Department of Economics, Wright State University, Dayton, Ohio 45435, USA.
J Safety Res. 2009;40(6):421-6. doi: 10.1016/j.jsr.2009.10.003. Epub 2009 Oct 30.
A state by year panel is analyzed to simultaneously explore the statistical correlation between state level traffic fatality rates and state level behavioral regulations regarding teen licensing, seat belt use, and driving under the influence (DUI) in a model that also controls for other correlates.
By including measures of all three of these policies, the estimated policy effects should not be overstated due to underspecification bias. The panel includes the 48 contiguous U.S. states for the time period from 1999 through 2003. State fatality rates are measured as fatalities per million miles traveled. Measures of state policies regarding traffic safety related behavior are based on information gathered by the Insurance Institute for Highway Safety. Estimates are calculated via a time fixed effects model that uses the double-log form to allow for interaction effects between the independent variables.
Least squares estimates indicate that, on average, more restrictive graduated teen licensing and DUI policies significantly reduce traffic fatality rates, while stricter seat belt enforcement policies have a statistically insignificant negative impact on fatality rates.
本文通过逐年的州面板分析，旨在探索州级交通死亡率与州级青少年驾驶许可、安全带使用和酒后驾车（DUI）行为规范之间的统计相关性，该模型还控制了其他相关因素。
通过包含这三项政策的措施，由于指定不足的偏差，估计的政策效果不应被夸大。该面板包括 1999 年至 2003 年期间的 48 个毗邻的美国州。州死亡率以每百万英里行驶的死亡人数来衡量。关于交通安全相关行为的州政策措施基于保险研究所高速公路安全收集的信息。通过使用双对数形式的时间固定效应模型计算估计值，以允许自变量之间的交互作用。
最小二乘法估计表明，平均而言，更严格的青少年分级驾驶许可和 DUI 政策显著降低了交通死亡率，而更严格的安全带执法政策对死亡率的影响具有统计学上的不显著的负面影响。
