Skip to main content
. Author manuscript; available in PMC: 2018 Feb 16.
Published in final edited form as: Annu Rev Public Health. 2017 Mar 20;38:351–370. doi: 10.1146/annurev-publhealth-031816-044208

Table 1.

Key requirements for a valid instrumental variable (see also Figure 4)

Requirement Description
Exclusion restriction Any effect of the instrument on health must be mediated by exposure to the policy of interest
Exogenous The instrument must be randomly distributed and uncorrelated with the unobserved or unmeasured characteristics of individuals
Meaningful effect The instrument must reliably predict the policy exposure (a strong instrument)
Monotonic The effect of the instrument on the policy exposure is not smaller than the exposure that would occur without the instrument
Stable unit treatment value The value of the policy variable and the relationship between the policy variable and health outcome in one individual must not be affected by variations in these factors among other individuals