Health Services Research (HSR) Methods
Print Page

A statistical technique for making causal inferences when one of more of the explanatory variables is thought to be correlated with the error term.

An IV must have two properties:
(1) It must be correlated with the suspected endogenous explanatory variable, preferably highly so.
(2) It must not be correlated with the error term, e.g., it must not affect the dependent variable in any way except through the endogenous explanatory variable. If a variable is a valid instrument, the coefficients on the explanatory variables obtained from IV estimation will be unbiased.

Further Reading
Wooldridge, J.M. 2003. Introductory Econometrics, Mason, OH: Thomas South-Western.

Kennedy, P. 1998. A Guide to Econometrics, Cambridge, MA: The MIT Press.

Newhouse, J.P. and M. McClellan. ''Econometrics in Outcomes Research: The Use of Instrumental Variables, Annual Review of Public Health.'' Vol. 19, 1998, pp.17-34.

Hernan, M.A. and J.M. Robins. ''Instruments for Causal Inference: An Epidemiologist's Dream?.'' Epidemiology Vol. 17, 2006, pp.360-372.