Building Causal Graphs for Applied Research

Casual graphs, particularly causal directed acyclic graphs (DAGs), are a fantastic tool to help researchers assess whether the required assumptions for estimating the strength of causal effects are reasonable in their analysis and data, and in helping design sensitivity or quantitative bias analyses, or estimating bounds, when those assumptions are not entirely reasonable. Despite the excellent and extensive literature on the mathematical, statistical, and philosophical underpinnings for causal DAGs, there is much less literature focused on practical guidance for creating causal DAGs from expert knowledge, evidence synthesis, data, or a combination of these.

Our Causal DAG working group aims to fill this gap by providing tools, guidance, training, and examples to help applied researchers use causal graphs in their work.

For a list of causal DAG-building resources, visit: DAG resources