|
Activity Number:
|
353
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Wednesday, August 6, 2008 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Social Statistics Section
|
| Abstract - #301754 |
|
Title:
|
Variations in Effect Estimation Techniques To Assess Graduated Driver
|
|
Author(s):
|
David M. Neyens*+ and Birsen Donmez and Linda Boyle+
|
|
Companies:
|
The University of Iowa and Massachusetts Institute of Technology and The University of Iowa
|
|
Address:
|
3131 Seamans Center, Iowa City, IA, 52242, 3131 Seamans Center, Iowa City, IA, 52242,
|
|
Keywords:
|
teen drivers ; crash data ; time-series models ; graduated licensing programs
|
|
Abstract:
|
The goal of Graduated Drivers Licensing programs is to provide teenage drivers the ability to gain driving experience while limiting their exposure to the most dangerous driving situations. The examination of these programs has been conducted using crash data and with various statistical techniques leading to very different estimates for GDL effectiveness. Failure to account for time and seasonal effects can result in overestimation of this public health intervention. This study uses data from Iowa as an example to demonstrate how various methods can have very different results. Time series models tend to provide the best estimate when compared to general linear models, relative risks, and rate ratios.
|
