Skip to main content
. 2014 Nov 19;11:18. doi: 10.1186/1742-7622-11-18

Figure 1.

Figure 1

Markov graph showing the effects of individual exposures (E) on their health outcome (O) in the context of observational studies. Both E and O can be affected by other variables, including (unknown) confounders (C). The existence of an Instrumental variable (IV) makes it possible to establish a causal relationship between E and O.