Safety Evaluation of J-turn Intersections in Missouri
-
2024-08-01
-
Details:
-
Creators:
-
Corporate Creators:
-
Corporate Contributors:
-
Subject/TRT Terms:
-
Publication/ Report Number:
-
Resource Type:
-
Geographical Coverage:
-
Edition:Final Report (January 2023-July 2024)
-
Corporate Publisher:
-
Abstract:The design of J-turn intersections has gained its prevalence in Missouri due to their demonstrated safety benefits. However, with the growing number of J-turns and the availability of more crash data, there is a renewed need to deepen the understanding of the safety performance of J-turns. This study presents a comprehensive safety evaluation of J-turn intersections, analyzing their effectiveness in reducing total and fatal & injury crashes using crash data from 47 J-turn intersections between 2005 and 2021. Employing a robust methodology including both comparison group and empirical Bayes analyses, this study assesses the impact of J-turns on crash reductions compared to traditional two-way stop-controlled intersections. Two methods were used because they had different tradeoffs such as data requirements, simplicity of implementation, and regression to the mean. The comparison group analysis revealed reductions of 46.6% in fatal and injury crashes and 44.4% in total crashes. Similarly, the empirical Bayes analysis supported these safety improvements, showing reductions of 51.4% in fatal and injury crashes and 40.3% in total crashes. Furthermore, crash frequency models developed for Missouri’s J-turn sites indicate that site characteristics such as loons, deceleration/acceleration lanes, and islands contribute to reduction in crashes. The study also includes detailed collision diagrams that outline crash locations and types at J-turn sites. The study findings provide insights and tools for MoDOT engineers as they consider J-turn design as a safety countermeasure at two-way stop-controlled intersections on rural highways.
-
Format:
-
Funding:
-
Collection(s):
-
Main Document Checksum:
-
Download URL:
-
File Type: