Policy Evaluation for Temporal and/or Spatial Dependent Experiments
This addresses the challenge of policy evaluation for technology companies conducting complex experiments with dependencies, though it appears incremental in extending existing causal inference methods to temporal/spatial contexts.
The paper tackles the problem of establishing causal links between policies and outcomes in experiments with temporal and/or spatial dependencies by proposing a novel Varying Coefficient Decision Process (VCDP) model. It decomposes the Average Treatment Effect into Direct and Indirect Effects, provides estimation and inference procedures, and validates the approach through simulations and real data analyses.
The aim of this paper is to establish a causal link between the policies implemented by technology companies and the outcomes they yield within intricate temporal and/or spatial dependent experiments. We propose a novel temporal/spatio-temporal Varying Coefficient Decision Process (VCDP) model, capable of effectively capturing the evolving treatment effects in situations characterized by temporal and/or spatial dependence. Our methodology encompasses the decomposition of the Average Treatment Effect (ATE) into the Direct Effect (DE) and the Indirect Effect (IE). We subsequently devise comprehensive procedures for estimating and making inferences about both DE and IE. Additionally, we provide a rigorous analysis of the statistical properties of these procedures, such as asymptotic power. To substantiate the effectiveness of our approach, we carry out extensive simulations and real data analyses.