|
|
TFHRC Home > Safety > Safety Publications > Safety Effects of Differential Speed Limits on Rural Interstate Highways
October 2005
FHWA-HRT-05-042
View PDF version (959 KB)
To view PDF files, you need the Acrobat® Reader®.
The Surface Transportation and Uniform Relocation Assistance Act, (STURAA) enacted on April 2, 1987, permitted individual States to raise rural interstate speed limits from the previously mandated national speed limit of 89 kilometers per hour (km/h) (55 miles per hour (mi/h)) to 105 km/h (65 mi/h) on rural interstate highways. Of those that changed their speed limits, some States raised the limits for passenger cars but not trucks while other States raised the limits for both passenger cars and trucks. The former category, with different speed limits for cars and trucks, is known as differential speed limits (DSL). The latter category, which mandates the same speed limits for cars and trucks, is known as uniform speed limits (USL). The 1995 repeal of the national maximum speed limit gave States additional flexibility in setting their limits, such that by 2002 several States had experimented with both DSL and USL.
This report compares the safety effects of USL for all vehicles as opposed to DSL for cars and heavy trucks. Detailed crash data, speed monitoring data, and traffic volumes were sought for rural interstate highways in 17 States for the period 1991 to 2000. The information and results of the study will be of particular interest to State traffic managers in making decisions about the application of USL or DSL in their highway systems.
Michael Trentacoste,
Director, Office of Safety Research and Development
Notice
This document is disseminated under the sponsorship of the U.S. Department of Transportation in the interest of information exchange. The U.S. Government assumes no liability for the use of the information contained in this document. This report does not constitute a standard, specification, or regulation.
The U.S. Government does not endorse products or manufacturers. Trademarks or manufacturers' names appear in this report only because they are considered essential to the objective of the document.
Quality Assurance Statement
The Federal Highway Administration (FHWA) provides high-quality information to serve Government, industry, and the public in a manner that promotes public understanding. Standards and policies are used to ensure and maximize the quality, objectivity, utility, and integrity of its information. FHWA periodically reviews quality issues and adjusts its programs and processes to ensure continuous quality improvement.
Technical Report Documentation Page
| 1. Report No FHWA-HRT-05-042 | 2. Government Accession No. N/A | 3. Recipient's Catalog No. N/A | |
| 4. Title and Subtitle THE SAFETY IMPACTS OF DIFFERENTIAL SPEED LIMITS ON RURAL INTERSTATE HIGHWAYS | 5. Report Date October 2005 | ||
| 6. Performing Organization Code N/A | |||
| 7. Authors(s) Nicholas J. Garber, John S. Miller, Bo Yuan and Xin Sun | 8. Performing Organization Report No. N/A | ||
| 9. Performing Organization Name and Address Virginia Transportation Research Council 530 Edgemont Road Charlottesville, VA 22903 | 10. Work Unit No. (TRAIS) N/A | ||
| 11. Contract or Grant No. VRC-000S(007) | |||
| 12. Sponsoring Agency Name and Address Office of Safety Federal Highway Administration 6300 Georgetown Pike McLean, VA 22101 | 13. Type of Report and Period Covered | ||
| 14. Sponsoring Agency Code | |||
| 15. Supplementary Notes Contracting Officer's Technical Representative: A. J. Nedzesky, HRDS-05 | |||
| 16. Abstract To compare the safety effects of a uniform speed limit (USL) for all vehicles as opposed to a differential speed limit (DSL) for cars and heavy trucks, detailed crash data, speed monitoring data, and traffic volumes were sought for rural interstate highways in 17 States for the period 1991 to 2000. Conventional statistical tests (analysis of variance, Tukey's test, and Dunnett's test) were used to study speed and crash rate changes in the four policy groups. A modified empirical Bayes formation was used to evaluate crash frequency changes without presuming a constant relationship between crashes and traffic volume. No consistent safety effects of DSL as opposed to USL were observed within the scope of the study. The mean speed, 85th percentile speed, median speed, and crash rates tended to increase over the 10-year period, regardless of whether a DSL or USL limit was employed. When all sites within a State were included in the analysis, temporal differences in these variables were often not significant. Further examination suggests that while these data do not show a distinction between DSL and USL safety impacts, the relationship between crashes and traffic volume cannot be generalized but instead varies by site within a single State. Because application of the modified empirical Bayes methodology suggested that crash risk increased for all four policy groups, a mathematical model that predicts sharp changes in crash rates based only on ADT does not appear valid at the statewide level. | |||
| 17. Key Words Differential Speed Limit, Universal Speed Limit, Truck Speed Limit, Speed Limit | 18. Distribution Statement No restrictions. This document is available to the Public through the National Technical Information Service; Springfield, VA 22161 | ||
| 19. Security Classif. (of this report) Unclassified | 20. Security Classif. (of this page) Unclassified | 21. No. of Pages 99 | 22. Price |
Form DOT F 1700.7 (8-72) Reproduction of completed page authorized
SI* (Modern Metric) Conversion Factors
LITERATURE REVIEW
Impact of DSL on Mean Speed
Impact of DSL on Speed Variance
Impact of DSL on Crashes
Modified Empirical Bayes Methodology
METHODOLOGY
Data Synthesis
Analysis of Speed Data
Analysis of Crash Data Using Conventional Statistical Approaches
Analysis of Crash Data Using the Empirical Bayes Technique
Conceptual Overview
Development of the Crash Estimation Model
Critical Assumptions of the Crash Estimation Model
Application of the Crash Estimation Model with the Before Data
Application of the Crash Estimation Model with the After Data
Quantifying the Safety Impact of the Speed Limit Change
RESULTS, DISCUSSION, AND LIMITATIONS
Vehicle Speeds
Mean Speeds: An Example of How the Data May Be Assessed
Graphical Overview of Changes in Speed Variance, 85th Percentile Speed, Median Speed, and Noncompliance Rates
Statistical Results of Changes in Speed Variance, 85th Percentile Speed, Median Speed, and Noncompliance Rates
Comparison of Six Interstate Highway Segments in Idaho
Discussion of Speed Impacts
Crash Rates (Analyzed by Conventional Methods)
Crashes (Analyzed by the Modified Empirical Bayes Method)
Virginia Crashes (DSL to USL)
Arkansas Crashes (USL to DSL)
Idaho Crashes (USL to DSL)
Crashes from the States of Arizona, Missouri, North Carolina, and Washington
Relating Speed and Crash Changes
STUDY LIMITATIONS
Caveats About the Use of the Empirical Bayes Method
General Caveats
CONCLUSIONS
Safety Impacts of DSL Versus USL
Methodological Findings
APPENDIX A. EXAMPLES OF DATA COLLECTION LETTERS AND PROCESSING
APPENDIX B. EXAMPLE OF A CLARIFYING DATA REQUEST LETTER
APPENDIX C. CONFIRMATION OF THE NEGATIVE BINOMIAL DISTRIBUTION TO CRASH DATA
Verification of the Poisson Distribution
Verification of the Negative Binomial Distribution
APPENDIX D. EFFECT OF CHANGING THE BASE YEAR IN Ci,y
Using the First Year as the Base Year
Using the Third Year as the Base Year
APPENDIX E. DETERMINING THE SAMPLE SIZE
Test for Significant Differences Using the Sample Size as the Number of Sites
Test for Significant Differences Using the Sample Size as the Number of Vehicles
Simulating the Speed Variances
Determining the Difference in Mean Speeds Needed to Show Significant Difference Based on the Number of Vehicles
Discussion
APPENDIX G. EXAMINATION OF THE EFFECTS OF ADT ON TOTAL CRASH RATES
Histograms of ADT Versus Total Crash Rate
Two-Way Analysis of Variance
APPENDIX H. EXAMPLE APPLICATION OF THE CRASH ESTIMATION MODEL TO THE AFTER DATA
Estimation of Expected Crash Frequency M1, M2,... My for the Before Period
Prediction of My+1, My+2,...My+Z for the After Period
Evaluation of Safety Effects of Changing the Speed Limit for This Particular Site
Evaluation of Safety Effects of Changing the Speed Limit for All Sites
LIST OF TABLES
Table 1. Accident proportions by speed limit, collision type, and vehicle involvement
Table 2. Overview of data availability for rural interstates from the various States
Table 3. Available speed data
Table 4. Available crash data for all sites
Table 5. Five potential models for total number of crashes for Virginia rural interstate highways
Table 6. Five models for total number of crashes on Arizona rural interstates
Table 7. Before/after mean speed comparisons from the ANOVA test
Table 8. Annual mean speed comparisons
Table 9. Longitudinal comparison of speed variables within the States
Table 10. Idaho speed limits
Table 11. ANOVA results of mean speed and 85th percentile speed in Idaho
Table 12. Statistical Tests for Significance in Crash Rates
Table 13. Virginia data for the before and after periods
Table 14. Crash estimation model parameters for Virginia data
Table 15. Total crashes for Virginia
Table 16. Virginia total fatal crashes
Table 17. Crash data for Arkansas
Table 18. Crash estimation model parameters for Arkansas data
Table 19. Arkansas rear-end crashes
Table 20. Total crashes for Arkansas
Table 21. Fatal crashes for Arkansas
Table 22. Crash data for Idaho
Table 23. Total crashes for Idaho
Table 24. Impact of speed limit changes according to the empirical Bayes formulation
Table 25. Crash increases and confidence intervals according to the empirical Bayes formulation
Table 26. T-Statistics for the empirical Bayes crash estimation models (before data)
Table 27. Poisson validation description and results using the total crashes at four test sites
Table 28. Negative binomial validation description and results
Table 29. Estimation of expected crashes using 1991 data as a base in the Ci,y ratio
Table 30. Estimation of expected crashes using 1993 data as a base in the Ci,y ratio
Table 31. Summary Idaho data from speed sampling sites
Table 32. Sample sizes required to achieve significant differences
Table 33. 95 Percent confidence intervals for the 85th percentile speed.*
Table 34. ANOVA variable definitions
Table 35. ANOVA Arizona results
Table 36. ANOVA Virginia results
Table 37. Before and after crash data for a single site
Table 38. Estimation results for the before years
Table 39. Prediction results for the after years
Table 40. Evaluation of the treatment for the example site
LIST OF FIGURES
Figure 1. Chart. Speed limits throughout the 1990s on rural interstate highways
Figure 2. Chart. Data analysis process flowchart
Figure 3. Equation. Crash rate
Figure 4. Chart. Fundamental steps of the empirical Bayes approach
Figure 5. Equation. Crash models for Virginia
Figure 6. Equation. Crash models for Washington
Figure 7. Chart. Comparison of crash estimation models for Virginia and Washington State based on 1991-1993 data
Figure 8. Equation. Expected mean value of crashes
Figure 9. Chart. Plot of goodness of fit for the crash estimation model versus ADT
Figure 10. Chart. Plot of goodness of fit for the crash estimation model versus length
Figure 11. Equation. Alternative crash estimation model
Figure 12. Equation. CEM for before years
Figure 13. Equation. Expected crash frequency m for period 1
Figure 14. Equation. Variance of expected crash frequency m for period 1
Figure 15. Equation. Expected crash frequency m for period y
Figure 16. Equation. Variance of expected crash frequency m for period y
Figure 17. Equation. Would-have-been crashes, had there been no speed limit change
Figure 18. Equation. Actual crashes, given that the speed limit did change
Figure 19. Equation. The difference between would-have-been and actual crashes
Figure 20. Equation. Variance for δ
Figure 21. Equation. Confidence intervals for δ
Figure 22. Equation. Reduction in the expected number of crashes
Figure 23. Equation. Ratio of actual to would-have-been crashes
Figure 24. Equation. Variance of ratio of actual to would-have-been crashes
Figure 25. Equation. Confidence intervals for θ
Figure 26. Chart. Mean speed for all vehicles
Figure 27. Chart. 85th Percentile speeds and median speeds
Figure 28. Chart. Median speed trends
Figure 29. Chart. Speed variance rates
Figure 30. Chart. Noncompliance rates
Figure 31. Chart. Total crash rates
Figure 32: Chart. Total truck-involved crash rates in Virginia interstate highways
Figure 33. Chart. Relationship between the Poisson and negative binomial distributions for crash frequencies
Figure 34. Chart. Comparison of Poisson distribution and actual crash distribution
Figure 35. Chart. comparison of negative binomial distribution and actual crash distribution (probability density function)
Figure 36. Equation. Crash frequency for year 1 as base year
Figure 37. Equation. Crash frequency for year 3 as base year
Figure 38. Equation. Expected value of crash count for year 1
Figure 39. Equation. Variance of expected value of crash count for year 1
Figure 40. Equation. Estimation of estimated values of crash counts for year 1
Figure 41. Equation. Variance of estimation of estimated values of crash counts for year 1
Figure 42. Equation. Expected value of crash count for year 3
Figure 43. Equation. Expected value of crash count, year 3 as base year
Figure 44. Equation. Variance of expected value of crash count, year 3 as base year
Figure 45. Equation. Statistically significant difference in mean speeds
Figure 46. Chart. Histogram based on random numbers
Figure 47. Equation. Formula to determine confidence intervals associated with mean speed
Figure 48. Equation. Example of formula in figure 47
Figure 49. Equation. Confidence interval for 85th percentile speed
Figure 50. Equation. Binomial distribution
Figure 51. Chart. Arizona total crash rate versus ADT
Figure 52. Chart. Virginia total crash rate versus ADT
Figure 53. Chart. Virginia total crash rate versus ADT
Figure 54. Equation. Crash estimation model
Figure 55. Equation. Mean of the estimate for 1991
Figure 56. Equation. Mean of the estimate for 1992
Figure 57. Equation. Mean of the estimate for 1993
Figure 58. Equation. Calculation for ratio before year y
Figure 59. Equation. Ratio before year 1991
Figure 60. Equation. Ratio before year 1992
Figure 61. Equation. Ratio before year 1993
Figure 62. Equation. Expected crash counts
Figure 63. Equation. Variance of the expected crash counts for year 1
Figure 64. Equation. Expected crash counts
Figure 65. Equation. Variance of expected crash counts
Figure 66. Equation. Application for 1991
Figure 67. Equation. Application for variance 1991
Figure 68. Equation. Application for 1992
Figure 69. Equation. Application for variance 1992
Figure 70. Equation. Application for 1993
Figure 71. Equation. Application for variance for 1993
Figure 72. Equation. Computation of E(m1,1995)
Figure 73. Equation. Computation of E(m1,1996)
Figure 74. Equation. Computation of C1,1995
Figure 75. Equation. Computation of C1,1996
Figure 76. Expected crash counts, year y
Figure 77. Variance of expected crash counts, year y
Figure 78. Equation. Expected crash counts, year 1995
Figure 79. Variance of expected crash counts, year 1995
Figure 80. Expected crash counts, year 1996
Figure 81. Variance of expected crash counts, year 1996
Figure 82. Equation. Total would-have-been crashes for a particular site
Figure 83. Equation. Total actual crashes for a particular site
Figure 84. Equation. Safety impact for a particular site
Figure 85. Equation. Ratio of actual to would-have-been crashes
Figure 86. Chart. Cumulative differences, by year, at the example site
Figure 87. Total would-have-been crashes
Figure 88. Total actual crashes
Figure 89. Safety impact
Figure 90. Variance of the difference between would-have-been crashes and actual crashes
Figure 91. Standard deviation of the difference between would-have-been crashes and actual crashes
Figure 92. Equation. Computation of the index of effectiveness
Figure 93. Equation. Variance of θ
Figure 94. Equation. Empirical confidence bounds
To compare the safety effects of a uniform speed limit (USL) for all vehicles as opposed to a differential speed limit (DSL) for cars and heavy trucks, detailed crash data, speed monitoring data, and traffic volumes were sought for rural interstate highways in 17 States for the period 1991 to 2000. Data from nine of those States were used such that they could be divided into four policy groups based on the type of speed limit employed during the period. These were maintenance of a uniform limit only, maintenance of a differential limit only, a change from a uniform to a differential limit, and a change from a differential to a uniform limit. Conventional statistical tests (analysis of variance, Tukey's test, and Dunnett's test) were used to study speed and crash rate changes in the four policy groups. A modified empirical Bayes formation was used to evaluate crash frequency changes without presuming a constant relationship between crashes and traffic volume.
No consistent safety effects of DSL as opposed to USL were observed within the scope of the study. The mean speed, 85th percentile speed, median speed, and crash rates tended to increase over the 10-year period, regardless of whether a DSL or USL limit was employed. When all sites within a State were included in the analysis, temporal differences in these variables were often not significant. Further examination suggests that while these data do not show a distinction between DSL and USL safety impacts, the relationship between crashes and traffic volume cannot be generalized but instead varies by site within a single State. Because application of the modified empirical Bayes methodology suggested that crash risk increased for all four policy groups, a mathematical model that predicts sharp changes in crash rates based only on ADT does not appear valid at the statewide level.
Any study that relies on historical data will be subject to the limitations of incomplete data sets, and to that extent, additional data collection may shed insights not available from an examination of 1990s data alone. Because the investigators believe that accurate mathematical models may require extensive calibration data, a future effort may be more productive if resources are focused on a small group of States over a period of several years, so that speed variance information and crash information may be obtained by individual roadway segment.
FHWA-HRT-05-042 |
 |