A0396
Title: Average and conditional inward and outward spillovers of one unit's treatment under network interference
Authors: Fei Fang - Yale University (United States) [presenting]
Laura Forastiere - Yale University (United States)
Edoardo Airoldi - Fox School of Business, Temple University (United States)
Abstract: In a connected social network, users may have varying levels of influence on others when they receive interventions. For example, giving an advertisement to a more influential person can have, on average, a greater impact on others' purchase decisions. Understanding and evaluating these effects can provide valuable insights for various applications, such as targeting strategies in marketing and behavioral interventions in public health. Under a partial interference assumption, influence effects are defined in two ways: i) the inward average spillover effect on a unit's outcome of a neighbors treatment, and ii) the outward average spillover of a unit's treatment on their neighbors outcomes. The comparison is investigated between the two causal effects in directed networks with different properties, including the conditions under which they are equivalent. Additionally, Horvitz-Thompson estimators are developed to assess both effects, on average and conditioning, on categorical covariates, and weighted least square estimators for these effects conditioning on continuous covariates. Design-based variance estimators are derived, and the consistency and asymptotic normality are established. Through simulations, the empirical performance of the proposed estimators is verified. Finally, the approach is employed to investigate the inward and outward average and conditional spillover effects of an information session on the adoption of weather insurance among rice farmers in China.
