Skip to content
flowss
Start freeChecking your session
Causal DAG templates

Biology — Causal DAG: Birth-Weight Paradox

Conditioning on a collider creates a false protective effect.

Causal DAGLife SciencesCausal DAG syntax docs
Template previewCausal DAG
Birth-weight paradox (collider bias)EMaternal smokingOInfant mortalityBLow birth weight (collider)UBirth defectsExposureOutcomeConfounderMediatorUnobservedThe exposure is unconfounded: every backdoor path is already blocked, so the crude association estimates the total effect. 4 variables, 4 edge…No adjustment is needed: the exposure is unconfounded in this diagram, so the crude association between `Maternal smoking` and `Infant mortality` already estimate…

Make it your own.

title "Birth-weight paradox (collider bias)"
exposure E "Maternal smoking"
outcome O "Infant mortality"
node B "Low birth weight (collider)"
unobserved U "Birth defects"
E -> B
U -> B
U -> O
E -> O