Uncertainty in assessment of individual exposure levels leads to bias, often, but not always, toward the null in estimates of health effects, and to underestimation of the variability of the estimates, leading to anticonservative p-values. In the absence of data on the uncertainty in individual...