Skip to contentUnited States Department of Transportation - Federal Highway Administration FHWA Home
Federal Highway Administration Research and Technology: Coordinating, Developing, and Delivering Highway Transportation Innovations

ELCSI-PFS: Overview | Seven Phases of ELCSI-PFS | Technical Advisory Committee Members | Technical Advisory Committee Meetings | Publications | Treatment Databases | NCHRP Project 17-25: Crash Reduction Factors forTraffic Engineering and Its Improvements |
NCHRP Project 17-35: Evaluation of Safety Strategies at Signalized Intersections

 

 

Evaluations of Low Cost Safety Improvements Pooled Fund Study

 

PPT version for Printing

Two–way Left–Turn Lanes on Two–lane Roads Results

Header image –Picture shows series of three scenarios: a vehicle on a meandering road, safety personnel at work, and a car that is very badly damaged after it appears to have collided into a telephone pole.
 
Craig Lyon, Persaud and Lyon, Inc

Overview

  • Background
  • Literature Review
  • Objective
  • Data Collection
  • Results
  • Economic Analysis
  • Conclusions
  • Four to Three Lane Conversions

Background on Strategy

  • Tried, moderate, long implementation
  • Evaluating 2 to 3 lane conversion
  • Two installation methods
  • IA study also evaluated 2–3 lane
 View alternative text

View alternative text

Literature Review

  • Highway Safety Design Model Draft (Chapter 8) provides Accident Modification Factor function for Two–way left turn lane
  • Accident rates for 2U and 3T California and Michigan roads (NCHRP Report 282)
  Accidents/million vehicle miles 2U Accidents/million vehicle miles 3T "Accident Modification Factor" ratio
Non–Intersection Commercial 2.39 1.56 0.65
Non–Intersection Residential 1.88 1.64 0.87
Un–signalized intersection Commercial 2.11 2.43 1.15
Un–signalized intersection Residential 2.88 1.91 0.66
Total– Commercial 4.50 3.99 0.89
Total– Residential 4.76 3.55 0.75
 

Objective

  • To estimate the safety effectiveness of installation of two–way left–turn lanes on two–lane roads
  • Target crashes
  • Questions of interest

Data Collection

View alternative text
  AR CA IL NC
Total Mileage 13.2 6.8 6.0 21.3
Crashes/site–year before 7.3 9.7 50.9 5.1
Crashes/site–year after 5.7 6.2 35.1 4.9
Injury crashes/site–year before 1.0 2.1 7.9 0.9
Injury crashes/site–year after 0.7 1.1 3.2 0.6
Rear–end crashes/ site–year before 2.5 4.1 28.3 1.9
Rear–end crashes/ site–year after 1.3 2.0 12.7 1.4

Aggregate Evaluation Results

States Percent reduction in Rear–end crashes Percent reduction in Injury crashes Percent reduction in Total crashes
AR 49.9 38.1 22.5
CA 49.4 27.5 34.1
IL 42.0 53.1 12.6
NC 21.7 –1.9 15.7
All 38.7 26.1 20.3

State Standard errorof
Rear–end crashes 		Standard error of
Injury crashes		 Standard error of
Total crashes
AR (7.3) (11.0) (5.8)
CA (7.3) (8.7) (5.7)
IL (7.6) (11.9) (7.3)
NC (7.7) (14.7) (4.8)
All (4.0) (6.8) (3.0)

Disaggregate Evaluation Results

Disaggregate Group Sites Estimate of percent reduction
(standard error)
Arkansas – rural 15 51.2 (7.1)
Arkansas – urban 10 3.8 (8.3)
California – rural 21 50.8 (5.7)
California – urban 10 –2.8 (13.4)
Illinois – rural 5 16.7 (10.5)
Illinois – urban 5 9.4 (10.0)
North Carolina – rural 38 27.3 (5.5)
North Carolina – urban 40 –5.0 (8.8)

Economic Analysis

Comparison of construction costs and accident savings for two–way left– turn lanes (TWLTL) projects

State Initial Initial converted to annual Cost per mile year of rear end accidents saved Low (unsignalized intersection) Cost per mile year of rear end accidents saved High non–intersection)
Arkansas $440,000 $31,882 $18,363 $41,657
California $500,000 $36,230 $25,697 $58,410
Illinois $1,780,000 $128,979 $124,715 $283,479
North Carolina $424,000 $30,733 $5,138 $11,680

Cost/mile calculation (based on a 7% discount rate and a 50 year life)

Conclusions

  • Significant reductions in total and rear–end crashes
  • Observed reductions in head–on, intersection and non–intersection crashes
  • Rural sites more effective than urban
  • Research needed to answer where urban treatments are most effective

Four to Three Lane Conversions

  • Four to Three Lane conversions studied under NCHRP 17–25
  • Iowa data for urban conversions
  • Previous study using HSIS data
  • Empirical Bayes Before–After study employed
  • Iowa
    • 15 treated sites
    • 296 reference sites
  • HSIS data
    • 30 sites in 8 cities from CA and WA
    • 51 reference sites
Results
Dataset Mean(θ) standard error
Iowa 0.777 0.028
HSIS 0.811 0.025
All 0.801 0.019

< Presentation 6 | Presentation 8 >