2rumble Strips

Safety Evaluation of Edge-Line
Rumble Stripes on Rural Two-Lane

Horizontal Curves
PUBLICATION NO. FHWA-HRT-17-069 DECEMBER 2017
Research, Development, and Technology

Turner-Fairbank Highway Research Center
6300 Georgetown Pike
McLean, VA 22101-2296
FOREWORD
The research documented in this report was conducted as part of the Federal Highway
Administration’s (FHWA) Evaluation of Low-Cost Safety Improvements Pooled Fund Study
(ELCSI-PFS). FHWA established this PFS in 2005 to conduct research on the effectiveness of
the safety improvements identified by the National Cooperative Highway Research Program
Report 500 guides as part of the implementation of the American Association of State Highway
and Transportation Officials Strategic Highway Safety Plan. The ELCSI-PFS studies provide a
crash modification factor and benefit–cost (B/C) economic analysis for each of the targeted
safety strategies identified as priorities by the pooled fund member States.
This study evaluated application of edge-line rumble stripes (ELRSs) on rural two-lane
horizontal curves. ELRSs are a variation of common shoulder rumble strips used to alert drowsy
or distracted drivers when they are leaving the travel lane to the right. ELRSs are installed with
the edge-line pavement marking placed directly over the rumble strip. Data were obtained at
treated rural two-lane horizontal curves in Kentucky and Ohio. The results for Kentucky indicate
statistically significant reductions for total, injury, run-off-road (ROR), and nighttime crashes.
The results for Ohio indicate statistically significant reductions for all crash types (i.e., total,
injury, ROR, nighttime, and nighttime ROR). The B/C analysis results suggest that this treatment
can be highly cost-effective. This report is intended for State departments of transportation,
transportation agencies, academics, researchers, and other practitioners.
Jonathan Porter, Ph.D.

Acting Director, Office of Safety
Research and Development
Notice
This document is disseminated under the sponsorship of the U.S. Department of Transportation
(USDOT) in the interest of information exchange. The U.S. Government assumes no liability for
the use of the information contained in this document.
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. 2. Government Accession No. 3. Recipient’s Catalog No.
FHWA-HRT-17-069
4. Title and Subtitle 5. Report Date
Safety Evaluation of Edge-Line Rumble Stripes on Rural Two-Lane December 2017
Horizontal Curves 6. Performing Organization Code
7. Author(s) 8. Performing Organization Report No.

Scott Himes, Frank Gross, Bhagwant Persaud, and Kimberly Eccles
9. Performing Organization Name and Address 10. Work Unit No. (TRAIS)
VHB
8300 Boone Boulevard, Suite 700 11. Contract or Grant No.
Vienna, VA 22182-2626 DTFH61-13-D-00001
12. Sponsoring Agency Name and Address 13. Type of Report and Period Covered
U.S. Department of Transportation Safety Evaluation; 2014–2016
Federal Highway Administration 14. Sponsoring Agency Code
1200 New Jersey Avenue, SE
Washington, DC 20590-3660
15. Supplementary Notes
The Federal Highway Administration (FHWA) Office of Safety Research and Development managed this study
under the Development of Crash Modification Factors program. The FHWA Office of Safety Research and
Development Contracting Officer’s Program and Task Manager was Roya Amjadi (HRDS-20).
16. Abstract
The Development of Crash Modification Factors (DCMF) program conducted safety evaluations of edge-line rumble
stripes (ELRSs) on rural two-lane horizontal curves for the Evaluation of Low-Cost Safety Improvements Pooled
Fund Study. This study evaluated the application of ELRSs on rural two-lane horizontal curves. ELRSs are a
variation of common shoulder rumble strips used to alert drowsy or distracted drivers when they are leaving the
travel lane to the right. ELRSs are installed with the edge-line pavement marking placed directly over the rumble
strip. Geometric, traffic, and crash data were obtained at treated rural two-lane horizontal curves in Kentucky and
Ohio. To account for potential selection bias and regression-to-the-mean, an empirical Bayes before–after analysis
was conducted using reference groups of untreated rural horizontal curves with similar characteristics to the treated
sites. The analysis also controlled for changes in traffic volumes over time and time trends in crash counts unrelated
to the treatment. Owing to a small sample for the reference group in Kentucky and a simultaneous statewide curve
warning sign upgrade program in Ohio, alternative reference sites were used to account for annual trends. The results
for Kentucky indicated statistically significant reductions for total, injury, run-off-road (ROR), and nighttime
crashes, with crash modification factors (CMFs) of 0.75, 0.64, 0.74, and 0.63, respectively. The results for Ohio
indicated statistically significant reductions for all crash types, with total, injury, ROR, nighttime, and nighttime
ROR CMFs of 0.79, 0.79, 0.78, 0.75, and 0.71, respectively. The two States’ results could not be combined because
of the statewide curve signing program in Ohio. It is important to note that all crash types considered in this research
excluded intersection-related and animal crashes. Benefit–cost (B/C) ratios were estimated to be 331:1 for Kentucky
and 477:1 for Ohio. If ELRSs were used as a curve-specific treatment, the B/C ratio would likely be much smaller
because of the higher installation cost; however, these results suggest that the treatment can be highly cost effective.
17. Key Words 18. Distribution Statement.
ELRS, two-lane rural roads, horizontal curves, distracted No restrictions. This document is available through the
driver, rumble strips, rumble stripes, low-cost, safety National Technical Information Service,
improvements, empirical Bayesian. Springfield, VA 22161.
http://www.ntis.gov
19. Security Classif. (of this report) 20. Security Classif. (of this page) 21. No. of Pages 22. Price
Unclassified Unclassified 61
Form DOT F 1700.7 (8-72) Reproduction of completed pages authorized.
ii
TABLE OF CONTENTS
EXECUTIVE SUMMARY ...........................................................................................................1

CHAPTER 1. INTRODUCTION .................................................................................................3
BACKGROUND ON STRATEGY ..........................................................................................3
BACKGROUND ON STUDY...................................................................................................4
LITERATURE REVIEW .........................................................................................................4
CHAPTER 2. OBJECTIVE ........................................................................................................15
CHAPTER 3. STUDY DESIGN .................................................................................................17
CHAPTER 4. METHODOLOGY ..............................................................................................23
CHAPTER 5. DATA COLLECTION........................................................................................25
KENTUCKY ............................................................................................................................25
Installation Data .................................................................................................................. 25
Reference Sites..................................................................................................................... 25
Roadway Data ..................................................................................................................... 26
Traffic Data ......................................................................................................................... 26
Crash Data ........................................................................................................................... 26
Treatment Cost Data .......................................................................................................... 26
OHIO.........................................................................................................................................27
Installation Data .................................................................................................................. 27
Reference Sites..................................................................................................................... 27
Roadway Data ..................................................................................................................... 27
Traffic Data ......................................................................................................................... 27
Crash Data ........................................................................................................................... 27
Treatment Cost Data .......................................................................................................... 28
DATA CHARACTERISTICS AND SUMMARY ................................................................28
CHAPTER 6. DEVELOPMENT OF SAFETY PERFORMANCE FUNCTIONS ...............31
CRASH SPILLOVER AND MIGRATION ..........................................................................31
SAFETY PERFORMANCE FUNCTIONS...........................................................................34
CHAPTER 7. BEFORE–AFTER EVALUATION RESULTS................................................37
AGGREGATE ANALYSIS ....................................................................................................37
DISAGGREGATE ANALYSIS OF OHIO DATA ...............................................................38
CHAPTER 8. ECONOMIC ANALYSIS ...................................................................................41
CHAPTER 9. SUMMARY AND CONCLUSIONS..................................................................43
APPENDIX: ADDITIONAL INSTALLATION DETAILS ....................................................45
RESPONSES FROM KENTUCKY .......................................................................................46
RESPONSES FROM OHIO ...................................................................................................46
ACKNOWLEDGEMENTS ........................................................................................................49
REFERENCES .............................................................................................................................51
iii
LIST OF FIGURES
Figure 1. Illustration. Examples of longitudinal rumble strip markings ......................................... 5

Figure 2. Illustration. Overview of rumble strip dimensions .......................................................... 6
Figure 3. Equation. Estimated change in safety............................................................................ 23
Figure 4. Equation. EB estimate of expected crashes ................................................................... 23
Figure 5. Equation. EB weight ...................................................................................................... 24
Figure 6. Equation. Index of effectiveness ................................................................................... 24
Figure 7. Equation. Standard deviation of index of effectiveness ................................................ 24
Figure 8. Equation. SPF model form for crash migration or spillover effects.............................. 31
Figure 9. Graph. Yearly multiplier by year for reference groups in Kentucky ............................ 32
Figure 10. Graph. Yearly multiplier by year for reference groups in Ohio .................................. 34
Figure 11. Equation. SPF model form for Kentucky and Ohio .................................................... 34
LIST OF TABLES
Table 1. Summary of studies on changes in target collision frequency from application of

SRSs .......................................................................................................................................... 8
Table 2. Rumble strip dimensions from Lyon et al. ..................................................................... 12
Table 3. Before-period crash rate assumptions ............................................................................. 18
Table 4. Minimum required before-period mile-years for treated sites—total crashes ................ 19
Table 5. Minimum required before-period mile-years for treated sites—injury crashes ............. 19
Table 6. Minimum required before-period mile-years for treated sites—ROR crashes ............... 20
Table 7. Minimum required before-period mile-years for treated sites—nighttime crashes ........ 20
Table 8. Minimum required before-period mile-years for treated sites—nighttime ROR
crashes ..................................................................................................................................... 21
Table 9. Analysis for crash effects in Kentucky ........................................................................... 22
Table 10. Analysis for crash effects in Ohio................................................................................. 22
Table 11. Kentucky treatment cost and service life data .............................................................. 26
Table 12. Ohio treatment cost data ............................................................................................... 28
Table 13. Definitions of crash types ............................................................................................. 28
Table 14. Data summary for treatment sites ................................................................................. 29
Table 15. Data summary for Kentucky reference sites ................................................................. 30
Table 16. Data summary for Ohio reference sites ........................................................................ 30
Table 17. Reference group SPF for total crashes—parameter estimates ...................................... 31
Table 18. Observed and predicted crashes for reference groups in Kentucky .............................. 32
Table 19. Observed and predicted crashes for reference groups in Ohio ..................................... 34
Table 20. SPFs by crash type ........................................................................................................ 35
Table 21. Time trend factors for predicted crashes....................................................................... 36
Table 22. Aggregate analysis results for Kentucky ...................................................................... 37
Table 23. Aggregate analysis results for Ohio .............................................................................. 37
Table 24. Ohio results by AADT .................................................................................................. 39
Table 25. Ohio results by before-period expected crash frequency.............................................. 39
Table 26. Recommended CMFs for ELRSs based on Kentucky data .......................................... 43
Table 27. Recommended CMFs for ELRSs and curve signage based on Ohio data .................... 43
iv
LIST OF ABBREVIATIONS
AADT average annual daily traffic

AASHTO American Association of State Highway and Transportation Officials
B/C benefit–cost
C-G comparison group
CLRS center-line rumble strip
CMF crash modification factor
DCMF Development of Crash Modification Factors (program)
EB empirical Bayes
ELCSI-PFS Evaluation of Low-Cost Safety Improvements Pooled Fund Study
ELRS edge-line rumble stripe
FHWA Federal Highway Administration
FTM flat thermoplastic edge marking
GIS geographic information system
HPMS Highway Performance Monitoring System
HSIS Highway Safety Information System
KABCO Scale used to represent injury severity in crash reporting (K is fatal injury, A is
incapacitating injury, B is non-incapacitating injury, C is possible injury, O is
property damage only)
KYTC Kentucky Transportation Cabinet
MUTCD Manual on Uniform Traffic Control Devices
NCHRP National Cooperative Highway Research Program
ODOT Ohio Department of Transportation
PFS pooled fund study
ROR run-off-road
SE standard error
SPF safety performance function
SRS shoulder rumble strip
SVROR single-vehicle run-off-road
v
EXECUTIVE SUMMARY
The Federal Highway Administration (FHWA) established the Development of Crash

Modification Factors (DCMF) program in 2012 to address highway safety research needs for
evaluating new and innovative safety strategies (improvements) by developing reliable
quantitative estimates of their effectiveness in reducing crashes. The ultimate goal of the DCMF
program is to save lives by identifying new safety strategies that effectively reduce crashes and
promote those strategies for nationwide implementation by providing measures of their safety
effectiveness and benefit–cost (B/C) ratios through research. State transportation departments
and other transportation agencies need to have objective measures for safety effectiveness and
B/C ratios before investing in new strategies for statewide safety improvements. Forty State
transportation departments have provided technical feedback on safety improvements to the
DCMF program and have implemented new safety improvements to facilitate evaluations. These
States are members of the Evaluation of Low-Cost Safety Improvements-Pooled Fund Study,
which functions under the DCMF program.
This study evaluated the application of edge-line rumble stripes (ELRSs) on rural two-lane
horizontal curves. ELRSs are a variation of common shoulder rumble strips (SRSs) used to alert
drowsy or distracted drivers when they leave the travel lane to the right. ELRSs are installed
where edge-line pavement markings would normally be placed, and the pavement marking is
installed directly over the rumble strip. In this way, the ELRSs are installed closer to the travel
lane than common SRSs. In addition, the vertical faces that are created within the milled rumble
strip to which pavement markings are applied have the effect of enhancing the visibility of the
edge line during nighttime and wet-weather conditions.
The project team obtained geometric, traffic, and crash data at treated rural horizontal curve
locations in Kentucky and Ohio. To account for potential selection bias and regression-to-the-
mean, the project team conducted an empirical Bayes (EB) before–after analysis using reference
groups of untreated rural horizontal curves with similar characteristics to the treated sites. The
analysis also controlled for changes in traffic volumes over time and time trends in crash counts
unrelated to the treatment. While the analysis focused on the safety effectiveness on horizontal
curves, the treatment applications were not limited only to horizontal curves. The treatment was
a corridor treatment applied to segments consisting of both tangents and curves; however, the
analysis removed horizontal tangents and only considered the effectiveness on curves.
The results for Kentucky indicated statistically significant reductions for total, injury, run-off-
road (ROR), and nighttime crashes at the 95-percent confidence level. Nighttime crashes had the
smallest crash modification factor (CMF), or the greatest reduction, with a value of 0.63. Total,
injury, and ROR crashes had CMFs of 0.75, 0.64, and 0.74, respectively. The CMF for nighttime
ROR crashes was 0.75 and was consistent with the same CMF from Ohio; however, it was
significant only at the 80-percent level, suggesting that sample size was the reason for the lack of
statistical significance at the 95-percent level. The results for Ohio indicated statistically
significant reductions for all crash types. Nighttime ROR crashes had the smallest CMF of 0.71.
Total, injury, ROR, and nighttime crashes had CMFs of 0.79, 0.79, 0.78, and 0.75, respectively.
The resulting Ohio installation CMFs reflected the installation of ELRSs on horizontal curves as
1
well as the impact of the statewide signing program. It is also important to note that all crash
types considered in this research excluded intersection-related and animal crashes.
A disaggregate analysis of the results indicated larger safety benefits for horizontal curves with
average annual daily traffic (AADT) greater than 4,000 for all crash types; however, the
differences by AADT were not statistically significant at the 95-percent confidence level. The
disaggregate analysis further indicated larger safety benefits for horizontal curves with a higher
before-period expected crash frequency. The results suggested no benefit for curves with low
before-period expected crash frequencies for all crash types. The difference in CMFs between
low before-period expected crash frequency and high before-period expected crash frequency
was statistically significant for total and ROR crashes. Due to correlation between variables,
caution should be used in interpreting and applying these disaggregate results; however, the
disaggregate analysis CMFs may be used to inform the process of prioritizing treatment sites for
ELRSs.
Estimated B/C ratios range from 189:1 to 467:1 for Kentucky and from 272:1 to 672:1 for Ohio.
On first inspection, the B/C ratios were larger than expected for an installation of this type.
However, the installations took place on corridors, while the analysis only looked at the safety
effects on horizontal curves. Horizontal curves have higher crash rates than overall corridors, and
the cost per mile of installation would not be representative for installations only on horizontal
curves. As a curve-specific treatment, the B/C ratio would likely be reduced owing to the higher
deployment cost for spot-specific installations. Regardless, these results suggest that the
treatment, even in its most expensive variation, can be highly cost effective.
2
CHAPTER 1. INTRODUCTION
This chapter presents background information on the strategy of using edge-line rumble stripes
(ELRSs), the goals of the study reported here, and a review of the existing literature on the use of
rumble strips.
BACKGROUND ON STRATEGY
ELRSs are a variation of the common shoulder rumble strips (SRSs) used to alert drowsy or
distracted drivers when they leave the travel lane to the right. SRSs and ELRSs both target run-
off-road (ROR) crashes. Key distinctions between SRSs and ELRSs are provided in the
following points:
• SRSs are provided on the shoulder between the pavement marking and the outside edge
of the pavement and can be defined by their offset from the edge-line pavement marking.
• ELRSs are installed where the edge-line pavement marking would normally be placed,
and the pavement marking is installed directly over the rumble strip. In this way, the
rumble stripes are installed closer to the travel lane than common SRSs. In addition,
vertical faces are created within the milled rumble strip to which pavement markings are
applied, thereby enhancing the visibility of the edge line during nighttime and wet-
weather conditions.
Application of ELRSs varies among States depending on climate and roadway surface type. In
colder areas, rumble strips are milled into the surface of the roadway, allowing them to be
snowplowable. In areas that do not receive snowfall, profiled thermoplastic pavement markings
can be used. In addition, profiled thermoplastic pavement markings have been used as an
alternative to milled ELRSs for roadways with a chip seal surface. The research in this study
focused on the safety effectiveness of milled ELRSs.
Several research studies have examined the use of SRSs; however, research into the performance
of ELRSs has been rare and has not been rigorously evaluated. In addition, milled rumble strips
have been installed on roadway segments consisting of both horizontal tangents and horizontal
curves. Installations on only horizontal curves have been uncommon, and therefore, safety
effectiveness evaluations have not focused on their effectiveness on horizontal curves specifically.
This study focused on the safety effectiveness of ELRSs on rural two-lane horizontal curves taken
from rumble strip installations that were not specific to horizontal curves.
Additional details concerning current practice with rumble strips can be found on the Federal
Highway Administration (FHWA) Rumble Strip Community of Practice Web page.(1) This site
provides a description of the three major types of rumble strips (milled, rolled (or formed), and
raised), detailed construction drawings, effectiveness estimates, interviews with users and other
experts, and other important material.
3
BACKGROUND ON STUDY
In 1997, the American Association of State Highway and Transportation Officials (AASHTO)
Standing Committee on Highway Traffic Safety, with the assistance of FHWA, the National
Highway Traffic Safety Administration, and the Transportation Research Board Committee on
Transportation Safety Management, met with safety experts in the field of driver, vehicle, and
highway issues from various organizations to develop a strategic plan for highway safety. These
participants developed 22 key emphasis areas that affect highway safety. The 22 emphasis areas
were published in the AASHTO Strategic Highway Safety Plan.(2)
The National Cooperative Highway Research Program (NCHRP) then published a series of
guides to advance the implementation of countermeasures targeted to reduce crashes and
injuries. Each guide addresses one of the key emphasis areas and includes an introduction to the
problem, a list of objectives for improving safety, and strategies for each objective. Each strategy
is designated as proven, tried, or experimental. Many of the strategies discussed in these guides
have not been rigorously evaluated; about 80 percent of the strategies are considered tried or
experimental.
In 2005, to support the implementation of the guides, FHWA organized a pooled fund study
(PFS) to evaluate low-cost safety strategies as part of this strategic highway safety effort. Over
time, the pooled fund has grown in size and, at the time of this study, included 40 States. The
PFS evaluates the safety effectiveness of several tried and experimental, low-cost safety
strategies through scientifically rigorous crash-based studies. FHWA selected the use of ELRSs
as a strategy to be evaluated as part of this effort.
LITERATURE REVIEW
The Manual on Uniform Traffic Control Devices (MUTCD) defines rumble strips as either
slightly raised or depressed road surfaces with a rough texture designed to provide a haptic alert
for inattentive drivers leaving the travel lane.(3) As shown in figure 1, rumble strips may be
installed either on the shoulders or center line on rural two-lane highways, and the edge line may
be adjacent to the rumble strip or overlapping with it, creating a rumble stripe.
4
A. Edge line not on rumble B. Edge line on rumble strip. C. Center line on rumble
strip. strip.
Source: FHWA.
Note: (A) Edge lines may be located alongside the rumble strip, (B) on the rumble strip, or (C) the center line
markings may also be located on a center line rumble strip. Arrows indicate direction of travel. Empty squares
indicate rumble strip.
Figure 1. Illustration. Examples of longitudinal rumble strip markings.(3)
Figure 2 provides an illustration of rumble strip dimensions, which are explained as follows:(4)
• A—Offset: Distance from the pavement marking (delineating the edge of the traveled
way) to the inside edge of the rumble strip.
• B—Length: Dimension of the strip that is perpendicular to the travel directions of the
roadway. This is often referred to as the transverse width of the rumble strip.
• C—Width: Dimension of the strip that is parallel to the travel direction of the roadway.
• D—Depth: Maximum distance from the surface of the roadway to the bottom of the
rumble strip.
• E—Spacing: Distance between adjacent rumble strips. It is most often measured from
the center of the strip to the center of the adjacent strip.
• F—Gap: Distance from the edge of the rumble strip to edge of rumble strip when there is
a break in the pattern. Gaps are commonly used to allow bicycles to cross the rumble
strip pattern, to allow passing vehicles to cross center-line rumble strips (CLRSs), and to
allow for turning movements at intersections and driveways.
5
Source: FHWA.
Figure 2. Illustration. Overview of rumble strip dimensions.(4)
The University Transportation Center for Alabama conducted an evaluation of ELRS markings
in terms of service life, lifecycle cost, and wet and dry visibility in comparison with flat
thermoplastic edge markings (FTMs).(5) The authors measured nighttime dry and wet
retroreflectivity at 16 1-mi FTM segments and 5 2-mi ELRS segments. They found that initial
dry retroreflectivity was similar between the two groups; however, ELRSs lost visibility at a
lower rate due to cumulative traffic passes. ELRSs were found to provide a higher dry
retroreflectivity and longer service life than FTMs under similar traffic conditions. For wet
retroreflectivity, the ELRSs had a higher initial value than FTMs, but the degradation could not
be compared between the two because of a lack of data for FTMs. The cost per mile for ELRSs
was $2,424 for a 5-year marking service life and an 8-year lifecycle.
A study by Miles et al. used video data to examine the impacts of CLRSs and ELRSs on passing
operations and lateral position on Texas highways.(6) After application of milled CLRSs on no-
passing and passing zones, the authors found no change in passing opportunities or the
percentage of vehicles that passed. However, center-line crossing time increased significantly,
and gap distance decreased significantly, irrespective of the speed of the data-recording vehicle.
For lateral position, vehicle placement shifted farther from the center line after implementation
of CLRSs. After implementation of ELRSs, researchers noted a decrease of about 50 percent in
shoulder encroachments as well as a significant reduction in other encroachments, including
inadvertent contact with the edge line.
6
Carlson et al. examined wet-night visibility of pavement markings using experimental drivers on
a closed rain tunnel.(7,8) The study tested nine different treatments in random orders and
measured perception distance for each sample location. The driver alerted the researcher when he
or she observed a marking and when the type could be determined. This research included testing
rumble stripes. The findings suggested there was little difference in detection distance between
flat thermoplastic lines and rumble strip lines at low rainfall rates. However, the detection
distance was 13 to 38 percent greater for rumble strip lines for medium and heavy rainfall rates.
The Mississippi Department of Transportation installed ELRSs on a portion of Interstate 59 with

generally encouraging results.(9) Preliminary data indicated that the strategy provided an
excellent audible alert, increased visual awareness of the travel lane, increased reflectivity, and
provided results similar to those for inverted profile striping, and Mississippi residents welcomed
the installation. However, some concerns associated with the rumble stripes arose, including
noise pollution and the potential for an increase in over-correcting and head-on crashes.
Torbic et al. summarized numerous studies on SRSs.(10) Table 1 summarizes their results of
many States’ studies (negative percentages indicate a decrease in crashes), and their NCHRP
report outlines several key findings.(10) The report notes that SRSs installed along freeways made
up the majority of the safety effectiveness evaluations and that only a small percentage of the
studies evaluated the safety effectiveness of nonfreeway installations. While the evaluations
generally focused on crash types most directly affected by rumble strip presence, such as single-
vehicle ROR (SVROR) crashes, several studies looked at the safety effect on total crashes.
Rumble strip application showed an average reduction of 36 percent in SVROR-type crashes,
with a range of 10 to 80 percent. The reduction of total crashes ranged from 13 to 33 percent,
with an average reduction of 21 percent.(10)
7
Table 1. Summary of studies on changes in target collision frequency from application of SRSs.(10)
Percent Change in
Target Collision
Type of Collision Frequency (Standard
State Type of Facility Targeted Deviation) Type of Analysis
Arizona(11) Interstate SVROR −80 Cross-sectional comparison
California(12) Interstate SVROR −49 Before–after with comparison sites
California(12) Interstate Total −19 Before–after with comparison sites
Connecticut(13) Limited-access SVROR −32 Before–after with comparison sites
roadways
Florida(11) — Fixed object −41 Naive before–after
Florida(11) — Ran-into-water −31 Naive before–after
Illinois and Freeways SVROR (total) −18 (± 6.8) Before–after with marked
California(14) comparison sites and a comparison
group
Illinois and Freeways SVROR (injury) −13 (±11.7) Before–after with marked
8

group
Illinois and Rural freeways SVROR (total) −21.1 (±10.2) Before–after with marked
group
Illinois and Rural freeways SVROR (injury) −7.3 (±15.5) Before–after with marked
group
Kansas (unpublished, Freeways SVROR −34 Unknown
cited in Stutts(15))
Maine (16) Rural freeways Total Inconclusive Before–after with comparison sites
Massachusetts — SVROR −42 Unknown
(unpublished, cited in
Stutts(15))
Michigan(17) — SVROR −39 Cross-sectional comparison
8
Percent Change in
Target Collision
Minnesota(18) Rural multilane Total −16 Naive before–after
divided highways
Minnesota(18) Rural multilane Injury −17 Naive before–after
divided highways
Minnesota(18) Rural multilane SVROR (total) −10 Naive before–after
divided highways
Minnesota(18) Rural multilane SVROR (injury) −22 Naive before–after
divided highways
Minnesota(18) Rural multilane Total −21 Before–after with comparison sites
divided highways
Minnesota(18) Rural multilane Injury −26 Before–after with comparison sites
divided highways
Minnesota(18) Rural multilane SVROR (total) −22 Before–after with comparison sites
9
divided highways
Minnesota(18) Rural multilane SVROR (injury) −51 Before–after with comparison sites
divided highways
Minnesota(19) Rural two-lane SVROR (total) −13 (8) Before–after EB analysis with
roads reference group
Minnesota(19) Rural two-lane SVROR (injury) −18 (12) Before–after EB analysis with
roads reference group
Montana(20) Interstate and SVROR −14 Before–after with comparison sites
primary highways
New Jersey — SVROR −34 Unknown
(unpublished, cited in
Stutts(15))
New York(21) Interstate parkway SVROR −65 to 70 Naive before–after
Pennsylvania (22)
Interstate SVROR −60 Naive before–after
Tennessee(23) Interstate SVROR −31 Unknown
Utah (24)
Interstate SVROR −27 Before–after with comparison sites
9
Percent Change in
Target Collision
Utah(24) Interstate Total −33 Before–after with comparison sites
Virginia(25) Rural freeways SVROR −52 Before–after with comparison sites
Washington(26) — Total −18 Naive before–after
Multistate(11) Rural freeways SVROR −20 Before–after with comparison sites
Note: This table is adapted from table 4 in Torbic et al. (2009).(10)
—Information was not available.
10
10
Of all the rumble strip crash reduction studies reviewed by Torbic et al., only one (Patel et al.)
specifically addresses rural two-lane roads.(19) That study focused on roads in Minnesota and
used the EB analysis approach, which is generally more accurate than alternative before–after
analysis types. Therefore, from among the listed sources, the Minnesota study appeared to
provide the most relevant and reliable indications of the potential safety effects of ELRSs. It
estimated a crash reduction of 13 percent (standard error (SE) = 8) for all SVROR crashes and
18 percent (SE = 12) for SVROR injury crashes. It should be noted, however, that these crash
reduction factors only applied to rural two-lane roads with an AADT greater than 4,000.(27)
Torbic et al. examined the safety effectiveness of SRSs on rural two-lane highways.(10) The EB
before–after results indicated no change in crashes after application of SRSs for total crashes and
fatal and injury crashes for combined data from Minnesota, Missouri, and Pennsylvania. The
results indicated a significant 16-percent decrease in SVROR crashes and a significant
36-percent decrease in SVROR fatal and injury crashes at combined sites. Additional analyses
indicated that Pennsylvania had a significant 24-percent reduction in total crashes, 44-percent
decrease in SVROR crashes, and 37-percent decrease in SVROR fatal and injury crashes. In
consideration of all analytical methods employed, Torbic et al. recommended the following
CMFs for SRSs on rural two-lane roads based on their research:(10)
• 0.84 for SVROR crashes.

• 0.64 for SVROR fatal and injury crashes.
In addition, Torbic et al. quantified the impact of SRS placement on safety, focusing on SVROR
fatal and injury crashes. Placement was defined as edge line and non-edge line, which were
compared with no rumble strips. ELRSs were defined as rumble strips with an offset distance of
0 to 8 inches, and non-ELRSs were defined as having an offset of 9 inches or more. For two-lane
rural roadways, there was no significant or practical difference between ELRSs and non-ELRSs.
Also, there was no evidence that suggested SRSs resulted in a reduction of SVROR crashes
involving heavy vehicles.(10)
Khan et al. evaluated the safety benefits of SRSs on rural two-lane highways in Idaho.(28) The
authors conducted an EB before–after analysis using data from 178.63 mi of data from treatment
sites. The results indicated a 14-percent reduction in ROR crashes. Further analysis indicated a
33-percent reduction in ROR crashes for sections with an AADT less than 1,000. In addition,
SRSs were most effective on horizontal tangents and horizontal curves with moderate curvature.
The study found that SRSs were most effective for paved shoulder widths of 3 ft or more.(28)
Potts et al. evaluated the safety impacts of wider pavement markings with both CLRSs and
ELRSs with resurfacing on rural two-lane highways in Missouri.(29) The EB analysis indicated
a significant 47.4-percent reduction in fatal and disabling injury crashes and a significant
38.3-percent reduction in fatal and all injury crashes. A benefit–cost (B/C) evaluation
indicated a B/C ratio of 35.6 for wide markings and both CLRSs and ELRSs with resurfacing
on rural two-lane roadways.(29)
Lyon et al. evaluated the safety impacts of combined SRSs and CLRSs using data from Kentucky,
Missouri, and Pennsylvania.(30) Kentucky data included SRSs and ELRSs, and the final data
included sites where SRSs/ELRSs and CLRSs were installed concurrently as part of a resurfacing
11
effort or where CLRSs had been installed as retrofits. Table 2 provides the dimensions of the
rumble strips implemented in each of the three States. Note that Pennsylvania had two typical
applications for CLRSs and an alternative design for bicycle-tolerable rumble strips.
Table 2. Rumble strip dimensions from Lyon et al.(30)

Width Length Depth Spacing
Location Type (inches) (inches) (inches) (inches)
Kentucky CLRS 7–7.5 12 1
/2–5/8 24
Kentucky SRS 7 ± 1/2 16 1
/2 ± 1/8 12 ± 1
Missouri CLRS 7 ± 1/2 12 7
/16 ± 1/16 12 and 24
Missouri SRS 7 ± 1/2 12 7
/16 ± 1/16 12
Pennsylvania CLRS 1 7 ± 1/2 16 1
/2 ± 1/16 24 and 48
Pennsylvania CLRS 2 7 ± 1/2 14–18 1
/2 ± 1/16 24
Pennsylvania ELRS 5 ± 1/2 6 1
/2 ± 1/16 7
Pennsylvania Bike-tolerable SRS1 5 ± 1/2 16 3
/8 ± 1/16 7
Pennsylvania Bike-tolerable SRS2 5 ± 1/2 16 3
/8 ± 1/16 6
1
Roadway’s posted speed limit was greater than or equal to 55 mi/h.
2
Roadway’s posted speed limit was less than 55 mi/h.
The EB analysis indicated the following significant CMFs for combined States:(30)
• 0.80 for total crashes (excluding intersection-related and animal crashes).

• 0.77 for fatal and injury crashes.
• 0.74 for ROR crashes.
• 0.63 for head-on crashes.
• 0.77 for sideswipe-opposite-direction crashes.
• 0.70 for head-on and sideswipe-opposite-direction crashes.
• 0.73 for ROR, head-on, and sideswipe-opposite-direction crashes.
Further disaggregate analyses by Lyon et al. indicated significant reductions in Kentucky and
Missouri but not in Pennsylvania.(30) The authors surmised that earlier installations (which were
used by Torbic et al.) were higher-crash locations, while more recently treated sites did not have
a high target crash issue (and therefore no safety benefit).(10) Additional analysis by Lyon et al.
indicated the following:(30)
• Larger reductions in ROR crashes for higher traffic volumes (greater than 3,200 AADT).
• Larger reductions in head-on and sideswipe-opposite-direction crashes for lower traffic

volumes (less than 9,200 AADT).
A B/C analysis found an estimated B/C ratio between 20.2 and 54.7 based on estimated service
lives of 7 to 12 years and estimated annual costs of $557 to $1,511/mi.(30)
Sayed et al. evaluated the safety effectiveness of CLRSs and SRSs alone and combined on rural
two-lane and four-lane divided highways in British Columbia using an EB before–after study
design.(31) The combined application on rural two-lane highways resulted in a 21.4-percent
12
reduction in off-road right, off-road left, and head-on collisions combined. For rural two-lane
highways, SRS applications resulted in a 26.1-percent reduction in off-road right collisions, and
CLRS applications resulted in a 29.3-percent reduction in off-road left and head-on collisions.(31)
Torbic et al. evaluated the effect of combined CLRSs and SRSs using data from approximately
80 mi of treated roadways in Mississippi.(32) The target crash types evaluated included SVROR
crashes left or right, sideswipe-opposite-direction crashes, and head-on crashes. Crash severities
evaluated individually included total crashes, fatal and injury crashes, and fatal and serious
injury crashes. The results of the EB before–after analysis indicated a significant 35-percent
reduction in total target crashes, significant 40-percent reduction in fatal and injury target
crashes, and an insignificant 12-percent increase in fatal and serious injury target crashes.(32)
Kay et al. evaluated the safety impacts of CLRSs and combined CLRSs and SRSs on rural two-
lane highways in Michigan.(33) The EB before–after analysis examined approximately 3,000 mi
of CLRS applications and 1,075 mi of combined CLRS and ELRS applications. The results for
CLRSs indicated the following significant reductions (K, A, B, C, and O refer to the KABCO
scale used to represent injury severity in crash reporting where K is fatal injury, A is
incapacitating injury, B is non-incapacitating injury, C is possible injury, and O is property
damage only):(33)
• 15.8 percent for total crashes.

• 27.3 percent for target crashes.
• 52.9 percent for target-wet pavement crashes.
• 1.4 percent for target-wintry pavement crashes.
• 42.8 percent for target-passing crashes.
• 28.8 percent for target-impaired driving crashes.
• 44.2 percent for target-K injury crashes.
• 32.0 percent for target-A injury crashes.
• 39.3 percent for target-B injury crashes.
• 27.9 percent for target-C injury crashes.
• 16.2 percent for target-O crashes.
The results for combined CLRSs and SRSs indicated the following significant reductions:(33)
• 17.2 percent for total crashes.

• 32.8 percent for target crashes.
• 55.6 percent for target-wet pavement crashes.
• 4.6 percent for target-wintry pavement crashes.
• 35.7 percent for target-passing crashes.
• 39.9 percent for target-impaired driving crashes.
• 51.4 percent for target-K injury crashes.
• 32.5 percent for target-A injury crashes.
• 53.7 percent for target-B injury crashes.
• 35.2 percent for target-C injury crashes.
• 28.5 percent for target-O crashes.
13
Target crashes were identified manually as crashes involving a vehicle crossing the center line of
the roadway.
Olson et al. conducted a before–after evaluation of combined CLRSs and SRSs on rural two-lane
highways in Washington.(34) The analyses compared simultaneous installations, installations
where CLRSs were later added to sections with SRSs, and installations where SRSs were later
added to sections with CLRSs. In addition, the authors analyzed composite conditions where
there were no rumble strips in the before period and conditions with both CLRSs and SRSs,
disregarding when they were installed.
For simultaneous installations, the application resulted in a 63.3-percent reduction in lane

departure crashes, a 65.4-percent reduction in crossover crashes, and a 61.4-percent reduction in
ROR right crashes. Installations were noted to be more effective at higher speeds and for sections
with shoulders greater than 4 ft.(34)
For sections where CLRSs were added to SRSs, the application resulted in a 64.7-percent
reduction in crossover crashes and an 8.5-percent increase in ROR right crashes, resulting in a
combined 44.6-percent reduction in lane-departure crashes. For sections where SRSs were added
to CLRSs, the application resulted in a 47-percent reduction in ROR right crashes and a
6.8-percent reduction in crossover crashes, resulting in a 37.2-percent reduction in lane-departure
crashes.(34)
The composite analysis indicated a 66-percent reduction in lane-departure crashes and a

56-percent reduction in fatal and serious injury crashes. The combined application was noted to
be slightly more effective for 11-ft lane widths than 12-ft lane widths.
Kubas et al. evaluated the safety effectiveness of CLRSs and SRSs and SRSs only on rural
two-lane highways in North Dakota.(35) The authors compared before- and after-crash rates to
estimate the effectiveness of rumble strip applications for various crash types. The installation of
CLRSs and SRSs resulted in a 2-percent decrease in total crashes, 45-percent decrease in fatal
crashes, 21-percent increase in injury crashes, 5-percent decrease in property damage only crashes,
and 29-percent decrease in ROR crashes based on a limited sample. The installation of SRSs
resulted in a 15-percent decrease in total crashes, 22-percent decrease in property damage only
crashes, and 97-percent increase in ROR crashes based on a limited sample. It should be noted that
no CMFs from this study received more than a two-star rating in the CMF Clearinghouse.(35)
14
CHAPTER 2. OBJECTIVE
The research described in this report examined the safety impacts of ELRSs on rural horizontal
curves in Kentucky and Ohio with the objective to estimate the safety effectiveness of this
strategy as measured by crash frequency. Excluding intersection-related and animal crashes, the
study included the following crash types:
• Total (all types and severities combined).

• Injury (K, A, B, and C injuries on the KABCO scale).
• ROR (all severities combined).
• Nighttime (including dusk and dawn; all severities combined).
• Nighttime ROR (including dusk and dawn; all severities combined).
A further objective was to address questions of interest including the following:
• Do effects vary by level of traffic volumes?

• Do effects vary by posted speed limit?
• Do effects vary by paved shoulder width?
Other questions included the following:
• Are crash migration effects evident?

• Are spillover effects evident?
The evaluation of overall effectiveness included the consideration of the installation costs and
crash savings in terms of the B/C ratio.
Meeting these objectives placed some special requirements on the data collection and analysis
tasks, including the need to do the following:
• Select a large enough sample size to detect, with statistical significance, what may be
small changes in safety for some crash types.
• Identify appropriate untreated reference sites.
• Properly account for changes in safety due to changes in traffic volume and other
nontreatment factors.
• Pool data from multiple jurisdictions to improve reliability of the results and facilitate
broader applicability of the products of the research.
15
CHAPTER 3. STUDY DESIGN
The study design involved a sample size analysis and prescription of needed data elements. The
sample size analysis assessed the size of sample required to statistically detect an expected
change in safety and also determined what changes in safety could be detected with likely
available sample sizes.
Sample size estimations required assumptions of the expected treatment effect and the average
crash rate at treatment sites prior to treatment. The project team calculated minimum and desired
sample sizes assuming a conventional before–after with comparison group (C-G) study design,
as described in Hauer, and a literature review of likely safety effects.(36) The sample size analysis
undertaken for this study addressed the size of sample required to statistically detect an expected
change in safety. The sample size estimates were conservative because the more robust EB
methodology was actually used in the before–after analysis rather than the C-G methodology.
Sample sizes were estimated for various assumptions of the likely annual crash rate in the before
period and likely safety effects of the strategy. Annual crash rates were assumed for five crash
types (i.e., total, injury, ROR, nighttime, and nighttime ROR) as shown in table 3. Intersection-
related and animal crashes were not included in these crash rates.
The horizontal curve site crash rates for the all and injury crash types were obtained directly
from Torbic et al. (rates A and B) and before-period data from Kentucky (rate C) and Ohio
(rate D).(10) The crash rates for Washington (rate A) and Minnesota (rate B) were selected in
particular because they represented the general upper and lower range of national crash rates. For
instance, estimated crash rates for sites in Pennsylvania and Missouri from the same NCHRP
report were 1.75 total crashes per mi/yr and 2.11 total crashes per mi/yr, respectively, which
were both within that range. The before-period crash rates for Washington and Minnesota were
used for planning purposes during the development of the study design, and the rates for
Kentucky and Ohio were provided to show the actual rates. The before-period rates for Kentucky
and Ohio were greater than those assumed during the planning stages, indicating that sufficient
sample sizes were more achievable.
The Washington and Minnesota crash rates for the ROR crash type were estimated by
multiplying the total crash rate by the ratio of ROR crashes to total crashes based on data from
Washington between 2001 and 2005. The nighttime crash rates were estimated by multiplying
the total crash rate by the ratio of nighttime collisions to total collisions based on 2008 Kentucky
crash data. The nighttime ROR crash rates were estimated by multiplying the nighttime crash
rate by the same ROR crashes ratio.
17
Table 3. Before-period crash rate assumptions.
Rate A Rate B Rate C Rate D

(Washington) (Minnesota) (Kentucky) (Ohio)
(crashes/mi/yr) (crashes/mi/yr) (crashes/mi/yr) (crashes/mi/yr)
Average Site Average Site Average Site Average Site
Crash Type Length 0.164 mi Length 0.142 mi Length 0.068 mi Length 0.073 mi
Total 3.37 0.84 3.37 4.12
Injury 1.52 0.31 1.25 1.69
ROR 0.94 0.23 2.04 3.15
Nighttime 1.15 0.29 0.97 1.54
Nighttime ROR 0.32 0.08 0.67 1.30
Table 4 through table 8 provide estimates of the required number of before- and after-period
mile-years for both the 90- and 95-percent confidence levels on horizontal curve sites by crash
type. The minimum sample indicated the level for which a study seemed worthwhile (i.e., it was
feasible to detect with the level of confidence the largest effect that might reasonably be
expected based on what was currently known about the strategy). These sample size calculations
were based on specific assumptions regarding the number of crashes per mile and years of
available data. Mile-years is the number of miles where the strategy was implemented multiplied
by the number of years of data before or after implementation. For example, if a strategy was
implemented at a 9-mi segment and data were available for the 3 years since implementation,
then a total of 27 mi-years of after-period data would be available for the study.
The sample size values recommended in this study are highlighted with an asterisk in
table 4 through table 8. These were selected based on the likeliness of obtaining the
estimated sample size as well as the anticipated effects of the treatment. As noted, the
sample size estimates provided are conservative in that the state-of-the-art EB
methodology proposed for the evaluations would require fewer sites than the less robust
conventional before–after study with a C-G that was assumed for the calculations.
Estimates can be predicted with greater confidence or a smaller reduction in crashes
would be detectable if there were more site-years of data available in the after period. The
same holds true if the actual data used for the analysis had a higher crash rate for the
before period than had been assumed.
18
Table 4. Minimum required before-period mile-years for treated sites—total crashes.
Expected
Percent
Reduction Rate A Rate B Rate C Rate D Rate A Rate B Rate C Rate D
in Crashes1 (Washington) (Minnesota) (Kentucky) (Ohio) (Washington) (Minnesota) (Kentucky) (Ohio)3
2 2 2 2 3 3 3
10 550 2,208 550 451 342 1,373 342 280

20 83* 332* 83* 68* 57* 230* 57* 47*
30 28 113 28 23 20 80 20 17
40 12 49 12 9 9 35 9 7
1
Assumes equal number of mile-years for treatment and comparison sites and equal length of before and after periods.
2
95-percent confidence level.
3
*Sample size values recommended in this study.
Table 5. Minimum required before-period mile-years for treated sites—injury crashes.

Expected
19
Percent
2 2 2 2 3 3 3
10 1,220 5,984 1,485 1,100 759 3,719 923 683

20 184* 900* 223* 165* 127* 623* 154* 114*
30 63 306 76 56 44 216 53 40
40 27 132 33 25 19 94 23 17
1
2
3
19
Table 6. Minimum required before-period mile-years for treated sites—ROR crashes.
Expected
Percent
2 2 2 2 3 3 3
10 1,973 8,065 910 589 1,227 5,013 565 366

20 297* 1,213* 137* 89* 205* 839* 95* 62*
30 101 413 47 31 71 291 33 22
40 44 178 20 13 31 126 15 10
1
2
3
Table 7. Minimum required before-period mile-years for treated sites—nighttime crashes.

Expected
20
Percent
in Crashes1 (Washington)2 (Minnesota)2 (Kentucky)2 (Ohio)2 (Washington)3 (Minnesota)3 (Kentucky)3 (Ohio)3
10 1,613 6,397 1,912 1,207 1,003 3,976 1,190 750
20 243* 962* 287* 181* 168* 666* 199* 125*
30 83 328 98 62 58 231 69 43
40 36 141 42 27 25 100 30 19
1
2
3
20
Table 8. Minimum required before-period mile-years for treated sites—nighttime ROR crashes.
Expected
Percent
in Crashes (Washington) (Minnesota) (Kentucky) (Ohio) (Washington) (Minnesota) (Kentucky) (Ohio)3
1 2 2 2 2 3 3 3
10 5,797 23,188 2,775 1,427 3,603 14,413 1,724 888

20 872 3,488 416 214 603 2,413 287 148
30 297* 1,188* 142* 73* 209* 838* 99* 51*
40 128 513 61 32 91 363 43 22
1
2
3
21
21
Following the data collection for both the before and after periods, the total mile-years of data
available was 90.38 for the before period and 34.36 for the after period in Kentucky. Ohio had
217.01 mi-yr for the before period and 120.99 mi-yr for the after period. The States are reported
separately because Ohio had additional statewide safety treatments (e.g., in-curve and advance
horizontal curve warning signage) applied at the same time as the ELRS installation. The
statistical accuracy attainable for a given sample size is described by the standard deviations of
the estimated percent change in safety. From this, one can estimate P-values for various sample
sizes and the expected change in safety for a given crash history. A set of such calculations is
shown in table 9 for Kentucky and table 10 for Ohio. The calculations were based on
methodology in Hauer.(36)
For the available data, the minimum percentage changes in crash frequency that could be
statistically detectable at 5- and 10-percent significance levels were estimated using the same
crash rates in table 3. The results indicate that the data should allow detection of the anticipated
crash reduction effects highlighted in table 4 through table 8 (i.e., 20-percent reductions for all
crash types except for nighttime ROR) in Ohio, if such an effect were present. It might be more
difficult to use the Kentucky data to detect the crash reduction effects highlighted in table 4
through table 8. However, as noted previously, the values were conservative because the EB
methodology requires fewer sites than a conventional before–after with C-G methodology.
Using these results, a decision was made to proceed with the evaluation using the data available
at the time.
Table 9. Analysis for crash effects in Kentucky.

90-Percent 95-Percent
Crash Type Confidence Level1 Confidence Level1
Total 20 25
Injury 30 35
ROR 25 30
Nighttime 35 40
Nighttime ROR 40 45
1
Minimum percent reduction detectable for crash rate assumption. Minimum
percent reduction is rounded to nearest 5 percent.
Note: Mile-years in before period = 90.38; mile-years in after period = 34.36.
Table 10. Analysis for crash effects in Ohio.

90-Percent 95-Percent
Crash Type Confidence Level1 Confidence Level1
Total 15 15
Injury 20 20
ROR 15 15
Nighttime 20 20
Nighttime ROR 20 25
1
Minimum percent reduction detectable for crash rate assumption. Minimum
percent reduction is rounded to nearest 5 percent.
Note: Mile-years in before period = 217.01; mile-years in after period = 120.99.
22
CHAPTER 4. METHODOLOGY
The EB methodology for observational before–after studies was used for the evaluation
conducted in this study. This methodology is considered rigorous in that it accounts for
regression-to-the-mean using a reference group of similar but untreated sites. In the process,
safety performance functions (SPFs) were used, which did the following:
• Overcame the difficulties of using crash rates in normalizing for volume differences
between the before and after periods.
• Accounted for time trends.
• Reduced the level of uncertainty in the estimates of safety effect.
• Properly accounted for differences in crash experience and reporting practice in

amalgamating data and results from diverse jurisdictions.
• Provided a foundation for developing guidelines for estimating the likely safety
consequences of a contemplated strategy.
In the EB approach, the change in safety for a given crash type at a site is given in figure 3.
 Safety = λ - π
Figure 3. Equation. Estimated change in safety.
Where:
λ = expected number of crashes that would have occurred in the after period without the
strategy.
π = number of reported crashes in the after period.
In estimating λ , the effects of regression-to-the-mean and changes in traffic volume were

explicitly accounted for using SPFs, relating crashes of different types to traffic flow and other
relevant factors for each jurisdiction based on untreated sites (reference sites). Annual SPF
multipliers were calibrated to account for temporal effects on safety (e.g., variation in weather,
demography, and crash reporting).
In the EB procedure, the SPF is used to first estimate the number of crashes that would be
expected in each year of the before period at locations with characteristics similar to the one
being analyzed (i.e., traffic volume and reference sites). The sum of these annual SPF estimates
(P) is then combined with the count of crashes (x) in the before period at a strategy site to obtain
an estimate of the expected number of crashes (m) before installation, as shown in figure 4.
m = w(P) + (1- w)(x) ,
Figure 4. Equation. EB estimate of expected crashes.
23
Where w is estimated from the mean and variance of the SPF estimate, which is shown in
figure 5.
1 ,
w=
1+ kP
Figure 5. Equation. EB weight.
Where k is constant for a given model.
k is estimated from the SPF calibration process with the use of a maximum likelihood procedure.
In that process, a negative binomial distributed error structure is assumed, with k being the
overdispersion parameter of this distribution.
A factor is then applied to m to account for the length of the after period and differences in traffic
volumes between the before and after periods. This factor is the sum of the annual SPF
predictions for the after period divided by P, the sum of these predictions for the before period.
The result, after applying this factor, is an estimate of λ . The procedure also produces an estimate
of the variance of λ .
The estimate of λ is then summed over all sites in a strategy group of interest (to obtain λ sum) and
compared with the count of crashes observed during the after period in that group (π sum). The
variance of λ is also summed over all sites in the strategy group.
Figure 6 illustrates the estimate of the index of effectiveness ( ).
 sum
 sum

Var( sum ) 
1  
  2
sum 
Figure 6. Equation. Index of effectiveness.
Figure 7 illustrates the standard deviation of  .
 Var ( sum ) Var ( sum ) 

 2   
  2
2sum 
StDev ( )  sum
2
 Var ( sum ) 
1  
  2 
 sum 
Figure 7. Equation. Standard deviation of index of effectiveness.
The percent change in crashes is calculated as 100(1   ); thus, a value of   0.7 with a
standard deviation of 0.12 indicates a 30-percent reduction in crashes with a standard deviation
of 12 percent.
24
CHAPTER 5. DATA COLLECTION
Kentucky and Ohio provided data containing locations and dates of the installation of ELRSs.
These States also provided roadway geometry, traffic volumes, and crash data for both
installation and reference sites. This chapter summarizes the data assembled for the analysis.
Additional details about the design, installation, and maintenance of ELRSs, as well as lessons
learned, can be found in the appendix of this report.
KENTUCKY
This section describes the installation data, reference sites, roadway data, traffic data, crash data,
and treatment cost data for Kentucky sites used in this evaluation.
Installation Data
The Kentucky Transportation Cabinet (KYTC) provided a list of roadway sections where ELRSs
were installed as part of resurfacing projects. The treatment consisted of adding the new
pavement surface, installing milled rumble strips, and painting the edge line on top of the strip.
KYTC resurfaced shoulders along with the travel lanes but did not widen shoulders as part of
this effort. KYTC installed the milled rumble strips with a standard 12-inch width and 1-inch
depth. Installations took place on corridors consisting of both tangents and horizontal curves.
Kentucky staff identified specific curves for treatment sites for this study using the geographic
information system (GIS) roadway curve inventory to select moderately to very sharp curves
(i.e., Highway Performance Monitoring System (HPMS) classes D, E, and F) within these treated
corridors. The final list of treated sites comprised 229 horizontal curves (15.6 mi) where KYTC
installed ELRSs.
Reference Sites
The treatment applied by KYTC had the potential for crash migration or spillover effects. Crash
migration occurs when there is “a transfer of crashes resulting from an improvement rather than
a reduction” (p. 4).(37) Spillover occurs when the safety benefits of a treatment extend to
untreated sites downstream of the treated site.
To detect crash migration and spillover effects, the project team used a two-stage approach
involving two reference groups. In the first stage, the project team selected one reference group
from a large sample of untreated sites less than 5 mi downstream of treated sites and a second
group from a limited sample of untreated sites located greater than 5 mi downstream of treated
sites. Two reference groups were selected in this way to observe potential crash migration or
spillover effects in the first reference group. According to a simulator study involving drowsy
drivers and rumble strips, signs of drowsiness return to drivers approximately 5 min after hitting
a rumble strip.(38) Assuming a 5-min drowsiness-relapse time and vehicle speed equal to or less
than 60 mi/h, the project team observed crash migration and spillover effects within 5 mi of a
treated site. By comparing the crash data of these two reference groups, the existence and
magnitude of the crash migration and spillover could be detected. If no crash migration or
spillover effects were detected, then the two reference groups were pooled together to form the
reference group for the EB method.
25
Roadway Data
Roadway data provided by Kentucky staff were in GIS shapefile format. The various road
characteristics (e.g., shoulder width) were contained in separate shapefiles, each segmented
differently. GIS files were obtained from the Kentucky Roadway Information and Data website.
The most useful file segmented the Kentucky road network into curve and tangent sections,
likely for HPMS purposes. Each curve was denoted along with its degree of curvature and
HPMS curve classification (A–F). Spatial joining was used to glean the available roadway
characteristic information: shoulder type and width, traffic volume, and degree of curvature. To
obtain many other roadway characteristics, such as area type, number of lanes, illumination,
rumble strip presence, and roadside hazard rating, the data collectors used Google® Maps™ and
Google® Street View™ imagery. To locate and view the curve in Google® Maps™, the
coordinates of the curve were extracted from the GIS map, imported to Microsoft® Excel, and
concatenated into a hyperlink that could be used to quickly find that location in Google®
Maps™.
Traffic Data
KYTC maintains traffic volume data in the GIS inventory files, specifically the Traffic Flow
(TF) shapefile. The project team obtained traffic data for the treatment and reference sites by
spatially joining the TF layer to the site layer to obtain the current and past years’ AADT values.
Crash Data
Crash data for Kentucky are publicly available on the Kentucky State Police’s crash data
website.(39) The project team used the following specifications for crash queries for each route:
• Crash query dates were from 1/1/2004 to 12/31/2012.
• A crash study area was defined for each horizontal curve, which consisted of up to
0.05 mi on each approach of the horizontal curve. If two curves were closer than 0.10 mi,
then the study area was defined as the midpoint between the curves.
• Separate files for collisions, units, and individuals were obtained for each site.
Treatment Cost Data
KYTC provided estimates of the costs and service lives of the treatments for use in conducting a
B/C analysis of the treatment (table 11).
Table 11. Kentucky treatment cost and service life data.

Initial Installation
Countermeasure Cost Maintenance Cost Service Life
Edge-line or shoulder $2,500/mi for rumble No additional 12–15 years for
rumble strips (installed strip, $305/mi for maintenance cost rumble strip,
as part of resurfacing) stripe 2 years for stripe
26
OHIO
This section describes the installation data, reference sites, roadway data, traffic data, crash data,
and treatment cost data for Ohio sites used in this evaluation.
Installation Data
The Ohio Department of Transportation (ODOT) provided a list of roadway sections where
ELRSs were installed on nonfreeway highways. The Highway Safety Information System
(HSIS) provided data identifying rural two-lane segments, and extents of horizontal curves. The
project team verified installations using ODOT’s video logs as well as Google® StreetviewTM.
Most installations were confirmed but several treatment sites were noted to have not received the
ELRS installation. The final list of treated sites comprised 579 horizontal curves (42.3 mi) where
ELRSs were installed.
Reference Sites
As with the Kentucky data, the project team used two reference groups in Ohio to account for
spillover or crash migration effects. The project team selected both reference groups (within 5 mi
of treatment sites and more than 5 mi downstream of treated sites) from the list of installation
sites that were not actually treated and were upstream/downstream of installation sites on the
same corridors. The final list of reference sites comprised 428 horizontal curves (26.1 mi).
Roadway Data
Roadway data were obtained from the HSIS for each study year. Requisite roadway data for
identifying study sites included functional classification, number of lanes, and extents of
horizontal curvature. Additional data included degree of curvature, posted speed limit, and
shoulder width. The project team used route, beginning milepost, and ending milepost to merge
roadway data with traffic data and crash data.
Traffic Data
The project team obtained traffic data from the HSIS for each study year. Traffic data were
obtained for the treatment and reference sites to obtain current and past years’ AADT values.
While data have not been collected every year, HSIS data included AADT information for each
year in the study.
Crash Data
The project team obtained Ohio crash data from the HSIS for each study year. The project team
used the following specifications for crash queries for each route:
• Crash query dates were from 1/1/2005 to 12/31/2013.
• A crash study area was defined for each horizontal curve, which consisted of up to
0.05 mi on each approach of the horizontal curve. If two curves were closer than 0.10 mi,
then the study area was defined as the midpoint between the curves.
27
Treatment Cost Data
ODOT provided estimates of the costs and service lives of the treatment for use in conducting a
B/C analysis of the treatment (table 12).
Table 12. Ohio treatment cost data.

Initial Installation
Countermeasure Cost Maintenance Cost Service Life
Edge-line or shoulder $850/mi for one side No additional Time until next
rumble strips (installed of roadway maintenance cost resurfacing
as part of resurfacing)
DATA CHARACTERISTICS AND SUMMARY
Table 13 defines the crash types used by each State. The project team attempted to make the
crash type definitions consistent. In all States, intersection-related and animal-related crashes
were excluded.
Table 13. Definitions of crash types.

Crash Types Kentucky Ohio
Total Identified as non-intersection and non- Identified as non-intersection and non-
ramp and excluding those where animal related
“Event Collision With” indicated an
animal or deer involvement
Injury Resulted in an injury or possible injury Resulted in a fatal injury,
incapacitating injury, non-
incapacitating injury, or possible injury
ROR “Event Collision With” indicates an “Accident Type” described as other
object off roadway was struck, and non-vehicle, fixed object, other object,
“Pre-Collision Action” is “avoiding other non-collision
object in roadway,” “going straight
ahead,” or “slowing or stopped”
Night Identified including “Dusk,” “Dawn,” Identified including “Dusk,” “Dawn,”
“Dark” (light), or “Dark (no light)” “Dark-No-Lights,” or “Dark-Lighted.”
Nighttime ROR Identified as being nighttime and ROR Identified as being nighttime and ROR
Table 14 provides summary information for the data collected for the treatment sites. The
information in table 14 should not be used to make simple before–after comparisons of crashes
per mile-year because it does not account for factors (other than the strategy) that might cause a
change in safety between the before and after periods. Such comparisons were properly done
with the EB analysis, as presented later. Table 15 and table 16 provide summary information for
the reference site data for Kentucky and Ohio, respectively. As discussed previously, separate
reference groups were established to identify potential spillover and crash migration effects.
28
Table 14. Data summary for treatment sites.
Variable Kentucky Ohio
Number of miles 15.59 42.25
Mile-years before 90.38 217.01
Mile-years after 34.36 120.99
Total crashes/mile/year before 3.37 4.12
Total crashes/mile/year after 2.50 3.35
Injury crashes/mile/year before 1.25 1.69
Injury crashes/mile/year after 0.76 1.36
ROR crashes/mile/year before 2.04 3.15
ROR crashes/mile/year after 1.46 2.52
Nighttime crashes/mile/year before 0.97 1.54
Nighttime crashes/mile/year after 0.61 1.19
Nighttime ROR crashes/mile/year before 0.67 1.30
Nighttime ROR crashes/mile/year after 0.49 0.94
AADT before Average 1,589 Average 2,784
Minimum 412 Minimum 240
Maximum 4,268 Maximum 15,670
AADT after Average 1,500 Average 2,659
Average paved shoulder width (ft) Average 1.36 Average 3.46
Minimum 0.00 Minimum 0.00
Maximum 2.00 Maximum 10.00
Average degree of curve Average 23.38 Average 10.03
29
Table 15. Data summary for Kentucky reference sites.
Variable <5 mi >5 mi
Mile-years 24.19 49.27
Total crashes/mile/year 5.13 1.12
Injury crashes/mile/year 1.53 0.43
ROR crashes/mile/year 2.94 0.57
Nighttime crashes/mile/year 1.24 0.28
Nighttime ROR crashes/mile/year 0.87 0.24
AADT Average 1,475 Average 403
Table 16. Data summary for Ohio reference sites.

Ohio
Variable
<5 mi >5 mi
Mile-years 76.05 159.03
Total crashes/mile/year 6.14 4.42
Injury crashes/mile/year 2.41 1.97
ROR crashes/mile/year 4.21 3.56
Nighttime crashes/mile/year 2.25 1.65
Nighttime ROR crashes/mile/year 1.66 1.44
AADT Average 3,383 Average 2,819
30
CHAPTER 6. DEVELOPMENT OF SAFETY PERFORMANCE FUNCTIONS
This chapter presents the SPFs developed for each crash type. The SPFs were used in the EB
methodology to estimate the safety effectiveness of this strategy.(36) Generalized linear modeling
was used to estimate model coefficients assuming a negative binomial error distribution, which is
consistent with the state of research in developing these models. In specifying a negative
binomial error structure, the dispersion parameter, k, was estimated iteratively from the model
and the data. For a given dataset, smaller values of k indicate relatively better models.
CRASH SPILLOVER AND MIGRATION
Before developing SPFs, the project team analyzed the separate reference groups to identify
potential crash migration and spillover effects. An SPF was developed using data from both
reference groups in order to develop yearly multipliers for each group. The form of the SPF is
provided in figure 8, with parameter estimates presented in table 17.
𝐶𝑟𝑎𝑠ℎ𝑒𝑠/𝑚𝑖𝑙𝑒/𝑦𝑒𝑎𝑟 = 𝑒 𝑎 ∗ 𝐴𝐴𝐷𝑇 𝑏 ∗ 𝐿𝑐 ∗ 𝑒 (𝑖𝑛𝑣𝑟𝑎𝑑𝑖𝑢𝑠 ∗𝑑+𝑟𝑖𝑔 ℎ𝑡_𝑠ℎ𝑜𝑢𝑙𝑑𝑒𝑟 ∗𝑒)
Figure 8. Equation. SPF model form for crash migration or spillover effects.
Where:
AADT = annual average daily traffic volume.
L = segment length (mi).
invradius = inverse of the horizontal curve radius (ft).
right_shoulder = right shoulder width (ft).
a, b, c, d, e = parameters estimated in the SPF calibration process.
k = overdispersion parameter of the model.
Table 17. Reference group SPF for total crashes—parameter estimates.

State a (SE) b (SE) c (SE) d (SE) e (SE) k
−9.063 1.288 0.391 27.738 −0.298
Kentucky 2.183
(0.843) (0.111) (0.219) (37.716) (0.159)
−5.226 0.634 0.390 206.883 −0.069
Ohio 1.022
(0.442) (0.046) (0.057) (47.521) (0.019)
Table 18 presents the observed crashes versus predicted crashes for each of the two reference
groups in Kentucky. Group 1 was the reference group more than 5 mi downstream, and group 2
was the reference group immediately downstream of the treatment sites. Group 1 also included
reference sites that were on different roadways than the treatment sites. The yearly factors were
the ratio of observed crashes to predicted crashes for the given group within the given year.
Spillover and crash migration effects would be apparent if the yearly factors became drastically
different between group 1 and group 2 after treatment application (2010 for most sites). These
effects would also become apparent if the yearly factor for group 2 increased or decreased
markedly after treatment application. However, neither of these scenarios appeared to be the
case. Figure 9 provides a graphical representation of the yearly factors from table 18.
31
Table 18. Observed and predicted crashes for reference groups in Kentucky.
Group 1 Group 2 Group 1 Group 2 Group 1 Group 2
Year Observed Observed Predicted Predicted Factor Factor
2004 5 9 6.072 12.806 0.823 0.703
2005 4 13 6.074 12.817 0.659 1.014
2006 5 14 6.077 12.827 0.823 1.091
2007 8 11 6.103 12.875 1.311 0.854
2008 5 12 6.166 12.928 0.811 0.928
2009 5 12 6.265 13.001 0.798 0.923
2010 7 14 6.368 13.065 1.099 1.072
2011 3 23 5.875 12.515 0.511 1.838
2012 13 16 5.446 12.007 2.387 1.333
Source: FHWA.
Figure 9. Graph. Yearly multiplier by year for reference groups in Kentucky.
The sample sizes for reference sites were too small to make a statistical observation of difference
between the groups (i.e., fewer than 20 crashes per year were observed). However, both groups
experienced an abnormally large number of total crashes in 2011 or 2012, and group 1 observed
an abnormally low number of crashes in 2011. Therefore, data for 2011 and 2012 were removed
for the reference sites for SPF development. Before-period data from treatment sites were
combined with reference-site data to bolster sample size, and the project team estimated an
interaction term for the pretreatment site indicator and AADT to determine whether there was a
difference in the effect of traffic volume at treatment sites in the before period versus at the
references sites. The results indicated that the reference data and pretreatment data could be
combined for SPF estimation.
32
In addition, because of the lack of data at the reference curves in this dataset, the project team
sought an alternative reference group for developing annual factors. The Kentucky reference data
from Lyon et al. provided a robust dataset for estimating annual factors for the before and after
periods from combined horizontal curves and tangents.(30) This reference set included sections
that were eligible for resurfacing but had not yet received resurfacing and had texturing; this was
consistent with the treatment group. The reference set also included those identified by Lyon et
al. to be “resurfacing effort” sites but excluded sites that had SRSs in the prior condition.(30) In
total, the reference group consisted of 401.21 roadway mi. The project team used the reference
segments to develop predicted crashes, which were compared with observed crashes to develop
annual factors. The after-period factor was 1.034 for 2009 installations, 0.993 for 2010
installations, and 0.982 for 2011 installations. Table 19 presents the observed crashes versus
predicted crashes for each of the two reference groups in Ohio. As with Kentucky, group 1
represents the reference group more than 5 mi downstream, and group 2 represents the reference
group immediately downstream of the treatment sites. Figure 10 provides a graphical
representation of the yearly factors from table 19. The annual factors show that the reference
sites, particularly those not adjacent to treatment sites, observed a substantial reduction,
beginning in 2011, relative to the before period. This possibility was expected for group 2 but not
for group 1.
The project team explored this finding further with statewide data for rural two-lane roadways
and for horizontal curves on rural two-lane highways. Consistently, the trends showed an
approximate 10- to 15-percent reduction in crashes beginning in 2011. The project team
contacted ODOT to gain further insight into this finding to determine why this systematic
reduction was observed statewide. ODOT noted that in 2010, the systemic program focused on
upgrading signage for horizontal curves statewide. This included upgrading hundreds of curves.
The upgraded signage included chevrons, curve ahead signs, and speed advisory signs, among
others. Therefore, the project team considered using horizontal tangent sections as a potential
reference group for developing annual factors to mitigate the impact of signage upgrades on the
overall findings. Spillover effects were observed for short tangents; the annual multipliers for
short tangents (i.e., tangents less than 0.5 mi in length) were found to match those of horizontal
curves within 1.0 percent for each installation year. Tangents longer than 1.0 mi were used as a
reference group for developing annual factors because no spillover effects were observed. The
reference segments were used to develop predicted crashes, which were compared with observed
crashes to develop annual factors. The after-period factor was 1.040 for 2010 installations, 1.043
for 2011 installations, and 1.038 for 2012 installations.
33
Table 19. Observed and predicted crashes for reference groups in Ohio.
Group 1 Group 2 Group 1 Group 2 Group 1 Group 2
Year Observed Observed Predicted Predicted Factor Factor
2005 81 49 84.82 46.24 0.954963 1.059689
2006 83 60 85.75 45.87 0.96793 1.308044
2007 72 38 84.47 45.37 0.852374 0.837558
2008 79 66 83.64 43.95 0.944524 1.501706
2009 87 49 81.15 44.11 1.072089 1.110859
2010 90 68 90.33 44.56 0.996347 1.526032
2011 58 52 90.33 44.61 0.64209 1.165658
2012 67 47 88.88 44.13 0.753825 1.065035
2013 86 38 89.71 43.88 0.958645 0.865998
Source: FHWA.
Figure 10. Graph. Yearly multiplier by year for reference groups in Ohio.
SAFETY PERFORMANCE FUNCTIONS
Figure 11 shows the form of the SPFs for combined reference groups, which are presented in
table 20.
Figure 11. Equation. SPF model form for Kentucky and Ohio.
Where:
posted = posted speed.
f = parameter estimated in the SPF calibration process.
34
The Kentucky SPF was estimated using treatment site before period data and 2004 to 2010
reference site data. The Ohio SPFs were estimated using reference site data from 2005 to 2010.
Table 20. SPFs by crash type.

Ohio Ohio
Kentucky Total ROR
Parameters Total (SE) (SE) (SE)
a −10.338 −6.962 −6.443
(0.573) (0.813) (0.927)
b 1.249 0.697 0.553
(0.072) (0.060) (0.066)
c 0.128 0.365 0.433
(0.100) (0.069) (0.078)
d N/A 236.022 282.222
(58.139) (63.957)
e N/A −0.063 −0.099
(0.024) (0.028)
f N/A 0.021 0.0319
(0.011) (0.013)
k 0.738 1.083 1.289
N/A = not applicable.
In addition, the project team considered crash sample size for reference sites in the development
of SPFs. In Kentucky, total crashes per year ranged from 65 to 77. Other crash types had smaller
sample sizes. The Highway Safety Manual recommends a minimum of 100 crashes per year to
produce reliable SPFs.(40) Therefore, an SPF was developed for total crashes, and proportion
factors relating other crash types to total crashes were used in place of separate SPFs. In Ohio,
total crashes ranged from 110 to 158. There were sufficient crashes to develop separate SPFs for
total crashes and ROR crashes. To relate other crash types to total crashes, proportion factors
were used in place of separate SPFs. The prediction from the SPF was multiplied by the
proportion factor to determine the number of predicted crashes of each specific crash type. The
following is a list of crash type proportions for Kentucky:
• Fatal and injury crashes = 0.370.

• ROR crashes = 0.603.
• Nighttime crashes = 0.289.
• Nighttime ROR crashes = 0.200.
The following is a list of crash type proportions for Ohio:
• Fatal and injury crashes = 0.403.

• Nighttime crashes = 0.373.
• Nighttime ROR crashes = 0.314.
In addition, observed crashes and predicted crashes were used to develop annual factors for time-
based trends at reference sites. Factors were used as multipliers for predicted crashes at treatment
35
sites in the after period. Factors greater than 1.00 indicate an increase in crashes at reference
sites, and factors less than 1.00 indicate a decrease in crashes at reference sites. Table 21
provides the annual factors based on total crashes in Kentucky and Ohio for each installation
year.
Table 21. Time trend factors for predicted crashes.

Kentucky Ohio
Year Total Total
2009 1.034 N/A
2010 0.993 1.040
2011 0.982 1.043
2012 N/A 1.038
N/A = not applicable.
36
CHAPTER 7. BEFORE–AFTER EVALUATION RESULTS
This chapter presents the results of the before–after evaluation, including aggregate analysis for
both Kentucky and Ohio and disaggregate analysis of the Ohio data. Disaggregate analysis of the
Kentucky data was not conducted because the sample size was too small.
AGGREGATE ANALYSIS
Table 22 provides the estimates of expected crashes in the after period without treatment, the
observed crashes in the after period, and the estimated CMF and its SE for all crash types
considered in Kentucky. Table 23 presents the results for Ohio.
Table 22. Aggregate analysis results for Kentucky.

Nighttime
Statistic Total Injury ROR Nighttime ROR
EB estimate of crashes expected 113.9 40.8 67.6 33.1 22.5
in the after period without
strategy
Count of crashes observed in the 86 26 50 21 17
after period
Estimate of CMF 0.75* 0.64* 0.74* 0.63* 0.75
SE of estimate of CMF 0.09 0.14 0.11 0.14 0.19
*Statistically significant results at the 95-percent confidence level.
Table 23. Aggregate analysis results for Ohio.

Nighttime
EB estimate of crashes expected 514.2 208.6 392.7 191.6 160.1
in the after period without
strategy
Count of crashes observed in the 405 165 305 144 114
after period
Estimate of CMF 0.79* 0.79* 0.78* 0.75* 0.71*
The results for Kentucky indicated statistically significant reductions for all crash types except
nighttime ROR crashes at the 95-percent confidence level. Nighttime crashes had the smallest
CMF (which translates to the greatest reduction) with a value of 0.63. Total, injury, and ROR
crashes had CMFs of 0.75, 0.64, and 0.74, respectively. The CMF for nighttime ROR crashes
was 0.75 and was consistent with the same CMF from Ohio; however, it was significant only at
the 80-percent level, suggesting that sample size was the reason for the lack of statistical
significance at the 95-percent confidence level. The CMFs were smaller than—but consistent
with—those found in the most comprehensive and reliable study of SRSs to date.(10) Based on a
before–after EB analysis, the project team found that milled SRSs had a crash reduction of
37
16 percent (SE = 8) for all SVROR crashes and 36 percent (SE = 10) for SVROR injury
crashes.(10) However, the analysis results for SRSs in Torbic et al. considered segments with both
horizontal tangents and curves; therefore, a direct comparison of results cannot be made.(10) It is
also important to remember that all crash types considered in this research excluded intersection-
related and animal crashes.
The results for Ohio indicated statistically significant reductions for all crash types. Nighttime
ROR crashes had the smallest CMF with a value of 0.71. Total, injury, ROR, and nighttime
crashes had CMFs of 0.79, 0.79, 0.78, and 0.75, respectively. As with the Kentucky results, the
CMFs were smaller than but consistent with those found in Torbic et al.(10) The resulting Ohio
installation CMFs reflected the installation of ELRSs on horizontal curves as well as the impact
of the statewide signing program.
A subset of both treatment and reference curves received sign upgrades, including chevrons,
curve ahead signs, and speed advisory signs, all of which target crash types (i.e., nighttime and
ROR) similar to those targeted by ELRSs but through a different mechanism (i.e., rumble strips
target distracted or drowsy drivers through a haptic alert). The initial set of reference sites
accounted for the impact of the signing upgrades. Additional analyses of the reference sites
indicated a spillover effect of the horizontal curve signing program on curves that did not receive
treatments as well as shorter tangents; therefore, longer tangent segments were used to determine
the expected trend in the after period had no treatment (i.e., the signing program or ELRS
installation) occurred. Owing to the spillover effects of the signing program, further analyses
involving curves that specifically received new or additional signs were not fruitful (i.e., the
resulting CMFs could not separate the effects of the signing program from those resulting from
ELRS installation).
DISAGGREGATE ANALYSIS OF OHIO DATA
The disaggregate analysis sought to identify those conditions under which the treatment was
most effective. Because ROR, nighttime, and nighttime ROR crashes were the focus of this
treatment, the project team focused on these crash types for the disaggregate analysis. In
addition, disaggregate results are presented for total crashes and fatal and injury crashes. The
data sample for Kentucky was too small to perform disaggregate analyses; therefore,
disaggregate analyses focused only on Ohio data.
Several variables were identified as being of interest and available for both States, including
degree of curve, posted speed limit, paved shoulder width, lane width, AADT, and before-period
expected crash frequency. Disaggregate results are provided by AADT in table 24 and before-
period expected crash frequency in table 25. The number of crashes in the after period is
presented for each CMF to indicate the sample size available. Several of the estimated CMFs
rely on small samples, especially for nighttime crashes and nighttime ROR crashes.
38
Table 24. Ohio results by AADT.
<4,000 <4,000 4,000+ 4,000+
Crash Type Observed CMF (SE) Observed CMF (SE)
Total 289 0.82* (0.06) 116 0.72* (0.08)
Injury 118 0.82* (0.08) 47 0.72* (0.12)
ROR 239 0.82* (0.06) 66 0.64* (0.09)
Nighttime 105 0.79* (0.08) 39 0.66* (0.12)
Nighttime ROR 88 0.78* (0.09) 26 0.54* (0.11)
Table 25. Ohio results by before-period expected crash frequency.

Greater Than Greater Than
Expected Less Than Less Than or Equal to or Equal to
Crash Value— Value— Value— Value—
Crash Type Frequency Observed CMF (SE) Observed CMF (SE)
Total 0.25 136 1.09 (0.11) 269 0.69* (0.05)
Injury 0.10 63 1.00 (0.14) 102 0.70* (0.08)
ROR 0.20 111 1.13 (0.12) 194 0.66* (0.05)
Nighttime 0.15 85 0.93 (0.11) 59 0.59* (0.08)
Nighttime
0.075 38 0.85 (0.15) 76 0.66* (0.08)
ROR
As shown in table 24, smaller CMFs (i.e., larger safety benefits) were found for all crash types
for sites with an AADT of 4,000 or more vehicles per day; however, the 95-percent confidence
intervals overlap for each crash type. At AADTs lower than 4,000 vehicles per day, for example,
an ROR crash CMF of 0.82 was estimated versus a CMF estimate of 0.64 for AADTs of
4,000 vehicles per day or greater. A similar difference was found for all other crash types. The
4,000 vehicles per day AADT cutoff is consistent with previous research by Patel et al. and
Lyon et al.(19,30)
For the before-period expected crash frequency, as shown in table 25, the project team found
larger safety benefits for all crash types for higher before-period expected crash frequency. The
95-percent confidence intervals did not overlap for total crashes and ROR crashes. Owing to the
differences in the frequencies of different crash types, the before-period expected crash
frequency cutoff varied for each crash type. For example, an ROR crash CMF of 1.13 was
estimated for horizontal curves with an ROR before-period expected crash frequency of less than
0.20 crashes/yr. This can be compared with a CMF of 0.66 for horizontal curves with 0.20 or
more before-period expected crashes/year. Note that the CMF of 1.13 for an ROR before-period
expected crash rate less than 0.20 is not statistically significant. Similar results were found for all
other crash types.
Caution should be used in interpreting and applying these disaggregate CMF results because of
correlation among variables and because they were not robust enough to develop crash
modification functions. A crash modification function is a formula used to compute the CMF for
a specific site as a function of its site-specific characteristics. For example, crash modification
39
functions would allow the estimation of CMFs for different levels of AADT and before-period
crash frequency. In addition, the disaggregate analysis results used the EB analysis data, which
include the effects of the statewide horizontal curve signing program. However, the disaggregate
analysis CMF results can be used to inform the process of prioritizing treatment sites for ELRSs.
For example, sites with a high proportion of ROR crashes and high AADTs could have high
priority for receiving this treatment because those are the sites likely to benefit the most.
40
CHAPTER 8. ECONOMIC ANALYSIS
The project team conducted an economic analysis to estimate the B/C ratio for this strategy on
rural two-lane horizontal curves. For the purposes of the economic analysis, the assumed
treatment was the application of ELRSs. The project team used the recommended CMFs of
0.75 for Kentucky and 0.79 for Ohio to estimate the benefit of this treatment strategy. The Ohio
results likely included the impact of additional sign upgrades, which were not provided or
considered in this analysis. In addition, the cost of pavement markings was not considered in the
study because these markings were already present on the roadway and the ELRSs did not affect
the lifespan of edge-line pavement markings. Treatment costs ranged from $1,700/mi for Ohio to
$2,500/mi for Kentucky. For Kentucky, service life was estimated as 12 to 15 years. For Ohio,
ODOT noted that the service life was as long as the pavement life, which was assumed as 7 to
10 years. A conservative value of 12 years was assumed for Kentucky and 7 years for Ohio.
The FHWA Office of Safety Research and Development suggested using the Office of
Management and Budget’s Circular A-4 to determine the conservative real discount rate of
7 percent.(41) This value was applied to calculate the annual cost of the treatment for 12- and 7-year
service lives in Kentucky and Ohio, respectively. With this information, the capital recovery factor
was computed to be 7.94 for a 12-year service life and 5.39 for a 7-year service life.
At the time of this report, the most recent FHWA mean comprehensive crash costs disaggregated
by crash severity, location type, and speed limit were based on 2001 dollar values.(42) The
2001 unit costs for property damage only and fatal and injury crashes from the FHWA report
($7,428 and $158,177, respectively) were multiplied by the ratio of the 2014 (when the analysis
was performed) value of a statistical life of $9.2 million to the 2001 value of a $3.8 million.(43,44)
The project team applied this ratio of 2.42 to the unit costs for property damage only and fatal
and injury crashes. The results were weighed by the frequencies of these two crash types in the
after period to obtain aggregate 2014 unit costs for total crashes for Kentucky and Ohio. The
resulting values were $128,268 and $166,603, respectively.
The total crash reduction was calculated by subtracting the actual crashes in the after period from
the expected crashes in the after period had the treatment not been implemented. The number of
crashes saved per mile-year was 0.812 in Kentucky and 0.913 in Ohio. These numbers were
obtained by dividing the total crash reduction by the number of after period mile-years per site.
The annual benefit (i.e., crash savings) of $104,165 in Kentucky and $150,368 in Ohio was the
product of the crash reduction per mile-year and the aggregate cost of a crash (all severities
combined). The B/C ratio was calculated as the ratio of the annual benefit per mile to the annual
cost per mile. The B/C ratio was estimated as 331:1 in Kentucky and 477:1 in Ohio. USDOT
recommended a sensitivity analysis be conducted by assuming values of a statistical life of 0.57
and 1.41 times the recommended 2014 values.(43) These factors were applied directly to the
estimate B/C ratios to obtain a range of 189:1 to 467:1 for Kentucky and 272:1 to 672:1 for
Ohio. On first inspection, the B/C ratios were larger than would reasonably be expected for an
installation of this type. However, the installations took place on corridors, while the analysis
only examined the safety effects on horizontal curves. Horizontal curves have higher crash rates
than overall corridors, and the cost per mile of installation would not be representative for
41
installations only on horizontal curves. As a curve-specific treatment, the B/C ratio would likely
be reduced owing to the higher installation cost; however, these results suggest that the
treatment can be highly cost effective.
42
CHAPTER 9. SUMMARY AND CONCLUSIONS
The objective of this study was to perform a rigorous before–after evaluation of the safety
effectiveness, as measured by crash frequency, of ELRSs applied on rural two-lane horizontal
curves. The study used data from Kentucky and Ohio to examine the effects for specific crash
types, including total, fatal and injury, ROR, nighttime, and nighttime ROR crashes. Crashes
occurring at or related to intersections, as well as animal-related crashes, were not included. Based
on the aggregate results, table 26 and table 27 show the recommended CMFs for the various crash
types. Note that the results for Kentucky were based on smaller sample sizes and that the results
for Ohio included the effects of a statewide horizontal curve warning sign upgrade program.
Table 26. Recommended CMFs for ELRSs based on Kentucky data.

Nighttime
Estimate of CMF 0.75* 0.64* 0.74* 0.63* 0.75
SE error of estimate of CMF 0.09 0.14 0.11 0.14 0.19
*Statistically significant results at the 95-percent confidence.
Table 27. Recommended CMFs for ELRSs and curve signage based on Ohio data.
Nighttime
Estimate of CMF 0.79* 0.79* 0.78* 0.75* 0.71*
*Statistically significant results at the 95-percent confidence.
To date, the most comprehensive and reliable study of SRSs was published by Torbic et al.(10)
Compared with the results of that study for ELRSs, the results of the current study suggest that
greater reductions in all crash types may be found by placing rumble strips on or near the edge
line for horizontal curves.
A disaggregate analysis of the results of the current study indicated that larger safety benefits
were found for horizontal curves with AADT greater than 4,000 for all crash types; however, the
differences by AADT were not statistically significant at the 95-percent confidence level. The
disaggregate analysis further indicated larger safety benefits for horizontal curves with a higher
before-period expected crash frequency. The results suggested no benefit for curves with low
before-period expected crash frequencies for all crash types. The difference in CMFs that were
dependent on before-period expected crash frequency were statistically significant for total and
ROR crashes. Caution should be used in interpreting and applying these disaggregate results;
however, the disaggregate analysis CMFs may be used in prioritizing treatment sites.
Estimated B/C ratios range from 189:1 to 467:1 for Kentucky and from 272:1 to 672:1 for Ohio.
On first inspection, the B/C ratios were larger than would reasonably be expected for an
installation of this type. However, the installations took place on corridors, while the analysis
only looked at the safety effects on horizontal curves. Horizontal curves have higher crash rates
than overall corridors, and the cost per mile of installation would not be representative for
installations only on horizontal curves. For a curve-specific treatment, the B/C ratio would
43
likely be reduced owing to the higher deployment cost for spot-specific installations.
Regardless, these results suggest that the treatment, even in its most expensive variation, can be
highly cost effective.
44
APPENDIX: ADDITIONAL INSTALLATION DETAILS
The following appendix presents additional details provided by Kentucky and Ohio. The States
were asked to provide responses to the following questions:
1. What was the “before-period” condition for the treatment sites with respect to center-line
and edge-line rumble strips? (No rumble strips? Center-line rumble strips only? Edge-line
rumble strips only? Or some combination of these?)
2. Do you know whether the treatment sites analyzed by this study were installed as a
RETROFIT, through RESURFACING, or by a combination of these?
3. The installation dates for the treatment sites on our list range from 2009 to 2011. What
type of rumble strip was installed at these treatment sites? (If there is more than one,
please indicate all that apply.)
a. Milled.
b. Rolled.
c. Formed.
d. Raised.
e. Other.
4. We would like to provide a summary of the rumble strip characteristics below. Would
you have any standard drawings—dated 2009/2010—that applied to all of the treatment
sites considered in this study? If not, would you be able to identify the following
characteristics for the edge-line rumble strips at the study sites?
a. Width.
b. Length.
c. Depth.
d. Spacing.
e. Pavement marking type.
f. Pavement marking width.
5. Were there any other requirements (e.g., minimum paved shoulder width, pavement
structure, etc.) for the installation of rumble strips at the study sites?
6. Were any other safety countermeasures (besides RESURFACING, if that was your
answer to no. 2 above) installed in conjunction with the rumble stripes at the treatment
sites evaluated by this study?
7. Please describe any notable challenges related to the rumble stripe installation and how
you overcame them.
8. Please describe any notable challenges related to the rumble stripe maintenance and how
you overcame them.
45
9. What lessons learned or recommendations would you share with another States interested
in the widespread application of edge-line rumble stripes?
RESPONSES FROM KENTUCKY
Kentucky staff responded to all questions. Their responses are listed in numeric order.
1. The “before condition” had no rumble strips.
2. The treatment sites analyzed by this study were a combination of RETROFIT and
RESURFACING.
3. The installations were all milled rumble strips.
4. The following dimensions were used for the ELRS based on the pavement cross section
in the ELRS Standard Drawings provided by Kentucky:
a. Width: 7 inches to 7.5 inches.

b. Length: 6, 8, and 12 inches, based on pavement width.
c. Depth: 0.5 to 0.625 inches.
d. Spacing: 12 inches center-to-center.
5. The pavement width was required to be at least 20 ft minimum (lanes and shoulders), and
the speed limit was 50 mi/h and greater.
6. No (that I am aware of). There may be sites that received updated signs but not in
conjunction with this treatment.
7. Communication with the field personnel through meetings to alleviate any

misconceptions and trepidation.
8. Minor issues with rumbles installed at the edge of the mat may cause accelerated
pavement edge degradation.
9. Communicate the intentions. Initially target overrepresented crash routes to convey safety
improvements. Provide experience from other public agencies as testimony. Leave room
for flexibility in design and implementation. Track crash statistics of comparative routes
that do not have rumbles to indicate missed opportunities for crash reductions.
RESPONSES FROM OHIO
Ohio staff responded to all questions. Their responses are listed in numeric order.
1. The “before condition” had no rumble strips.
2. They were installed on routes meeting the minimum requirements (i.e., shoulder width,
acceptable PCR….and not just resurfacing). They may have been added to district
46
pavement marking contracts as well as resurfacing in the year or two when they were
installed through the systematic program.
3. The installations were all milled rumble strips.
4. The following dimensions were used for the ELRS based on the pavement cross section
in the ELRS Standard Drawings provided by Kentucky:
a. Width: 5 inches ± 0.5 inches.

b. Length: 6, 10, and 16 inches, based on shoulder width.
c. Depth: 0.375 inches.
d. Spacing: 12 inches center-to-center.
5. The treatment sites were installed as a combination of adequate shoulder (2 ft or greater),

acceptable pavement condition (PCR rating of 80 or higher), minimum lane widths (11
ft), two-lane routes outside of urban areas and locations where the speed limit is 45 mi/h
or greater.
6. No. Sites were gathered if they met the criteria shown in question 5.
7. After surveying a few of the ODOT Districts, no notable challenges were encountered
upon installation.
8. After surveying a few of the ODOT Districts, no notable challenges were mentioned in
regard to the maintenance of the rumble stripes.
9. The recommendation would be to have some sort of policy set that governs where they
will/can be installed. We all know that they provide a safety benefit; however, we are still
selective of where they can be placed depending on shoulder width and pavement
condition. The goal is to uphold the condition of the roadway system as best we can
while continuing to improve safety.
47
ACKNOWLEDGEMENTS
This report was prepared for the FHWA Office of Safety Research and Development under
Contract DTFH61-13-D-00001. The FHWA Program and Task Manager for this project was
Roya Amjadi.
The project team gratefully acknowledges the participation of the following organizations for
their assistance in this study:
• Kentucky Transportation Cabinet.

• Ohio Department of Transportation.
49
REFERENCES
1. Federal Highway Administration. (2017). Rumble Strips Community of Practice, Federal

Highway Administration, Washington, DC. Available online: https://safety.fhwa.dot.gov/
roadway_dept/pavement/rumble_strips/, last accessed May 12, 2017.
2. American Association of State Highway and Transportation Officials. (2005). AASHTO

Strategic Highway Safety Plan, American Association of State Highway Transportation
Officials, Washington, DC.
3. Federal Highway Administration. (2009). Manual on Uniform Traffic Control Devices for
Streets and Highways, Federal Highway Administration, Washington, DC.
4. Federal Highway Administration. (2015). Rumble Strips and Rumble Stripes, Federal
Highway Administration, Washington, DC. Available online: http://safety.fhwa.dot.gov/
roadway_dept/ pavement/rumble_strips/design-and-construction.cfm, last accessed
December 23, 2015.
5. Lindly, J. and Narci, A. (2006). Evaluation of Rumble Stripe Markings, UTCA Report 04405,
University Transportation Center for Alabama, University of Alabama, Tuscaloosa, AL.
6. Miles, J.D., Carlson, P.J., Pratt, M.P., and Thompson, T.D. (2005). Traffic Operational
Impacts of Transverse, Centerline, and Edgeline Rumble Strips, Report No. FHWA/TX-05/0-
4472-2, Texas Department of Transportation, Austin, TX.
7. Carlson, P.J., Miles, J.D., Pratt, M.P., and Pike, A.M. (2005). Evaluation of Wet-Weather
Pavement Markings: First Year Report, Report No. FHWA/TX-06/0-5008-1, Texas
Department of Transportation, Austin, TX.
8. Carlson, P.J., Miles, J.D., Pike, A.M., and Park, E.S. (2007) Evaluation of Wet-Weather and
Contrast Pavement Marking Applications: Final Report, Report No. FHWA/TX-07/0-5008-
2, Texas Department of Transportation, Austin, Texas.
9. Willis, J. and Dean, W. (2004). Mississippi’s Rumble Stripe Experience, 83rd Annual
Meeting of the Transportation Research Board, Washington, DC.
10. Torbic, D., Hutton, J., Bokenkroger, C., Bauer, K., Harwood, D., Gilmore, D., Dunn, J.,
Ronchetto, J., Donnell, E., Sommer III, H., Garvey, P., Persaud, B., and Lyon, C. (2009).
Guidance for the Design and Application of Shoulder and Centerline Rumble Strips, National
Cooperative Highway Research Board Report 641, Transportation Research Board,
Washington, DC.
11. Ligon, C.M., Carter, E.C., Joost, D.B., and Wolman, W.F. (1985). Effects of Shoulder
Textured Treatments on Safety, Report No. FHWA/RD-85/027, Federal Highway
Administration, Washington, DC.
51
12. Chaudoin, J.H. and Nelson, G. (1985). Interstate Routes 15 and 40 Shoulder Rumble Strips,
Report No. Caltrans-08-85-1, California Department of Transportation, Sacramento, CA.
13. Annino, J.M. (2003). Rumble Strips in Connecticut: A Before/After Analysis of Safety
Benefits, Connecticut Department of Transportation, Newington, CT.
14. Griffith, M.S. (1999). “Safety Evaluation of Rolled-in Continuous Shoulder Rumble Strips
Installed on Freeways,” Transportation Research Record, 1665, pp. 28–34, Transportation
Research Board, Washington, DC.
15. Stutts, J.C. (2000). NCHRP Synthesis of Highway Practice 287: Sleep Deprivation
Countermeasures for Motorists Safety, Transportation Research Board, Washington, DC.
16. Gårder, P. and Alexander, J. (1995). Continued Research on Continuous Rumble Strips:
Final Report, Technical Report 94-4, Maine Department of Transportation, Augusta, ME.
17. Morena, D.A. (2003). The Nature and Severity of Drift-Off Road Crashes on Michigan
Freeways, and the Effectiveness of Various Shoulder Rumble Strip Designs, 82nd Annual
Meeting of the Transportation Research Board, Washington, DC.
18. Carrasco, O., McFadden, J., Chandhok, P., and Patel, R. (2004). Evaluation of the
Effectiveness of Shoulder Rumble Strips on Rural Multi-Lane Divided Highways in
Minnesota, 83rd Annual Meeting of the Transportation Research Board, Washington, DC.
19. Patel, R.B., Council, F.M., and Griffith, M.S. (2007). Estimating the Safety Benefits of
Shoulder Rumble Strips on Two-Lane Rural Highways in Minnesota: An Empirical Bayes
Observational Before–After Study, 86th Annual Meeting of the Transportation Research
Board, Washington, DC.
20. Marvin, R.R. and Clark, D.J. (2003). An Evaluation of Shoulder Rumble Strips in Montana,
Montana Department of Transportation, Helena, MT.
21. Perrillo, K. (1998). The Effectiveness and Use of Continuous Shoulder Rumble Strips,
Federal Highway Administration, Albany, NY. Available online: https://safety.fhwa.dot.gov/
roadway_dept/research/conts_rumble/conts_rumble.pdf, last accessed May 12, 2017.
22. Hickey, Jr., J.J. (1997). “Shoulder Rumble Strip Effectiveness, Drift-Off-Road Accident
Reductions on the Pennsylvania Turnpike,” Transportation Research Record, 1573, pp. 105–
109, Transportation Research Board, Washington, DC.
23. Tennessee Department of Transportation. (2005). Statewide Installation of Shoulder Rumble

Strips on Tennessee’s Interstates, Tennessee Department of Transportation, Nashville, TN.
24. Cheng, E.Y.C., Gonzalez, E., and Christensen, M.O. (1994). “Application and Evaluation of
Rumble Strips on Highways,” Utah Department of Transportation, Compendium of Technical
Papers, 64th ITE Annual Meeting, Dallas, TX.
52
25. Chen, C., Darko, E.O., and Richardson, T.N. (2003) “Optimal Continuous Shoulder Rumble
Strips and the Effects on Highway Safety and the Economy,” ITE Journal, 73(5), Institute of
Transportation Engineers, Washington, DC.
26. Harwood, D.W. (1993). NCHRP Synthesis of Highway Practice 191: Use of Rumble Strips to
Enhance Safety, Transportation Research Board, Washington, DC.
27. Bahar, G., Masliah, M., Wolff, R., and Park, P. (2008). Desktop Reference for Crash
Reduction Factors, Report No. FHWA-SA-08-011, Federal Highway Administration,
Washington, DC.
28. Khan, M., Abdel-Rahim, A., and Williams, C. (2014). “Potential Crash Reduction Benefits
of Shoulder Rumble Strips in Two-Lane Rural Highways,” 93rd Annual Meeting of the
Transportation Research Board, Washington, DC.
29. Potts, I.B., Harwood, D.W., Bokenkroger, C.D., and Knoshaug, M.M. (2011). Benefit/Cost
Evaluation of MoDOT’s Total Striping and Delineation Program: Phase II, Report No. CM
12-002, Missouri Department of Transportation, Jefferson City, MO.
30. Lyon, C., Persaud, B., and Eccles, K. (2015). Safety Evaluation of Centerline Plus Shoulder
Rumble Strips, Report No. FHWA-HRT-15-048, Federal Highway Administration,
Washington, DC.
31. Sayed, T., DeLeur, P., and Pump, J. (2010). Impact of Rumble Strips on Collision Reduction
on Highways in British Columbia, Canada: Comprehensive Before-and-After Safety Study.
Transportation Research Record, 2148, pp. 9–15, Transportation Research Board,
Washington, DC.
32. Torbic, D., Bauer, K., Hutton, J., and Campbell, J. (2014). Delta Region Transportation
Development Program: Rural Safety Innovation Program Evaluation, Report No. FHWA-
SA-14-029, Federal Highway Administration, Washington, DC.
33. Kay, J., Savolainen, P.T., Gates, T.J., Datta, T.K., Finkelman, J., and Hamadeh, B. (2015).
Safety Impacts of a Statewide Centerline Rumble Strip Installation Program, Paper No. 15-
0179, 94th Annual Meeting of the Transportation Research Board, Washington, DC.
34. Olson, D., Sujka, M., and Manchas, B. (2013). Performance Analysis of Centerline and
Shoulder Rumble Strips Installed in Combination in Washington State, Report No. WA-RD
799.1, Office of Research and Library Services, Washington State Department of
Transportation, Olympia, WA.
35. Kubas, A., Kayabas, P., Vachal, K., and Berwick, M. (2013). Rumble Strips in North Dakota:
A Comparison of Road Segments, Safety, and Crash Patterns, prepared for North Dakota
Department of Transportation and Federal Highway Administration, North Dakota State
University, Fargo, ND.
53
36. Hauer, E. (1997). Observational Before–After Studies in Road Safety—Estimating the Effect
of Highway and Traffic Engineering Measures on Road Safety, Elsevier Science,
Incorporated, Amsterdam, Netherlands.
37. Bonneson, J. and Lord, D. (2005). Role and Application of Accident Modification Factors in
the Highway Design Process, Report No. FHWA/TX-05/0-4703-2, Texas Department of
Transportation, Austin, TX.
38. Anund, A., Kecklund, G., Vadeby, A., Hjälmdahl, M., and Akerstedt, T. (2008). “The
Alerting Effect of Hitting a Rumble Strip—A Simulator Study with Sleepy Drivers,”
Accident Analysis and Prevention, 40, pp. 1,970–1,976, Elsevier Science, Incorporated,
Amsterdam, Netherlands.
39. Kentucky State Police. (2017). Kentucky Collision Analysis for the Public, Kentucky State
Police, Frankford, KY. Available at http://crashinformationky.org. Last accessed February 1,
2017.
40. American Association of State Highway and Transportation Officials. (2010). Highway
Safety Manual, American Association of State Highway and Transportation Officials,
Washington, DC.
41. Office of Management and Budget. (2003). Circular A-4: Regulatory Analysis, Office of
Management and Budget, Washington, DC.
42. Council, F., Zaloshnja, E., Miller, T., and Persaud, B. (2005). Crash Cost Estimates by
Maximum Police-Reported Injury Severity Within Selected Crash Geometries, Report No.
FHWA-HRT-05-051, Federal Highway Administration, Washington, DC.
43. U.S. Department of Transportation. (2014). Guidance on Treatment of the Value of a

Statistical Life (VSL) in U.S. Department of Transportation Analyses—2014 Adjustment,
Memo dated June 13, 2014. Available online:
http://www.dot.gov/sites/dot.gov/files/docs/VSL_Guidance_ 2014.pdf, last accessed
November 12, 2014.
44. Zaloshnja, E., Miller, T., Council, F., and Persaud, B. (2006). “Crash Costs in the United
States by Crash Geometry,” Accident Analysis and Prevention, 38(4), pp. 644–651, Elsevier
Science, Incorporated, Amsterdam, Netherlands.
54
HRDS-20/12-17(200)E

2rumble Strips

Uploaded by

Copyright:

Available Formats

2rumble Strips

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

2rumble Strips

Uploaded by

Copyright:

Available Formats

Safety Evaluation of Edge-Line

Rumble Stripes on Rural Two-Lane

Research, Development, and Technology

Jonathan Porter, Ph.D.

The U.S. Government does not endorse products or manufacturers. Trademarks or

Quality Assurance Statement

7. Author(s) 8. Performing Organization Report No.

EXECUTIVE SUMMARY ...........................................................................................................1

Figure 1. Illustration. Examples of longitudinal rumble strip markings ......................................... 5

Table 1. Summary of studies on changes in target collision frequency from application of

AADT average annual daily traffic

The Federal Highway Administration (FHWA) established the Development of Crash

Figure 2. Illustration. Overview of rumble strip dimensions.(4)

The Mississippi Department of Transportation installed ELRSs on a portion of Interstate 59 with

California(14) comparison sites and a comparison

• 0.84 for SVROR crashes.

Table 2. Rumble strip dimensions from Lyon et al.(30)

• 0.80 for total crashes (excluding intersection-related and animal crashes).

• Larger reductions in head-on and sideswipe-opposite-direction crashes for lower traffic

• 15.8 percent for total crashes.

• 17.2 percent for total crashes.

For simultaneous installations, the application resulted in a 63.3-percent reduction in lane

The composite analysis indicated a 66-percent reduction in lane-departure crashes and a

• Total (all types and severities combined).

A further objective was to address questions of interest including the following:

• Do effects vary by level of traffic volumes?

Other questions included the following:

• Are crash migration effects evident?

• Identify appropriate untreated reference sites.

Rate A Rate B Rate C Rate D

10 550 2,208 550 451 342 1,373 342 280

Table 5. Minimum required before-period mile-years for treated sites—injury crashes.

10 1,220 5,984 1,485 1,100 759 3,719 923 683

10 1,973 8,065 910 589 1,227 5,013 565 366

Table 7. Minimum required before-period mile-years for treated sites—nighttime crashes.

10 5,797 23,188 2,775 1,427 3,603 14,413 1,724 888

Table 9. Analysis for crash effects in Kentucky.

Table 10. Analysis for crash effects in Ohio.

• Accounted for time trends.

• Reduced the level of uncertainty in the estimates of safety effect.

• Properly accounted for differences in crash experience and reporting practice in

In estimating λ , the effects of regression-to-the-mean and changes in traffic volume were

m = w(P) + (1- w)(x) ,

Figure 4. Equation. EB estimate of expected crashes.

Figure 6 illustrates the estimate of the index of effectiveness ( ).

Figure 7 illustrates the standard deviation of  .

 Var ( sum ) Var ( sum ) 

• Crash query dates were from 1/1/2004 to 12/31/2012.

Treatment Cost Data

Table 11. Kentucky treatment cost and service life data.

• Crash query dates were from 1/1/2005 to 12/31/2013.

Table 12. Ohio treatment cost data.

DATA CHARACTERISTICS AND SUMMARY

Table 13. Definitions of crash types.

Table 16. Data summary for Ohio reference sites.

CRASH SPILLOVER AND MIGRATION

𝐶𝑟𝑎𝑠ℎ𝑒𝑠/𝑚𝑖𝑙𝑒/𝑦𝑒𝑎𝑟 = 𝑒 𝑎 ∗ 𝐴𝐴𝐷𝑇 𝑏 ∗ 𝐿𝑐 ∗ 𝑒 (𝑖𝑛𝑣𝑟𝑎𝑑𝑖𝑢𝑠 ∗𝑑+𝑟𝑖𝑔 ℎ𝑡_𝑠ℎ𝑜𝑢𝑙𝑑𝑒𝑟 ∗𝑒)

Table 17. Reference group SPF for total crashes—parameter estimates.

Figure 9. Graph. Yearly multiplier by year for reference groups in Kentucky.