Meta-Analysis of Graduated Driver Licensing Laws: Effectiveness of Specific Program Components [Traffic Tech]

NHTSA - Behavioral Safety Researchhttps://doi.org/10.21949/15257842015PDFBriefUnited States. Department of Transportation. National Highway Traffic Safety AdministrationUnited States. Department of Transportation. National Highway Traffic Safety AdministrationUnited StatesReviewsHighway safetyPeriodicalsResearch HubHighway safetyPeriodicalsReviewsDOT HS 812 212Graduated driver licensing (GDL) programs in the United States do not represent a single homogeneous intervention; rather, they contain different combinations and variations of program components. Programs vary by the duration of each stage of the GDL process, age restrictions imposed at each stage, nighttime and passenger restrictions, and supervised driving requirements. While research has suggested GDL is effective at reducing young driver crashes, results are not as clear about which specific program component variations are associated with the largest crash reductions. Studies on the effectiveness of GDL implementations have produced a substantial amount of data and analysis results. NHTSA used these data to conduct a meta-analysis in which the results of existing research were pooled to look for effects revealed by their combined power. As part of the meta-analysis process, researchers first screened studies of GDL programs for relevance and quality. Researchers then coded separate rate ratio effect sizes from the 14 selected studies to estimate the impact of overall GDL programs and variations of program components (referred to as GDL component \u201ccalibrations\u201d from here forward) on 15, 16, 17, 18, 19, and 20 year-olds\u2019 total, injury, and fatal crash outcomes. The full technical report explains the meta-analysis methodology more completely together with all of the study results. This Traffic Tech focuses on the effectiveness of GDL component calibrations for reducing per capita crash rates for 16- and 17-year-olds when at least two effect sizes were available for a given analysis.130