TTE.ac
Course

Reading and Drawing Causal Diagrams

Slide 1 of 8

Understanding Paths in DAGs

What is a Path?

A path is any sequence of edges connecting two nodes in a DAG, regardless of the direction of the arrows. Understanding paths is fundamental to reading causal information from graphs because paths transmit statistical association—and our goal is to distinguish causal paths from non-causal ones.

Formal Definition

A path between two nodes X and Y is a sequence of distinct nodes starting with X and ending with Y such that every consecutive pair of nodes is connected by an edge (in either direction).

Types of Paths

There are two critically important types of paths in causal inference:

Path TypeDefinitionCausal Interpretation
Directed (Causal) PathAll arrows point in the same direction from cause to effectRepresents actual causal influence
Backdoor PathA path that starts with an arrow into the treatmentRepresents non-causal association (confounding)

Example: The Classic Confounding Triangle

Consider the DAG: A ← L → Y with also A → Y

  • The path A → Y is a directed path (causal)
  • The path A ← L → Y is a backdoor path (non-causal)

The observed association between A and Y is a mixture of both paths. Our goal in causal inference is to isolate the causal path by blocking all backdoor paths.

From What If

"A backdoor path from treatment A to outcome Y is any path that starts with an arrow pointing into A."

Source: What If (Hernán & Robins), Chapter 6

Use arrow keys to navigate