Causal graphs Flashcards
1
Q
Mediator
A
Effect of X on Y is mediated by variable Z
2
Q
Confounder
A
X and Y have a common cause
3
Q
Collider
A
X and Y have a common effect Z
4
Q
Closing the backdoor
A
causal inference that passes through a confounder is a backdoor path. This needs to be conditioned. If closed they’re d-seperated
5
Q
DAG graph
A
Specifying assumptions and identify confounders
6
Q
Purification principle
A
The more variables are controlled for, the more accurate the estimation of effect. NOT TRUE. Overcorrection could lead to underestimation and collider bias
7
Q
Collider bias
A
Controlling for common effects will create a relation that doesn’t exist
