Session T56: Physics of Human Social Systems

Science on the Web: How networks bias academic communication online

Most academics are promoting their work online. At the same time, the public, journalists, and interested governments increasingly turn to the Web for scientific information. It thus becomes ever more critical that we better understand the dynamics of online science dissemination networks. My talk presents our latest results about (1) how scientific publications spread on various types of online platforms, losing essential information; (2) how gender and ethnic inequalities impact the coverage of scholarship; and (3) how subsequently retracted articles receive more attention online. Our findings highlight crucial biases in the online sharing of science. They inform efforts to close gaps in scholars' success and curb the online spread of science-related misinformation.   

What's the point of this conference? Modeling the physics of team formation at meetings

How do scientific teams form? While this question has traditionally been the subject of qualitative study, new data sets have allowed us to develop and test quantitative models. Motivated by the process of chemical catalysis, we present a differential equation based model for the generation of new collaborations at conferences, quantify its performance, and compare it with other candidate models. We further examine the impact of conference modality---in-person or virtual---on model performance and on more general properties of scientific social networks.   

Forecasting U.S. elections using compartmental models of infection

Election dynamics are a rich complex system, and forecasting U.S. elections is a high-stakes problem with many sources of subjectivity and uncertainty. In this talk, we take a dynamical-systems perspective on election forecasting, with the goal of helping to shed light on the forecast process and raising questions for future work. By adapting a Susceptible-Infected-Susceptible model to account for interactions between voters in different states, we show how to combine a compartmental-modeling approach with polling data to produce forecasts of senatorial, gubernatorial, and presidential elections at the state level. Our results for the last two decades of U.S. elections are largely in agreement with those of popular analysts. We use our modeling framework to determine how weighting polling data by polling organization affects our forecasts, and we explore how our forecast accuracy changes in time in the months leading up to each election.