David Gibson, University of Pennsylvania
Many social processes and outcomes hinge on interaction—episodes of face-to-face or technologically mediated communication, in the course of which information is exchanged, opinions are compared, attitudes and emotional states are diffused, power is exercised, decisions are made, and (in the case of face-to-face [or more intimate] contact) pathogens are spread. The study of social interaction encompasses the study of the conditions under which interaction occurs, the network structure of encounters, the internal workings of an encounter, and the consequences of encounters. On the frontier of the study of social interaction are efforts to bridge long-standing methodological and theoretical divides, data collection and statistical innovations that have pushed forward our ability to infer causation and to model relational interdependencies, and attempts to exploit new data collection technologies.
The session will open with introductory remarks by David Gibson, whose work bridges structural and sequential approaches to the study of interaction. Amine Ouazad will then present research on peer effects on student test scores, exemplifying how natural experiments have revolutionized causal inference, as well as the use of statistical models to contend with the symmetrical nature of social influence. Finally, James Holland Jones will present his research on sexual networks in Malawi, which uses new exponential random graph models to study the emergence of higher-level network features from local partner selection propensities, and simulations to study the consequences of those features for disease transmission.