Opening the Blackbox of Treatment Interference: Tracing Treatment Diffusion through Network Analysis

Abstract

Causal inference under treatment interference is a challenging but important problem. Past studies usually make strong assumptions on the structure of treatment interference in order to estimate causal treatment effects while accounting for the effect of treatment interference. In this article, we view treatment diffusion as a concrete form of treatment interference that is prevalent in social settings and also as an outcome of central interest. Specifically, we analyze data from a smoking prevention intervention conducted with 4,094 students in six middle schools in China. We measure treatment interference by tracing how the distributed intervention brochures are shared by students, which provides information to construct the so-called treatment diffusion networks. Besides providing descriptive analyses, we use exponential random graph models to model the treatment diffusion networks in order to reveal covariates and network processes that significantly correlate with treatment diffusion. We show that the findings provide an empirical basis to evaluate previous assumptions on the structure of treatment interference, are informative for imputing treatment diffusion data that is crucial for making causal inference under treatment interference, and shed light on how to improve designs of future interventions that aim to optimize treatment diffusion.

Causal inference in collaboration networks using propensity score methods

“…A revealing example of how information spreads through networks and thereby affects non-treated subjects comes from a smoking-prevention intervention study. Regarding interference itself as the outcome of interest, the study showed how friendship ties significantly increased the log odds of receiving information from an intervention brochure (An and VanderWeele 2019). Whilst interference is a nuisance in many research settings when it violates the SUTVA, it is here the outcome of interest.…”

Causal Effects with Hidden Treatment Diffusion on Observed or Partially Observed Networks

“…Therefore, some researchers have developed novel methodologies to account for spillovers in a more general setting, where units interfere according to the links observed over a network (general or network interference) both in experimental settings (Aronow and Samii, 2017;Athey et al, 2018;Leung, 2020b;Bargagli Stoffi et al, 2020;Imai et al, 2020) and in observational studies (Ogburn et al, 2017;Sofrygin and van der Laan, 2017;Forastiere et al, 2018Forastiere et al, , 2020bTortù et al, 2020). These estimators tackle the problem of interference without distinguishing between the specific underlying mechanisms (An and VanderWeele, 2019). In the presence of treatment diffusion, these methods would estimate the causal effects of the initial treatment assignments, including the spillover effects from the treatment assigned to other units.…”

“…Very few works explicitly deal with treatment diffusion An and VanderWeele, 2019). However, the existing works on treatment diffusion assume to have perfect knowledge on both the baseline social network describing relations among units and the actual treatment diffusion network describing who passes the treatment to whom over time.…”

“…A wide variety of policy evaluation studies in the field of social sciences can be affected by the presence of a treatment diffusion process An and VanderWeele, 2019). For instance, the empirical setting proposed by Lattarulo et al (2017) studies the effect of different types of school incentives aimed at encouraging students to attend cultural events and the intervention of interest is a promotional video, which can be tangibly spread among scholars through virtual or real social connections.…”

“…One of the goals of policy evaluation studies is to assess the causal effect of an intervention (Rubin, 1974;Imbens and Rubin, 2015). However, in a wide variety of real world scenarios, the treatment of interest can be diffused among units An and VanderWeele, 2019). This phenomenon can occur when individuals interact with one another (Liu et al, 2017) and the intervention under evaluation is the distribution of an object or product transferable by nature (e.g., flyers, booklets, videos).…”