Collider (epidemiology)

In epidemiology, a collider is a variable which is the effect of two (or more) other variables. It is known as collider because, in graphical models, the other variables (which may not be associated) lead to the collider in a way that their arrow heads appear to collide on the same node that is the collider^[1] e.g.

$M \rightarrow P \leftarrow L$

The result of having a collider in the path is that a collider blocks^[2] the association between two variables, thus the two variables M and L with a collider P in between are independent of each other because the association between them is blocked^{[clarification needed]} by the collider variable P.

See also

• Causality
• Directed acyclic graph
• Path analysis

References

1. ^ Hernan, Miguel A; Robins, James M (2010), Causal inference, Chapman & Hall/CRC monographs on statistics & applied probability, CRC, p. 70, ISBN 1420076167 
2. ^ Greenland, Sander; Pearl, Judea; Robins, James M (January 1999), "Causal Diagrams for Epidemiologic Research", Epidemiology 10.1: 37–48, ISSN 10443983, OCLC 484244020, http://www.epidemiology.ch/history/PDF%20bg/Greenland,%20Pearl%20and%20Robins%201999%20causal%20diagrams%20for%20epidemiologic%20research.pdf