[Principles, construction methods, and rules of single world intervention graphs]

The potential outcomes framework (POF) and structural causal models are two of the most popular methodological approaches to causal inference. Each framework has its advantages and limitations, and because of the different perspectives on analyzing problems, they have long remained separate. By introducing potential outcome variables into causal graphs, the single world intervention graph (SWIG) innovatively combines the POF with structural causal models, addressing the lack of intuitive causal inference in the POF and the limitations of traditional directed acyclic graphs in expressing potential outcome variables and conditional independence. This article systematically introduces the basic principles, construction of SWIG, graphical rules, and through an introduction to various SWIG research templates, provides useful tools and standardized guidance for researchers conducting causal inference studies.