CPR Working Paper Series No. 49
On the Measurement of Job Risk
in Hedonic Wage Models
Dan A. Black and Thomas J. Kniesner
March 2003
(Revised from January 2003)
We examine the incidence, form, and research consequences of measurement error in measure of fatal injury risk in U.S. workplaces using both BLS and NIOSH data. We find evidence of substantial measurement errors in the fatality risk researchers attach to individual workers when estimating the implicit price of risk and the value of a statistical life. We first examine possible classical attenuation bias in the fatality risk coefficient. However, because we also find non-classical measurement error that differs across multiple risk measures and is not independent of other regressors, more complex statistical procedures than a standard instrumental variables estimator need be applied to obtain statistically improved estimates of wage-fatality risk tradeoffs.
A revised version of this paper appears in The Journal of Risk and Uncertainty, Volume 27(3), "On the Measurement of Job Risk in Hedonic Wage Models."
