Dagitty

Dagitty is a free, open-source web-based tool for drawing and analyzing causal directed acyclic graphs (DAGs). It is widely used in epidemiology, social sciences, and economics to support causal inference from observational data.

Users construct graphs by defining variables as nodes and causal relationships as directed edges. Dagitty implements formal criteria from causal inference, notably Pearl's back-door criterion and d-separation, to help researchers identify covariate adjustment sets that block confounding pathways between an exposure and an outcome. It can determine whether the causal effect is identifiable given the graph and can generate minimal or alternative adjustment sets.

The interface provides interactive editing, labeling, and validation features. Dagitty also supports exporting graphs in Dagitty's own syntax and in standard formats such as GraphML or JSON, enabling interoperability with other software. It can generate code snippets for the R package dagitty, allowing users to reproduce analyses within R and to extract recommended adjustment sets programmatically.

Dagitty is maintained as an open-source project with online documentation and example graphs. It is commonly used for study design planning, causal diagram exploration, and teaching causal inference concepts.

Limitations include dependence on the correctness of the specified DAG: unmeasured confounding, selection bias, measurement error, or model misspecification are not resolved by the tool. DAGs require acyclicity, so feedback loops are not represented. Results are only as robust as the underlying assumptions encoded in the graph.