Indiana University Network Science Institute

Abstract: The COVID-19 pandemic reduced natural opportunities for spontaneous meetings due to government regulations that enforced social distancing and people's decision to avoid getting coronavirus from their close contacts.  Given that the main channel for COVID-19 spread is in-person contacts, it is crucial to figure out who is likely to get COVID and how to contact them through remote or in-person channels.  Given this background, this project examines the impacts of these massive social changes on with whom and how people discuss their "important matters" and how they perceive others' COVID infection status in their core discussion networks. To answer these questions, we conduct a nationwide online daily survey through the Lucid Marketplace from April 10, 2020, to April 5, 2021, across 51 states with 100 daily completes (total N=36,300).  First, we compare the COVID-19 networks to those previously collected in eight national core discussion network surveys conducted from 1985 to 2016.  We observed remarkable network stability during COVID-19; American core discussion networks remained intact, both with respect to size and relationship composition.  In contrast to the robust nature of core networks, racial homophily among kin ties and political homophily among non-kin ties sharply increased.  We further show that Americans adaptively switched their mode of communication in response to growing concerns over COVID-19 and that the use of remote channels was associated with increases in political homophily. Overall, these results suggest that the COVID-19 pandemic may bring Americans remotely together but only with like-minded, deepening social divides and increasing health inequalities in American society.

Bio: I am an Assistant Professor of Sociology at Indiana University. I received my Ph.D. from the Department of Sociology at Columbia University with the Robert Merton Award for Best Dissertation. I have been using causal inference, network analysis, and machine learning methods to study social divisions, political polarization, and social determinants of health. My current projects examine: 1) the impact of COVID-19 on Americans’ social ties using large-scale egocentric network surveys, 2) social determinants (demographic homogeneity, religion, state policy) of suicide and opioid overdose using medical claims data and administrative data, 3) political polarization, hate speech and cross-ideological interactions using large-scale Facebook forum data, 4) the causes and consequences of cultural changes using survey experiments, 5) the evolution of terrorist networks using agent-based simulations, among others.