The Graphical Identification for Total Effects by using Surrogate Variables
This work addresses a specific challenge in causal inference for researchers, offering incremental improvements in graphical methods for identifiability.
The paper tackles the problem of identifying total effects in causal graphs when treatment or response variables are unobservable, by providing graphical criteria using surrogate variables to determine identifiability from graph structure.
Consider the case where cause-effect relationships between variables can be described as a directed acyclic graph and the corresponding linear structural equation model. This paper provides graphical identifiability criteria for total effects by using surrogate variables in the case where it is difficult to observe a treatment/response variable. The results enable us to judge from graph structure whether a total effect can be identified through the observation of surrogate variables.