This award funds research in economic theory. The project seeks to improve our understanding of how social networks shape the spread of opinions, products, and ideas. The core objective is to turn insights from theoretical models into practical guidance on how to conduct targeted seeding, how to design regulations for social media, and how to advertise new products with uncertain quality. The award funds three projects. The first project will develop a framework to study how best to target individuals based on their network positions in order to spread a new idea or innovation. The second project will model the distinctive features of social media markets in an effort to identify inefficiencies and propose potential regulatory solutions. The third project will study the interplay between word-of-mouth advertising and how consumers learn about product quality. Managing the spread of information and misinformation has become a core concern in recent years, and results from these projects will help regulators and business leaders make more effective decisions in this domain. Methods developed in the process may also prove useful in other applications in which individual choices affect a contagion process?for instance, modeling the spread of disease while accounting for changes in human behavior that occur as it spreads. The research therefore has the potential to advance our nation?s capacity to meet current and future challenges.
These projects contribute to a broader agenda on how to leverage social spillovers to spread innovations and how to manage these spillovers to mitigate the spread of misinformation. The first project will lead to a new approach to optimal targeting in order to disseminate new ideas and innovations. The second will provide a framework to evaluate proposed reg- ulations on social media. The third will help us understand how word-of-mouth marketing conveys information about product quality. While several recent papers demonstrate the efficacy of many network-based interventions, this agenda faces challenges because network data are typically noisy and expensive to gather, and determining optimal policies is computationally hard. Building on methodology that the PI introduced in a previous paper, these projects leverage new results on random graphs to obtain a computationally tractable model that makes predictions based on distributional information about the network structure.
