduce a new research tool called Yourfeed. Yourfeed is a digital environment that mirrors the attentional context of social media, and tracks both dwell time and engagement for each piece of content. Together, these contribu- tions provide empirical evidence for how attention operates on social media, and highlights a suite of design interventions to promote high-quality interactions with these platforms.