DataTerminology.comNEW
A graphical condition identifying causal effects via mediators when back-door adjustment is blocked by unobserved confounders.
DESCRIPTION
The front-door criterion identifies causal effects using a mediator that fully transmits the treatment’s effect and is unaffected by hidden confounders of treatment and outcome. It requires that all back-door paths from treatment to mediator are blocked and that, conditional on mediator, treatment has no additional effect on outcome through unblocked paths. Front-door adjustment leverages do-calculus to compute causal effects despite unobserved confounding. It complements back-door and instrumental variables in causal identification strategies.
ALTERNATIVE DEFINITIONS
Wikipedia
The front-door criterion is a graphical criterion for identifying causal effects using an intermediate variable (mediator) under certain conditions.
Highlights identification via a mediator and conditions for validity.
UCLA Causal Lab
We introduce the front-door criterion, a method for identifying causal effects in the presence of unmeasured confounding through a mediating variable.
Emphasizes handling unmeasured confounding by exploiting mediators.
RELATED TERMS
Directed Acyclic Graph (Causal) Do-Calculus Mediator Confounding Identification Strategy Treatment Effect Structural Equation Model Potential Outcomes Framework
COMPETING TERMS
Backdoor Criterion Instrumental Variable
KNOWLEDGE AREAS
TARGET AUDIENCE
Data Scientist AI Engineer Data Analyst Information Analyst
EXTERNAL SOURCES
- Wikipedia - Front-door criterion
- UCLA Causal Lab - Pearl’s technical report on front-door
- Hernán & Robins - Textbook explanation and examples
TOP QUESTIONS
When should the Front-Door Criterion be used?
Use it when back-door adjustment is impossible due to unmeasured confounding, but a valid mediator satisfies front-door assumptions.
Category: Requirements
What assumptions underpin the Front-Door Criterion?
Treatment affects the mediator; mediator fully transmits effect; mediator is not confounded with treatment or outcome given conditions.
Category: Overview
How does the Front-Door Criterion differ from the Back-Door Criterion?
Back-door blocks confounding with observed covariates, while front-door uses a mediator to identify effects despite hidden confounders.
Category: Alternatives
What is a common mistake with the Front-Door Criterion?
Assuming any intermediate variable suffices; the mediator must meet specific independence and exclusion conditions.
Category: Mistakes
Can the Front-Door Criterion be tested empirically?
Assumptions are largely untestable directly; triangulation, domain knowledge, and sensitivity analyses support plausibility.
Category: Risks
How is the Front-Door Criterion applied?
Estimate treatment-to-mediator and mediator-to-outcome effects, then combine per do-calculus to recover the causal effect.
Category: Example
What risks arise from violating Front-Door assumptions?
Bias and incorrect effect estimates occur, potentially worse than naive adjustment; careful validation is essential.
Category: Risks
Is Instrumental Variable analysis similar to Front-Door?
Both address confounding; IV uses a valid instrument, front-door uses a mediator with stronger structural assumptions.
Category: Opinion
How do you document Front-Door analyses?
Provide DAGs, justify assumptions, report sensitivity checks, and detail estimation steps and robustness results.
Category: Management