Cascade-based Randomization for Inferring Causal Effects under Diffusion Interference
This work addresses causal inference challenges for researchers and practitioners in network data analysis, offering an incremental improvement over existing cluster-based methods by specifically handling cascade interference.
The paper tackles the problem of biased causal effect estimation in network experiments due to interference that propagates in cascades, and it proposes a cascade-based randomization method that reduces estimation error by leveraging knowledge of cascade seed nodes, outperforming state-of-the-art approaches in experiments on real-world and synthetic datasets.
The presence of interference, where the outcome of an individual may depend on the treatment assignment and behavior of neighboring nodes, can lead to biased causal effect estimation. Current approaches to network experiment design focus on limiting interference through cluster-based randomization, in which clusters are identified using graph clustering, and cluster randomization dictates the node assignment to treatment and control. However, cluster-based randomization approaches perform poorly when interference propagates in cascades, whereby the response of individuals to treatment propagates to their multi-hop neighbors. When we have knowledge of the cascade seed nodes, we can leverage this interference structure to mitigate the resulting causal effect estimation bias. With this goal, we propose a cascade-based network experiment design that initiates treatment assignment from the cascade seed node and propagates the assignment to their multi-hop neighbors to limit interference during cascade growth and thereby reduce the overall causal effect estimation error. Our extensive experiments on real-world and synthetic datasets demonstrate that our proposed framework outperforms the existing state-of-the-art approaches in estimating causal effects in network data.