The letter by Desapriya et al concerns 2 issues: exposure measurement and unmeasured variables. We appreciate the opportunity to clarify them. We did not use vehicle miles traveled as our exposure measure but instead based our analysis on population of 16-year-olds, because our focus was the net benefit of graduated driver licensing (GDL) in reducing motor vehicle mortality among 16-year-olds. We did acknowledge in our discussion section1(p61) that, in the absence of exposure data, it was impossible to determine the extent to which the observed GDL-associated reduction in fatal crashes was attributable to decreased exposure to driving or decreased crash-instance density.

Confounding by unmeasured variables is a concern inherent in all observational studies, including the studies cited by Desapriya et al. Traffic safety can be affected by numerous factors related to the environment, the driver, and the vehicle. There is no way other than a randomized, experimental design to control...

You do not currently have access to this content.