DAG conditional separation explorer
Check graphical separation for a small directed acyclic graph using ancestral moralization.
Your entries stay in this page’s memory and are not submitted to us. Downloads are generated locally. Reloading or leaving the page clears your entries. Do not enter patient information.
Use simple node labels with letters, digits or underscores; at most 20 nodes and 60 edges. List source and target first, then conditioning nodes separated by commas, or a hyphen for none. Endpoints cannot be conditioned on.
The starting values are hypothetical. Results update as you edit. Use aggregate or synthetic inputs only.
Calculation results
- Graphical result
- D-connected given the specified set.
- Conditioned nodes
- Z
- Relevant ancestral nodes
- 4
- Causal limitation
- Graph assumptions are supplied by the user and are not validated by this result.
How to interpret the output
The algorithm keeps ancestors of source, target and conditioning set, connects co-parents, drops arrow direction and removes conditioned nodes before testing connectivity. D-separation is a graphical implication under the stated DAG; it does not verify causal assumptions, identify every valid adjustment set or prove absence of confounding.
Method referenceSource checked 2026-11-06. Read the reproducible calculator validation report.