Estimating Social Effects with Randomized and Observational Network Data

Birds of a feather flock together. That is, people’s behaviors often mirror those of their friends. But how can we tell if friends’ behaviors are shaped by mutual influence or whether people are simply drawn together because of pre-existing similarities—or perhaps both?

In the article “Estimating Social Effects with Randomized and Observational Network Data,” Principal Julian Chan and his co-authors introduce an innovative approach to estimate these two competing effects. They applied the method to a data set of student interactions they collected in Hong Kong, where some schools used a lottery for seating assignments, thus offering a natural experiment to explore interactions.

Read the article here.